Backarc Basins

Backarc Basins
Tectonics and Magmatism

Backarc Basins
Edited by
Brian Tay/or
University o.fHawaii at Manoa
Honolulu. Hawaii
Springer Science+Business Media, LLC

Library of Congress Cataloging-in-Publication Data
Backarc basins tectonics and magmatism / edited by Brian Taylor.

p. cm.
Includes bibliographical references and index.
ISBN 978-1-4613-5747-6 ISBN 978-1-4615-1843-3 (eBook)
DOI 10.1007/978-1-4615-1843-3
1. Back-arc basins. 2. Magmatism. 3. Geology, Structural.
I. Taylor, Brian, 1953-
OE511.2.B33 1995
551.8--dc20 95-2949
CIP
ISBN 978-1-4613-5747-6
© 1995 Springer Science+Business Media New York

Originally published by Plenum Press, New York in 1995
Softcover reprint of the hardcover 1st edition 1995
All rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any
means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written
permission from the Publisher
Contributors
Andrey A. Andreyev • Institute of Marine Geology and Geophysics, Yuzhno-Sakha-

linsk, Sakhalin 693002, Russia
Jean-Marie Auzende • IFREMER/CB, 29280 Plouzane, France. Present address:

ORSTOM, BPA5, Noumea, New Caledonia
Peter F. Barker • British Antarctic Survey, National Environment Research Council,

Cambridge CB3 OET, England
Herve Bellon • CNRS URA 1278-GDR "GEDO" 910, Universite de Bretagne Occi-
dentale, 29287 Brest, France
J. W. Cole • Department of Geology, University of Canterbury, Christchurch, New

Zealand
Joseph Cotten • CNRS URA 1278-GDR "GEDO" 910, Universite de Bretagne Occi-
dentale, 29287 Brest, France
D. J. Darby • Institute of Geological and Nuclear Sciences, Lower Hutt, New Zealand
Kathleen A. Devaney • Department of Geological Sciences, University of Texas at El

Paso, El Paso, Texas 79968
Jean-Philippe Eissen • Antenne ORSTOM-IFREMERlCB, 29280 Plouzane, France
Martin R. Fisk • College of Oceanic and Atmospheric Sciences, Oregon State Univer-
sity, Corvallis, Oregon 97331-5503
Patricia Fryer • Hawaii ~nstitute of Geophysics and Planetology, School of Ocean

and Earth Sciences and Technology, University of Hawaii at Manoa, Honolulu, Hawaii
96822
J. A. Gamble • Department of Geology, Research School of Earth Sciences, Victoria

University of Wellington, Wellington, New Zealand
Martine Gerard • ORSTOM, 93143 Bondy, France
Helios S. Gnibidenko • Shirshov Institute of Oceanology, Russian Academy of Sci-

ences, Moscow 117851, Russia. Present address: Geodynamics Research Institute, Texas
A&M University, College Station, Texas 77843
v
vi CONTRIBUTORS
Elena V. Gretskaya • Institute of Marine Geology and Geophysics, Yuzhno-Sakha-

linsk, Sakhalin 693002, Russia
James W. Hawkins, Jr. • Geological Research Division, Scripps Institution of Ocean-

ography, La Jolla, California 92093-0220
Thomas W. C. Hilde • Geodynamics Research Institute, Texas A&M University,

College Station, Texas 77843
Shu-Kun Hsu • IFREMER, Centre de Brest, 29280 Plouzane, France
Jun-ichiro Ishibashi • Laboratory for Earthquake Chemistry, Faculty of Science, Uni-

versity of Tokyo, Bunkyo-ku, Tokyo 113, Japan
Jean-Louis Joron • Groupe des Sciences de la Terre, Laboratoire Pierre-SUe, CEN

Saclay, 91191 Gif sur Yvette, France
Shigeru Kasuga • Hydrographic Department, Maritime Safety Agency, Chuo-ku,

Tokyo 104, Japan
Randall A. Keller • College of Oceanic and Atmospheric Sciences, Oregon State

University, Corvallis, Oregon 97331-5503
Kazuo Kobayashi • Japan Marine Science and Technology Center, Yokosuka 237,
Japan
Lawrence A. Lawver • Institute for Geophysics, University of Texas at Austin, Austin,

Texas 78759-8397
Char-Shine Liu • Institute of Oceanography, National Taiwan University, Taipei,

Taiwan, Republic of China
Patrick Maillet • ORSTOM, Centre de Brest-GDR "GEDO" 910, 29280 Plouzane,

France
Kathleen M. Marsaglia • Department of Geological Sciences, University of Texas

at EI Paso, EI Paso, Texas 79968
Setsuya Nakada • Kyushu University, Fukuoka 812, Japan
Kyoko Okino • Hydrographic Department, Maritime Safety Agency, Chuo-ku, Tokyo

104, Japan
Bernard Pelletier • ORSTOM, Noumea, New Caledonia
Alain Person • Laboratoire de Geologie des Bassins Sedimentaires, Universite Pierre

et Marie Curie, 75252 Paris, France
Richard C. Price • La Trobe University, Bundoora, Victoria 3083, Australia
Etienne Ruellan • CNRS Sophia-Antipolis, 06560 Valbonne, France

CONTRIBUTORS vii
Chuen-Tien Shyu • Institute of Oceanography, National Taiwan University, Taipei,

Taiwan, Republic of China
Jean-Claude Sibuet • IFREMER, Centre de Brest, 29280 Plouzane, France
T. A. Stern • Research School of Earth Sciences, Victoria University of Wellington,

Wellington, New Zealand
Jorge A. Strelin • Departamento de Ciencias de la Tierra, Instituto Antartico Argen-

tino, Buenos Aires, Argentina
Kensaku Tamaki • Ocean Research Institute, University of Tokyo, Nakano-ku, Tokyo

164, Japan
Tetsuro Urabe • Geological Survey of Japan, Tsukuba 305, Japan
I. C. Wright • New Zealand Oceanographic Institute, National Institute of Water and

Atmospheric Research, Wellington, New Zealand
Preface
This is the first book to provide a comprehensive treatment of the tectonics and magmatism
of backarc basins, from their initial rift stage to mature spreading. It focuses on the young
backarc basins of the circum-Pacific, where the volcano-tectonic processes are best studied
because they are still active. Twelve chapters describe the Taupo-Havre-Lau system, the
North Fiji Basin, the New Hebrides, Mariana, Bransfield, and Okinawa troughs, the East
Scotia and Japan Seas, and the Shikoku and Kuril Basins. Two chapters provide syntheses
of hydrothermal activity related to arc-backarc magmatism and tectonic and magmatic
controls on basin sedimentation. The authors of each chapter are the recognized experts for
their area/topic.
The last ten years have seen fundamental changes in our understanding of the
tectonics, volcanism, hydrothermal circulation, and sedimentation in backarc basins. This
has resulted from exploration and detailed studies using swath bathymetry and sidescan
acoustic imagery, multichannel seismics, manned submersibles and remotely operated
vehicles, and deep ocean drilling. The studies in this volume summarize these new findings
and use them to build improved models of the dynamic processes occurring in backarc
basins. Below I preview some of the key themes woven throughout the chapters in this
book, together with my own observations.
VOLCANO-TECTONIC SETTING
Backarc basins are situated above mantle exhibiting strong seismic wave attenuation,
that is in turn above or beyond Wadati-Benioff seismic zones that trace the descent of
oceanic lithosphere from bathymetric trenches into the mantle. Not all subduction zones
have associated active arc and/or backarc volcanism, sometimes because the above wedge
of mantle is missing and/or because there is compressional coupling between the subduct-
ing and overriding plates. The formation of backarc basins involves the splitting of volcanic
arcs, initially forming rifts (elongate depressions bounded by faults) and subsequently
backarc spreading centers. They may form within continental or intraoceanic arcs.
CAUSE AND LOCATION
The trench axes of all subduction systems with active backarc basins are observed to
migrate seawards in a hotspot reference frame. Whether the sinking of the subducting plate
beneath the overriding plate is a cause or an effect of backarc opening remains debated, but
presently a majority favor the former. Depending on its other boundary conditions, the
upper plate (being horizontally coupled by suction to the downgoing plate) is pulled
ix
x PREFACE
seawards and stretched. Periodically the stretching is sufficient to pull the weakest part of
the upper plate apart. This occurs where the plate is both hottest and has the thickest crust
(the stronger mantle is thinnest), namely within about 50 kilometers of the volcanic front.
Sometimes this is predominantly on the trench (forearc) side (e.g., Lau Basin), sometimes
in the backarc (e.g., New Hebrides and Okinawa troughs), and sometimes variably on both
sides along strike (e.g., Izu-Bonin-Mariana system).
INITIAL RIFTING
The formation of backarc basins involves extensive rifting of the preexisting arc
massifs. The initial rift basins subside along a zigzag pattern (in plan view) of border faults,
many of which dip at low angles (20-50°). Transfer zones that link opposing master faults
and/or rift flank uplifts further subdivide the rifts into segments along strike. The rifts are
often better developed between the arc volcanoes: magmatism, rather than stretching and
subsidence, often fills the opening adjacent to arc volcanoes. Explosive, high-silica volca-
nism commonly reaches peak intensities during the initial rifting (as documented in the
Japan, Izu-Bonin-Mariana, and Lau-Colville-New Zealand arcs). Nearby calderas shed
voluminous pumiceous sediments into rapidly filled rift basins.
RIFT MAGMAS
Magmas intrude along faults into the rifted arc crust of the basins, sometimes produc-
ing lineated magnetic anomalies. The magmas may be derived from the suprasubduction
zone mantle arc source and/or from pressure-release partial melting of the shallow mantle:
both arc-like and mid-ocean ridge basalt (MORB)-like compositions are observed tempo-
rally and spatially juxtaposed. Nevertheless, slightly wet but otherwise MORB-like lavas
erupt in even the earliest rifts, very proximal (-10 km) to arc volcanoes. Isotope data
indicate that the mantle source region may be replenished by trench ward flow of aestheno-
sphere.
CONTINUED STRETCHING
As continued stretching widens the basins, the zone of greatest subsidence and
intrusion migrates with the collapsing arc margin border faults. The result is an asymmetric
basin in cross section, with a wider zone of stretched and intruded crust on the proto-
remnant arc side whether the rifting initiated in the forearc or in the backarc. In the former
case, there may be no active frontal arc, the proto-remnant arc may remain active during
initial rifting, and half-graben depocenters within the rift basin may be largely isolated from
margin sediments.
INITIAL SPREADING
After continued stretching and intrusion for several million years has produced a rift
basin up to 200 kilometers wide, the shallow and deep sources of mantle partial melts
PREFACE xi
become horizontally separated and an organized spreading center forms, fed by the shallow
mantle. This evolution is spatially transgressive: just as rifting migrates laterally, spreading
is localized in some areas first (commonly along a basin-bounding transform fault) and then
propagates longitudinally into adjacent rifting areas. Eventually the suprasubduction zone
arc magmas establish a new volcanic front along the rifted edge of the old one (the South
Sandwich arc is actually built on backarc basin oceanic crust). However, in several but not
all (Mariana) systems, arc activity declined markedly at the start of subsequent backarc
spreading (or even terminated in the Izu-Bonin-Shikoku Basin case).
MATURE SPREADING
The recognition that backarc basins have an extensive rift stage prior to spreading has
helped reconcile some of their initially perceived differences to mid-ocean spreading
centers. The study of mature systems in the northern Lau, North Fiji, Manus, central
Mariana, and East Scotia basins has confirmed first-order similarities in backarc spreading
parameters (rates, morphotectonics, magnetization, crustal structure, and geochemistry) to
those of mid-ocean ridges. Nevertheless there remain important differences, particularly in
the diversity of magma, hydrothermal fluid, and sediment types, and possibly in the thermal
and crustal structures, associated with their suprasubduction zone setting.
HYDROTHERMAL ACTIVITY
One sixth of the known sites of seafloor hydrothermal activity has been discovered in
backarc basins. Silica, sulfate, and sulfide chimneys venting clear, white, and black solu-
tions at temperatures of 200° to 400°C have been found on all the active backarc spreading
centers studied (Lau, North Fiji, Manus, and Mariana). These fluids have low pH (2 to 5),
have evidence of direct interaction with magma, and typically contain more zinc, barium,
lead, cadmium, and arsenic than their mid-ocean ridge equivalents. Hydrothermal deposits
discovered in backarc rifts such as the Okinawa Trough and Sumisu Rift are a modern
analog of the Kuroko-type volcanic-hosted massive sulfide (VMS) deposits that formed in
the middle Miocene Hokuroku Basin during the last phase of Japan Sea opening. Like all
sites of major VMS ore bodies, and unlike mid-ocean ridge deposits, they occur in
association with subduction-related volcanism including silica-rich lavas as well as basalt.
Acknowledgments
Diane Henderson of the SOEST Publications Group was of great assistance in
coordinating, reformatting and copyediting the papers. I thank all the authors for contribut-
ing their work. The rigorous manuscript reviews by those listed below significantly im-
proved the volume: Dallas Abbott, Hugh Bibby, Thomas Brocher, Glenn Brown, Anthony
Crawford, Anthony Ewart, Fred Hochstaedter, Kevin Johnson, Lynn Johnson, Donna
Jurdy, David Kemp, Loren Kroenke, Robert Larter, Lawrence Lawver, Neil Lundberg,
Alex Malahoff, Fernando Martinez, James Natland, Richard Price, Ian Smith, Robert Stern,
and Michael Underwood.
Brian Taylor
Contents
Chapter 1
Taupo Volcanic Zone and Central Volcanic Region: Backarc Structures of
North Island, New Zealand
J. W. Cole, D. J. Darby, and T. A. Stern
ABSTRACT ...................................................... .
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. Plate Tectonic Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
3. Taupo-Hikurangi Arc-Trench System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
4. Distinction between Central Volcanic Region and Taupo Volcanic Zone .... 3
5. Earlier Volcanic Arcs ............................................. 4
6. Offshore Bay of Plenty ........................................... 5
7. Taupo Volcanic Zone Chronology ................................... 6
8. Central Volcanic Region Structure .................................. 7
9. Geophysical Data Bearing on Crustal and Upper Mantle Structure, Heat
Flux, and Kinematics ............................................. 7
9.1. Earthquake Seismology ....................................... 8
9.2. Explosion Seismology ........................................ 9
9.3. Interpretation of the 7.4-7.5 km/s Layer ......................... 9
9.4. Gravity Data .................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
9.5. Heat Flux .................................................. 11
10. Geodetic Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
10.1. Large-Scale Horizontal Deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
10.2. Intermediate-Scale Horizontal Deformation ...................... 14
10.3. Vertical Deformation ........................................ 15
10.4. Active Surface Faulting ...................................... 15
11. Focal Mechanisms ............................................... 16
12. Kinematics ..................................................... 17
13. Magma Types and Genesis ........................................ 17
13.1. High-AI Basalts ............................................ 17
13.2. Basaltic Andesite, Andesite, and Dacite ......................... 20
13.3. Rhyolites and Ignimbrites .................................... 21
14. Discussion........... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
15. Summary and Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
xiii
xiv CONTENTS
Chapter 2
The Southern Havre Trough: Geological Structure and Magma Petrogenesis
of an Active Backarc Rift Complex
J. A. Gamble and /. C. Wright
ABSTRACT ....................................................... 29
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2. Regional Structure and Tectonics .................................... 32
2.1. General ..................................................... 32
2.2. Southern Havre-Kennadec Backarc System ....................... 32
2.3. Taupo-Havre Continental-Oceanic Transition ...................... 33
3. Structure and Morphology of the 35°30'S-37°S Sector .................. 34
3.1. General ..................................................... 34
3.2. Outer Rift Graben. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.3. Inner Rift Graben ............................................. 37
3.4. Rift Volcanism ............................................... 38
3.5. Rift Flank Structure and Volcanism .............................. 38
3.6. Seismicity ................................................... 39
3.7. Heat Flow ................................................... 40
3.8. Volcanic Front ............................................... 40
4. Structure and Morphology of the 33°S-34°S Sector ............... . . . . . . 41
4.1. Rift System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.2. Volcanic Front ............................................... 41
5. Tectonics of the Southern Havre Trough .............................. 42
5.1. Rifting Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
5.2. Age and Rate of Extension ..................................... 44
6. Geochemistry .................................................... 44
6.1. General ..................................................... 44
6.2. Kennadec Arc . . . . . .. . . .. . . . . . . .. . .. . . . . . . . . . . . . . .. . . . . . . . . .. . 44
6.3. Southern Havre Trough-Ngatoro Rift System ...................... 45
6.4. Offshore TVZ Volcanic Front and Backarc ........................ 48
7. Petrogenesis ..................................................... 50
7.1. General ..................................................... 50
7.2. Taupo-Havre-Lau System ..................................... 51
8. Synthesis ........................................................ 54
Acknowledgments ................................................ 58
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Chapter 3
The Geology of the Lau Basin
James W. Hawkins, Jr.
ABSTRACT ....................................................... 63
1. Backarc Basins ................................................... 64
CONTENTS xv
2. Geologic Setting of the Lau Basin ................................... 67

2.1. Tonga Ridge ................................................. 67
2.2. Tonga Trench ................................................ 68
2.3. Lau Ridge ................................................... 69
2.4. Lau Basin-An Overview ...................................... 69
2.5. The New View of Lau Basin Evolution ........................... 72
3. Lau Basin Morphologic Provinces ................................... 73
3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.2. Eastern and Central Lau Spreading Centers ........................ 74
3.2.1 Central and Eastern Lau Spreading Centers ................... 74
3.2.2. Mangatolu Triple Junction ................................ 75
3.2.3 Northwestern Lau Spreading Center ......................... 75
3.3. Western Extensional Basin. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.4. Peggy Ridge and the Northern Basin ............................. 76
4. Petrologic Discussion of the Neovolcanic Zones ........................ 77
4.1. SSZ Mantle Influence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.2. Eastern and Central Lau Spreading Centers ........................ 78
4.2.1. Introduction ............................................ 78
4.2.2. Petrography-ELSC and CLSC ............................ 84
4.2.3. Petrology and Geochemistry-ELSC and CLSC .............. 85
4.3. Valu Fa Ridge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.3.1. Introduction ............................................ 93
4.3.2. Petrography-VFR ...................................... 93
4.3.3. Geochemistry-VFR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.3.4. Petrogenesis-VFR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.4. Mangatolu Triple Junction ............... . . . . . . . . . . . . . . . . . . . . . . . 95
4.4.1. Introduction ............................................ 95
4.4.2. Petrography and Petrology ................................ 96
5. Western Extensional Basin. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
5.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
5.2. Old Crust ................................................... 98
5.3. MORB-like Drill Sites ......................................... 99
5.4. Arc1ike Drill Sites ............................................ 102
6. The Peggy Ridge and Seamounts of the Northern Basin . . . . . . . . . . . . . . . . . . 103
6.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6.2. Geology of Peggy Ridge ....................................... 103
6.3. Geology of Donna Seamount. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
6.4. Geology of Rochambeau Bank .................................. 106
6.5. Geology of Niuafo'ou Island. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
6.6. Geology of Zephyr Shoal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
7. Lau Basin Petrologic Evolution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
7.1. Introduction........................... . . . . . . . . . . . . . . . . . . . . . . . 107
7.2. Nature of the Magma Source. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
7.3. The Immobile Elements-HFSE and REE . . . . . . . . . . . . .. .. . . . . . . . . . 108
7.4. Titanium and Vanadium ........ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
7.5. Alkalis and Alkaline Earth Elements ............................. 111
7.6. Evidence for a Signature from Subducted Sediment ................. 112
xvi CONTENTS
8. Discussion of the Isotope Data ..................................... 113

8.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
8.2. Isotope Data ................................................ 114
9. Hydrothermal Activity ............................................ 118
9.1. Introduction................................................. 118
9.2. Hydrothermal Sediments ...................................... 118
9.3. Hydrothermal Vents, Crusts, and Massive Sulfides ................. 119
10. Backarc Basin Sedimentation ...................................... 120
11. Regional Geologic History. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
12. Summary and a Model for Lau Basin Evolution ....................... 123
Acknowledgments ............................................... 129
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 129
Chapter 4
The North Fiji Basin: Geology, Structure, and Geodynamic Evolution
Jean-Marie Auzende, Bernard Pelletier, and Jean-Philippe Eissen
ABSTRACT ....................................................... 139

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
2. Magnetism and Paleomagnetism .................................... 155
3. Seismicity ...................................................... 157
4. Heat Flow Data ................................................. 159
5. Rock Geochemistry .............................................. 160
6. Sulfide Deposits ................................................. 162
7. Water Sampling ................................................. 164
8. Discussion and Conclusion ........................................ 165
8.1. Hectokilometric, Decakilometric, and Kilometric Ridge Segmentation 165
8.2. Unusual Tectonic Features ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
8.3. Petrology and Geochemistry ................................... 166
8.4. Hydrothermal Activity: Water Chemistry and Sulfides .............. 167
8.5. Evolution of the Basin ........................................ 168
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Chapter 5
Tectonics, Magmatism, and Evolution of the New Hebrides Backarc Troughs
(Southwest Pacific)
Patrick Maillet, Etienne Ruellan, Martine Gerard, Alain Person, Herve Bellon,
Joseph Cotten, Jean-Louis Joron, Setsuya Nakada, and Richard C. Price
ABSTRACT ....................................................... 177

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
CONTENTS xvii
2. Geological and Tectonic Framework of the New Hebrides Island Arc and the
North Fiji Basin .................................................. 178
2.1. The New Hebrides Island Arc ................................... 179
2.1.1. Western Belt ........................................... 179
2.1.2. Eastern Belt ............................................ 181
2.1.3. Central Chain. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
2.2. The North Fiji Basin .......................................... 182
3. The New Hebrides Backarc Troughs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
3.1. Previous Work and Recent Investigations. . . . . . . . . . . . . . . . . . . . . . . . . . 183
3.2. Structure and Tectonics ........................................ 187
3.2.1. The Jean-Charcot Troughs ................................ 187
3.2.2. The Coriolis Troughs .................................... 188
3.2.3. The Aoba Basins ........................................ 192
3.2.4. The Vanikoro-Torres Basin ............................... 192
3.3. Volcanic Petrology, Geochronology, and Geochemistry .............. 194
3.3.1. Volcanic Petrology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
3.3.2. Geochronology.......................................... 196
3.3.3. Geochemistry........................................... 196
3.4. Backarc Hydrothermal Activity and Ferromanganese Crusts . . . . . . . . . .. 202
3.4.1. Backarc Hydrothermal Activity ............................ 202
3.4.2. Ferromanganese Crusts ................................... 222
4. Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 225
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 230
Chapter 6
Geology of the Mariana Trough
Patricia Fryer
ABSTRACT ....................................................... 237

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 238
2. Structure of the Mariana Trough ..................................... 241
2.1. Initiation of Rifting ........................................... 241
2.2. From Rifting to Spreading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 246
2.3. The Central Mariana Trough .................................... 249
2.4. Extension in the Southern Mariana Trough ........................ 256
3. Volcanism of the Mariana Trough. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 260
3.1. The Volcanic Front and Cross-Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 260
3.2. Intrabasin Faults and the Spreading Ridges ........................ 263
4. Hydrothermal Activity ............................................. 264
4.1. Cross-Chains and Volcanic Front ................................ 264
4.2. Central Spreading Basin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 265
4.3. Southern Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 266
5. Petrology and Petrogenesis ......................................... 267
xviii CONTENTS
5.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 267

5.2. Northern Rifting Basin and the Kasuga Cross-Chain . . . . . . . . . . . . . . . .. 268
5.3. Central Spreading Basin. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 270
5.4. Southern Platform. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 271
6. Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 272
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 274
Chapter 7
Tectonic Framework of the East Scotia Sea
Peter F. Barker
ABSTRACT ....................................................... 281

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 282
2. Regional Context of East Scotia Sea Development ...................... 283
2.1. Early Exploration ............................................. 283
2.2. Marine Geophysics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 283
2.3. Neotectonics: Regional Context. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 284
2.4. Neotectonics: The Subduction Zone .............................. 286
2.5. Scotia Sea Reconstructions ..................................... 287
2.6. East Scotia Sea Development and Structure ..................... . .. 289
3. New or Enlarged Data Sets ......................................... 290
3.1. Bathymetry .................................................. 290
3.2. Shipboard and Geostat GM Free-Air Gravity . . . . . . . . . . . . . . . . . . . . . .. 290
3.3. Magnetic Anomalies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 293
3.4. Seismic Reflection Profiles ..................................... 293
3.5. Onshore Geology and Dredged Rocks ............................ 293
4. Data Interpretation ................................................ 295
4.1. Context: The South Sandwich Island Arc and Trench ........... . . . .. 295
4.1.1. South Sandwich Island Arc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 295
4.1.2. South Sandwich Trench .................................. 296
4.2. Character of the Backarc ....................................... 298
4.3. Magnetic Anomaly Identifications, Spreading Rates, Asymmetry . . . . . .. 299
4.4. The Ridge Crest and Spreading Center Chemistry . . . . . . . . . . . . . . . . . . . 300
4.5. Western Flank: Older Anomalies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
4.6. Southern Province: Ridge Crest Collision and Rifting . . . . . . . . . . . . . . .. 302
4.6.1. Backarc Spreading History ................................ 303
4.6.2. Ridge Crest-Trench Collision ............................. 304
4.7. Eastern Flank: Volcanic Arc, Forearc, and Trench ................ . .. 306
4.8. Reconstructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 308
5. Conclusions................................................... . .. 310
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 312
CONTENTS xix
Chapter 8
Bransfield Strait, Antarctic Peninsula: Active Extension behind a Dead Arc
Lawrence A. Lawver, Randall A. Keller, Martin R. Fisk, and Jorge A. Strelin
ABSTRACT ....................................................... 315

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 316
2. Regional Tectonic Setting ......................................... 317
3. Tectonic Setting of Bransfield Strait ................................. 320
4. Earthquakes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 322
5. Seismic Refraction ............................................... 324
6. Seismic Reflection ............................................... 325
7. Magnetics ...................................................... 327
8. Heat Flow ...................................................... 327
9. Hydrothermal Activity ............................................ 328
10. Petrology and Geochemistry ....................................... 328
10.1. Regional .................................................. 328
10.1.1. Antarctic Peninsula Arc ............................... 328
10.1.2. Bransfield Strait .................................... " 330
10.1.3. James Ross Island Volcanic Group. . . . . . . . . . . . . . . . . . . . . .. 331
10.104. Temporal Comparison ................................. 333
10.1.5. Across-Arc Comparison ............................... 333
10.2. Intra-Bransfield Strait Comparison ............................. 336
10.2.1. On Axis versus off Axis ............................... 336
10.2.2. Submarine versus Subaerial ............................ 336
10.2.3. Along Axis ........................................ " 337
II. Conclusions ............ ;........................................ 337
References .................................................... " 338
Chapter 9
Structural and Kinematic Evolutions of the Okinawa Trough Backarc Basin
Jean-Claude Sibuet, Shu-Kun Hsu, Chuen-Tien Shyu, and Char-Shine Liu
ABSTRACT ....................................................... 343

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 344
2. Tectonic Setting of the Okinawa Trough ............................. 346
2.1. Northern Okinawa Trough ..................................... 346
2.2. Southern Okinawa Trough ..................................... 347
2.3. Middle Okinawa Trough ...................................... 349
3. Extensional Tectonics in the Okinawa Trough ......................... 350
3.1. Identification of Two Families of Faults ........................ " 351
3.2. Spatial Distribution of the Two Families of Faults ................. 355
3.3. Quantification of the Spatial Distribution of the Two Families of Faults 358
xx CONTENTS
4. Kinematics of the Okinawa Trough Opening ........................... 362

4.1. Determination of the Poles of Rotation Corresponding to the Early
Pleistocene and Recent Tectonic Phases ........................... 362
4.2. Determination of the Total Amount of Extension in the Okinawa Trough 364
4.3. Detennination of Parameters of the Total Rotation .................. 368
4.4. Trial to Quantify the Amount of Extension Corresponding to the Two
Last Tensional Phases ......................................... 370
4.5. Parameters of Rotations for the Three Tensional Phases .............. 370
5. Newly Compiled Bathymetric and Magnetic Data in the Southern Okinawa
Trough.......................................................... 372
5.1. New Bathymetric Map. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 372
5.2. New Magnetic Map ........................................... 373
5.3. Discussion............................................. . . . . .. 374
6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 376
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 377
Chapter 10
Shikoku Basin and Its Margins
Kazuo Kobayashi, Shigeru Kasuga, and Kyoko Okino
ABSTRACT ....................................................... 381

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
2. Topography of the Basin and Its Margins . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 384
2.1. General Bathymetric Features and Basement Topography . . . . . . . . . . . .. 384
2.2. The Southern Border of the Basin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 386
2.3. Axial Seamount Chain. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 386
2.4. Nishi-Shichito Ridge-The Eastern Border of the Basin ............. 388
2.5. Kyushu-Palau Ridge-The Western Border of the Basin . . . . . . . . . . . .. 388
3. Gravity Anomalies and Isostatic Compensation of the Basin and Its Margins 389
4. Stratigraphy of the Shikoku Basin .................................. " 389
5. Petrology and Ages of the Basin and Margins. . . . . . . . . . . . . . . . . . . . . . . . .. 393
6. Magnetic Anomalies and Spreading History of the Shikoku Basin . . . . . . . . .. 394
6.1. Rifting of the Shikoku Basin .................................... 396
6.2. Early Opening of the Shikoku Basin ............................ " 398
6.3. Spreading Phase from Anomaly 6B to Anomaly 6 .................. 398
6.4. Last Stage of Spreading and Eruption of the Kinan Seamounts ........ 398
6.5. Postspreading Off-Ridge Volcanism in the Shikoku Basin ............ 398
7. Uplift and Subsidence of the Basin and Its Margins ..................... 399
8. Interaction of the Shikoku Basin with Southwest Japan and Nankai Trough 400
9. Summary, Discussion, and Further Problems. . . . . . . . . . . . . . . . . . . . . . . . . .. 401
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 402
CONTENTS xxi
Chapter 11
Opening Tectonics of the Japan Sea
Kensaku Tamaki
ABSTRACT ....................................................... 407

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 407
2. Topography ........................ . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 408
3. The Japan and Yamato Basins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 410
4. Extension of Island-Arc Crust ....................................... 411
5. Formation Age ................................................... 413
6. Model of Opening of the Japan Sea .................................. 414
7. Common Process of Backarc Basin Opening ........................... 416
8. Conclusions........................ . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 418
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . .. 419
Chapter 12
Kuril (South Okhotsk) Backarc Basin
Helios S. Gnibidenko, Thomas W. C. Hilde, Elena V. Gretskaya,
and Andrey A. Andreyev
ABSTRACT ....................................................... 421

1. Introduction .................................................... " 421
2. Bathymetry ...................................................... 423
3. Geophysical Fields ................................................ 427
3.1. Magnetic Anomalies ......................................... " 427
3.2. Gravimetry .................................................. 428
3.3. Heat Flow ................................................... 428
3.4. Seismicity and Stress Field ..................................... 429
4. Tectonics ........................................................ 431
4.1. Crustal and Lithospheric Structure ............................... 431
4.2. Basement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 435
4.3. Composition ................................................. 435
4.4. Age ........................................................ 437
4.5. Structural Features ............................................ 438
5. Basin Sedimentary Fill ........................................... " 439
5.1. Seismic Stratigraphy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 439
5.2. Thickness and Distribution ..................................... 440
5.3. Composition and Age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 440
6. Holocene Sedimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 440
7. Discussion: Structural Evolution and Genesis .......................... 443
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 446
xxii CONTENTS
Chapter 13
Hydrothermal Activity Related to Arc-Backarc Magmatism in the Western
Pacific
lun-ichiro Ishibashi and Tetsuro Urabe
ABSTRACT ....................................................... 451

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 452
2. Tectonic Setting and Associated Volcanism ............................ 452
3. Case Studies ..................................................... 458
3.1. North Fiji Basin .............................................. 458
3.2. Manus Basin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 460
3.3. Lau Basin ................................................... 462
3.4. Mariana Trough .............................................. 463
3.5. Izu-Bonin (Ogasawara) Arc .................................... 466
3.6. Okinawa Trough . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 468
4. Discussion........ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 471
4.1. Volcanism and Hydrothermal Activity ............................ 471
4.2. Time Scale of Volcanic and Hydrothermal Events. . . . . . . . . . . . . . . . . .. 472
4.3. Hydrothermal Mineralization in Relation to Ancient Analogs. . . . . . . . .. 473
4.4. Mineralogy, Geochemistry, and Sulfur Isotopes of the Deposits. . . . . . .. 474
4.5. Chemical Composition of the Hydrothermal Fluid. . . . . . . . . . . . . . . . . .. 475
4.6. Gas Geochemistry of the Hydrothermal Fluid ...................... 481
4.7. Relationship of Fluid Chemistry to Seafloor Mineralization ........... 483
5. Summary........................................................ 483
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 485
Chapter 14
Tectonic and Magmatic Controls on Backarc Basin Sedimentation: The
Mariana Region Reexamined
Kathleen M. Marsaglia and Kathleen A. Devaney
ABSTRACT ....................................................... 497

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 497
2. Characteristics of and Controls on Intraoceanic Backarc Basin Sedimentation 499
3. Mariana Drilling Results ........................................... 501
4. Previous Petrographic Studies of Backarc Basin Sand . . . . . . . . . . . . . . . . . . .. 503
5. Methods ........ ~. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 504
6. Petrographic Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 505
7. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 506
7.1. Spatial and Temporal Trends in Sand Composition across the Mariana
Trough...................................................... 506
7.2. Relationship between Lava Geochemistry and Detrital Modes ......... 511
7.3. Comparison of Mariana Sand Composition to That of Other Backarc
Basins ...................................................... 512
CONTENTS xxiii
7.4. Implications of Forearc versus Backarc Rifting on Sedimentary Models 516

8. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 517
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 517
Index............................................................. 521
Backarc Basins
1
Taupo Volcanic Zone and Central

Volcanic Region
Backarc Structures of North Island,
New Zealand
J. W Cole, D. J. Darby, and T. A. Stern
ABSTRACT
Volcanism in North Island, New Zealand, is related to subduction within the Taupo-
Hikurangi arc-trench system. The central volcanic region (CVR) is a wedge-shaped basin,
approximately 4 m.y. old, which is defined by geophysical parameters. It corresponds to an
area of low density, low-velocity volcanic rocks and anomalous seismic properties within
the upper mantle. Seismic data also provide evidence for a 7.4-7.5 km/s layer at the
unusually shallow depth of about 15 km beneath CVR; this material is inferred to be par-
tially melted upper mantle. The youngest and active part of CVR is the Taupo volcanic zone
(TVZ), the currently active backarc basin ofthe subduction system where volcanism is <2
m.y. old. TVZ is an area of high-convective-heat output (4 x 109 W). This output is equiva-
lent to an average heat flow of about 800 mW/m2 , about 13 times greater than the conti-
nental norm and one of the highest on record for a backarc basin. Rates of extension in the
currently active TVZ are about 7 mm/yr in the north and 18 mm/yr in the south, derived
from shear strain rates varying between 0.18 and 0.5 x 1O- 6 /yr shear. Subsidence rates are
1-2 mm/yr. Focal mechanisms from recent small earthquakes are inconclusive, but most
appear to be transcurrent. The larger 1987 Edgecumbe (M=6.3) earthquake, in the Bay of
Plenty, had a normal focal mechanism. There are three volcanic assemblages in TVZ: high-
Al basalt, basaltic andesite-andesite-dacite, and rhyolite-ignimbrite, in order of increasing
volume. High-AI basalt and rhyolite occur in a bimodal assemblage associated with caldera
structures in the extensional backarc basin, while the basaltic andesite-andesite-dacite
assemblage occurs principally at the northern and southern ends of TVZ and in a narrow
frontal arc along the eastern margin. The andesitic assemblage is undoubtedly slab related
and was formed in a multistage process involving dehydration of the slab, anatexis of the
J. W. Cole • Department of Geology, University of Canterbury, Christchurch, New Zealand. D. J.

Darby • Institute of Geological and Nuclear Sciences, Lower Hutt, New Zealand. T. A. Stern • Re-
search School of Earth Sciences, Victoria University of Wellington, Wellington, New Zealand.
Backarc Basins: Tectonics and Magmatism, edited by Brian Taylor, Plenum Press, New York, 1995.
2 J. W. COLE et al.
overlying mantle, fractionation of mineral phases, and minor crustal assimilation. The
high-AI basalts are considered to be derived from a hot mantle wedge above the subduction
zone with variation due to fractionation and minor assimilation. Origin of the rhyolites and
ignimbrites is more controversial. They are most likely to be the result of either melting of
earlier-formed igneous rocks or sedimentary/metamorphic rocks of a restricted composi-
tion from within the crust, perhaps with a minor upper mantle source component, or partial
melting of a mantle-derived source with significant crustal contamination.
1. INTRODUCTION
The CVRffVZ offers a rare opportunity to study backarc processes from land-based
observations. Most backarc spreading areas are covered by ocean, which makes direct
observations of geology, volcanism, and earthquake seismology difficult. In particular,
reliable heat flow data from oceanic backarc basins are notoriously difficult to acquire
(Sclater et aI., 1980).
The results of a variety of techniques are described in this study including seismic
tomography, geodesy, geochemical analysis, and integrated heat flux. In this chapter three
workers from what at times have been seemingly disparate subdisciplines of the earth
sciences-geochemistry, geodesy, and geophysics-have attempted to produce a common
synthesis of knowledge on the CVRffVZ. During this synthesis one of the problems that
arose was the differing inferences drawn from geophysics and geology about what the
principal volcano-tectonic unit is in central North Island. This is a question we address
but do not finally resolve. In general, however, we believe that effective progress will be
made in learning about such as areas the CVRffVZ by such a cross-disciplinary approach.
2. PLATE TECTONIC SETTING
New Zealand lies at the boundary between the Pacific and Australian plates, and in the
northern part of New Zealand the plate boundary zone is dominated by subduction of
oceanic crust of the Pacific plate beneath continental crust of the Australian plate (Fig. 1.1).
The relative instantaneous pole of rotation for these two plates is sited at 62°S, 174°E by
Chase (1978), and rotation is estimated to be l.27°/m.y. As a result, subduction is oblique
under North Island (Fig. 1.1) to form the Taupo-Hikurangi arc-trench system (Cole and
Lewis, 1981). Subduction becomes progressively more oblique to the south until, in South
Island, continental crust of the Chatham Rise is being obliquely obducted onto continental
crust of the Australian plate in a zone of transpression about 10 km wide. Much of the
movement in this zone occurs as reverse-dextral movement on the Alpine Fault (Fig. 1.1).
3. TAUPO-HIKURANGI ARC-TRENCH SYSTEM
The Taupo-Hikurangi arc-trench system extends from the Hikurangi Trough off the
east coast of North Island, New Zealand, to the TVZ in central North Island (Fig. 1.1). The
Hikurangi Trough is slightly offset in terms of bathymetry from the Kermadec Trench to the
north (Carter, 1980) and merges southward into the Alpine Fault system.
BACKARC STRUCTURES OF NORTH ISLAND 3
FIGURE 1.1. Location of Pacific-Australian

plate boundary in the New Zealand region. Stip-
pled area represents continental crust around
N.z. Arrows show motion of Pacific plate rela-
tive to Australian plate (after Cole, 1990). Rates
are from Walcott (1987).
To the west of the trough is an accretionary prism, up to 150 kIn wide, which becomes
progressively older from east to west (Lewis, 1980; Cole and Lewis, 1981). The western part
of the accretionary prism is a frontal ridge of Upper Paleozoic-Mesozoic graywacke-
argillite which is undergoing rapid uplift (e.g., Lamb and Vella, 1987) and is cut by a major
zone of dextral strike-slip faults-the North Island shear belt (Fig. 1.1).
4. DISTINCTION BETWEEN CENTRAL VOLCANIC REGION AND TAUPO

VOLCANIC ZONE
The CVR is a wedge-shaped basin of predominantly Quaternary rhyolitic volcanism.

Calhaem (1973) noted that the northwestern boundary of CVR corresponded to the location
of a number oflow-potash andesites K-Ar dated at 4 Ma (although some ofthese rocks are
now known to be altered, so the error limits of the dates are high). The TVZ is the eastern
part of CVR and represents that portion «2 Ma) that is currently volcanically active. The
CVR and TVZ share a common eastern boundary with the present active volcanic front
or arc.
The CVR is the principal volcanic basin structure defined by seismic and gravity
parameters in central North Island (Fig. 1.2) and corresponds to an area oflow-density, low-
velocity volcanics that occupy both the eastern side of the Coromandel volcanic zone and
TVZ. The western boundary of the TVZ within the CVR is equivocal and is drawn
somewhat arbitrarily on Fig. 1.3 to correspond to (i) the westernmost limit of active faulting
in central North Island, (ii) the apparent position of the active volcanic front at 2 Ma, and
(iii) a marked change in structural trend in offshore Bay of Plenty. The CVR is associated
with the latter stages of a volcanic arc that has been in existence for at least 22 m.y. (see
4 J. W. COLE at al.
III Cole - alkaline andesite
~ Centrol Volcanic Reoion

[J Quaternary sediment.
~ Pliocene sediment,
D Miocene sediments
~ Greywacke
• Recently active andesite
vo6conoes
FIGURE 1.2. Map of the northern part of the North

30· Island showing the location of the central volcanic
region and the distribution of andesites, graywacke
and principal sedimentary basins.
following). The TVZ corresponds to a change in orientation of the plate boundary resulting
from opening of the Havre Trough in the past 4 m.y.
5. EARLIER VOLCANIC ARCS
Within Northland, outcrops of calc-alkaline andesite and basalt volcanism are found
interspersed among older volcanics of an obducted ophiolite sequence (Brothers and
Delaloye, 1982) and Permian-Triassic basement graywacke (Fig. 1.2). A number of workers
have noted an apparent migration of the low-potash andesites southeastward through the
117~ •
White Is
FIGURE 1.3. Map of the Taupo volcanic zone showing

location of probable calderas 0, Rotorua; 2, Haroharo; 3,
Kapenga; 4, Reporoa; 5, Mangakino; 6, Wbakamaru; 7,
Oruanui; 8, Taupo); high-AI basalts (triangles) and
andesite-dacite localities (filled circles) mentioned in the
a. High AI
~ r,~~~~~C) basalt text. (Ka, Kawerau; Ed, Edgecumbe; Wa, Waiotapu; Br,
O~km o Bock-orc
• Andesite
ond dacites
Broadlands; Ng, Ngatamariki; Ro, Rotokawa; Rp, Rolles
Peak; Wk, Wairakei; Ta, Tauhara).
Northland-Auckland peninsula with time (e.g., Kear, 1959; Hatherton, 1969). With the
availability of K-Ar dates of these andesites, various plate reconstructions have been
developed that either take the plate boundary during the Tertiary as being adjacent and
parallel to Northland (e.g., Calhaem, 1973; Cole and Lewis, 1981; Walcott; 1987, Cole 1990)
or that a substantially shallower dipping Pacific plate extended beneath Northland from the
current plate boundary (Brothers and Delaloye, 1982).
Regardless of what the exact orientation for the plate boundary was, two points should
be noted. First, the Northland-Waikato area has been subjected to up to 22 m.y. of
volcanism and, presumably, thermal weakening of the lithosphere; second, there has been a
more rapid migration of the andesite arc across the CVR than within Northland.
Two distinct volcanic arcs can be recognized in Northland (Fig. 1.4): a western
Waitakere arc comprising tholeiitic basalt to medium-K andesite erupted between 22 and 15
Ma (Hayward, 1979); geophysical anomalies and offshore drilling indicate that much ofthe
arc is offshore. A second arc (Coromandel volcanic zone) extends from the eastern side of
Northland through Great Barrier Island to Coromandel. The earliest activity was submarine
with the eruption of medium-K andesites, comparable in age to those of the Waitakere arc,
but volcanism continued in the Coromandel volcanic zone after the Waitakere arc ceased
activity. About 10 Ma, rhyolitic volcanism began on the east side of Coromandel volcanic
zone to form the Whitianga Group (Skinner, 1986), with numerous rhyolite domes and
extensive ignimbrites, some associated with calderas.
6. OFFSHORE BAY OF PLENTY
The western boundary of the TVZ is perhaps best illustrated in offshore Bay of Plenty
where a structure 40-50 km wide bounded by inward-facing normal fault zones is evident
(Wright, 1992). The western margin of the extensional basin is marked by the Tauranga
e Miocene andesites
M MagnetiC anomoly
\ "
'-,
,
\
FIGURE 1.4. Map of Northland and Coromandel ~Okm
showing location of Miocene (22-12 Ma) volcanic
arcs.
6 J. W. COLE at al.
...
.Ii :-. i~ 1'.j~IJ;y
Ma""Q v,\ll "
~ 'i : ~ JI/
'i ,/
Iolond\l\"p/ / ~
~ &_ili /0~~JI'
~~'
;;hi"li~
'"
.Tau,a"", /
Island iZ:/r1 .~~
0/);:::-
CVR / _uhara (Whale) Island
___
/ TVZ
or 0 om 20 FIGURE 1.5. Faulting in offshore
~_=~~---L ...L_..J..:::i2~:£=--...I___] Bay of Plenty (after Wright, 1992).
fault zone and the eastern margin by the White Island fault zone (Fig. 1.5). Within the basin
is a frontal nonvolcanic graben, a volcanic (arc) ridge, and a volcanic (back-arc) basin.
7. TAUPO VOLCANIC ZONE CHRONOLOGY
The commencement of volcanism in TVZ is difficult to establish. The first recognized

activity that can be regarded as part of TVZ is the formation of the Mangakino caldera. This
is sited at the intersection between the southern limit of the Coromandel volcanic zone and
TVZ (Fig. 1.3). It was active from 1.6 to 0.9 Ma (Table I) and was the site of nine
voluminous (> 1100 km3) ignimbrites erupted between 1.62 and 1.51 Ma (Ngaroma and
ignimbrites B and C) and 1.23 to 0.91 Ma (Ongatiti, Ahuroa, Rocky Hill, Kaahu, Marshall,
and Waioraka ignimbrites; Briggs et aI., 1993) together with two phreatomagmatic deposits
(units D and E; Wilson, 1986).
Silicic volcanics of the central part of the TVZ are associated with seven main centers
and associated calderas (e.g., Healy, 1962; Wilson et al., 1984; Cole 1990; Nairn et al.,
1994) (Fig. 1.3). The earliest activity is probably now completely buried by later units.
Wilson et al. (1984) recognized the Kapenga caldera as one of the oldest, and possibly the
source of the 0.71 Ma Waiotapu ignimbrite (Table I). Activity then moved to the SW with
the eruption of the Whakamaru Group ignimbrites -0.35 Ma (Wilson et al., 1986). These
were probably erupted over a significant time span and filled parts of the earlier Mangakino
and Kapenga structures. Since this time most rhyolitic activity has been from the Okataina,
Reporoa, and Taupo calderas on the E side of the TVZ (Fig 1.3), but the Mamaku ignimbrite
(-0.22 Ma) is likely to have come from the Rotorua caldera (Fig 1.3).
The last 0.20 m.y. of activity has been dominated by rhyolite dome formation,
explosive eruptions, and the eruption of ignimbrites (e.g., Rotoiti, Oruanui, and Taupo)
mainly from the Okataina or Taupo centers (Fig 1.3). Interspersed between these events has
been the eruption of minor «1% of total volume of TVZ volcanics) high-AI basalt,
commonly from fissures.
The most complete sequence of andesitic lavas occurs in the Tongariro volcanic center
at the southern end (Fig 1.3), where eruptive activity spans at least 0.25 m.y. (Hackett and
TABLE I
Stratigraphy of Ignimbrites from the Major Calderas, with Approximate Ages of Eruptiona
Ma Mangakino TaupolWhakamaru Kapenga/Reporoa RotoruaJOkataina
Taupo ig 0.0018 Rotoiti 0.065

Oruanui ig 0.0226 Mamaku 0.22
Kaingaroa 0.23
0.2 Rangitaiki } Pokai ? Matahina 0.28
0.32
Whakamaru Chimp ?
Te Kopia 0.34
0.4
Waiotapu 0.71
0.6
Marshall ig 0.91 ± 0.02
0.8 Kaahu ig 0.92 ± 0.07
Rocky Hill ig 0.97 ± 0.02
1.0
Aharoa ig 1.19 ± 0.03
1.2
Ongatiti ig 1.23 ± 0.02
1.4
Ignimbrite B 1.51 ± 0.02
1.6 Ngaroma ig 1.60 ± 0.03
aData from Briggs et al., 1993; Houghton et aI., 1995.
Houghton, 1989). Andesites have also been intersected by drilling at Rotokawa, Nga-
tamariki, Kawerau, and Waiotapu, and relationships to dated ignimbrite units suggest they
may be up to 0.7 Ma.
8. CENTRAL VOLCANIC REGION STRUCTURE
Drilling for exploration and exploitation of geothermal power within the TVZ over the
past 20 years has provided a great deal of stratigraphic data. About 10 holes have been
drilled to depths of more than 2 km, and these show predominantly rhyolitic material in the
form of ignimbrite sheets within the upper 1500 m or so, and then an increasing amount of
dense volcanic rocks, particularly andesite, below 1500 m. A west-east cross section of
drill-hole stratigraphy across the TVZ is given by Stern (1987). Mesozoic graywacke has
been encountered at the base of some holes that are within 6 km of the eastern boundary of
the TVZ, but no sedimentary basement rocks have been encountered in drill holes else-
where within the TVZ. The deepest hole to date (at Ngatamariki) terminated in a quartz
diorite at a depth of about 2.8 km (Wood, 1986), while other deep holes on the eastern side
of the TVZ have terminated in andesite (e.g., Browne, 1971; Browne et al., 1992).
9. GEOPHYSICAL DATA BEARING ON CRUSTAL AND UPPER MANTLE

STRUCTURE, HEAT FLUX, AND KINEMATICS
Geophysical data from central North Island have provided a basis from which the
deeper crust and upper mantle conditions can be interpreted. Together, these data indicate
8 J. W. COLE et al.
It .. ;
~:.~:: ------'"
Heat flow
Imex I
1:800 1
mW~
'-----------
o A A'
50 -------.
----- ", Bouguer gravity
[
o "...
...... _- ...',......., ...... "
-50 '\~
'... "" ...
-100 mgels 9 1~O km .. , ...... _ ....' A'
A 3.0 ~. Crustal structure
o
6.2-6.4 kml.
r.. : 1 ¥k:.1
, - 7TTTTTT - ... ,
20 ?... ...
--- TTrrrTrT"" 7.4-7.5kml.?,
7.6 -1.9 kmla \ ...
... ""TTT'TTr - - - -
40km 8.1 kmlo
o
B B'
100 Seismicity distribution
o, ,
l00km FIGURE 1.6. A cartoon cross-section of
central North Island showing crustal and up-
200km
LOCALlTY~
DIAGRAM per mantle structure and the inferred position
of the subducted Pacific Plate (after Smith et
...
A al.. 1989). Seismic velocities from refraction
• surveys of Stem and Davey (1987) and Stem
et al.• (1987).
an area of grossly anomalous crustal and upper mantle conditions. A cross section of central
North Island, which includes the CVR, is shown in Fig 1.6. Information on this cross section
will be discussed in the following sections.
9.1. Earthquake Seismology

Seismologists have known for some time that earthquake waves that traverse the upper
mantle beneath central North Island are subjected to severe attenuation (Eiby, 1958;
Mooney, 1970) and travel at anomalously slow velocities (Haines, 1979). For example,
Haines's partition ofthe upper mantle beneath North Island shows an area beneath the CVR
to have upper mantle P-wave velocities (Pn velocity) of 7.4 ± 0.1 km/s, while more normal
Pn velocities of 8.1-8.5 km/s are found to the east.
Seismic intensity data have been inverted to determine the attenuation parameter Q (a
measure of the amount of seismic energy lost as a wave passes through a medium; Fowler,
1990) in the mantle beneath North Island (Satake and Hashida, 1989). Figure 1.7 shows the
results for Q in two depth ranges, 0-30 km and 30-80 km. Dominating both maps is a
low-Q region that occupies the CVR and spreads out into the Coromandel volcanic zone.
LA YER 1 ( 0-30km) LAYER 2 (30-80km)

J."
.
-0.. 0.1 •• 1
-0.', ••. ~ • ~J 0.'

-•. ~ ','1\... •...' -
~ •. o.• ~ft. o
-D.'
\
~
(':0'"
'\
'-'
••• 1.0 !O.I -0.5 N~';!~ \;.~ -•.•
'-'-.-O},/
. . . . . r -•.•'.,
\
.•.• -•.• ',\... ,.O'L"".)' -0.' • I I
-1.lr, ,5.1
L (S• •I .\ ~
I
-•.• _LO..1.
.•.• i~~~) •. ~/ /.~.(~ I .•
'~O.'
I
•.•
/,:,>-st'O..
rI.I~~\~. . -=~/.,0"1.1
,.o -I..
-1.2
-0.' -0 •• .1 0.1 ;0'" -0.$ -D.' 1.1 ..... -0 •• -· ..
/
-t'.--".....I. . . ,.
\.~.,
.-'-'
o'
,./ / "
0 ......... • -D.. -1.1 -I. 1.5
-... -...
-D.'
--::... (0 .• 7 '!:'L
-0.'./ ( ' I
~ ".1 - ' -'.1 H -•. ,~
-0.' -•.• H -0. s -•.• ,,)0.. -0.' -3.0 -S.I -S.O -1.1 -2.&
-D.S -0.1 .. 0.' }O.I -0.1 --:0.• ".1 -'.1 -'.1 }-•. , ... 1
d.• ~I -0.' -D.' -0.'-. -0.' ~I -'.1 -'.1 . ,.•"
-0.' ~ .•-1":;
) -,
•.•d_. -•.;/'. S ( 1.,-' j'" ~•. ,~•.• t .•.• /
-0.' -•. S / •. ,~ fl .•~.,- •.•.) I.' -I.J -2.S -0.1 ~
r--' -loS ( . \ 1.7 '-,-:.: (Q)
42'S -0.1 D.~ ..... -0.' \ ~~.;.~.
(Q)
1711' 178'E
FIGURE 1.7. Attenuation structure (Q) determined by the inversion of felt intensities (Satake and Hashida,
1989). Q at two different depth ranges are shown. The numbers are deviations from an initial value of attenuation
coefficient and are in units of 10- 2 so,; negative values are weak attenuation (high Q) and positive values are
strong attenuation (low Q).
This probably represents areas of pervasive partial melt in the lower crust and upper mantle
(Haines, 1979).
9.2. Explosion Seismology

Seismic refraction, wide-angle reflection, and near-vertical reflection surveys con-
ducted during the past 10 years have provided information on velocities and structures
within the crust and depth to the upper mantle. Using shot points in the Bay of Plenty and
Lake Taupo, Stem and Davey (1987) conducted a first-order crustal refraction survey, using
analog microearthquake seismographs. They reported a crustal velocity model for the CVR
that features a 3.0-6.1 km/s crust overlying a layer of velocity 7.4-7.5 km/s at a depth of
15±2 Ian (Fig 1.6). A recent offshore seismic reflection profile reported by Davey (1993)
found a band of subhorizontal reflectors at 5-6 s of two-way travel time. For an average
crustal velocity of, say, 5.9 km/s, these reflectors correspond to a depth range of 14-18 Ian
and are thus broadly consistent with the top of the 7.4-7.5 km/s layer found in the refraction
experiment of Stem and Davey (1987) described above.
9.3. Interpretation of the 7.4-7.5 kmls Layer

Seismic velocities of 7.4-7.5 km/s are ambiguous in that they may be interpreted as
either lower crustal or upper mantle in origin. For example, velocities of 7.3-7.4 km/s have
been detected at a depth of only 9 Ian beneath Fiordland, New Zealand (Davey and
Broadbent, 1980), where they have been interpreted as uplifted lower crust. However, low-
Pn velocities of 7.4-7.9 km/s are almost invariably associated with tectonically active areas
of high heat flow (e.g., Eaton, 1984; KRISP working party, 1991). Therefore the 7.4-7.5
10 J. W. COLE at al.
Wellington Northland Central Volcanic Region
" I •I' I '

'"
o
o
tectonic activity - high heat llow

-------~
III
Continental
-~'=;';;"-'='------ --- -- --- ---------------
••Iamtc 'Ielocl•• In k...,.
FIGURE 1.8. Crustal structures from different locations around the world illustrating a transition from
continental-type crustal structure through tectonically active types to oceanic-type crust (after Garrick, 1968;
Berckhemer et al., 1975; KRISP working party, 1991; Stem et al., 1987; Stem and Davey, 1987). Dark gray pattern
represents lower crustal velocities of 6.4-7.0 km/s, and it is notable that this layer is not present in the CVR.
km/s layer beneath the CVR is likely to be the upper mantle, albeit in a partially molten
state (see Black and Braille, 1982, for discussion on relationship between Po velocities and
heat flow). Figure 1.8 summarizes crustal structures from around the world and puts the
proposed crustal structure model for the CVR in a global context.
9.4. Gravity Data

About 8000 gravity observations have been made within and adjacent to the CVR and
these data are displayed on 1:250,000 maps of the Bouguer and isostatic gravity anomaly
field (Reilly, 1972). An attempt to separate numerically the regional and residual compo-
nents of the gravity field of the CVR was made by Stern (1979). Figure 1.9 shows the
resulting residual anomalies. The principal feature of this map is that almost the whole
CVR is dominated by a background residual low of - 300 I.lNlkg, with some areas as low as
-600 f.LNlkg, and other areas where the residual anomalies rise to zero (on the Bay of
Plenty coast where a graywacke outcrop exists). Rogan (1982) has made a three-
dimensional interpretation of the residual anomalies with a simple constant-density con-
trast between the volcanic rocks and basement. Interpretations by Stern (1986) have been
made along three lines (location shown on Fig 1.9) that are controlled by seismic refraction
data: one west-east line through the large gravity low at Mangakino (40 km northwest of
Lake Taupo); a north-south line from the graywacke outcrop near the Bay of Plenty coast;
and a west-east line through the gravity low just to the north of Lake Taupo. For the gravity
lows near Lake Taupo and Mangakino, negative-mass anomalies are inferred deeper than
the 4.9-5.5 km/s seismic basement. These anomalies may represent partially cooled rhyo-
litic intrusions or low-density pyroclastic deposits lying beneath higher-density ignimbrite
units where the ignimbrite effectively constitutes basement (Stern, 1986).
&4Y OF PUNTY
eonlou' ~... so &IN/kg
[EJ < -100 Al'Ukg
o, 20km
I
FIGURE 1.9. Residual gravity anomalies for the CVR (after Stern, 1979). The three profiles (A-A', B-B', and
C-C') are those interpreted by Stern (1986). Contour interval = 50 N/kg (10 N/kg = 1 mgal).
9.5. Heat Flux

Heat output from the CVR is almost totally expressed in the discharge of hot water and
steam from hydrothermal systems within the TVZ, as shown in Fig 1.10 (Studt and
Thompson, 1969). Natural heat output from the individual geothermal systems has been
measured by a variety of methods (e.g., Dawson and Dickinson, 1970; Bibby et al., 1984;
Calhaem, 1973). Natural heat output data shown in Fig 1.10 are from Calhaem (1973) and
Bibby et al. (1984) and sum to about 4 x 109 W. This figure is a minimum, as other "hidden"
geothermal fields similar to Mokai (Bibby et aI., 1984) may yet be discovered.
Taking the recharge zone for these geothermal areas to be about 5000 km2 (Stem,
1987), the equivalent heat flow is then about 800 mW/m2, a value about 13 times greater
than the continental norm and, as far as we are aware, one of the highest reported for any
backarc basin. This may be because most backarc basins are oceanic and therefore covered
by a shallow ocean that effectively masks the true heat output, thus resulting in an
12 J. W. COLE et al.
CONVECTIVE HEAT OUTPUT (MW)
While I. . .
0,
• Recently active andesite/dacite volcanoes
• Geothermal fieki and heat output (MW)
o low-temperature and/or old

geothermal field
FIGURE 1.10. Distribution and heat output

(in MW) of geothermal fields within the
CVR (after Stem, 1987). Total natural heat
output in shaded area from south end of
Lake Taupo to the Bay of Plenty coast is
010203040 about 4 x 109 W. Shaded region represents
I , ! I ,
our best estimate of the region where con-

176"E
vective heat transfer is occurring at present.
underestimate of the heat flow (Sclater et aI., 1980). Nevertheless, an average heat flow of
800 mW/m2 is ofthe same magnitude as areas like Yellowstone caldera in the United States
(Morgan et aI., 1977) and spreading centers like Iceland (Palmason and Saemundsson,
1974) and, in general, is indicative of a tectonic regime where a large-scale convective
transfer of heat is taking place.
10. GEODETIC DATA
10.1. Large-Scale Horizontal Deformation

Five analyses of first-order triangulation have been undertaken to estimate large-scale
deformation across North Island, particularly across the CVR and the TVZ (Sissons, 1979;
Adams, 1984; Walcott, 1984, 1987; Crook and Hannah, 1989; Reilly, 1990), using the
method of Bibby (1973, 1981, 1982).
A first-order North Island triangulation across the TVZ was established between 1923
and 1936 by the predecessor of the New Zealand Department of Survey and Land Informa-
tion (DOSLI; Lee, 1978). A narrow belt of this first-order network was resurveyed and
densified, with triangulation and trilateration completed by DOSLI in 1976 (Bevin et al.,
1984). Because of problems with scale calibration of the electronic distance meter (EDM)
instruments used for the trilateration, Reilly (1990) used these data only as line ratios.
Sissons (1979) estimated 7 mm/yr secular extension across the TVZ from a compari-
son of these triangulation observations, partitioning the set of reobserved stations according
to the boundaries of the TVZ (Fig 1.11). Shear rates were negligible for his networks BPI
and BP3 to the west and east of the TVZ. In contrast, the shear rate is highly significant for
network BP2, which spans the TVZ. Apart from one station to the southwest, this network
lies between the Rotorua-Tarawera region and the Bay of Plenty coast. The shear-rate
component, relatively extensional at azimuth 1300 and perpendicular to the dominant 0400
strike of faults in the zone, is O.l8±O.06 (1 s.e.) x 1O-6/yr. Under the assumption that there is
no length change parallel to the faults, this component would represent actual extension
across the 40-km-wide zone of 7.2±2.4 (1 s.e.) mm/yr.
Adams (1984) compared the 1920s survey with both the trilateration and triangulation
of the 1976 survey. His results show a uniform strain rate across most of North Island, with a
marked increase in rate, but little change in orientation, near the coast at East Cape. This
study indicated a shear strain component corresponding to an extension perpendicular to
the faulting trend, assuming there is no parallel length change, of typically 0.09±O.03 (1
s.e.) x 1O-6/yr. Careful reading of Sissons (1979) and Adams (1984) appears to preclude the
possibility that the factor-of-2 difference between the two results is due to confusion
between tensor and engineering definitions of shear magnitude, but this nevertheless still
seems the most likely explanation.
Only triangulation measurements appear to have been used by Walcott (1984), who
reported shear results of O.l5±O.03 (1 s.e.) x 1O-6/yr across the entire CVR, zero in the
eastern Bay of Plenty, and a much larger value near East Cape. Reilly (1990) used both
direction and distance measurements throughout New Zealand and determined horizontal
Vl
o
&,
BPI
t
'"
FIGURE 1.11. Shear strain results from

Sissons (1979) (BPI, BP2, and BP3) and
from Darby and Williams (1991) (NTP).
The shear strain is plotted as a bar whose
half-length is proportional to the maximum
engineering shear and with the orientation
of maximum relative extension. Confi- Vl
dence regions of one standard error for o
these values are shown. Lettered stars show 0'4 0'2 0'0 0'2 0'4
locations of earthquakes referred to in the S'tra;n' rat~ (pp'm/Y;)
text: Wi, Wairakei; Wu, Waiotapu; M, o ~
Matata; E, Edgecumbe; T, Te Aroha; P, , 'Ge'agr~ph:c s~al~ (k;") , ,
Paeroa. 176' 0 E 177' 0 E

0.10 m.crorad.an/year
o FIGURE 1.12. Shear strain results, redrafted
- ---------
from Reilly (1990). The bars represent the axis of
..... .....----. maximum relative contraction, the length being
o
proportional to the tensor shear (half the engi-
o /
------
neering shear). The sector symbols give the limits
\ of two standard deviations in both magnitude and
100
direction; circles denote results less than two
kilometres
standard deviations.
derivatives, up to the third order, of station horizontal velocities to describe the variation of
strain (Fig 1.12). In his Raglan to East Cape transect, he found the rate of shear steadily
increasing from zero in the west, to a maximum of 0.30±0.08 (1 s.e.) x 1O- 6/yr in the east
with a rate in CVR of about 0.1 x 1O- 6/yr, consistent with Walcott's estimate. Walcott (1987)
considered this to be a typical smoothed value for the Bay of Plenty region; he multiplied
it by the 120-km maximum width of the CVR over which it was estimated to obtain an
extension rate of 12 mm/yr, but this value obviously depends strongly on the assigned width
of CVR. Finally, Crook and Hannah (1988, 1989) estimated shear of 0.18±O.11 x 1O- 6/yr
(1 s.e.), relatively extensional at azimuth 150°, for the period 1951-1977 by comparing one
quadrilateral of the first-order network spanning the TVZ at the Bay of Plenty coast with
lower-order triangulation. This result is very similar to that of Sissons (1979).
All these results are for networks across the northern part of the CVR/TVZ, within 40
to 70 km of the coast. There are no large-scale results farther south, but global positioning
system (GPS) surveys of the TVZ were completed in 1990 and 1991 with a view to their
comparison with triangulations of the 1920s, 1950s, and 1970s (Blick et ai., 1992) and
analysis of these data is still in progress.
10.2. Intermediate-Scale Horizontal Deformation

Comprehensive second- and third-order triangulation was established by DOSLI
between 1949 and 1955 in the Bay of Plenty, Rotorua, Taupo, and Raglan districts (Wil-
liams, 1989). A survey across the Taupo fault belt immediately north of Lake Taupo was
completed in 1986 comprising triangulation, trilateration, and vertical angle measurements
(Williams 1987). From these surveys, Darby and Williams (1991) estimated uniform shear
ofO.5±O.1 x 1O- 6/yr (1 s.e.), relatively extensional at azimuth 14±29° (1 s.e.) immediately to
the north of Lake Taupo for the period 1950-1986 (Fig 1.11). Investigation of variable strain
across the network showed a west-to-east difference in magnitude from 0.4 x 1O- 6/yr to
1.0 x 1O- 6/yr and in relatively extensional orientation from 102° to 170°. Integration of the
uniform shear result across the 40-km width of the TVZ yields 18±5 mm/yr extension
perpendicular to the faults.
10.3. Vertical Deformation

Lake-leveling observations on Lake Taupo have recorded vertical deformation from
1985 to the present (Otway and Sherburn, 1994) (Fig 1.13). The main results are a long-term
subsidence centered on the Taupo fault belt and relatively-short-term variations most
noticeable outside the fault belt. While the subsidence is undoubtedly related to the
intermediate-scale extension discussed, Otway and Sherburn (1994) show there is no
obvious relationship of the short-term variations with fluctuations in the low-level seis-
micity in the region. This is in contrast to the uplift preceding the Taupo earthquake swarms
in 1983 and the subsidence during them (Otway, 1986).
1004. Active Surface Faulting

In the late Pleistocene normal faulting was dominant and formed the Taupo fault belt
from Ruapehu and Tongariro to Lake Taupo, the Paeroa Range, Tarawera, and Whakatane
(Fig 1.14). North of Lake Taupo, this belt cuts across the trend of the older Hauraki graben
that contains the mid-rift Kerepehi Fault (Fig 1.14). Faulting is dense southwest of the
Okataina volcanic center but less so to the northeast. The belt is about 20 km wide with an
overall strike of 040°, swinging to 055° in the northeast (Healy et aI., 1964).
The Taupo fault belt has been active in the late Quaternary with movements dated
from pyroclastic deposits of known age (Nairn, 1976). Southwest of Okataina volcanic
center, the faults have moved repeatedly during the last 50 k.y. with many having moved
between 13 and 11 ka. Some displacements postdate the 1.85-ka Taupo eruption (Nairn and
Hull, 1985). There were also historical surface ruptures of up to 3 m in the 1922 Taupo
earthquake sequence (Grange, 1932; Sissons, 1979; Grindley and Hull, 1986) and of 50 mm
in the 1983 Taupo earthquake swarms (Otway, 1986). Northeast of Okataina volcanic center
the faults displace the eroded surface of Rotoiti breccia (50 ka), whereas others displace the
9-ka Rotoma ash (Nairn, 1981). Ruptures of up to 2.7 m occurred with the 1987 Edgecumbe
earthquake (Beanland et aI., 1990).
Rates of faulting are best known within the Whakatane graben, which is the expression
of the TVZ at the coast between Matata and Whakatane. Nairn and Beanland (1989) infer
vertical faulting rates exceed 1.9 mm/yr from the offset of 0.28 Ma Matahina ignimbrite and
point out that this is consistent with their inference of 1-2 mm/yr subsidence within TVZ
and 1 mm/yr uplift rates outside. Horizontal extension rates are poorly known. Beanland
-10
S -20
El
-30
..c: -40
<l
-50
0 5 10 15
kilometers
FIGURE 1.13. Profiles of apparent height change of Taupo fault belt stations relative to an origin station from
1986 to 1991 (after Otway and Sherburn, 1994); KA, Kawakawa; 00, Omoho; TK, Te Kauwae; KH, Kinloch; TA,
Te Itarata; WO, Whakaipo; WI, Waikarariki; KO, Kaiapo; MP, Mine Point; RA, Rangatira.
Q
'--_----"50 km
White
Island
FIGURE 1.14. Active faulting in the central

North Island. (Source: Institute of Geological and
Nuclear Sciences 1:250,000 digital compilation).
et al. (1990) argue that if 2-3 mm/yr vertical displacement is assumed to occur at both the
eastern and western margins on 45° dipping faults, then the horizontal extension would be
4-6 mm/yr. This rate would be a minimum because it does not include extension on
smaller-scale structures within the graben.
11. FOCAL MECHANISMS
Seismological studies in the CVR and TVZ have not lessened the difficulty of
interpreting focal mechanisms. Despite the preponderance of normal fault traces, for
smaller-magnitude events normal faulting focal mechanisms have generally been absent,
both for the less reliable composite kind associated with swarms and microearthquakes and
the more reliable single events.
Transcurrent mechanisms have been more commonly observed: for example, NE
dextral fault mechanisms in a microseismic survey of the Wairakei geothermal field (Hunt
and Latter, 1982); possibly NE sinistral for the Waiotapu M=5.l earthquake of 1983 (Smith
et aI., 1984); both NE dextral and NE sinistral for the Taupo swarms, reaching M=3.7 and
M=4.3, in 1983 (Webb et aI., 1986); and NE dextral with a normal component for the Matata
M=5.4 earthquake of 1977 (Richardson, 1989).
The clearest, predominantly normal focal mechanism was for the 1987 Edgecumbe
M=6.3 earthquake (Anderson and Webb, 1989). Other evidence is less compelling. The
1972 Te Aroha M=5.3 earthquake may have had a normal focal mechanism but with
different mechanisms for the aftershocks (Adams et al., 1972). The fault strike was possibly
north but is very poorly constrained (T. Webb, personal communication, 1990). Also, small
events following the 1982 June M=4.2 Paeroa earthquake had first motions consistent with a
normal mechanism (Smith et aI., 1984). It is unfortunate that the only other available
seismic studies, of the 1964-65 Taupo swarms that reached M=4.6 (Gibowicz, 1973) and
of a microseismic survey from the Taupo fault belt to the Kaingaroa plateau (Evison et aI.,
1976), were not able to provide focal mechanism solutions.
12. KINEMATICS
An unusually diverse collection of independent data bears upon crustal kinematics of

the CVR. These data include the apparent migration of low-potash andesites across the
CVR (Fig. 1.15a,b), geodetic ally determined strain rates (Sissons, 1979; Walcott, 1984;
Darby and Williams, 1991), and paleomagnetic data from Tertiary sediments to the east of
the CVR (Walcott, 1984; Wright and Walcott, 1986). These data all point to a surprisingly
consistent, late Tertiary, kinematic evolution of central North Island (summarized in Stem,
1987). Included in this evolution are rotations of about 6°/m.y. over the past 4 m.y. and
translations, or spreading, of 7-18 mmlyr over the same period. Both the rotation and
translation contribute to the development of a wedge-shaped CVR. What makes this
evolutionary solution particularly compelling is that the data sets previously described are
totally independent and pertain to time scales from 20 m.y. to 30 yr. Similar evolutionary
schemes for backarc extension, which also involve wedge-shaped rotations and transla-
tions, have been proposed elsewhere (Otofuji et ai., 1985).
Darby and Williams (1991) point out that while all the geodetic results are broadly
consistent with each other, Sissons's, Adams's, Walcott's, and Reilly's results provide no
distinction in rate of deformation between the narrow TVZ and the broader CVR. They
generally found little change in orientation of the strain field across the CVR. Furthermore,
they show no marked western boundary to the CVR, which may have been expected from
Adams's and Walcott's network subdivisions if not from Reilly's polynomial smoothing.
This may be due to the inability of the typically 40-km station spacing to resolve variations
in strain at smaller distance scales. Darby and Williams's result of a significantly higher
strain may be due to a local maximum of a regionally varying strain, or it may be influenced
by systematic effects arising from the poor geometry of the network and the relatively large
number of eccentric stations used in the later survey compared with the earlier survey. In
general, it is unfortunate that strain rates are integrated by some investigators over different
distances to derive widening rates and then become widely quoted without the associated
distances.
13. MAGMA TYPES AND GENESIS
The TVZ is characterized by three distinct groups of eruptives: high-AI basalt, basaltic
andesite-andesite-dacite, and rhyolite, in order of increasing volume. The high-AI basalt
and rhyolite occur in a bimodal assemblage associated with caldera structures (see Fig. 1.3)
in the backarc part of the TVZ, while the basaltic andesite-andesite-dacite assemblage
occurs at the surface, particularly at the northern (White Island) and southern (Tongariro)
ends of the TVZ and in a narrow frontal arc along the eastern margin (see Fig. 1.3).
13.1. High-AI Basalts

High-AI basalts have erupted sporadically at 15 locations in the TVZ (Houghton et ai.,
1987) (see Fig. 1.3), ranging from single monogenetic cones (e.g., 10hnsons Road) to fissure
eruptions (e.g., Tarawera, 1886) and phreatomagmatic tuff rings (e.g., Kaiapo and Acacia
Bay). The basalts range from aphyric to coarsely porphyritic (Cole, 1973) with phenocrysts
of olivine + plagioclase ± clinopyroxene. A summary of mineralogy is given in Gamble
et ai. (1990). Chemically the basalts are high-alumina (HAB) as defined by Crawford et ai.
o, ,5qkm
WHITE ISLAND
AO'''''
II Greywacke
1m] Geothermal region
• Zero geothermal gradient eite

A··glow-potash ande.lte & age In My
a ,~ Downflow area
e
i 5
t 3 mm/y
••
J
'a
4
C
'0 3
•
•
III
.c 2
:.::
o+---~----,---~--~~----~--~~~~
140 120 100 eo eo 40 20 0
b Perpendlculer dle.enc. from 0 My boundery (km)
FIGURE 1.15. (a) Distribution of geothermal phenomena and low-potash andesites within and adjacent to the
CVR. Solid lines indicate possible positions of the active volcanic front at times indicated in My. (b) Plot of KJ AI
ages versus perpendicular distance from volcanic front (0 Ma boundary) for low-potash andesites (modified from
Stem, 1987). The rate of 21 ± 3, mmlyr is the apparent average rate of migration in a southeasterly direction.
(1987) with 100 Mg/(Mg + Fe) values between 56 and 72, and CaO/Alp3 ratios between
0.59 and 0.70. They have a low abundance of K20 (Fig. 1.16) and high-field-strength
elements, but show lower Ti/Zr, higher TiIV and Ti/Sc ratios, and generally higher Zr
abundance than MORB (Gamble et ai., 1993). They are light rare-earth enriched (CelYb =
2-3) with flat heavy rare-earth patterns. Isotopically they define an almost linear array of
decreasing Nd- and increasing Sr-isotopic ratios with Kakuki basalt (87Srf8 6 Sr = 0.70388
BASALTIC
BASALT ANDESITE DACITE RHYOLITE
ANDESITE
4
Wt
HiOh-AI
basalt>: - - Low-K
48 52 56 60 64 68 72 76
Wt % Si02
FIGURE 1.16. Weight percent K20 versus weight percent Si02 for volcanic rocks of TVZ. Classification is from
Le Maitre (1989), fields from Graham et al. (1992).
and ENd = +5.1) isotopically the most primitive (Fig. 1.17). Kakuki also has the least
radiogenic Pb-isotope signature (Graham et aI., 1992), while Tarawera basalt has the least
radiogenic signature (Fig. 1.18). The latter is, however, riddled with small rhyolitic inclu-
sions, and despite best efforts to avoid these in analysis a crustal component is likely to be
reflected in isotope values.
Gamble et al. (1993) conclude that the TVZ basalts contain both crustal and slab-
derived subduction signatures. They consider that the slab component comes from a
relatively fertile hot mantle and may well reflect the youth of the TVZ magma system.
Ascent of the magmas was hindered by the low-density continental crust, and variation is
caused largely by fractionation of olivine ± plagioclase, followed by clinopyroxene and
Fe-Ti oxides in the more evolved lavas.
+8
+7
+6
RB -2
0
-.
+5 TORLESSE
KB
TERRANE
ENd
+4 70400 roeoo "71200 71600
87 5r/ 86 5r
0
+3
+2 TVZ
Type 6
ANDESITE
+1
TVZ
RHYOLITE -
IGNIMBRITE
0
70350 70400 70450 70500 70550 ·70600
e7Sr / 86 Sr
FIGURE 1.17. 87Srf8 6 Sr versus eNd for TVZ rocks (after Graham and Cole, 1991). "Mantle Array" from Nohda
(1984). KB, Kakuki basalt; RB, Ruapehu basalt, RCB, Red Crater basalt; WB, Waimarino basalt.
--- ------
(
....--
t/
IS-60
HiOh-AI
baSOI: eT
RP e
ffiK
18·~ 18-8S
FIGURE 1.18. 207PbJ204Pb versus 206PbJ204Pb for TVZ rocks. Field for each volcanic rock type and Waipara/
Torlesse metasediments from Graham et al.(l992). KB, Kakuki basalt; RB, Ruapehu basalt; T, Tarawera basalt;
RCB, Red Crater basalt; RP RoUes Peak basaltic andesite; WI, White Island basaltic andesite.
13.2. Basaltic Andesite, Andesite, and Dacite

The most mafic members of this association occur on Ruapehu, Red Crater (Tongariro
volcanic center), and Waimarino (SW of Lake Taupo). The latter is a magnesian quartz
tholeiite which is porphyritic with 16% olivine (Fo 90) and 7% clinopyroxene and has high
MgO, Cr, and Ni (Graham and Hackett, 1987). However, it has low eNd, high 87Srf8 6Sr (Fig.
1.17), and moderately radiogenic 206PbP04Pb (Fig. 1.18), reflecting crustal contamination.
The two basaltic andesites from the Tongariro volcanic center have lower Al 20 3 contents
than HAB and contain plagioclase, orthopyroxene, and rare crustal xenoliths. The
Tongariro basaltic andesite has comparable 87Srf8 6Sr ratios to the HAB (Graham et at.,
1992), but Ruapehu basaltic andesite has higher eNd (Fig. 1.17) and Red Crater basaltic
andesite higher 207PbP04Pb (Fig. 1.18).
Most of the volume of the Tongariro volcanic center is composed of medium-K
andesite (Fig. 1.16). Of the two massifs in the center, Ruapehu has been most extensively
studied geochemically. Graham and Hackett (1987) identified six main types; a seventh
type was added by Patterson and Graham (1988). These are characterized by differing
phenocryst assemblages, major and trace element chemistries (Fig. 1.19) and 87Srf8 6Sr
ratios (Fig. 1.17). Andesites also occured in the 1977 eruption of White Island (Cole and
Graham, 1989; Graham and Cole, 1991), where Mg-rich bombs and blocks were erupted
(Fig. 1.19), at Kawerau (Browne, 1978), Waiotapu (Hedenquist, 1983), Broadlands (Browne,
1971; Wood, 1983), Rotokawa and Ngatamariki (Browne et al., 1992), Rolles Peak (Graham
and Worthington, 1988), and Wairakei (Grindley, 1965); see Fig. 1.3 for locations.
Dacites are volumetrically minor in the TVZ and are of two types (Reid and Cole,
1983). Only one type (Type A; Graham et at., 1992) appears directly related to the basaltic
2 2 - . - - - - - - - - - - . - - - - - -_ _--,-10
FeO T
20 ( - - - , Whor."
'- .x:
18 Ruapehu
'............... ",
, 6
~
WI%
1,3 -7
(§J J Wt %
FeO'T)
16
White ,/ /
•
14
,~/
10 MgO CoO 10
5
- ...... White Is
" -~
'-,-
'-
',
Wt% Ruapehu
" \
\ Wt %
MgO 1-4,6,7
CoO
FIGURE 1.19. Major element plots

for andesites of Tongariro volcanic
center and White Island. Fields from
O+-~,-,-,-~~-r~-.-_r~_,_.~_.--Lo
Graham and Cole (1991). Solid circles 50 54 58 62 50 54 58 62 66
are 1977 eruptives from White Island. Wt % 5,0 2
andesite-andesite-dacite association. Dacites of this type are medium K (Fig. 1.16) and
occur in the Tongariro volcanic center and White Island. They are characterized by slightly
higher 87Srj86Sr and ENd than those of Tongariro (Fig. 1.17) but comparable 207Pbf204 Pb
(Fig. 1.18). The second type (Type B; Graham et at., 1992) occurs at Tauhara, Broadlands,
Waiotapu, Edgecumbe, and Whale Island (Fig. 1.3). These are most likely to be products of
mixing between a RoUes Peak (type) andesite and rhyolite (Graham and Worthington,
1988).
The origin of the basaltic andesite-andesite-dacite (Type A) association has been
discussed by a number of authors (Cole 1978, 1990; Graham and Hackett, 1987; Graham
and Cole, 1991, Graham et aI., 1992; Browne et aI., 1992). Most agree that the assemblage
was formed by a multistage process involving (1) sub solidus slab dehydration fluxing the
overlying mantle wedge, (2) anatexis of the asthenosphere or subcontinental lithosphere
of the mantle wedge to produce low-AI basalt, and (3) fractionation of plagioc1ase-
orthopyroxene/olivine-augite-magnetite (POAM) mineral phases (15-55%) with minor
subarc crustal assimilation (1-30%, particularly by Torlesse metasediment; Graham et at.,
1992) (Fig. 1.18) or crustal accumulation. Both Type 5 in the Tongariro volcanic center and
the RoUes Peak andesite (with high Sr and lower 87S r /86S r ratios (Fig. 1.17) must have come
from a different mantle source to the bulk of the lavas. RoUes Peak also has lower
207Pbf204 Pb (Fig. 1.18).
13.3. Rhyolites and Ignimbrites

The silicic volcanics are voluminous (12,000 km3 ; Cole, 1979), representing at least
80% of the total eruptives of the TVZ (Hochstein et aI., 1993). The rhyolites range from an
anhydrous assemblage of plagioclase, pyroxene, magnetite, and ilmenite to a strongly

porphyritic assemblage of quartz, plagioclase, calcic-hornblende, biotite, magnetite, and
ilmenite. Details are given in Ewart et al. (1976) and Cole (1979). Chemically they range
from 70% to 79% Si02 with 2-4% KzO (Fig. 1.16) and NazO > KzO in most rocks.
Isotopically there is a range of 87Sr/86Sr ratios from 0.70500 to 0.70650 (Fig. 1.17), whereas
206Pbf204 Pb ratios (Fig. 1.18) are remarkably uniform (Graham et aI., 1992). There is no
systematic difference between rhyolites and ignimbrites nor over time in the stratigraphic
sequence.
The origin of the silicic volcanics is much more controversial. Ewart and Stipp (1968),
Cole (1981), and Reid (1983) considered them a product of partial melting of subvolcanic
basement, particularly Waipapa Group metasediments. This origin was consistent with
general chemistry, 87Sr/86Sr ratios (Fig. 1.17), and 207Pbf204 Pb ratios (Fig. 1.18). However,
Blattner and Reid (1982) pointed out that such an origin is incompatible with low 8 180
values and Conrad et al. (1988) showed by experimental studies that the mildly per-
aluminous Waipapa metasediments were unlikely to produce the predominantly diopside-
normative rhyolites. Graham et al. (1992) showed that the Pb-isotope data were consistent
with a derivation from primary mantle-derived basaltic liquids by assimilation-fractional
crystallization (AFC); see Fig. 1.18), but he had serious reservations because of bimodality
and relative magma volumes. Most recently Briggs et al. (1993), studying the Mangakino
center, have suggested a multiple origin including partial melting of a crustal source similar
to western (Waipapa) graywacke or its metamorphic equivalent, melting of a plagioclase-
rich plutonic or metamorphic source similar to anorthosite or trondjemite, and a minor
upper mantle source component.
14. DISCUSSION
A question that has puzzled geophysicists for a long time is, what is the heat source
for the TVZ (e.g., Evison et al., 1976; Studt and Thompson, 1969)? If we make the
reasonable assumption that the present-day natural heat output is representative, and not a
local perturbation in time, and we take note of geological evidence (Grindley, 1965) that the
present-day geothermal fields may have lifetimes of, say, 1 m.y., then a major source of heat
is required.
Stem (1987) proposed one possible model whereby all the heat is provided by cooling
igneous rocks within the TVZ. He found that for a spreading rate of at least 16 mrnlyr, and
with spreading being accompanied by the intrusion of new igneous rock, then a steady-state
4 x 109 W heat output can be maintained if the full 13 km (15 km less the top 2 km of
volcanic ash, etc.) of crust is made of igneous rocks.
Other solutions are possible. For example, Hochstein et al. (1993) argue that about half
the heat output may be supplied by heat released during plastic deformation. Another
alternative may be heat from the mantle being transferred by meteoric water that con-
vects to depths of 13 km or so and then efficiently transfers heat from the upper mantle to the
surface.
In order to account for the preponderance of normal fault traces in the near absence of
normal focal mechanisms, Smith and Webb (1986) proposed a crustal model consisting of a
complex system of conjugate transcurrent faults joined by normal faults. The absence of
surface transcurrent faults was attributed to the inability of the weak 2-km-thick pyroclastic
surface layer to transmit transcurrent movement. Webb et al. (1986) added the provision
that the normal faulting is aseismic, perhaps because of high fluid pressure reducing the
normal stress, and noted that neither temporal changes in the stresses nor a chaotic stress
distribution near the apex of the region of active widening could be ignored.
Darby and Williams (1991) countered these explanations by arguing that a lowered
normal stress component over a region would affect normal and transcurrent faulting
equally, that a peculiar strength anisotropy of the weak pyroclastic layer would be required
to favor dip-parallel over strike-parallel shear failure, and that normal faulting did occur
during the 1987 Edgecumbe (M=6.3) earthquake, albeit after the papers of Smith and Webb
(1986) and Webb et al. (1986) had appeared.
In consideration of smaller earthquakes down to M=5, there is general agreement with
the findings of Smith and Webb (1986) and Otway and Sherburn (1994) that there does not
appear to be a simple relationship between the lower-magnitude seismicity data and surface
deformation as manifested by faulting or by geodetic measurement. However, the largest
historical earthquakes in the TVZ (viz. 1922 Taupo and 1987 Edgecumbe) had major
components of normal faulting consistent with geological expectations and the geodetic
results reviewed here.
Kinematic data suggest that the CVR has been spreading for at least 4 m.y., with most
of the recent evolution in the TVZ. Shear rates are negligible in the older part of the CVR
(BPI in Fig. 1.11) but highly significant in TVZ, leading Cole (1990) to suggest that the TVZ
may be transtensional with current orientations of the fault belts controlled by a combina-
tion of dextral shear in the basement-northeastern continuation of the North Island shear
belt beneath the TVZ and back-arc extension. This possibility is, however, only likely to be
resolved with greater GPS station density and information on both dilational and rotational
components of strain tensor.
Origin of the voluminous ignimbrites is a major problem in TVZ. It seems likely that
most may be a result of remelting earlier formed igneous rocks (e.g., Graham et al., 1992) or
sedimentary/metamorphic rocks of a restricted composition together with a small mantle
component (e.g., Briggs et aI., 1993). Some may be a result offractional crystallization of a
mantle source with significant crustal contamination (e.g., McCulloch et al., 1994). Cole
(1990) noted that rhyolitic centers only occur within a restricted area (between 38°
and 39°S) in the TVZ (see Fig. 1.3), while back-arc extension is clearly taking place
throughout the zone. It may thus be that it is only in this area where suitable sources occur
that would partially melt to form rhyolitic magmas. Cole (1990) has suggested that if these
are igneous rocks they may be a SE extension of the Coromandel volcanic zone.
Some of these possibilities should be explored by a combined geophysical-geochemical
study of subvolcanic features in the TVZ. In particular, the advent of a new generation of
digital seismological equipment with enhanced dynamic range should provide us with a
higher resolution of the crust and upper mantle than has been possible in the past.
15. SUMMARY AND CONCLUSIONS
1. The CVR is the principal volcanic basin structure defined by seismic and gravity parameters
in central North Island, and the TVZ represents the eastern part of the CVR that is currently
volcanically and tectonically active.
2. Activity in the TVZ began with eruptions from the Mangakino caldera about 1.6 m.y. ago,
while activity in the main part of the TVZ has taken place within the past 0.7 m.y.
3. Earthquake seismology indicates that the mantle and lower crust beneath the CVR are highly
attenuative and is probably in a state of partial melt. A combination of earthquake and
explosion seismology data indicates a crustal thickness of IS±2 km for much of the CVR.
4. CVR is dominated by low gravity «300 I-"N/kg) with areas -600 I-"N/kg representing, in
part, cooled or partially cooled rhyolitic magma. It has an average heat flow of 800 mW/m2,
about 13 times greater than the continental norm and one of the highest on record for a
backarc basin. This may be due to high rates of extension.
S. Rates of extension in the currently active TVZ vary from about 7 mmlyr with a shear of
0.18±O.1l x 1O-6yr (1 s.e.) in the Bay of Plenty to about 18 mmlyr and a shear of O.5±O.l
(1 s.e.) in the Taupo fault belt. Vertical faulting rates indicate 1-2 mmlyr subsidence.
6. Focal mechanisms have been difficult to interpret. Most appear to have a large transcurrent
component, although the 1987 Edgecumbe earthquake was normal.
7. The TVZ volcanism is characterized by three distinct types of eruptives: high-AI basalt;
basaltic andesite, andesite, and dacite; and rhyolite and ignimbrites, in order of increasing
volume.
8. The high-AI basalts come from a relatively fertile hot mantle, with magmas rising into the
crust, fractionating, and assimilating small amounts of crust.
9. The basaltic andesite-andesite-dacite assemblage, of which most ofthe near-surface volume
is in the Tongariro volcanic center, was formed by a multistage process involving subsolidus
slab dehydration fluxing the overlying mantle wedge, anatexis of the mantle wedge, frac-
tionation, and minor crustal assimilation.
10. The rhyolites and ignimbrites form the greatest volume (-12,000 km3) of the TVZ volcanics.
Their origin is debatable, but they may have a multiple origin, including partial melting of
crustal rocks with a small mantle component or fractionation of a mantle source with crustal
contamination.
11. The simplest explanation of the extreme heat output is by the cooling of magma and the
transfer of heat to the surface geothermal fields by meteoric water. It is not clear, however, at
what depth this heat transfer takes place.
REFERENCES
Adams, D. A. 1984. Some methods of analysis of geodetic data and their applications to the measurement of
crustal deformation, M.Sc. thesis, Victoria University of Wellington.
Adams, R. D., Muir, M. G., and Kean, R. J. 1972. Te Aroha earthquake, 9 January 1972, Bull. N.Z. Nat. Soc.
Earthquake Eng. 5(2):54-58.
Anderson, H., and Webb, T. 1989. The rupture process of the 1987 Edgecumbe earthquake, New Zealand, N.z. J.
Geol. Geophys. 32(1):43-52.
Beanland, S., Blick, G. H., and Darby, D. J. 1990. Normal faulting in a back arc basin: geological and geodetic
characteristics of the 1987 Edgecumbe earthquake, New Zealand, J. Geophys. Res. 95(B4):4693-4707.
Berckhemer, H., Baier, B., Bartelsen, H., Behle, A., Burkhardt, A., Gebrande, H., Makris, J., Menzel, H., Miller,
H., and Vees, R. 1975. Deep seismic soundings in the Afar region and on the highland of Ethiopia, in Afar
Depression of Ethiopia, Vol. I (A. Pilger and A. Roster, eds.), pp. 89-107, E. Schweizer, Velagbuchhandl,
Stuttgart.
Bevin, A. J., Otway, P. M., and Wood, P. R. 1984. Geodetic monitoring of crustal deformation in New Zealand, in
An Introduction to Recent Crustal Movements of New Zealand (R. I. Walcott, ed.), pp. 13-60, Royal Society
of New Zealand Misc. Ser. 7.
Bibby, H. M. 1973. The reduction of geodetic survey data for the detection of earth deformation, DSIR Geophysics
Division Report 84.
Bibby, H. M. 1981. Geodetically detennined strain across the southern end of the Tonga-Kennadec-Hikurangi
subduction zone, Geophys. J. R. Astron. Soc. 66:513-533.
Bibby, H. M. 1982. Unbiased estimate of strain from triangulation data using the method of simultaneous
reduction, Tectonophysics 82(1-2):161-174.
Bibby, H. M., Dawson, G. B., Rayner, H. H., Stagpoole, V. M., and Graham, D. J. 1984. The structure of the Mokai
geothennal field based on geophysical observations, J. Volcanol. Geotherm. Res. 20:1-20.
Black, P. R., and Braille, L. W. 1982. Pn velocity and cooling of continental lithosphere, J. Geophys. Res.
87:10557-10563.
Blattner, P., and Reid, F. W. 1982. The origin of the lavas and ignimbrites of the Taupo volcanic zone, New Zealand
in the light of oxygen isotope data, Geochim. Cosmochim. Acta 46:1417-1429.
Blick, G. H., Darby, D. 1., Meertens, C. M., Otway, P. M., Perin, B., Rocken, c., and Scott, B. J. 1992. Crustal
defonnation surveys in the Taupo Volcanic Zone and central North Island using GPS measurements, N.Z
Surveyor 33(281):299-312.
Briggs, R M., Gifford, M. G., Moyle, A. R, Taylor, S. R, Nonnan, M. D., Houghton, B. F., and Wilson, C. J. N.
1993. Geochemical zoning and eruptive mixing in ignimbrites from Mangakino caldera, Taupo volcanic
zone, New Zealand, J. Volcanol. Geotherm. Res. 56:175-203.
Brothers, R N., and Delaloye, M. 1982. Obducted ophiolites of North Island, New Zealand: origin, age,
emplacement and tectonic implications for Tertiary and Quaternary volcanicity, N.Z J. Geol. Geophys.
25:257-274.
Browne, P. R L. 1971. Petrological logs of drillholes, Broadlands Geothennal Field, N.Z. Geol. Surv. Rep. 52.
Browne, P. R L. 1978. Petrological logs of drillholes, Broadlands Geothennal Field, N.Z Geol. Surv. Rep. 84.
Browne, P. R. L., Graham, I. 1., Parker, R J., and Wood, C. P. 1992. Subsurface andesite lavas and plutonic rocks in
the Rotokawa and Ngatamariki geothennal systems, Taupo volcanic zone, New Zealand, J. Volcanol.
Geotherm. Res. 51:199-215.
Calhaem, I. M. 1973. Heat flow measurements under some lakes in North Island, New Zealand, unpublished Ph.D.
thesis, Victoria University of Wellington, Wellington.
Carter, L. 1980. New Zealand Region Bathymetry 1:6,000,000, 2nd ed., New Zealand Oceanographic Inst. Chart.
Misc. Series 15.
Chase, C. G. 1978. Plate kinematics: the Americas, East Africa and the rest of the world, Earth Planet Sci. Lett.
37:355-348.
Cole, J. W. 1973. High-alumina basalts of Taupo volcanic zone, New Zealand, Lithos 6:53-64.
Cole, 1. W. 1978. Andesites of the Tongariro volcanic center, North Island, New Zealand, J. Volcanol. Geotherm.
Res. 3:121-153.
Cole, 1. W. 1979. Structure, petrology and genesis of Cenozoic volcanism, Taupo volcanic zone, New Zealand-a
review, N.Z J. Geol. Geophys. 22:631-657.
Cole, J. W. 1981. Genesis oflavas of the Taupo volcanic zone, North Island, New Zealand, J. Volcanol. Geotherm.
Res. 10:317-337.
Cole, J. W. 1990. Structural control and origin of volcanism in the Taupo volcanic zone, New Zealand, Bull.
Volcanol. 52:445-459.
Cole, J. w., and Graham, 1.1.1989. Petrology of strombolian and phreatomagmatic ejecta from the 1976-82, White
Island eruption sequence, in The 1976-82 Eruption Sequence at White Island Volcano (Whakaarl), Bay of
Plenty, New Zealand, N.Z Geol. Surv. Bull. 103:61-68.
Cole, J. W., and Lewis, K. B. 1981. Evolution of the Taupo-Hikurangi subduction system, Tectonophysics 72:
1-21.
Conrad, W. K., Nicholls, I. A., and Wall, V. 1. 1988. Water-saturated and under-saturated melting of metaluminous
and peraluminous crustal compositions at 10 kb: Evidence for the origin of silicic magmas in the Taupo
volcanic zone, New Zealand and other occurrences, J. Petrol. 29:765-803.
Crawford, A. 1., Falloon, T. J., and Eggins, S. 1987. The origin of island arc high-alumina basalts, Contrib.
Mineral. Petrol. 97:417-430.
Crook, C. N., and Hannah, J. 1988. Regional horizontal defonnation associated with the March 2, 1987,
Edgecumbe earthquake, New Zealand, Geophys. Res. Lett. 15(4):361-364.
Crook, C. N., and Hannah, 1. 1989. Regional horizontal defonnation associated with the 1987 Edgecumbe
earthquake, Bay of Plenty, New Zealand-an introduction, N.Z J. Geol. Geophys. 32(1):93-98.
Darby D. J., and Williams, R. O. 1991. A new geodetic estimate of defonnation in the central volcanic region of the
North Island, New Zealand, N.Z J. Geol. Geophys. 34:127-136.
Davey, F. J. 1993. Crustal seismic reflection measurements over a continental backarc basin, North Island, New
Zealand, in Programme & Abstracts, Geophysical Symposium, N.Z. Geophysical Society, Victoria Univer-
sity of Wellington, abstract.
Davey, E J., and Broadbent, M. 1980. Seismic refraction measurements in Fiordland, southwest New Zealand,
N.Z J. Geol. Geophys. 23:395-406.
Dawson, G. B., and Dickinson, D. J. 1970. Heat flow studies in thermal areas of the North Island of New Zealand,
Geothermics Spec. Issue 2:466-473.
Eaton, G. P. 1984. The Miocene Great Basin of western North America as an extending backarc region,
Tectonophysics 102:275-295.
Eiby, G. A 1958. The structure of New Zealand from seismic evidence, Geol. Rundsch. 47:647-662.
Evison, E E, Robinson, R, and Arabasz, W. J. 1976. Microearthquakes, geothermal activity, and structure, central
North Island, New Zealand, N.Z J. Geol. Geophys. 19(5):625-637.
Ewart, A, Hildreth, w., and Carmichael, I. S. E. 1976. Quaternary acid magma in New Zealand, Contrib. Mineral.
Petrol. 51:1-27.
Ewart, A., and Stipp, J. J. 1968. Petrogenesis of the volcanic rocks of the central North Island, New Zealand as
indicated by a study of 87Srt8 2Sr ratios and Sr, Rb, U, and Th abundances, Geochim. Cosmochim. Acta
32:699-736.
Fowler, C. M. R 1990. The Solid Earth-an Introduction to Solid Earth Geophysics, Cambridge University Press.
Gamble, J. A., Smith, I. E. M., Graham, I. J., Kokelaar, B. P., Cole, J. w., Houghton, B. E, and Wilson, C. J. N.
1990. The petrology, phase relations and tectonic setting of basalts from the Taupo Volcanic Zone, New
Zealand and the Kermadec Island arc-Havre Trough, S.W. Pacific, J. Volcanol. Geotherm. Res. 43:253-270.
Gamble, J. A., Smith, I. E. M., McCulloch, M. T., Graham, I. J., and Kokelaar, B. P. 1993. The geochemistry and
petrogenesis of basalts from the Taupo volcanic zone and Kermadec Island arc, S.W. Pacific, J. Volcanol.
Garrick, R A 1968. A reinterpretation of the Wellington crustal refraction profile (letter), N.Z J. Geol. Geophys.
11:1280-1294.
Gibowicz, S. J. 1973. Variation of frequency-magnitude relationship during Taupo earthquake swarm of 1964-65,
N.Z J. Geol. Geophys. 16(1):18-51.
Graham, I. J., and Cole, J. W. 1991. Petrogenesis of andesites and dacites of White Island Volcano, Bay of Plenty,
New Zealand in the light of new geochemical and isotopic data, N.Z J. Geol. Geophys. 103:650-661.
Graham, I. J., Gulson, B. L., Hedenquist, J. W., and Mizon, K. 1992. Petrogenesis of Late Cenozoic volcanic rocks
from the Taupo volcanic zone, New Zealand, in the light of new lead isotope data, Geochim. Cosmochim.
Acta 56:2797-2819.
Graham, I. J., and Hackett, W. R 1987. Petrology of calc-alkaline lavas from Ruapehu Volcano and associated
vents, Taupo volcanic zone, New Zealand, J. Petrol. 28:531-567.
Graham, I. J., and Worthington, T.1. 1988. Petrogenesis of Tauhara Dacite (Taupo volcanic zone, New Zealand)-
Evidence for mixing between high-alumina andesite and rhyolite, J. Volcanol. Geotherm. Res. 35:279-294.
Grange, L.1. 1932. Taupo earthquakes, 1922. Rents and faults formed during earthquake of 1922, in Taupo district,
N.Z J. Sci. Technol. 14(3):139-141.
Grindley, G. W. 1965. The geology structure and exploitation of the Wairakei geothermal field, Taupo New
Zealand, N.Z Geol. Surv. Bull. no. 75.
Grindley, G. W., and Hull, A. G. 1986. Historical Taupo earthquakes and earth deformation, in Recent Crustal
Movements of the Pacific Region, (w. I. Reilly; B. E. Harford, eds.), R. Soc. N.Z Bull. 24:173-186.
Hackett, W. R., and Houghton, B. E 1989. A facies model for a Quaternary andesitic composite volcano: Ruapehu,
New Zealand, Bull. Volcanol. 51:51-68.
Haines, A. J.1979. Seismic wave velocities in the uppermost mantle beneath New Zealand, N.Z J. Geol. Geophys.
22:245-257.
Hatherton, T. 1969. The geophysical significance of calc-alkaline andesites in New Zealand, N.Z J. Geol.
Geophys. 12:436-459.
Hayward, B., 1979, Eruptive history of the Early to Mid-Miocene Waitakere volcanic arc and paleogeography of
the Waitemata Basin, northern New Zealand, J. R. Soc. N.Z 9:297-320.
Healy, J. 1962 Structure and volcanism in the Taupo volcanic zone, New Zealand, in Crust of the Pacific Basin, pp.
151-157, Geophys. Monogr. Ser., Vol. 6, American Geophysical Union, Washington, DC.
Healy, J., Schofield, J. c., and Thompson, B. N. 1964. Sheet 5, Rotorua (1st ed.), Geological Map of New Zealand
1:250,000, New Zealand Department of Scientific and Industrial Research.
Hedenquist, J. W. 1983. Waiotapu, New Zealand: the geochemical evolution and mineralisation of an active
hydrothermal system, Ph.D. thesis, University of Auckland.
Hochstein, M. P., Smith, I. E. M., Regenauer-Lieb, K., and Ehara, S. 1993. Geochemistry and heat transfer
processes in Quaternary rhyolitic systems of the Taupo volcanic zone, New Zealand, Tectonophysics
223:213-237.
Houghton, B. E, Wilson, C. J. N., Lloyd, E. E, Gamble, J. A., and Kokelaar, B. P. 1987. A catalogue of basaltic
deposits within the central Taupo volcanic zone, N.Z. Geol. Survey Rec. 18:95-101.
Houghton, B. E, Wilson, C. J. N., McWilliams, M. D., Lanphere, M. A., Weaver, S. D., Briggs, R. M. and Pringle,
M. S. 1995. Chronology and dynamics of large silicic magmatic system: Central Taupo Volcanic Zone, New
Zealand, Geology.
Hunt, T. M., and Latter, J. H. 1982. A survey of seismic activity near Wairakei geothermal field, New Zealand, 1.
Volcanol. Geotherm. Res. 14(3-4):319-334.
Kear, D. 1959. Stratigraphy of New Zealand's Cenozoic volcanism northwest of the volcanic belt, N.Z. Geol.
Geophys. 2:578-589.
KRISP working party. 1991. Large scale variation in lithsopheric structure along and across the Kenya Rift, Nature
354:223-227.
Lamb, S. H., and Vella, P. 1987. The last million years of deformation in part of the New Zealand plate boundary
zone, 1. Struct. Geol. 9:877-891.
Le Maitre, R. W. (ed.). 1989. A Classification of Igneous Rocks and Glossary of Terms, Blackwell, Oxford.
Lee, L. P. 1978. First-Order Geodetic Triangulation ofNew Zealand 1909-49 and 1973-74, Department of Lands
and Survey Technical Ser. No. 1.
Lewis, K. B. 1980. Quaternary sedimentation on the Hikurangi oblique-subduction and transform margin, New
Zealand, in Sedimentation on Oblique Slip Mobile Zones (P. E Ballance and H. G. Reading, eds.), Int. Assoc.
Sedimentology. Spec. Pub!. 4:171-189.
McCulloch, M. T., Kyser, T. K., Woodhead, 1., and Kinsley, L. 1994. Pb-Sr-Nd-O isotopic constraints on the origin
of rhyolites from the Taupo volcanic zone of New Zealand: evidence for assimilation followed by fractiona-
tion of basalt, Contrib. Mineral. Petrol. 115:303-312.
Mooney, H. M. 1970. Upper mantle inhomogeneity beneath New Zealand: seismic evidence, 1. Geophys. Res.
75:285-309.
Morgan, P., Blackwell, D. D., and Spafford, R. E. 1977. Heat flow measurements in Yellowstone lake and the
thermal structure of Yellowstone caldera, 1. Geophys. Res. 82:3719-3732.
Nairn, I. A 1976. Late Quaternary faulting in the Taupo volcanic zone, in Excursion Guide No. 55A and 55C (S.
Nathan, comp.), 25th Int. Geological Congr.
Nairn, I. A. 1981. Some studies of the geology, volcanic history, and geothermal resources of the Okataina
volcanic center, Taupo volcanic zone, New Zealand, Ph.D. thesis, Victoria University of Wellington.
Nairn, I. A, and Beanland, S. 1989. Geological setting of the 1987 Edgecumbe earthquake, New Zealand, N.Z. 1.
Geol. Geophys. 32(1):1-13.
Nairn, I. A, and Hull, A G. 1985. Post-18oo years B.P. displacements on the Paeroa fault zone, Taupo volcanic
zone, DSIR, N.z. Geol. Surv. Rec. 8:135-142.
Nairn, I. A., Wood, C. P., and Bailey, R. A 1994. The Reporoa caldera, Taupo volcanic zone: source of the
Kaingaroa ignimbrites, Bull. Volcanol. 56:529-537.
Nohda, A. 1984. Classification of island arcs by Nd-Sr isotopic data, Geochem. 1. 18:1-1.9.
Otofuji, Y., Matsuda, T., and Nohda, S. 1985. Opening mode of the Japan Sea inferred from the paleomagnetism of
the Japan arc, Nature 317:603-604.
Otway, P. M. 1986. Vertical deformation associated with the Taupo earthquake swarm, June 1983, R. Soc. N.z.
Bull. 24:187-200.
Otway, P. M., and Sherburn, S. Vertical deformation and shallow seismicity around Lake Taupo, New Zealand,
1985-90. N.z. 1. Geol. Geophys. 37:195-200.
Palmason, G., and Saemundsson, K. 1974. Iceland in relation to the mid-Atlantic ridge. Annu. Rev. Earth Planet.
Sci. 2:25-50.
Patterson, D. B., and Graham, I. 1. 1988. Petrogenesis of andesitic lavas from Mangatepopo valley and Upper
Tama Lake, Tongariro volcanic center, New Zealand, 1. Volcanol. Geotherm. Res. 35:17-29.
Reid, E W 1983. Origin of the rhyolitic rocks of the Taupo volcanic zone, 1. Volcanol. Geotherm. Res. 15:
315-338.
Reid, E W. and Cole, 1. W. 1983. Origin of dacites of the Taupo volcanic zone, New Zealand, 1. Volcano Geotherm.
Res. 18:191-214.
Reilly, W I. 1972. The New Zealand gravity map series, N.Z. Geol. Geophys. 15:3-15.
Reilly, WI. 1990. Horizontal crustal deformation on the Hikurangi margin, N.z. 1. Geol. Geophys. 33(2):393-400.
Richardson, W. P. 1989. The Matata earthquake of 1977 May 31: a recent event near Edgecumbe, Bay of Plenty,
New Zealand, N.Z. J. Geol. Geophys. 32(1):17-30.
Rogan, M.1982. A geophysical study of the Taupo volcanic zone, New Zealand, J. Geophys. Res. 87:4073-4088.
Satake, K., and Hashida, T. 1989. Three dimensional attenuation structure beneath the North Island, New Zealand,
Sc1ater,1. G. Jaupart, c., and Galson, D. 1980. The heat flow through oceanic and continental crust and heat loss of
the earth, Rev. Geophys. Space Phys. 18:269-312.
Sissons, B. A. 1979. The horizontal kinematics of the North Island of New Zealand, Ph.D. thesis, Victoria
University of Wellington.
Skinner, D. N. B. 1986. Neogene volcanism of the Hauraki volcanic region, in Late Cenozoic Volcanism in New
Zealand (I. E. M. Smith, ed.), R. Soc. N.Z. Bull. 23:21-47.
Smith, E. G. C., Scott, B. 1., and Latter, J. H. 1984. The Waiotapu earthquake of 1983, December 14, Bull. N.Z. Nat.
Soc. Earthquake Eng. 17(4):272-279.
Smith, E. G. C., Stem, T. A., and Reyners, M. 1989. Subduction and backarc activity at the Hikurangi convergent
margin, New Zealand, Pure Appl. Geophys. 129:385-409.
Smith, E. G. C., and Webb, T. H. 1986. The seismicity and related deformation of the central volcanic region,
North Island, New Zealand, in Late Cenozoic Volcanism in New Zealand (I. E. M. Smith, ed.), R. Soc. NZ.
Bull. 23:112-133.
Stem, T. A. 1986. Geophysical studies of the upper crust within the central volcanic region, New Zealand, in Late
Cenozoic Volcanism in New Zealand, (I. E. M. Smith, ed.), R. Soc. N.Z. Bull. 23:92-111.
Stem, T. A. 1987. Asymmetric backarc spreading, heatflux and structure associated with the central volcanic
region of New Zealand, Earth Planet. Sci. Lett. 85:265-276.
Stem, T. A. 1979. Regional and residual gravity fields, central North Island, New Zealand, N.Z. J. Geol. Geophys.
22:479-485.
Stem, T. A., and Davey, F. J. 1987. A seismic investigation of crustal and upper mantle structure within the central
volcanic region of New Zealand, NZ. J. Geol. Geophys. 30:217-231.
Stem, T. A., Smith, E. G. c., Davey, F. J., and Muirhead, K. H. 1987. Crust and upper mantle structure of the
northwestern North Island, New Zealand, Geophys. J. R. Astronom. Soc. 91:913-936.
Studt, F. E., and Thompson, G. E. K.1969. Geothermal heat flow in the North Island of New Zealand, N.Z. J. Geol.
Geophys. 12:673-683.
Walcott, R. I. 1984. The kinematics of the plate boundary zone through New Zealand: a comparison of short- and
long-term deformations, Geophys. J. R. Astronom. Soc. 79:613-633.
Walcott, R. I. 1987. Geodetic strain and the deformational history of the North Island of New Zealand during the
late Cainozoic, Phil. Trans. R. Soc. London A 321:163-181.
Webb, T. H., Ferris, B. G., and Harris, J. S. 1986. The Lake Taupo, New Zealand, earthquake swarms of 1983,
N.Z. J. Geol. Geophys. 29(4):377-389.
Williams, R. O. 1987. The June 1986, Survey of the North Taupo regional monitoring pattern, Department of
Scientific and Industrial Research, N.Z. Geo!. Surv. Rep. EDS111.
Williams, R. O. 1989. Horizontal angle survey data, 1949-1954, acquired for studies of contemporary deformation
of the Taupo belt, N.Z. Geol. Surv. Rep. EDSI22.
Wilson, C. J. N. 1986. Reconnaissance stratigraphy and volcanology of ignimbrites from Mangakino Volcano, in
Late Cenozoic Volcanism in New Zealand (I. E. M. Smith, ed.), R. Soc. NZ. Bull. 23:179-193.
Wilson, C. J. N., Houghton, B. F., and Lloyd, E. F. 1986. Volcanic history and evolution of the Maroa-Taupo area,
central North Island, in Late Cenozoic Volcanism in New Zealand (I. E. M. Smith, ed.), R. Soc. NZ. Bull.
23:194-223.
Wilson, C. J. N., Rogan, A. M., and Smith, I. E. M. 1984. Caldera volcanoes of the Taupo volcanic zone, New
Zealand, J. Geophys. Res. 89:8463-8484.
Wood, C. P. 1983. Petrological logs of drillholes BI 26, to Br 40: Broadlands Geothermal Field, NZ. Geol. Surv.
Rep. 108.
Wood, C. P. 1986. Stratigraphy and petrology of N.M.4-Ngatamariki geothermal field. Unpublished report of
the N.Z. Geo!. Survey, DSIR. Lower Hutt, New Zealand.
Wright, I. C. 1992. Shallow structure and active tectonism of an offshore continental backarc spreading system: the
Taupo volcanic zone New Zealand, Mar. Geol. 103:287-309.
Wright, I. c., and Walcott, R. I. 1986. Large tectonic rotation of part of New Zealand in the last 5 Ma, Earth Planet.
Sci. Lett. 80:348-352.
2
The Southern Havre Trough

Geological Structure and
Magma Petrogenesis of an
Active Backarc Rift Complex
J. A. Gamble and I. C. Wright
ABSTRACT
Associated with Pacific-Australian plate convergence, the Havre Trough-Kermadec arc-

backarc system (SW Pacific) is an archetypal example of an "island arc" plate boundary.
Swath data from the southern Havre Trough between 33-34°S and 35°30' -37°S reveal that
within an active, young backarc basin highly complex and heterogeneous tectonic and
magmatic processes can be associated with a tectonically "simple" subduction system. The
compressed across arc-backarc length scale would appear to be a significant control on the
complexity of the tectonism and magmatism. A near-contiguous, segmented axial rift
system, flanked by marginal rifts and subbasins, and horst ridges form the dominant
backarc morphology. Arc edifices of the present frontal arc lie within the "backarc region,"
sited 15-25 km west of the Kermadec frontal ridge. At its southern extremity, the axial rift
system is propagating southward toward New Zealand, penetrating the continental-oceanic
arc crustal boundary. Contrary to previous models, the entire Havre Trough is now
interpreted to be undergoing backarc rifting prior to true spreading, with the attendant
tectonic fabric dextrally oblique to the bounding Colville and Kermadec ridges. "Pseudo-
linear" magnetic anomalies are interpreted to result from the generally irregular spatial and
temporal emplacement of sheeted lava and dike intrusives between older arc basement.
Significant magma source heterogeneity is observed both along and orthogonal to the
frontal Kermadec arc. Basalts from the northern and southern sectors of the Kermadec arc
define two distinctive isotopic arrays, with the latter overlapping with primitive basalts
from the onshore Taupo Volcanic Zone. Across the arc-backarc system, lavas within the
axial rift system have a marked spatial asymmetry and heterogeneity, with mid-ocean ridge
basalts (MORB)-like and arc basalts being separated by as little as 15 km.
J. A. Gamble • Department of Geology, Research School of Earth Sciences, Victoria University of Welling-
ton, Wellington, New Zealand. I. C. Wright • New Zealand Oceanographic Institute, National Institute of
Water and Atmospheric Research, Wellington, New Zealand.
29
30 J. A. GAMBLE AND I. C. WRIGHT
1. INTRODUCTION
The Havre Trough-Kermadec arc-backarc system associated with Pacific-Australian

plate convergence (Fig. 2.1), along with its contiguous northern and southern extensions
(the Lau Basin and Taupo Volcanic Zone (TVZ», is an archetypal example of an "island
are" plate boundary. Indeed the Havre Trough-Kermadec system formed an original
example of a convergent plate boundary from which the concept of backarc basin evolution
was initially developed over 20 years ago (e.g., Karig, 1970a, 1974; Packham and Falvey,
1971). In spite of its early recognition as a young, active backarc basin the nature and spatial
distribution of present-day tectonic and magmatic processes within the Havre Trough are
proving now to be not well known. Varying interpretations of Havre Trough backarc tecton-
ism and magmatism, based on limited data, have been proposed (e.g., Karig, 1970a; Mala-
hoff et aI., 1982; Taylor and Kamer, 1983). Segments of the Havre Trough backarc system,
now relatively well studied, reveal that tectonic and magmatic processes are highly com-
plex and variable, even at length scales of -15 km. Such heterogeneity is significant within a
backarc basin linked to a tectonically "simple" subduction system where the plate bound-
ary is linear, convergence is more or less orthogonal to it, and subduction polarity has not, if
ever, reversed from the present westward directed convergence within the last 30-40 m.y.
The implications of these observations are important as we unravel from the rock record the
processes of volcanism and extension within, and tectonic settings of, backarc arc systems.
NORTH
FIJI
BASIN
AUSTRALIAN
PLATE
SOUTH
FIGURE 2.1. Australian-Pacific "island

arc" plate boundary and associated contig-
uous Lau-Havre-Taupo backarc system.
Boxes I, 2, and 3 refer to sections of this
backarc system described by, respectively,
Hawkins (Chapter 3 this volume), this chap-
ter, and Cole et al. (Chapter I this volume).
Open arrows give the rate of relative plate
convergence in mm yc 1 from Minster and
Jordan (1978).
THE SOUTHERN HAVRE TROUGH 31
SOUTH
FIJI
BASIN
FIGURE 2.2. Regional bathymetry

(in meters) and tectonic structure of
the Kermadec-Hikurangi system and
associated Havre-Taupo backarc sys-
tem. Position and depth of the sub-
ducted slab are from Pelletier and
Dupont (1990). Dashed line is the in-
ferred position at about 32°S of
changes in the morphology of the
Kermadec subduction system (see text
for discussion). Box outlines A and B
show locations of Figs. 2.3 and 2.9.
Shaded inserts show the two sectors
of the backarc system mapped with
swath data.
Within the last six years two sectors of the southern Havre Trough (SHT) have been
relatively well surveyed with modem swath data (Fig. 2.2): a segment between 33° and
34 oS (Caress, 1991) and the southernmost sector of the Havre Trough between 35°30' and
37°S (Seafloor Surveys International, 1990, 1993; Wright, 1990; Wright et aI., 1990; Black-
more and Wright, 1995). Both sectors appear to display a range of typical and atypical
tectonic morphologies for at least the SHT. The southernmost sector is particularly impor-
tant because (1) it in part forms the transition from an oceanic backarc-arc system into the
continental Taupo Volcanic Zone of North Island, New Zealand (Karig, 1970b; Gamble et
al. 1993b; Cole et al., Chapter 1 this volume), and (2) seafloor photographs and geochemical
data are available from this segment (Gamble et aI., 1993b, 1994; Wright, 1993a, 1994) from
which to characterize SHT volcanism and hence relate the concomitant processes of
tectonism and magmatism. This chapter provides a synopsis of these new data from the
SHT, so as to document the complexity and heterogeneity of tectonic and magmatic
processes within an active, young, tectonically "simple," backarc basin.
2. REGIONAL STRUCTURE AND TECTONICS
2.1. General
The regional structure and tectonics of the submarine Lau and Havre segments of the
nearly 2700-kIn-long contiguous Tonga-to-New Zealand backarc system, although display-
ing variations along strike, are typical of active backarc basins. The Lau-Havre system is
characterized by variable but relatively high heat flow (Sclater et al., 1972; Watanabe et al.,
1977), extremely complex seafloor morphology (Caress, 1991; Wright, 1993a), minimal or
absent sediment cover (Wright, 1993b), heterogeneous and spatially dispersed volcanism
(Gamble et al., 1993b; Hawkins et al., 1994), and extensive shallow seismicity (Sykes et al.,
1969; Pelletier and Louat, 1989). The first-order variation in the backarc structure, along the
strike of the Lau-Havre system, is essentially a function of distance along the plate
boundary from the Australian-Pacific rotation pole located at 62°S, 174°E. Increasing
distance from the rotation pole northward along the plate boundary naturally coincides with
an increase in the rate of plate convergence (Fig. 2.1). Between 35°S and 200 S the rate
increases from about 53 to 90 mm yr- 1 (Minster and Jordan, 1978; de Mets et al., 1990);
estimates of the rate do vary by some 15%, depending on whether Tonga-Kermadec
earthquake-slip vectors are incorporated into the plate motion models. This northward
increase in the rate of convergence is also matched by a general northward increase in the
rate of backarc opening. The latter is shown by present-day spreading occurring in the Lau
Basin to the north (Parson and Hawkins, 1994) and rifting to the south within the Havre
Trough (Wright, 1993a).
Pelletier (1990) and Pelletier and Dupont (1990) have, using reconnaissance data,
recognized clear differences in the morphology, depth, and position of the subduction
trench and backarc region, north and south of 32°S, which they relate to an increase in the
length of the subducted slab north of 32°S (Fig. 2.2). Thus, along the northern Havre-
Kermadec sector, tectonic erosion of the inner subduction trench wall, a relatively simple,
thickly sedimented backarc region as shallow as 2500 m coincides with subducted slab
seismic activity to a depth of 550-650 kIn. To the south, accretion along the inner
subduction trench wall, and a structurally complex backarc region as deep as 3000 to 4000
m, coincides with subducted slab seismic activity to a depth of 250 kIn. Limited swath data
substantiate that the changes in structural morphology, at least in the backarc region
(Caress, 1991), occur near 32°S.
2.2. Southern Havre-Kermadec Backarc-Arc System
The SHT (south of 32°S) is, on average, 120 kIn wide between the 2000-m isobaths of
the parallel and flanking Colville and Kermadec ridges. Normal faults mark the boundaries
between the trough and the inner margins of both ridges (Karig, 1970a; Caress, 1991;
Wright, 1994). Both ridges are characterized by basalt-andesitic arc volcanism (Cole, 1986;
Gamble et al., 1990) and are interpreted to have been a single magmatic arc prior to opening
of the Havre Trough (Karig, 1970a; Wright, 1993a). Present-day arc volcanism, including
eruptions at submarine and subaerial volcanoes, generally occurs within the vicinity of the
Kermadec Ridge (Fig. 2.2; Cole, 1986; Dupont, 1988), although the precise position of the
volcanic front varies along the plate boundary (Wright, 1994). Volcanism, however, may
persist at the southern end of the Colville Ridge (Wright et al., 1990; Seafloor Surveys
International, 1993).
The intratrough tectonic fabric comprises a series of dextral en echelon basement
ridges and rift grabens that are generally 20-30° oblique to the trend of the bounding
Colville and Kermadec ridges (Wright et al., 1990; Caress, 1991). Slip vectors from shallow
seismic events within the southern Havre Trough have near identical oblique azimuths
(Pelletier and Louat, 1989). At its southern extremity, the SHT is marked by the 3300-
m-deep Ngatoro rift graben system (Fig. 2.3), which in part forms a major bathymetric
reentrant impinging on the northeast New Zealand continental slope (Cole and Lewis, 1981;
Lewis and Pantin, 1984).
2.3. Taupo-Havre Continental-Oceanic Transition

The transition between the oceanic and continental segments, recognizable from
bathymetric, magnetic anomaly, and petrologic data forms a near-linear boundary along the
northeast New Zealand continental margin (Fig. 2.3; Wright, 1993a). Marked changes in
both the amplitude and spatial fabric of magnetic anomalies occur at the boundary (Davey
and Robinson, 1978; Malahoff et ai., 1982; Lewis and Pantin, 1984; Wright, 1992). Mag-
netic and gravity modeling indicate that the Raukumara Plain to the northeast of the crustal
boundary is underlain by oceanic crust (Gillies and Davey, 1986). Similarly, Gamble et al.
(1990, 1993b, 1994) have documented significant differences in the compositional range,
mineralogy, and petrologic nature of backarc rift and arc-related volcanic rocks across this
FIGURE 2.3. Synoptic bathymetry (meters)

and structure of the southern Kermadec-Havre
Trough region. Box outlines C and D show loca-
tions of Figs. 2.4 and 2.8. NB, Ngatoro Basin;
NR, Ngatoro Ridge; VMFS, Vening Meinesz
fracture zone.
boundary. These relations have been interpreted to reflect the different crustal settings
of the volcanism.
Two features of the oceanic-continental transition are pertinent to the regional struc-
ture and tectonism of the SHT. First, this crustal boundary is remarkably distinct and linear
(Fig. 2.3) and appears to coincide with the inferred position of the Vening Meinesz fracture
zone (van der Linden, 1967). Hence, the present oceanic arc-continental boundary may be
an inherited feature resulting from transform motion along the outer edge of the New
Zealand continental margin during Oligocene backarc spreading of the South Fiji Basin
(Davey, 1982). Second, the sole deviation from the linearity ofthe crustal boundary is the
tectonic reentrant coinciding with the southern Ngatoro graben basin (Fig. 2.3). This
reentrant is interpreted to be attentuated continental crust, with morphological and geo-
chemical characteristics transitional between normal continental crust and oceanic arc crust
(Gamble et al., 1993b, 1994), and marks the southern limit of SHT rifting that is propagating
into the New Zealand continental margin (Wright, 1993a).
The continental segment of the backarc (the TVZ) extends southward for some 370
km into North Island, New Zealand. The axes of oceanic SHT and continental TVZ backarc
extension are not continuous (Fig. 2.3), having a left lateral offset of 4S-S0 km (Lewis and
Pantin, 1984; Wright et al., 1990). A series of dextrally oblique and en echelon faults form
the surficial expression of this offset (Wright, 1992). The offshore TVZ comprises a 40-4S-
km-wide volcano-tectonic structure consisting of a volcanic arc ridge (the Ngatoro Ridge)
flanked by rift grabens (Wright, 1992). Volcanism within the offshore TVZ is widely
distributed and has a range of compositions, including silicic dome complexes (Gamble
et aI., 1993b).
3. STRUCTURE AND MORPHOLOGY OF THE 3S030'S-37'S SECTOR
3.1. General
The 3S030'S-37°S sector comprises a morphologically heterogeneous SO-6S-km-
wide zone of complexly arranged basement ridges and troughs, split by an axial rift system
(Fig. 2.4; Wright, 1993a). The characteristic morphology of the rift system is a series of
segmented half-grabens bounded by master normal faults. Each half-graben typically has
the constituent elements of (1) an outer rift graben with associated outer escarpments, (2) a
series of generally downstepping fault block ridges and terraces, and (3) an inner rift graben
(Fig. 2.S).
3.2. Outer Rift Graben

The outer rift graben forms a IS-2S-km-wide, segmented and en echelon rift system
that trends 04So. The morphology of the rift system and shallow seismic reflection data
(Figs. 2.4a and 2.S) suggest that the opposing outer graben escarpments comprise a parallel
pair of inward-facing master and antithetic normal faults. These master and antithetic fault
escarpments typically have relief of 800-1000 m and IS0-200 m, respectively. In plan, the
outer escarpments have an en echelon and zigzag arrangement with both the master and
antithetic fault segments trending either 030° or 060° (Fig. 2.6). Generally, the fault
segments are right-stepping, with both fault azimuths appearing to have similar strike
lengths.
38'30'
Raukumara
Plain
I(EY
" " Mast... fault

/ ' Fau~
37"·
178' 178"30, A
FIGURE 2.4. (a) Bathymetry (100-m contour interval) of the southern Kermadec Ridge and Havre Trough, and offshore Taupo Volcanic Zone region (after Blackmore and Wright, 1995).
SNR, CNR, NNR, and NC refer to the southern, central and northern Ngatoro rifts, and Ngatoro Canyon, respectively. Lines H4 and H7 locate single-channel seismic reflection profiles
shown in Fig. 2.5. (b) GLORIA mosaic of the southern Kermadec Ridge and Havre Trough, and offshore Taupo Volcanic Zone region. Bright zones are areas of high acoustic reflectivity.
NW SE
-
2
-Fd
11~
Fd
Jlt
I¢o A>dIII SHT RIft o..ben ~
I...
..
3
-
[Una H4 ! 0 101<m
I I
5
...- Fd Fa
11~ It
NW I¢:> AldaI SHT Rlft Graben c¢I SE
I
l-
I-
~
I-
~----------------------------------------~--------~5
FIGURE 2.5. Single-channel air-gun seismic reflection profiles showing the tectonic morphology of the Ngatoro
rift system, southern Havre Trough (after Wright, 1993a). TWTT=two-way travel time (s).
....
,
FIGURE 2.6. Synoptic geological-

tectonic interpretation of the Ngatoro
rift system, southern Havre Trough,
showing the rift graben, graben fill sedi-
••
ments, faults, volcanic vents, and arc
stratovolcanoes (after Wright, 1993a).
SNR, CNR, and NNR, refer to the .
•01 ...
Nt IIlIIMIcIno
~
southern, central and northern Ngatoro
rifts, respectively. "-
Between the outer antithetic fault escarpment and the inner rift, fault block terraces
and ridges form a heterogeneous structure of down stepping normal faults interspersed with
elongate ridges and depressions (Figs. 2.4a and 2.5). Faults bounding individual blocks
comprise linear or near-linear scarps with relief of 50-100 m and are longitudinally
continuous over distances of at least 6-10 km. Fault terraces are back-tilted and infilled
to form a descending series of ponded sediment dams.
The structural fabric of the SHT rift system is disrupted by "transfer offsets," which
partition the half-grabens into 25-30-km-Iong, partially linked contiguous rift segments
(Fig. 2.6). These transfer offsets disrupt the spatial arrangement of both the outer and inner
rift grabens. The surficial morphology of the rifts suggests that the asymmetry of the half-
grabens varies between rift segments along the strike of the system, alternating at the
transfer offsets.
3.3. Inner Rift Graben

At the southern extremity of the SHT rift system, the three inner rift grabens (the
southern, central, and northern Ngatoro rifts) are each 24-26 km long and 10-12 km wide.
At the "transfer offsets" the inner rift grabens are connected by narrow (2-5 km wide)
"interrift corridors" that consist of bathymetric sills or steps with relief of 100-200 m
coinciding with sediment-free basement highs.
Bathymetric and geologic long-range inclined asdic (GLORIA) data (Figs. 2.4 and
2.6) highlight the pronounced V-shape of the southern Ngatoro rift basin, which penetrates
the continental edge of New Zealand by some 25 km (Wright, 1993a). The progressive
southwestward narrowing of the rift appears to be accommodated by the antithetic faults
altering their orientation, while the opposing master fault retains its linearity. At the head of
the graben, the rift escarpments progressively converge to become a series of parallel
normal faults that subsequently merge with the tectonically controlled, 500-m-wide
Ngatoro canyon.
Along the length of the southern rift floor is a 3-km-wide, 150-m-high volcanic ridge
capped with conical knolls (Blackmore and Wright, 1995). This ridge, comprising in part
glassy basalts (Gamble et ai., 1993b) and largely devoid of sediment cover, forms the main
recent constructional volcanic feature within at least the three segments of the Ngatoro rift
system (Wright, 1993a).
Landward of the propagating southern Ngatoro rift, a narrow linear structural "fur-
row" comprising inward-facing active normal faults is identifable from single-channel
seismic reflection data. This structural furrow (Fig. 2.3) is inferred to form a line of
incipient continental rifting (Cole and Lewis, 1981; Wright, 1992), extending southwest-
ward to the basalt-pantellerite volcano of Mayor Island (Ewart, 1968; Cole, 1978; Houghton
et ai., 1992). Farther southwestward in North Island, New Zealand, across the projected line
of rifting, geodetic data indicate considerable extension around 4 mm yc 1
The central and northern Ngatoro rifts (Fig. 2.6) form the main, generally flat-floored
and partially sediment infilled, grabens of the Ngatoro segment of the SHT system.
Although the morphology of these rifts appears typical of SHT rifts in general (Blackmore
and Wright, 1995), the significant sediment infilling of up to 600 m (Wright, 1993b) is not.
The latter is interpreted to be a function of sediment supply, due to a high flux of terrigenous
detritus from the New Zealand landmass via the Ngatoro Canyon system (Wright et al.,
1990). Isopachs of graben fill clearly define an underlying structural basement comprising
elongate ridges and troughs with relief of 200 m (Wright, 1993a) that are interpreted to mark
the position of subsurface synthetic and antithetic faults. Most of the extension within these
rifts appears to be accommodated by the bounding master and antithetic faults. Some active
deformation does, however, occur on the intrarift synthetic and antithetic faults, as evinced
by the development of seafloor scarps on the sediment-flooded rift floor (Seafloor Surveys
International, 1990).
3.4. Rift Volcanism

Recent volcanism within the Ngatoro rift system is generally not associated with
youthful constructional terrains on the rift graben floor. The major exception is the axial
ridge of the southern rift (Fig. 2.6). Likewise, farther to the north (between 35°35' and
35°50'S) constructional volcanic features within the half-graben rift (Fig. 2.4a) are minor
or absent (Seafloor Surveys International, 1993). Sites of "neovolcanism" occur along both
the master and antithetic fault escarpments of at least the Ngatoro rifts. Photographic and
rock dredge samples show that the rift escarpment volcanism comprises lobate and ropy
pillow flows of basalt with fresh 4-5-mm-thick glassy rinds (Gamble et ai., 1993b; Wright,
1993a).
3.5. Rift Flank Structure and Volcanism

Flanking the axial rift system, the intratrough region encompasses a heterogeneous
terrain of isolated knolls and seamounts and elongate ridges and depressions. The latter
have axes parallel or subparallel to the axial rift system (Fig. 2.4). Apparent morphological
differences occur between the two terrains flanking the rift system (Wright, 1993a),
although these have yet to be quantified. Distant from the effects of hemipelagic sedimenta-
tion derived from New Zealand, significant areas of the flanking terrains consist of exposed
basement rock (Caress, 1991; Seafloor Surveys International, 1993). Side-scan imagery
shows clearly that the flanking margins, at least those of the 35°35' to 35°50' S rift segment,
comprise major fields of extrusive volcanism (Seafloor Surveys International, 1993), which
are possibly active. Like the structural fabric, this extrusive volcanism also appears to vary
between the respective margins flanking the rift, being more extensive on the flank nearest
to the volcanic front and subduction trench.
3.6. Seismicity
Regional compilations of New Zealand shallow earthquakes during the period 1956-
1987 (Hatherton, 1980; Reyners, 1989) show that the southern extremity of the SHT rift
system is a seismically active region (Fig. 2.7). As noted by Reyners (1989), most of this
seismicity occurred during two earthquake swarms in 1974 and 1984-1985. The site of the
earlier swarm, lying along the western margin of the southern Ngatoro rift, is clearly
associated with active extensional tectonism of the rift system (Wright, 1993a). Further, the
concentration of the 1974 swarm along the western margin of the southern rift is consistent
with the interpreted position of the master fault on the western flank for that rift segment. To
the north of 36°30' S the apparent decrease in seismicity largely reflects the limit of onshore
instrument coverage, although the seismicity that is recorded is almost entirely restricted to
the axial rift system. Elsewhere intratrough shallow seismicity (e.g., Pelletier and Louat,
1989) cannot be readily related to the tectonic structure of the backarc, because ocean
bottom seismometers have yet to be deployed within the Havre Trough.
FIGURE 2.7. Compilation of shallow seismicity (ML 4.0 and depth <40 Ian) lying within the limits of the New
Zealand Seismograph Network for the period January 1964-September 1991 within the Ngatoro rift system (after
Wright, 1993a).
The 1984-1985 swarm was more extensive than the 1974 swarm (Reyners, 1989) and
coincides with the northern terminus of the offshore TVZ (Fig. 2.7; Wright, 1993a). Seis-
mological evidence indicates that the 1984-1985 events are the result of normal and strike-
slip faulting with a consistent T-axis striking northeast-southwest, parallel with SHT
structure.
3.7. Heat Flow

Although the precise magnitude of heat flow within submarine backarc basins is
difficult to establish (Sclater et al., 1980), reconnaissance measurements within and near the
Ngatoro rift system reveal a variable but generally high heat flow (Whiteford, 1990, 1992).
Outside the rift graben mean values from both flanks lie within the range of 106-114 mW
m- 2, whereas within the central Ngatoro rift segment the mean value is 86 mW m- 2 • Heat
flow within the onshore TVZ segment of the backarc basin is about 800 mW m- 2 (Stem,
1987; Cole et aI., Chapter 1 this volume).
3.B. Volcanic Front

Like the northern Kermadec Ridge where active subaerial island volcanoes (e.g.,
Raoul, Macauley, and Curtis) lie along the ridge axis (Fig. 2.2), it has been generally
FIGURE 2.8. Bathymetry (100-m contour interval) and GLORIA sonograph of the southern Kerrnadec Ridge-
Havre Trough (35°-36°S) (after Wright, 1994). Signal polarity is the same as Fig. 2.4.
assumed that the southern Kennadec volcanic front coincides with the frontal ridge (e.g.,
Dupont, 1988). Within the 35°-37°S sector, the Kennadec volcanic front is defined by
seven discrete submarine arc volcanoes (Figs. 2.4 and 2.8): Rumble II, III, IV, V, Silent II,
Tangaroa, and Clark (Kibblewhite and Denham, 1967; Wright, 1994). To the south the
volcanic front continues with Whakatane volcano sited at the northern extremity of the
TVZ (Wright, 1992; Blackmore and Wright, 1995). All seven of the edifices of the volcanic
front lie within the eastern confines of the Havre Trough, some 15-25 km west of the
southern Kennadec frontal ridge axis (Wright, 1994), and 35-45 km east of the axial
backarc rift graben system (Wright, 1993a). Along the same latitudinal sector of the
subduction system, constructional arc volcanoes are not identified on the Kennadec Ridge.
The temporal history of southern Kennadec arc volcanism is largely unknown; however,
very limited data suggest that arc volcanism associated with the present 35-37°S sector of
the arc front extends back to least 0.77 Ma (Wright, 1994). These arc volcanoes continue to
be the sites of recent volcanism with fresh lavas devoid of sediment cover observed on most
volcanoes (Gamble et aI., 1993b; Blackmore and Wright, 1995), and modern submarine
eruptions were recorded from Rumble III (Latter and Hall, 1986). Hence, the late Quater-
nary to modern southern Kennadec volcanic front lies within the "backarc" region, west of
the Kennadec frontal ridge, and thus postdates the early phases of backarc widening.
4. STRUCTURE AND MORPHOLOGY OF THE 33°S-34°S SECTOR
4.1. Rift System

Data from the only other swath-mapped sector (between 33° and 34°S) within the
Havre Trough (Fig. 2.9) indicate that the axial, en echelon rift graben morphology recogni-
zable between 35°40' and 37°S is the characteristic structural fabric for at least the SHT. A
series of short, segmented en echelon grabens, which are oblique by some 20° to the
bounding arc ridges (Fig. 2.9), are clearly identifiable from GLORIA and Sea Beam data
(Caress, 1991). Like the segment to the south these grabens are 15-35 km in length, 6-10
km wide, have depths of 3500-4000 m, and are generally 500-1000 m deeper than the outer
rift flanks. As with the 35°40' -37°S sector the morphology of the grabens is dominated by
extensional faulting with no obvious constructional volcanic features within the area
mapped by Sea Beam. The individual rift segments are typically offset by 14-30 km and
linked, like those to the south, by "transfer offsets" comprising 100-300-m-high, but nar-
row, topographic sills. Sparse single-channel seismic reflection data (Caress, 1991) indicate
the rifts are probably asymmetric half-grabens like those to the south. The only apparent
difference between the two swath-mapped sectors is the degree of sediment infilling of the
rift grabens, which is interpreted to be a function of sediment supply (Wright, 1993a).
4.2. Volcanic Front

As in the southern sector, the arc stratovolcanoes of the 33-34°S sector lies within the
"backarc" region, sited some 25 km west of the Kennadec Ridge (Wright, 1994). GLORIA
(Caress, 1991) and conventional bathymetric data (Wright, 1994) show that the volcanic
front is marked by three major (as yet unnamed) volcanoes (Fig. 2.9), each with a construc-
tional volume around 100 km3• Similarly, these data show no evidence of late Quaternary
constructional arc stratovolcanoes on the attendant section of the Kennadec frontal ridge.
FIGURE 2.9. Bathymetry (at 500-m interval) of the Kermadec Ridge, Havre Trough, and Colville Ridge between
32°S and 35°S based on swath data (Caress, 1991) and conventional soundings (Wright, 1994). Insert box and
dashed lines mark, respectively, the swath survey area and positions of the four grabens described by Caress (199\).
5. TECTONISM OF THE SOUTHERN HAVRE TROUGH
5.1. Rifting Model

The Havre Trough, until recently, has been almost exclusively interpreted as a site of
intraoceanic spreading with the accretion of new oceanic crust along a spreading center
system (Malahoff et aI. , 1982; Caress, 1991; Wright, 1992, 1993b). More recently Wright
(1993a) has interpreted the shallow structure and surficial morphology of, at least, the
SHT as evidence of rifting, with its concomitant attenuation by rift block development (Fig.
2.10). In particular, the recognition of a segmented axial rift system, with its constituent
asymmetric half-grabens along an apparently near-contiguous 440-km section of the SHT
(between at least 33°S and 37°S), is considered to be convincing evidence for rifting rather
than spreading. The shallow structure and surficial morphology of the SHT are almost
identical to that documented within the Izu-Bonin backarc system (Taylor et aI., 1990,
1991). Multichannel crustal seismic reflection studies and Ocean Drilling Program (ODP)
drill-hole data (Taylor et aI., 1991; Klaus et aI., 1992; Taylor, 1992) provide conclusive
evidence for rifting within the Izu-Bonin system. The absence of ODP drill-hole data and
crustal seismic reflection or refraction data from the Havre Trough means the nature of the
inferred rift structure and associated magmatism at subcrustal depths is unknown. How-
ever, by analogy with the Izu-Bonin and western Lau Basin backarc basins (Hawkins et aI.,
1994; Hawkins, Chapter 3 this volume), it is inferred that variable volumes of pillow and
FIGURE 2.10. Schematic block model interpretation showing the southern Havre Trough rift morphology with
attendant sheeted pillow lava and dike intrusives. Overlay I shows a schematic magnetization model of the rifted
structure with magnetic inter-block intrusives emplaced between rift blocks of older low-magnetization arc crust.
Overlay 2 shows the resultant magnetic anomaly profiles along arbitrary survey lines which form "pseudo-linear
magnetic anomalies" (after Wright. 1993a).
sheet lava flows and sheeted dikes are emplaced within thinned, pervasively faulted Havre
Trough subarc crust (Fig. 2.10; Wright, 1993a).
Further evidence for rifting within the Havre Trough comes from the recognition that
crustal spreading in the Lau Basin, along the Valu Fa spreading ridge, has propagated
southward only as far as -23°S (Parson et al., 1990; Weidicke and Collier, 1993). Hence,
crustal spreading has yet to propagate southward into the Havre Trough. The termination of
crustal spreading within the southern region of the Lau Basin is consistent with a marked
increase in the rate of backarc widening from about 8 to 80 mm yc 1 at this latitude
(Pelletier and Louat, 1989). Thus, Lau Basin crustal spreading is propagating southward at
110 mm yc 1 (Parson and Hawkins, 1994) toward the Havre Trough, while 1000 km to the
south, rifting in the Havre Trough is propagating into the already extensional TVZ of
continental New Zealand. The extremely high heat flow within the onshore TVZ has been
taken as evidence for true crustal spreading (Stem, 1987) but may be better explained as
reflecting massive emplacement of arc and rift intrusives within continental crust.
A corollary of the rifting model is that the magnetic anomalies, identified within the
SHT by aeromagnetic survey (Malahoff et aI., 1982), do not record the accretion of new
backarc crust associated with spreading. The SHT magnetic anomaly sequence forms a
variable and, in part, low-fidelity record, with both irregular and apparently symmetrical
anomaly profiles. Much of the displacement and irregularity of the anomalies was origi-
nally attributed to the presence of transform faults (Malahoff et aI., 1982).
Alternatively, Wright (1993a) has interpreted the low-fidelity SHT magnetic anomaly
sequence as recording the emplacement of magnetic sheeted pillow lavas and dike intru-
sives between older, less magnetized, crustal arc rift blocks (Fig. 2.10). Hence, the magnetic
anomalies are interpreted as "pseudolinear magnetic anomalies," resulting from the gener-
ally irregular spatial and temporal emplacement of subcrustal intrusives between these rift
block segments (which are <35 km in strike length) flanking the axial rift graben. The
magnetic anomaly pattern is complicated further by the location of the volcanic arc
stratovolcanoes within the "backarc" region (Fig. 2.10). The proposed diffuse, low-fidelity
anomaly sequence resulting from this rifting model is argued by Wright (1993a) to be more
consistent with the observed SHT anomaly data than with crustal spreading.
5.2. Age and Rate of Extension

Interpretation of the SHT magnetic anomalies as recording phenomena other than the
accretion of oceanic crust necessitates that the age and rate of basin widening be resolved
from data other than magnetic anomaly sequences. Previously the anomaly sequences have
been interpreted as recording a whole spreading rate and age of basin opening as, respec-
tively, either 54 mm yr- 1 and -2 Ma (Malahoff et aI., 1982) or 20 mm yc 1 and 4 Ma
(Korsch and Wellman, 1988).
Three sets of independent proxy data are available, however, to constrain the timing of
basin initiation and the subsequent rate of extension (Wright, 1993a). These data are (1) K-
Ar dating of eastward-migrating North Island intracontinental arc volcanism (Stem, 1987),
(2) onshore North Island geodetic retriangulation (e.g., Walcott, 1984; Darby and Williams,
1991), and (3) slip vectors of present-day plate motion (e.g., Pelletier and Louat, 1989).
These data provide a first approximation for the timing of basin opening and subsequent
rate of widening for the 35° -37°S segment of the Havre Trough, which Wright (1993a) has
taken as 5 Ma and 15-20 mm yc1,respectively. The estimate of 5 Ma for the initiation of
SHT rifting is consistent with a 10-5 Ma age for the earliest phases of rifting within the Lau
Basin to the north (Parson and Hawkins, 1994) and concordant with a model of southward-
directed propagation of rifting and spreading along the Lau-Havre backarc system.
6. GEOCHEMISTRY
6.1. General
Geochemical data for volcanic rocks from the oceanic Kermadec arc have been re-
ported by Brothers (1967) and Ewart et aZ. (1977), and more recently by Ewart and Hawkes-
worth (1987), Smith and Brothers (1988), and Gamble et at. (1990, 1993a, b, 1994). For the
backarc region, geochemical data are sparse and presently confined to a pillow basalt from
34°28'S 178°52'E (Gamble et aI., 1990, 1993a) and a suite of pillow lavas from the eastern
and western flanks of the Ngatoro rift system between 36° and 37°S (Gamble et aI., 1993b,
1994). For the continental setting of New Zealand, geochemical data on basalts and
andesites are more numerous from the onshore TVZ (e.g., Cole, 1986; Graham and Hackett,
1987; Gamble et aI., 1990; Graham and Cole, 1991) and its northward offshore extension in
the Bay of Plenty (e.g., Gamble et al., 1993b, 1994). Gamble et al. (1993b, 1994) have
commented on the contrasting relative abundances of volcanic rocks between the oceanic
and offshore continental sectors, with the former dominated by basalt and basaltic andesite
and the latter by a spectrum of compositions extending from basalt and andesite to dacite
and rhyolite. In this section we summarize briefly the major geochemical features of
basaltic volcanism along and orthogonal to the arc front in the "oceanic" and "continen-
tal" settings. Three main geochemical provinces are defined by (1) the oceanic Kermadec
arc, (2) the oceanic Havre Trough backarc complex, and (3) the offshore continental arc
(defined by the N gatoro ridge-Whakatane volcano) and the continental backarc region.
The following discussion concentrates on the basalts as they are the magma type common
to all the provinces.
6.2. Kermadec Arc

Representative chemical analyses of basaltic rocks from the northern subaerial Ker-
madec arc volcanoes of Raoul, Macauley, Curtis, and L'Esperance and the southern
submarine volcanoes of Rumble II, III, and IV at the southern termination of the oceanic
segment are presented in Table I. The total alkalis + silica diagram (Fig. 2.11a; LeBas et aI.,
1986) and plot of Ti02 versus Nap (Fig. 2.11b) serve to identify some principal chemical
differences between basalts from the two sectors. For example, Kermadec arc basalts show
low total alkali contents (Fig. 2.11a) overlapping with the basalt fields from the offshore and
onshore TVZ, which are very similar and extend to higher-alkali contents, particularly
NazO (Fig. 2.11b). Incompatible multielement plots for representative Kermadec arc basalts
(Fig. 2.12a,b), normalized to MORB (Pearce, 1983), bear the distinctive hallmark of
subduction-related magmas with high-alkali earth and metal and depleted high-field-
strength (HFS) element abundances relative to MORB.
The Sr- and Nd-isotope measurements on Kermadec arc basalts (Ewart and Hawkes-
worth, 1987; Gamble et al., 1993a, 1994, unpublished data), shown on a standard covaria-
tion plot (Fig. 2.13), define two clearly discernible fields for the respective northern and
southern segments of the Kermadec arc. The two fields have overlapping Nd isotopes but
distinctive Sr isotopes. A single basalt from the flanks of the Rumble IV volcano is
exceptional, having a Sr-isotope signature lower than other southern Kermadec basalts and
being more characteristic of northern Kermadec lavas.
6.3. Southern Havre Trough-Ngatoro Rift System

Chemical analyses of basaltic rocks from the SHT are contained in Table II. On the
total alkalis + silica diagram (Fig. 2.11a) basalts from the SHT show higher total alkalis,
largely a reflection of their higher NazO, and define a vertical array at more or less constant
Si02 content. Similarly, these rocks show higher Ti0 2 and NazO contents (Fig. 2.11b) than
basalts from the Kermadec arc and both the onshore and offshore segments of the TVZ. On
incompatible multielement plots, basalts with >7% MgO (Fig. 2.12c) show subtle enrich-
ments of the alkali earths and alkali metals (note the vertical scale), together with HFS
elements Ce, P, and Zr relative to MORB.
A significant aspect of SHT backarc volcanism is the marked spatial asymmetry and
heterogeneity of the erupted basalts, as evinced by the Ngatoro rift graben lavas (Gamble
TABLE I
Representative Chemical Analyses of Basalts from the Kermadec Arca
Sample no. AU23396 AU14782 AU37546 AU37548 AUl4837

Field no. KA4 KA2 KA 10 KA 11 KA 15
Locality Raoul I Herald I Macauley I Macauley I L' Esperance
Si02 50.78 48.76 48.36 48.39 51.64

Ti02 0.77 0.79 0.52 0.52 1.01
AIP3 14.01 17.75 17.37 17.34 18.18
Fe20 3 1.35 1.39 1.20 1.20 1.57
FeO 9.01 9.24 8.98 7.98 10.46
MnO 0.25 0.19 0.17 0.18 0.24
MgO 11.98 7.33 8.06 8.01 4.97
CaO 9.64 11.44 14.19 14.15 9.27
Nap 1.32 1.75 1.25 1.23 1.33
K20 0.17 0.12 0.16 0.16 0.30
PP5 0.12 0.08 0.16 0.04 0.06
LO! 0.50 1.28 0.04 0.12 0.47

Total 99.90 100.12 100.46 99.32 99.47
Mg# 0.72 0.61 0.67 0.66 0.48
Sc 37 47 44 44 39
V 275 355 300 297 388
Cr 25 90 155 151 17
Ni 28 41 50 54 19
Cu 130 150 120 121 47
Zn 88 92 65 66 109
Ga 16 16 15 15 17
Rb 4 2 2 2 8
Sr 173 161 194 193 220
Y 24 19 11 11 21
Zr 32 30 20 18 37
Nb 2 1
Ba 105 33 48 47 115
La 3 4.00 7.50 7.60 3.30
Ce 7 2.93 3.85 6.19 9.69
Pb 2 2 2 3
Th 0.5 0.8 0.84 0.4
U 0.3 0.45 0.46 0.2
87Srf8 6Sr 0.703452 ± 14 0.703405 ± 15 0.703446 ± 14 0.703444 ± 20 0.703936 ± 18
143Ndl l44Nd 0.513029 ± 8 0.513055 ± 10 0.513030 ± 6 0.513020 ± 6 0.512986 ± 5
eps Nd +7.4 +7.8 +7.4 +7.2 +6.5
TABLE I
( Continued)
AU36982 AU36986 VUW11600 VUW11606 VUW11607 VUW11595

KA 16 KA 17 X 16211 X16811A X16811B X 16911
Rumble II Rumble III Rumble IV Rumble IV Rumble IV Rumble IV
50.83 51.69 50.88 52.03 50.25 49.64

0.61 0.71 0.87 0.82 0.65 0.82
13.70 15.19 18.30 16.56 15.40 16.96
1.22 1.32 1.15 0.97 1.11 0.95
8.13 8.81 7.68 6.46 7.37 6.34
0.18 0.20 0.17 0.16 0.14 0.14
9.22 6.47 5.08 7.27 8.31 9.10
12.97 11.08 11.12 10.55 12.52 12.22
1.45 2.42 2.62 2.96 2.11 2.78
0.53 0.47 0.45 0.59 0.45 0.25
0.15 0.11 0.11 0.20 0.06 0.09
0.26 0.25 0.77 1.00 1.14 0.00

99.25 98.72 99.20 99.57 99.51 99.29
0.69 0.59 0.54 0.67 0.67 0.72
41 37 30 29 38 30
315 325 300 290 278 196
327 172 64 24 355 345
69 50 32 23 90 125
133 142 90 118 95 66
69 90 76 96 68 55
12 15 19 17 14 15
9 6 8 9 8 7
353 229 288 263 239 222
15 21 20 20 14 17
32 48 57 62 40 59
2 2 2
263 230 187 206 235 123
7.45 6.71 5 4 3 5
15.90 12.40 16 18 12 15
2 2 7 6 5 5
1.29 0.70 3
0.45 0.42
0.703966 ± 17 0.704090 ± 13 0.703976 ± 12 0.704251 ± 13 0.704270 ± 13 0.703348 ± 11

0.512919 ± 5 0.513032 ± 5 0.513000 ±6 0.513014 ± 3 0.512989 ± 12 0.513050 ± 7
+5.2 +7.4 +7.1 +7.3 +6.8 +8.0
aAll major element and trace element XRF Analyses were done at the Analytical Facility of Victoria University, Wellington
(Gamble et al., 1993a, b, 1994). Isotope analyses were done in the Research School of Earth Science, Australian National
University. Details of the analytical methods used are in Gamble et al. (1993a). La, Ce, Th, and U values quoted at less than whole
number significance were determined by INNA (data from Gamble et aI., 1993a). Dashes mean element concentration was less
than detection limit.
7.-------------~--~-M-~--.~--_----~
_.-dadte )
dad_ and rhyoll... •
io
6 Ruapehu
trend (Graham .. _ )
N BuIn ~/
5 ..~-= T[OU8h •
N
~
+
oN 3 1VZ_..
z"
(Gamble et aI. 1990)
2
I
~ 64 Kermadec Me _ ..
~ (Gambleetall990)
a
0
40 50 60 70
- - Si02-----+
FIGURE 2.11. Offshore Bay of Plenty rocks com-

pared to TVZ and Kermadecs. (a). Total alkali-silica
r...
(TAS diagram, Le Bas et ai., 1986) for volcanic rocks
from TVZ, Kermadec arc, Havre Trough, Ngatoro
Basin (SHT). The field of Ruapehu basalts-dacites
.
0
Z
(from Graham and Hackett, 1987) is shown for com-
parison. Filled squares: Bay of Plenty offshore basalts
I
and andesites to rhyolites. Open squares: Ngatoro ba-
sin basalts. Large cross: Havre Trough basalt. Open
triangle: Kermadec arc basalts. Small crosses: TVZ
Kermadec AIc basalts. (b). Na~p- Ti02 variation diagram for basalts
b from TVZ, Kermadec arc, Ngatoro Basin, Havre
0
0 Trough and offshore Bay of Plenty. Symbols as in
- - Ti02 ----+ (a), above.
et aI., 1993b; Wright, 1993a). Basalts with higher Ti0 2 and Nap contents are restricted to
the western escarpment of the Ngatoro rift (that farthest away from the subduction and
volcanic front), whereas basalts from the eastern escarpment are geochemically indis-
tinguishable from the Kermadec arc lavas (Fig. 2.14). This heterogeneity occurs over a
length scale of 15-20 krn (the width of the rift graben) and appears to have a nearly perfect
compositional separation in this best-surveyed segment of the SHT.
Sr- and Nd-isotope measurements of basalts from the SHT, including the western
Ngatoro rift lavas (Gamble et aI., 1993a, 1994; Fig. 2.13), are similar both to many MORB
basalts (Ito et aI., 1987) and backarc basalts (BABB) from the southern Lau Basin (Hergt
and Hawkesworth, 1994). The western SHT rift basalts are appreciably less radiogenic
(higher Nd- and lower Sr-isotope ratios) than basalts from the Kermadec arc or TVZ. The
only exception is the basalt from the floor of the southern Ngatoro rift, which plots in the
TVZ array, consistent with its transitional setting in the attenuated continental crust of
the propagating reentrant.
6.4 Offshore TVZ Volcanic Front and Backarc

Basalts from the Ngatoro ridge-Whakatane volcano section of the offshore TVZ
segment (Table III) are very similar to onshore TVZ basalts with considerable overlap (Fig.
2.11). Incompatible multielement plots show the distinctive signature of subduction-related
magmas (Fig. 2.12d,e).
'00 '00 o KA-4 Raoul

o 18111
• KA-2 Herald
• 16211 o KA-'O McAuley
• KA-,5 Espenonc
'0
• KA-'7_11
'0
0.'
a b
0.0' -'---,---,....,,.,.--,----,-,,---.-,---,---r-r-r-,-,....,--,--
'0 '00 o 191/3

o 153/1
• 204118
o 190/1
.201/2
'0
• 19611A
6 196119
0.5
0.'
c d
0.' --' -"-,-,."-,-,--.,-,-,-,--,-,-,-,-,.,,---.,--.,--.,-,--,-,-,-,-,."---.,-
'0 '00 o TVZl13

• TVZl14
o TVlJ'5
• TVZ/4
'0
0.5
f
0.' -L-,---,,---..--,---r-r-r.-..-r-r-r--,-rll'--r- 0.' -'--,---,..-,--..---,---r-r-r.--..-r-r-r--,-,-,-,--
FIGURE 2_12_ MORB nonnalized multielement diagrams for basalts from tbe Kennadec-New Zealand subduc-
tion system. Nonnalizing values from Pearce (1983). (a) Basalts from tbe vicinity of Rumble IV volcano at tbe
southern end of the Kennadec arc. (b) Basalts from Raoul Island, Herald Island, Macauley Island, L'Esperance
rock in tbe northern Kennadec Island arc and Rumble II toward the soutb. (c) Basalts from tbe Ngatoro Basin, note
scale of vertical axis. (d) Basalts from Ngatoro Ridge area, offshore Bay of Plenty. (e) Basalts from Aldennan
Trough, offshore Bay of Plenty, note scale of vertical axis. (f) Basalts from Taupo Volcanic Zone, New Zealand.
All analyses from Gamble et al. (1993a, 1994).
Isotope ratios of Sr and Nd are distinct from those of the Kermadec arc and the Havre
Trough backarc rift system, showing higher-Sr isotopes and lower-Nd isotopes_ Further-
more, for a given Nd-isotope ratio the offshore basalts are generally more radiogenic in Sr
than the onshore TVZ or Taranaki basalts_
Further, andesites, dacites, and rhyolites are appreciably more abundant from the
offshore TVZ region than from the oceanic sector to the north. Offshore andesites to
rhyolites overlap with the onshore Ruapehu andesite-rhyolite lineage (Graham and Hack-
ett, 1987) and are broadly similar to the general onshore TVZ andesite-dacite suite (Cole,
1979), with the exception that none of the felsic volcanics sampled in the offshore region
approach the high Si02 contents (>75%) of younger «20 ka) rhyolites from onshore TVZ.
0.51315
0.51310
adec Arc
0.51305
0.51300
It3N d/1"Nd ocean
MORB
0.51295
Basin
0.51290
0.51285 Taupo Volcanic Zone
0.51280
Taranaki
o
0.51275
0.7025 0.703 0.7035 0.704 0.7045 0.705 0.7055
87Sr f6Sr
FIGURE 2.13. Sr- and Nd-isotope covariation diagram for basalts from the Tonga-New Zealand convergent
plate boundary. Open triangles: Ngatoro and Havre Trough basalts, note how sample 185/1 falls in the field ofTVZ
basalts. Open circles: Kermadec arc basalts, subdivided into northern and southern Kermadecs. Open diamonds:
TVZ basalts, RB is basalt from Ruapehu. Filled diamonds: basalts from Ngatoro Ridge, offshore Bay of Plenty.
Crosses: Taranaki (Egmont) young basalts (Price et aI., 1992). Fields for Pacific MORB, Indian Ocean MORB,
Lau Basin basalts, and Tofua arc lavas are from Ito et al. (1987), Klein et al. (1991), and Hergt and Hawkesworth
(1994).
7. PETROGENESIS
7.1. General
Subduction zones mark the sites of recycling lithosphere into the deep earth and as
such have been influential in the development of models of plate tectonics, crustal growth
(Taylor and McLennan, 1985), and fluid transfer into the convecting mantle. Indeed, this
latter process has been the focus of much recent geochemical work on volcanic arcs (e.g.,
Gill, 1981; Ellam and Hawkesworth, 1988; Morris et at., 1990; McCulloch and Gamble,
1991; Hawkesworth et at., 1993). Geochemical studies of volcanic rocks from the Tonga-
Kermadec island arc system have been important in evolving models of arc magma
petrogenesis (e.g., Ewart et at., 1977; Ewart and Hawkesworth, 1987). Extension of this
simple oceanic island arc system southward into the continental TVZ in New Zealand
introduces the additional quanta of continental crust and further concomitant complexities
to unravel (Wilson, 1989).
The coupling of backarc basins to volcanic arcs was first recognized in the margins of
the Pacific Ocean basin (Karig, 1974). Petrological study of these regions of extension
recognized the broadly MORB-like geochemical characteristics of the basaltic rocks (e.g.,
Hawkins, 1976; Saunders and Tarney, 1979, 1991; Hawkins and Melchior, 1985; Sinton and
Fryer, 1987). More recent studies have furnished an abundance of new geochemical data
from the western Pacific backarc basins (e.g., Hochstaedter et aI., 1990a, b; Loock et ai.,
1990; Hergt and Hawkesworth, 1994; Hergt and Farley, 1994). From these data, models
have begun to emerge linking both arc and backarc basin magma sources (e.g., Woodhead
et ai., 1993; Gamble et aI., 1994) into a dynamic model of melt generation and extraction.
7.2. Taupo-Havre-Lau System

Whole rock and (where available) coexisting glass data for basalts from TVZ and the
Kermadec arc (Gamble et aI., 1990, 1993a), Taranaki (Price et aI., 1992), offshore TVZ,
southern Havre Trough (PPTUW/5) (Gamble et aI., 1990; 1993b, 1994), and the Lau Basin
(Sunkel, 1990) are shown in Fig. 2.15 in a CaO-MgO-Alz03-Si02 (CMAS) pseudoter-
nary projection from plagioclase (Walker et aI., 1979). The majority of TVZ, Kermadec arc,
and Lau Basin basalts cluster toward the evolved (hypersthene-quartz normative) end ofthe
l-atm cotectic as does the evolved basalt from the floor ofthe southern Ngatoro rift. Many
of these lavas are strongly porphyritic, and their phase relations have been complicated by
processes of mixing, assimilation, and storage in crustal magma chambers (cf. Gamble
et al., 1990). Young basalts from Egmont volcano (Price et aI., 1992) and basalts from the
Havre Trough-Ngatoro Basin and offshore TVZ define parallel lineages from mildly
nepheline normative compositions toward the l-atm cotectic. Possible interpretations of
these data are that they reflect either (a) different degrees of melting of comparable sources
or (b) melting of progressively more refractory sources (higher normative olivine). In each
case the primary magmas evolve into the l-atm cotectic by olivine + plagioclase ±
clinopyroxene fractionation. The implication here is that the trajectories are source related
and that the bulk of basalts plotting in the shaded region (Fig. 2.15) have their primary
characteristics obliterated by secondary modification effects such as crystal fractionation
and mixing.
In an attempt to see through these effects Gamble et ai. (1993a) used plots of
incompatible HFS elements and ratios (e.g., Ti/Zr versus Zr); these elements are relatively
insoluble in aqueous fluids or silicic melts, which are commonly thought to be slab derived,
yet sensitive to mantle wedge source mineralogy and degree of melting. This approach was
later amplified and extended to volcanic arc-backarc basin systems in the circum-Pacific
region (Woodhead et ai., 1993). The observed systematic relationships between HFS
elements were modeled in terms of multistage melting of a primary MORB source. It was
concluded that ratios such as TilZr and VITi in backarc basin basalts could be produced by
partial melting of a source similar to that of MORB, but that many arc basalts required a
source with higher Ti/Zr and VITi, which could be produced by prior melt extraction in the
backarc basin regime. This model, moreover, was in keeping with current geodynamic
models of mantle flow into the wedge comer region (e.g., Spiegelman and McKenzie,
1987).
The Sr- and Nd-isotope data provide additional clues to significant magma source
heterogeneity both along and across the 3000 km of the Tonga-Kermadec-New Zealand
arc system. Fields for arc and backarc basalts from the entire region, together with fields for
Pacific and Indian Ocean MORB, are given in Fig. 2.13. We note the following points:
(1) Basalts from the southern Havre Trough and western Ngatoro rift are broadly MORB-
like, span the fields of Pacific and Indian Ocean MORB, and, in general, are similar to
TABLE /I
Representative Chemical Analyses of Basalts from the Southern Havre TroughD
Sample no. VUW11574 VUW11578 VUW11579 VUWl1580 VUW11581

Field no. X153/1 X 15411 X154/5 X15811 X15812
Locality Ngatoro Basin Ngatoro Basin Ngatoro Basin Ngatoro Basin Ngatoro Basin
Si02 49.99 50.47 50.43 51.05 50.19

Ti0 2 1.14 1.66 1.62 1.43 0.91
AIP3 16.52 16.27 16.18 16.10 16.86
Fe 20 3 1.02 1.19 1.19 1.14 0.93
FeO 6.76 7.90 7.95 7.58 6.23
MnO 0.15 0.18 0.18 0.16 0.13
MgO 8.06 6.89 7.06 6.92 7.66
CaO 10.60 9.31 9.36 9.45 12.61
Nap 3.33 4.39 4.25 3.73 2.93
K20 0.40 0.52 0.47 0.51 0.24
P 20 5 0.13 0.18 0.18 0.15 0.07
LOI 1.19 0.28 0.50 1.13 0.58
Total 99.29 99.24 99.37 99.35 99.34
Mg# 68.00 60.90 61.30 61.90 68.70
Sc 33 29 31 30 34
V 234 286 282 257 209
Cr 363 209 217 293 280
Ni 128 90 96 100 86
Cu 52 43 45 49 74
Zn 67 79 78 81 62
Ga 17 20 20 19 17
Rb 7 7 6 7 4
Sr 217 213 210 173 190
Y 25 33 34 32 19
Zr 97 119 119 110 60
Nb 3 3 3 4 1
Ba 82 95 88 63 58
La 4 6 5 5 4
Ce 16 23 28 17 10
Pb 5 4 5 5 6
Th
U
87Srf8 6Sr 0.703090 ± 14 0.703103 ± 14 0.702917 ± 10
143ND/I44Nd 0.513076 ± 7 0.513086 ± 8 0.513105 ± 7
eps Nd +8.5 +8.7 +9.1
basalts from the Lau Basin to the north (Hergt and Hawkesworth, 1994). (2) Basalts from
the northern and southern sectors of the Kermadec arc define two distinctive arrays
distinguished by generally higher Sr in the southern Kermadec arc. The data field for the
active Tofua arc of the Tonga-Lau Basin arc-backarc system (Ewart and Hawkesworth,
1987) spans both the northern and southern Kermadec fields. Perhaps a more detailed study
of the isotope systematics ofthis arc segment would tum up distinctive fields as recognized
along the Kermadecs. (3) Basalts from the southern Kermadec arc overlap with primitive
TABLE /I
(Continued)
VUW11582 VUW11583 VUW11599 VUW11616 VUW33441

X158/3 X158/4 X160/l X185/1 PPTUW/5
Ngatoro Basin Ngatoro Basin Ngatoro Basin Indentor Havre T
51.70 50.74 50.09 51.51 50.66

1.45 1.45 1.45 0.92 1.49
16.05 16.15 15.92 17.68 16.09
1.12 1.14 1.21 1.14 1.11
7.45 7.59 8.07 7.87 7.37
0.16 0.16 0.17 0.16 0.16
7.29 7.03 7.04 4.78 7.49
9.69 9.42 9.95 10.83 10.26
3.76 3.75 3.35 2.81 3.95
0.41 0.42 0.56 0.54 0.29
0.15 0.15 0.16 0.12 0.18
0.16 1.35 1.35 0.71 0.54
99.39 99.35 99.32 99.07 99.59
63.30 62.30 60.90 52.00 66.00
28 30 32 32 34
248 255 311 209 246
280 300 310 222 280
103 104 120 65 97
40 48 60 64 58
77 80 79 113 78
18 19 18 17 16
5 6 10 10 3
168 172 209 263 174
30 31 30 22 35
106 107 116 113 127
3 2 4 4 2
64 66 75 196 23
5 5 5 11 6.4
17 17 22 25 21.4
5 6 5 8 3
2 2 2 0.72
0.41
0.702922 ± 15 0.703052 ± 10 0.704267 ± 14 0.702556 ± 30
0.513084 ± 26 0.513097 ± 6 0.512845 ± 2 0.513129±8
+8.7 +9.0 +4.0 +9.3
aAnalytical details are outlined in Table I. Details of sample locations are contained in Gamble et af. (1993a.b).
basalts from the TVZ, and the latter define an array toward higher-Sr and lower-Nd
isotopes, which Gamble et al. (1993a) interpreted in terms of contamination due to
interaction with New Zealand continental lithosphere. (4) Recent data (Gamble et al., 1994)
on basalts from offshore TVZ are displaced to the right of the TVZ array, showing higher-
Sr isotopes at equivalent Nd. These analyses are all for samples from the offshore arc front,
delineated by the Ngatoro Ridge-Whakatane volcano lineament. Interestingly, the TVZ
sample most closely approximating these data is a basalt from Ruapehu (RB in Fig. 2.15)
5
• Western rill flank
Backarc
[J Eastern rill flank
•• Basin
4 D. Arc Stratovolcano
• Basalt
i • RmBasalt ~
0
(34'28'8)
• •
-------------------.
3 [J
zIi
[J [J
's.
D.
"Island Arc· [J
2
Basalt
1.8
Backarc
".
1.8
III
Basin
1.4 Basalt
i
-------~:-------------~-------------
1.2
0 1.0
1=
0.8 D. D D D.
D
0.6 "Island Arc"
Basalt FIGURE 2.14. Variation diagram of Ti0 2 and Nap plot-
0.4
4 6 8 10 ted against MgO content for southern Havre Trough-
MgO (wt%) Ngatoro rift basalts.
defining the volcanic front in TVZ. This observation led Gamble et al. (1994) to suggest that
the source region for arc-front lavas might be particularly susceptible to fluxing by slab-
derived fluids as a function of previous melt extraction events. Clearly this observation has
important implications for the evolution of magma sources in arcs with coupled backarc
basins and may be especially important in systems, such as TVZ and southern Havre
Trough, where the subduction zone length scales have compressed the arc-backarc dimen-
sions.
8. SYNTHESIS
Modem geological studies (e.g., Taylor et aI., 1992; Hawkins et aI., 1994) have shown
that backarc basins are the sites of an extremely complex and heterogeneous range of
extensional tectonic and magmatic processes. The SHT, as shown here, is no exception.
Such heterogeneity occurring within a young rifting backarc basin associated with a
tectonically "simple" subduction system is significant.
One of the more notable observations from the SHT is the compositional hetero-
geneity of the rift volcanism over short length scales (Gamble et al., 1993b, 1994; Wright,
1993a). The emplacement of MORB-like or BABB along one rift escarpment while "arc"-
like basalts are emplaced only 15 km away along the opposite rift escarpment, apparently
without transitional compositions, requires special tectonic and magmatic conditions. A
two-stage process is envisaged whereby variable fluxing of variably depleted mantle wedge
sources produces heterogeneous primary melts (Gamble et ai., 1994) which subsequently
ascend and are emplaced into attenuated, subarc crust that is pervasively disrupted by rift
block faults (Gamble et aI., 1993b). These faults apparently control the final emplacement
of the lavas, with the concomitant proximity of markedly different basalt compositions. The
geometry of the SHT subcrustal faults is, however, unknown. The interpretation that the
surficial morphology of the rift system appears to comprise a series of asymmetric half-
grabens (Fig. 2.4a), and crustal studies within other back-arc basins (e.g., Taylor et aI.,
1991), would suggest that the crustal structure beneath the SHT rifts consists of low-angle
detachment faults. However, as noted by Wright (1993a), high-angle normal faults would
more easily penetrate the crust to tap heterogeneous magma melts and then act as conduits
for ascending magma than would low-angle detachment faults. It is possible though that the
SHT rift system has evolved to a stage where high-angle antithetic and synthetic faults are
the dominant tectonic structure and they now control subarc rift magmatism. In the past,
similar compositional heterogeneity recognized within the rock record would most likely
be interpreted as separate "terranes" rather than as constituent margins of the same rift
system.
The association of BABB within the rifting SHT, as in the backarc rifts of the Izu-
Bonin system (Fryer et aI., 1990; Hochstaedter et ai., 1990a), and early history of the Lau
Basin (Hawkins et aI., 1994) demonstrate that MORB-like magmatism can occur within
pre spreading rift grabens of young backarc basins. These data support the contention of
Taylor and Karner (1983) and Fryer et al. (1990) that early models ofthe temporal evolution
of backarc magmatism, whereby BABB occur only in mature spreading backarc basins, are
not universally valid.
A further significant aspect of SHT tectonic and magmatic evolution is the position of
the present southern Kermadec arc front. Two observations are noteworthy. The first is that
the position of the present volcanic front sited within the "backarc" region (Figs. 2.8-2.10)
indicates that the spatial relationship of the arc and backarc is nebulous (Wright, 1994). A
similar indistinct spatial relationship of the arc front and regions of extension and subsi-
dence occurs within the TVZ to the south (Gamble et aI., 1990; Wright, 1992). The second
observation reiterates an earlier conclusion about fine-scale compositional heterogeneity of
the arc-backarc volcanism. The Kermadec arc stratovolcanoes, and concomitant basaltic-
andesite arc volcanism, lie within 30-40 km of the SHT rift graben erupting BABB
(Wright, 1993a; Gamble et aI., 1993b). Again the implications of these observations for
interpreting the tectonic settings and subduction polarity of backarc basins within the rock
record are significant.
The essential observation from the SHT is that the influence of the geometry of
subduction is fundamental. It would appear that spatial parameters of a subduction system
(e.g., the volume of the underlying mantle wedge, distance between the rift axis and arc
front, distance between the arc front and subduction front, rate of subduction, rate of
backarc widening, rate of trench rollback, length and terminal depth of the subducted slab)
are as important, or more important, than the temporal evolution of a backarc basin. When
these parameters are shortened, as in a proximal arc-backarc system like the TVZ and SHT,
their influence may be especially important.
TABLE 11/
Representative Chemical Analyses of Basalts from Offshore TVZa
Sample no. VUW11626 VUWl1622 VUWl1628 VUW 11 628 VUW11630 VUWl1633

Field no. X19l/3 XI9011 X196/1A X1961lB X20112 X204IlB
Locality W. Alderman T Ngatoro R. Ngatoro R. Ngatoro R. Ngatoro R. Ngatoro R.
Si02 49.14 50.55 49.57 49.96 48.41 48.41

Ti02 1.25 0.80 0.99 0.89 0.93 0.87
AI 20 3 18.18 17.77 19.15 19.10 20.01 19.23
Fe20 3 1.46 1.30 1.28 1.31 1.20 1.33
FeO 9.74 8.67 8.51 8.73 8.00 8.86
MnO 0.21 0.19 0.18 0.18 0.15 0.17
MgO 4.83 5.75 3.94 3.91 5.17 5.44
CaO 10.70 11.34 12.05 11.88 12.49 12.34
Nap 2.78 2.33 2.48 2.39 2.26 2.13
~O 0.39 0.45 0.41 0.37 0.37 0.27
P20 S 0.15 om 0.08 0.08 0.10 om
LOI 0.35 0.09 0.19 0.32 0.39 0.19
Total 99.18 99.31 98.83 99.12 99.48 99.31
Mg# 46.9 54.2 45.2 44.4 53.5 52.2
Sc 38 38 47 43 35 41
V 429 315 425 357 320 334
Cr 25 35 15 11 58 17
Ni 12 20 5 4 22 9
Cu 43 102 40 35 72 50
Zn 104 85 108 73 83 87
Ga 21 17 20 19 20 19
Rb 8 11 5 5 7 5
Sr 320 257 320 315 350 276
Y 22 20 19 18 19 17
Zr 57 52 51 50 47 40
Nb I 2 2 3 2
Ba 140 200 229 221 165 168
La 6 4 6 5 7 3
Ce 22 18 20 18 19 17
Pb 5 7 5 6 6 6
Tb 1 4
U 0.4 0.4 0.4 0.4 0.4 0.4
87Srf8 6Sr 0.705128 ±1 0.705128 ±I 0.705132 ±I 0.705297 ±1
2 4 0 4
143Ndl l44 Nd 0.512908 ±9 0.512843 ±7 0.512824 ±3 0.512196 ±7
eps Nd +5.3 +4.0 +3.6 +3.7
TABLE 11/
( Continued)
VUWl1613 VUWl1585 VUW11613 VUWl1627 VUW11652 VUW11653

X 17411 X146/1 X17412 X194/1 A-002 B-002
So Colville So Colville So Colville Raukumara PI VoIcano1og VoIcano1og
48.29 47.05 49.20 53.62 51.33 51.05

0.94 1.05 0.97 0.76 0.81 0.81
16.54 17.89 16.26 17.85 17.90 18.05
1.36 1.20 lAO 1.08 1.10 1.13
9.09 8.06 9.36 7.18 7.34 7.54
0.44 0.20 0.30 0.16 0.15 0.15
6.37 7.51 6.52 4.98 4.97 5.02
11.51 11.66 11.70 9.64 10.35 10.39
2.62 2.71 2.73 2.59 2.28 2.30
0.29 0.31 0.27 0.56 0.82 0.80
0.07 0.18 0.06 0.11 0.25 0.26
1.30 1.61 0.21 0.61 1.34 1048
98.82 99.25 98.98 99.14 99.64 99.98
56.5 6204 5504 55.3 55.7 54.3
43 36 42 31 35 37
333 292 345 264 257 265
58 173 57 45 41 41
83 69 53 16 11 10
139 46 130 35 26 18
102 99 99 79 85 83
18 17 19 18 20 19
6 3 5 11 27 26
259 299 258 252 305 300
14 23 16 19 23 22
40 75 38 76 94 93
4 0 3 4 4
126 95 133 279 312 293
2 10 6 6 10 11
13 35 12 15 25 24
7 8 5 7 9 7
1 2 5 5
0.3 0.3 004 0.2 0.2 0.2
"Analytical details are outlined in Table I. Details of sample locations are contained in Gamble et al. (l993a,b).
FIGURE 2.15. CaO-MgO-AI 20 3-Si02

(CMAS) pseudo-ternary projection from pla-
gioclase onto the plane Di-Ol-Qtz (Walker et
aI., 1979) for basalts and coexisting glasses
(where available) from the Kermadec-New
Zealand arc-backarc system. Cotectics for 1
atm, 15 kb, and 20 kb are shown. TD =thermal
divide on l-atm cotectic. Fields for Egmont
lavas (Price et al., 1992), offshore TVZ, Lau
Basin (Sunkel, 1990), and the majority of Ker-
madec arc and TVZ basalts are indicated.
Acknowledgments
Thanks are due to the officers and crew of the RV Rapuhia, RRS Charles Darwin, and
RV Akademik Lavrenteyev for their forbearance while at sea. Special thanks are due to the
Australian Overseas Telecommunications Commission (OTC) for allowing use of the
PacRimEast SYS09 cable route survey data. ICW thanks the Royal Society (UK) and the
Royal Society of New Zealand for financial assistance toward study leave at IOSDL (UK),
where much of the manuscript was completed. I. Smith, J. Baker, R. Christie, and J.
Woodhead provided much useful discussion. K. Majorhazi and P. Bennett produced the
figures. This work was funded by New Zealand Foundation Research Science Technology
contracts 93-VIC-8308 (JAG) and 93-WAR-32-384 (ICW).
REFERENCES
Blackmore, N. A., and Wright, I. C. 1995. Southern Kermadec Volcanoes. 1:400,000, New Zealand Oceanographic
Institute, Miscellaneous Ser. No. 71, National Institute of Water and Atmospheric Research, Wellington.
Brothers, R. N. 1967. Andesite from Rumble II volcano, Kermadec Ridge, southwest Pacific, Bull. Volcanol.
31:17-19.
Caress, D. W. 1991. Structural trends and backarc extension in the Havre Trough, Geophys. Res. Lett. 18:853-856.
Cole, J. W. 1978. Tectonic setting of Mayor Island volcano (Note), N.Z 1. Geol. Geophys. 21:645-647.
Cole, J. W.1979. Structure, petrology and genesis of Cenozoic volcanism, Taupo volcanic zone, New Zealand-a
review, N.Z 1. Geol. Geophys. 22:631-657.
Cole, J. W. 1986. Distribution and tectonic setting of late Cenozoic volcanism in New Zealand. in: Late Cenozoic
Volcanism in New Zealand (I. E. M. Smith, ed.), Roy. Soc. N.Z Bull. 23:7-20.
Cole,1. w., and Lewis, K. B. 1981. Evolution of the Taupo-Hikurangi subduction system, Tectonophysics 72:1-21.
Darby, D. 1., and Williams, R. 0. 1991. A new geodetic estimate of deformation in the central volcanic region of
the North Island, New Zealand, N.Z 1. Geol. Geophys. 34:127-136.
Davey, F. 1. 1982. The structure of the South Fiji Basin, Tectonophysics 87:185-241.
Davey, F. J., and Robinson, A. G. 1978. Cook (1st ed.) Magnetic Total Force Anomaly Map Oceanic Series
1:1,000,000, New Zealand Department of Scientific and Industrial Research, Wellington.
de Mets, C., Gordon, R. G., and Stein, S. 1990. Current plate motions, Geophys. 1. Int. 101:425-478.
Dupont, 1.1988. The Tonga and Kermadec ridges, in The Ocean Basins and Margins, Vol. 7B, The Pacific Ocean
(A. E. M. Nairn, E G. Stehli, and S. Uyeda, eds.), pp. 375-409, Plenum Press, New York.
Ellam, R. M., and Hawkesworth, C. 1. 1988. Elemental and isotope variations in subduction related basalts:
evidence for a three component model, Contrib. Mineral. Petrol. 98:72-80.
Ewart, A. 1968. Geochemistry of the pantellerites of Mayor Island, New Zealand, Contrib. Mineral. Petrol.
17:116-140.
Ewart, A., Brothers, R. N. and Mateen, A. 1977. An outline of the geology and geochemistry and the possible
petrogenetic evolution of the volcanic rocks of the Tonga-Kermadec-New Zealand island arc, J. Volcanol.
Ewart, A., and Hawkesworth, C. 1.1987. The Pleistocene to Recent Tonga-Kermadec arc lavas: interpretation of
new isotope and rare earth data in terms of a depleted source model, J. Petrol. 28:495-530.
Fryer, P., Taylor, B., Langmuir, C. H., and Hochstaedter, A. G. 1990. Petrology and geochemistry oflavas from the
Sumisu and Torishima backarc rifts, Earth Planet. Sci. Lett. 100:161-178.
Gamble,1. A., Smith, I. E. M., Graham, I. 1., Kokelaar, B. P., Cole, 1. w., Houghton, B. E, and Wilson, C. 1. N.
1990. The petrology, phase relations and tectonic setting of basalts from the Taupo volcanic zone,
New Zealand, and the Kermadec Island arc-Havre Trough, S.W. Pacific, J. Volcanol. Geotherm. Res. 43:
235-270.
Gamble, 1. A., Smith, I. E. M., McCulloch, M. T., Graham, I. 1., and Kokelaar, B. P., 1993a. The geochemistry and
petrogenesis of basalts from the Taupo volcanic zone and Kermadec Island arc, S.W. Pacific, J. Volcanol.
Gamble,1. A., Wright, I. c., and Baker, 1. A. 1993b. Seafloor geology and petrology ofthe oceanic to continental
transition zone of the Kermadec-Havre-Taupo Volcanic arc system, New Zealand, N.z. J. Geol. Geophys.
36:417-435.
Gamble, 1. A., Wright, I. c., Woodhead, 1. D., and McCulloch, M. T. 1994, Arc and backarc geochemistry in the
southern Kermadec arc-Ngatoro Basin and offshore Taupo volcanic zone, in Volcanism Associated with
Extension at Consuming Plate Margins (1. L. Smellie, ed.), pp. 193-212, Geol. Soc. Lond. Spec. Publ. 81.
Gill, J. W. 1981. Orogenic Andesites and Plate Tectonics, Berlin, Springer-Verlag.
Gillies, P. N., and Davey, E 1.1986. Seismic reflection and refraction studies of the Raukumara forearc basin, New
Zealand, N.Z. J. Geol. Geophys. 29:391-403.
Graham, I. 1., and Cole, 1. W. 1991. Petrogenesis of andesites and dacites of White Island volcano, Bay of Plenty,
New Zealand, in the light of new geochemical and isotopic data, N.z. J. Geol. Geophys. 34:303-315.
Graham, I. J., and Hackett, W. R. 1987. Petrology of calc-alkaline lavas from Ruapehu volcano and related vents,
Taupo volcanic zone, New Zealand, J. Petrol. 28:531-567.
Hatherton, T. H. 1980. Shallow seismicity in New Zealand 1956-75, J. Roy. Soc. N.z. 10:19-25.
Hawkesworth, C. 1., Gallagher, K., Hergt, 1. M., and McDermott, E 1993. Trace element fractionation processes in
the generation of island arc basalts, Phil. Trans. R. Soc. London, A 342:179-191.
Hawkins, 1. W. 1976. Petrology and geochemistry of basaltic rocks of the Lau Basin, Earth Planet. Sci. Lett.
28:283-297.
Hawkins, 1. W., and Melchior, 1. T. 1985. Petrology ofthe Mariana Trough and Lau Basin basalts, J. Geophys. Res.
90:11,431-11,468.
Hawkins, J. W., Parson, L. M., Allan, 1., et al. 1994. Introduction to the Scientific Results of Leg 135: Lau Basin-
Tonga Ridge Drilling Transect Proc. ODP, Sci. Results, 135, Ocean Drilling Program, College Station, TX,
pp.3-5.
Hergt, 1. M. and Farley, K. 1994. Major, trace element, and isotope (Pb, Sr and Nd) variations in site 834 basalts:
Implications for the initiation of backarc opening, in Proc. ODp, Sci. Results, 135 (1. W. Hawkins, L. M.
Parson, J. Allan, et aI., eds.), pp. 471-485, Ocean Drilling Program, College Station, TX.
Hergt, 1. M. and Hawkesworth, C. 1. 1994. The Pb, Sr and Nd isotopic evolution of the Lau Basin: Implications for
mantle dynamics during backarc opening, in Proc. ODp, Sci. Results, 135 (1. W. Hawkins, L. M. Parson, 1.
Allan et aI., eds.), pp. 505-517, Ocean Drilling Program, College Station, TX.
Hochstaedter, A. G., Gill, 1. B., Kusakabe, M., Newman, S., Pringle, M., Taylor, 8., and Fryer, P. 1990a. Volcanism
in the Sumisu Rift. I: Major element, volatile, and stable isotope geochemistry, Earth Planet. Sci. Lett.
100:179-194.
Hochstaedter, A. G., Gill, 1. B., and Morris, 1. D. 1990b. Volcanism in the Sumisu Rift. II: Subduction and non-
subduction related components, Earth Planet. Sci. Lett. 100:283-297.
Houghton, B. E, Weaver, S. D., Wilson, C. 1. N., and Lanphere, M. A. 1992. Evolution of a Quaternary peralkaline
volcano: Mayor Island, New Zealand, J. Volcanol. Geotherm. Res. 51:217-236.
Ito, E., White, W. M., and Copel, C. 1987. The 0, Sr, Nd, and Pb isotope composition of MORB, Chem Geol.
62:157-176.
Karig, D. E. 1970a. Ridges and basins of the Tonga-Kermadec Island arc system, 1. Geophys. Res. 75:239-254.
Karig, D. E. 1970b. Kermadec arc-New Zealand tectonic confluence, N.Z 1. Geol. Geophys. 13:21-29.
Karig, D. E. 1974. Eolution of arc systems in the western Pacific, Ann. Rev. Earth Planet. Sci. 2:51-75.
Kibblewhite, A. c., and Denham, R. N. 1967. The bathymetry and total magnetic field of the South Kermadec
Ridge seamounts, N.Z 1. Sci. 10:53-67.
Klaus, A., Taylor, B., Moore, G. F., Murakami, F., and Okamura, Y. 1992. Back-arc rifting in the Izu-Bonin island
arc: Structural evolution of Hachijo and Aoga rifts, The Island Arc 1:16-31.
Klein, E. M., Langmuir, C. H., and Staudigel, H. 1991. Geochemistry of basalts from the southeast Indian Ridge, 1.
Geophys. Res. 92:8089-8115.
Korsch, R. 1. and Wellman, H. W. 1988. The geological evolution of New Zealand and the New Zealand region, in
The Ocean Basins and Margins, Vol. 7B, The Pacific Ocean (A. E. M. Nairn, F. G. Stehli, and S. Uyeda, eds.),
pp. 411-482, Plenum Press, New York.
Latter, 1. H., and Hall, L. 1986. Rumble III volcano, Bull. Global Volcano Network 11(7):15.
Le Bas, M. J., Le Maitre, R. W., Streckeisen, A., Zanettin, B.1986. A classification of igneous rocks based upon
the alkali-silica diagram, 1. Petrol. 27:745-750.
Lewis, K. B., and Pantin, H. M. 1984. Intersection of a marginal basin with a continent: Structure and sediments
of the Bay of Plenty, New Zealand, in Marginal Basin Geology: Volcanic and Associated Sedimentary and
Tectonic Processes in Modern and Ancient Marginal Basins (B. P. Kokelaar and M. F. Howells, eds.), pp.
121-135, Geol. Soc. Lond. Spec. Publ. 16.
Loock, G., McDonough, W. L., Goldstein, S. L. and Hofmann, A. W. 1990. Isotopic compositions of volcanic
glasses from the Lau Basin, Mar. Mining 9:235-245.
Malahoff, A., Feden, R. H., and Fleming, H. S. 1982. Magnetic anomalies and tectonic fabric of marginal basins
north of New Zealand, 1. Geophys. Res. 87:4109-4125.
McCulloch, M. T. and Gamble, 1. A. 1991. Geochemical and geodynamical constraints on subduction zone
magmatism, Earth Planet. Sci. Lett. 102:358-374.
Minster, J. B., and Jordan, T. H. 1978. Present day plate motions, 1. Geophys. Res. 83:5331-5354.
Morris, 1. D., Leeman, W. P., and Tera, F. 1990. The subducted component in island arc lavas: Constraints from Be
isotopes and B-Be systematics, Nature 344:31-36.
Packham, G. H., and Falvey, D. A. 1971. An hypothesis for the formation of marginal seas in the western Pacific,
Parson, L. M., and Hawkins, J. W. 1994. Two-stage ridge propagation and the geological history of the Lau backarc
basin, in Proc. ODp, Sci. Results, 135 (1. W. Hawkins, L. M. Parson, J. Allan et al., eds.), pp. 819-828, Ocean
Drilling Program, College Station, TX.
Parson, L. M., Pearce, 1. A., Murton, B. 1., Hodkinson, R. A, Bloomer, S., Ernewein, M., Hugget, Q. J., Miller, S.,
Johnson, L., Rodda, P., and Helu, S. 1990. Role of ridge jumps and ridge propagation in the tectonic evolution
of the Lau Basin backarc basin, southwest Pacific, Geology 18:470-473.
Pearce, J. A. 1983. Role of sub-continental lithosphere in magma genesis at active continental margins, in
Continental Basalts and Mantle Xenoliths (c. 1. Hawkesworth and M. J. Norry, eds.), pp. 231-249, Shiva,
Nantwich.
Pelletier, B. 1990. Tectonic erosion, accretion, backarc extension and slab length along the Kermadec subduction
zone (Abstract), p. 65, Fifth Circum-Pacific Energy and Mineral Resources Conf, Honolulu.
Pelletier, B., and Dupont, J. 1990. Tectonic erosion, accretion, extension arriere-arc et longuer du plan de
subduction Ie long de la marge active des Kermadec, Pacifique Sud-Ouest, C.R. Acad. Sci. Paris
310(11):11657 -11664.
Pelletier, B., and Louat, R. 1989. Seismotectonics and present-day relative plate motions in the Tonga-Lau and
Kermadec-Havre region, Tectonophysics 165:237-250.
Price, R. c., McCulloch, M. T., Smith, I. E. M., and Stewart, R. B. 1992. Pb-Nd-Sr isotopic compositions and trace
element characteristics of young volcanic rocks from Egmont volcano, New Zealand and comparisons with
basalts and andesites from the Taupo volcanic zone, Geochim. Cosmochim. Acta 56:941-953.
Reyners, M. 1989. New Zealand seismicity 1964-1987: An interpretation, N.Z 1. Geol. Geophys. 32:307-315.
Saunders, A. D., and Tarney, J. 1979. The geochemistry of basalts from a backarc spreading centre in the east
Scotia Sea, Geochim. Cosmochim. Acta 43:555-572.
Saunders, A. D. and Tarney, 1. 1991. Back-arc basins, in Oceanic Basalts (P. A. Floyd, ed.), pp. 219-263, Blackie,
Glasgow and London.
Sclater,1. G., Hawkins, 1. W., Mammerickx, J., and Chase, C. G. 1972. Crustal extension between the Tonga and
Lau ridges: petrologic and geophysical evidence, Geol. Soc. Am. Bull. 83:505-518.
Sclater, J. G., Jaupart, c., and Galson, D. 1980. The heat flow through oceanic and continental crust and heat loss of
the earth, Rev. Geophys. Space Phys. 18:261-312.
Seafloor Surveys International. 1990. PacRimEast Submarine Telecommunications Cable Route Survey, Auck-
land to Oahu, Vol. I (Charts 1-23), Seattle.
Seafloor Surveys International. 1993. PacRimEast Route Diversion Survey. Detailed Survey Charts of the
Seafloor, Vol. II (Charts 21-40), Seattle.
Sinton, J., and Fryer, P. 1987. Mariana Trough lavas from 18°N: Implications for the origin of backarc basin
basalts, J. Geophys. Res. 92:12,782-802.
Smith, I. E. M., and Brothers, R. N. 1988. Petrology of Rumble seamounts, southern Kermadec Ridge, southwest
Pacific, Bull. Volcanol. 50:139-147.
Spiegelman, M., and McKenzie, D. P. 1987. Simple 2-D models for melt extraction at mid-ocean ridges and island
arcs, Earth Planet. Sci. Lett. 83:137-152.
Stem, T. A. 1987. Asymmetric backarc spreading, heat flux and structure associated with the central volcanic
region of New Zealand, Earth Planet. Sci. Lett. 85:265-276.
Sunkel, G. 1990. Origin of petrological and geochemical variations of Lau Basin lavas (SW Pacific), Mar. Mining
9:205-234.
Sykes, L. R., Issacks, B., and Oliver, J. 1969. Spatial distribution of deep and shallow earthquakes of small
magnitude in the Fiji-Tonga region, Bull. Seismol. Soc. Am. 59:1093-1113.
Taylor, B. 1992. Rifting and volcanic-tectonic evolution of the Izu-Bonin-Mariana arc. in Proc. ODp, Sci.
Results, 126 (B. Taylor, K. Fujioka, et aI., eds.), pp. 627-651, Ocean Drilling Program, College Station, TX.
Taylor, B., Brown, G. R., Fryer, P., Gill, J., Hochstaedter, E, Hotta, H., Langmuir, c., Leinen, M., Nishimura, A.,
and Urabe, T. 1990. Alvin-SeaBeam studies of the Sumisu rift, Izu-Bonin arc, Earth Planet. Sci. Lett.
100:127-147.
Taylor, B., Fujioka, K., et al. 1992. Proc. ODP, Sci. Results, 126, Ocean Drilling Program, College Station, TX.
Taylor, B., and Kamer, G. D. 1983. On the evolution of marginal basins, Rev. Geophys. Space Phys. 21:1727-1741.
Taylor, B., Klaus, A., Brown, G. R., and Moore, G. E 1991. Structural development of Sumisu Rift, Izu-Bonin arc,
J. Geophys. Res. 96:16,113-16,129.
Taylor, S. R., and McLennan S. M. 1985. The Continental Crust: Its Composition and Evolution, Blackwell
Scientific, Oxford.
van der Linden, W. J. M. 1967. Structural relationships in the Tasman Sea and southwest Pacific Ocean, N.Z. J.
Geol. Geophys. 10:1250-1301.
Walcott, R. I. 1984. The kinematics of the plate boundary zone through New Zealand: A comparison of short and
long-term deformations, Geophys. J. Roy. Astron. Soc. 79:613-633.
Walker, D., Shibata, T., and De Long, S. E. 1979. Abyssal tholeiites from the Oceanographer Fracture Zone. II:
phase equilibria amd mixing, Contrib. Mineral. Petrol. 70:111-125.
Watanabe, T., Langseth, M. G., and Anderson, R. N., 1977. Heatflow in backarc basins of the western Pacific, in
Island Arcs, Deep Sea Trenches and Back-Arc Basins (M. Talwani and W. C. Pitman III, eds.), Vol. I, pp.
137-167, Maurice Ewing Series, American Geophysical Union, Washington, DC.
Weidicke, M., and Collier, J. 1993. Morphology of the Valu Fa spreading ridge in the southern Lau Basin, J.
Geophys. Res. 98:11,769-11,782.
Whiteford, P. C. 1990. Heat flow measurements in the Bay of Plenty, New Zealand, Report of the February 1988
Vulkanolog Cruise, Research Report No. 221, Geophysics Division, New Zealand Department of Scientific
and Industrial Research, Wellington.
Whiteford, P. C. 1992. Heat flow measurements in and near the Ngatoro Basin, Bay of Plenty, New Zealand,
Report of the February 1990 Vulkanolog Cruise, Research Report No. 240, Geophysics Division, New
Zealand Department of Scientific and Industrial Research. Wellington.
Wilson, M. 1989. Igneous Petrogenesis: A Global Tectonic Approach, Unwin Hyman, London.
Woodhead, J. D., Eggins, S., and Gamble, 1. A. 1993. High field strength and transition element systematics in
island arc and back-arc basin basalts: Evidence for multi-stage melt extraction and ultra-depleted mantle
wedge, Earth Planet. Sci. Lett. 114:491-504.
Wright, I. C. 1990. Bay of Plenty-Southern Havre Trough Physiography, 1:400,000, New Zealand Oceanographic
Institute Misc. Ser. No. 68, New Zealand Department of Scientific and Industrial Research, Wellington.
Wright, I. C. 1992. Shallow structure and active tectonism of an offshore continental spreading system: The Taupo
volcanic zone, New Zealand, Mar. Geol. 103:287-309.
Wright, I. C. 1993a. Pre-spread rifting and heterogeneous volcanism in the southern Havre Trough backarc basin,
Mar. Geo/. 113:179-200.
Wright, I. C. 1993b. Southern Havre Trough-Bay of Plenty (New Zealand): Structure and seismic stratigraphy of
an active backarc basin complex, in South Pacific Sedimentary Basins. 2. Sedimentary Basins of the World (P.
F. Ballance, ed.), pp. 195-211, Elsevier, Amsterdam.
Wright, I. C. 1994. Nature and tectonic setting of the southern Kermadec submarine arc volcanoes: An overview,
Mar. Geo/. 118:217-236.
Wright, I. c., Carter, L., and Lewis, K. B. 1990. GLORIA survey of the oceanic-continental transition of the
Havre-Taupo backarc basin, Geo-Marine Lett. 10:59-67.
3
The Geology of the Lau Basin

James VII. Hawkins, Jr.
ABSTRACT
The Lau Basin comprises oceanic crust that separates the remnant Lau Ridge volcanic arc
from the active Tofua arc. The basin is situated above mantle exhibiting strong seismic
wave attenuation; it overlies the west dipping seismic zone of the Tonga Trench subduction
system and is presently opening at rates as high as 1.6 cm/yr. Seaward rollback of the trench
axis, coupled with upwelling high temperature mantle diapirs, is proposed as the main
driving force to cause crustal extension in this region of oceanic plate convergence. The
trapezoidal shape of the basin suggests that it has opened progressively from north to south;
the trace of the Louisville seamount chain may help constrain the southern apex of the
basin. The basin has opened in two tectonic styles. Initially, beginning at about 6 Ma,
attenuation and rifting of the forearc or outer part of the Lau Ridge formed a series of half-
grabens that were partly sedimented and received basaltic flows. This extension and
magmatism overlapped with Lau Ridge arc volcanism. The second stage of opening (about
5.5-5 Ma) was promoted by a southward-propagating rift system that formed new crust by
seafloor spreading. A second propagator (about 1.5 Ma) has overtaken the first and forms an
overlapping ridge system. A small three-limbed spreading system in the northeastern basin
forms a triple junction. A fourth spreading system is recognized in the northwestern part
of the basin.
In spite of the suprasubduction zone (SSZ) setting and the proximity to volcanic arcs,
the basin is dominated by mid-ocean ridge basalt (MORB)-like crust. Isotope data indicate
that both Pacific and Indian MORB-source mantle have been involved; Pacific source
formed some of the oldest crust, the Indian source feeds the modem spreading ridges.
Helium-isotope data suggest a Samoan "plume" component is important in generating
seamounts of the northern part of the basin. Lau Basin crust exhibits a general depletion in
high-field-strength elements, and some samples show varied enrichment in large-ionic-
radius lithophile elements, but the source must be like the MORB-source in most respects.
However, mixing of MORB-like and arc1ike magmas has been important in the basin's
history. Locally, crust is transitional to arc compositions but this occurs largely in areas of
older crust or where rifting first encountered older crust/mantle. Heterogeneity of the SSZ
James W. Hawkins, Jr. • Geological Research Division, Scripps Institution of Oceanography, La Jolla,
California 92093-0220.
63
64 JAMES W. HAWKINS, JR.
mantle is indicated by the presence of intrabasinal arc-composition volcanoes erupted

nearby to ridges where MORB crust was forming. The data suggest that a wave of arc-
composition volcanism migrated across the basin as it opened; presently it forms the Tofua
arc, which is a relatively young feature. The variety of mantle sources involved, as
expressed in the range of chemical signatures in the crustal rocks, points to the complexity
of backarc magmas. They reflect the complexity of SSZ mantle. No simple definition of
backarc basin basalt is possible.
1. BACKARC BASINS
The western Pacific Ocean basin is rimmed by an array of island-arc systems and their
related trenches. Alfred Wegener (1929) proposed that
the island arcs, and particularly the eastern Asiatic ones, are marginal chains which were detached
from continental masses, when the latter drifted westwards and remained fast in the old seafloor,
which was solidified to great depths. Between the arcs and the continental margins later, still-liquid
areas of seafloor were exposed as windows.
These "windows" of seafloor are the backarc basins. Most of them are relatively shallow
(typically less than 3000 m deep) regions of ocean crust that are considerably younger than
the seafloor being subducted at the adjacent oceanic trenches. They are situated between
inactive, remnant volcanic arcs and the active volcanic arcs that form in response to the
subduction of oceanic lithosphere. The Lau Basin (Fig. 3.1) is a well-studied, classic
example.
This discussion summarizes the geology of the Lau Basin as an interim report on our
understanding of its evolution. The discussion focuses mainly on the tectonic setting and
the petrology of Lau Basin crust and draws on the wealth of information on these topics that
has been acquired in the last 25 years. Major new insights have been gained as a result of
geological long-range inclined asdic (GLORIA) imagery (Parson et aI., 1990) and data
from six holes drilled in the western part of the Lau Basin on Ocean Drilling Project (ODP)
Leg 135 (Parson et aI., 1992; Hawkins et aI., 1994).
Several of our new findings may be summarized as follows:
1. The Lau Basin has fonned by two tectonic styles of opening. Initially the western part of the
basin fonned by crustal extension and rifting-not by seafloor spreading. The extension was
accompanied by magmatism that partly filled rift basins.
2. The second phase of opening was by seafloor spreading. Spreading developed on propagating
rifts that started from a transfonn fault boundary.
3. Basaltic magmatism in the rift basins and on the propagating rifts was contemporary with arc
activity on the (present) remnant arc.
4. A triple junction is fonning new crust in the northeastern part of the basin.
5. As the basin opened by extension, arc-composition seamounts fonned within the basin close
by to MORB-like eruptions.
6. The basin is in a suprasubduction zone setting, yet the new crust is dominated by MORB-like
magma compositions. Both an Indian mantle MORB source and a Pacific mantle MORB
source are recognized with isotope data.
These new views on the geological evolution of the Lau Basin are useful in under-
standing other (SSZ) convergent plate margins with arc-backarc basin systems. Our better
understanding of the petrologic and tectonic complexity of these systems also gives insight
GEOLOGY OF THE LAU BASIN 65
FIGURE 3.1. Lau Basin and Tonga Trench arc system showing locations of Leg 135 Sites 834-841 and Deep Sea
Drilling Project (DSDP) Site 203. Major features shown include sites of modem volcanism in the Lau Basin:
Mangatolu triple junction (MTJ) (Hawkins, 1989; Nilsson et aI., 1989; Nilsson, 1993); central Lau spreading center
(CLSC), eastern Lau spreading center (ELSC), relay zone (RZ), and Valu Fa Ridge (VF) (Parson et aI., 1990; von
Stackelberg and von Rad, 1990). Islands and shoals are Ata (A), Capricorn seamount (C), Donna seamount (D),
'Eua (E), Lakemba (L), Metis Shoal (M), Niuataputopu (N), Niuafo'ou (NF), Rochambeau Bank (R), Tongatapu
(T), Tafahi (TA), Upolu, Western Samoa (U), Vava'u (V), Zephyr Shoal (Z). Contour interval in kilometers.
to the type of oceanic lithosphere we find preserved in ophiolites; namely, most of them
probably come from SSZ settings.
Although backarc-arc-forearc-trench systems are situated in zones of lithosphere
shortening, between plates with opposing relative motion, abundant evidence exists that
these systems are loci of crustal extension and formation of new crust as illustrated in Fig.
3.2. Ocean Drilling Program (ODP) studies have provided good evidence that the "new"
crust of backarc basins has formed contemporaneously with the volcanic island arcs that
bound them (Hawkins et aI., 1991; Fryer and Pearce, 1992; Taylor, 1992; Taylor et al., 1992;
Hawkins et aI., 1994). Seafloor spreading has been an important factor in their evolution,
but lithosphere extension and rifting has also been of major importance (Hawkins, 1994;
Parson and Hawkins, 1994). Magmas that form the volcanic arcs and the crust of backarc
basins form in SSZ settings as a result of fractional melting of upwelling mantle diapirs.
These melts are largely derived from the mantle wedge lying above the subduction zone-
66 JAMES W HAWKINS, JR.
~------------------.~ CONVERGENCE
EXTENSION ---
- TRENCH ROLLBACK
MAGMA TYPES SSZ MAGMATISM COMPONENTS

FA = Forearc Boninite (early) 1. Multiply depleted mantla
Arc Tholeiite (later) 2, Subducted oceanic lithosphere
ARC = Arc Tholeiitic Series 3, Dehydration fluids
SASS = Arc·like to MORB·like Ba.ans 4. SubduCled sediments
5. Counterflow 01 "enriched" mantie
FIGURE 3.2. Schematic cross section of an intraoceanic convergent margin showing the remnant are, backarc
basin, active are, and forearc (FA). Sites of magmatic activity are shown by arrows and shaded crescentic patterns.
Numbers refer to potential components that may contribute to SSZ magmatism.
that is, the SSZ mantle (Pearce et aI., 1984). Additional contributions may come from the
counterflow of mantle into the SSZ region and from the subducted lithosphere plate (e.g.,
Tatsumi et al., 1986; Takazawa et aI., 1992). The result is new crust having varied
enrichments in low-partition-coefficient elements relative to crust formed at oceanic
spreading centers.
More than 40 years after Wegener's (1929) classic paper, several nearly contemporary
papers by Karig (1970, 1971), Packham and Falvey (1971), Sleep and Toksoz (1971), and
Moberly (1972) presented the first discussions of the geometry and tectonic setting of
western Pacific backarc basin-arc-trench systems. The authors proposed that new crust had
been formed in a zone of extension between an inactive remnant arc and an active volcanic
island arc; a kinematic relation between the subduction process and the backarc basin
extension was postulated, but neither the mechanism nor the evolution were well under-
stood.
Various attempts to model and explain backarc extension have been made. These may
be broadly separated into models invoking mantle diapirism, induced aesthenospheric
convection, and global plate kinematics. Taylor and Kamer (1983) discussed these general
models and concluded that none were adequate. Chase (1978) proposed that motion of the
upper plate away from the trench resulted in backarc extension, whereas Hynes and Mott
(1985) attributed extension to seaward migration of the subducted plate. Sleep and Toksoz
(1971) called on induced asthenospheric convection above the subduction zone. Karig
(1971) proposed mantle diapirism triggered by the subducted slab. Rollback of the trench
axis (Elsasser, 1971) is rejected by some authors (e.g., Hynes and Mott, 1985) but is used to
explain the observations by others (e.g., Chase, 1978; Carlson and Melia, 1984).
The geologic evidence shows a relation between subduction, crustal extension, and the
production of backarc magmas. The MORB-like nature of backarc magmagenesis requires
mantle upwelling under the backarc basins. The upwelling and lithosphere extension are
either a cause or the result of seaward rollback of the trench (Elsasser, 1971; Uyeda and
Kanamori, 1979); the Tonga Trench may be migrating eastward at up to 10 crn/yr (Carlson
and Melia, 1984). Furlong et al. (1982) proposed that an increase in subduction velocity
may promote backarc spreading. Ribe (1989) presented a model showing how backarc
spreading, regardless of how it started, would induce mantle flow above the subduction
zone. He concluded that this helps account for the observed distribution of distinctive
magma types.
In this discussion I will use a model in which crustal extension above the subduction
zone is a consequence of trench rollback coupled with mantle diapirism induced by
subduction. I propose that mantle counterflow above the subducting lithosphere plate and
concurrent mantle upwelling probably are the main driving forces behind backarc basin
evolution (e.g., Hawkins et aI., 1984). Unloading of the crust by extension and thinning
promotes the rise and decompression melting of mantle diapirs (e.g., McKenzie and Bickel,
1988).
2. GEOLOGIC SETTING OF THE LAU BASIN

Intraoceanic convergent plate margins are dynamic systems that evolve by rifting,
extension, and magmatism above the subduction zone. The geology of all parts of the
system needs to be understood. The Lau Basin is but one element in the convergent margin
system that includes the inactive Lau Ridge remnant volcanic arc and the Tonga Ridge
(Figs. 3.1 and 3.2). A brief summary of the geology of these crustal segments follows.
2.1. Tonga Ridge

The Tonga Ridge extends for more than 1100 km; its southern end joins the Kermadec
Ridge at the mutual intersection with the Louisville seamount chain. The Tonga Ridge
comprises two belts of seamounts, shoals, atoll reefs, and islands that separate the 2- to
3-km-deep Lau Basin from the 1O.5-km-deep Tonga Trench. The western belt is formed of
seamounts, shoals, and volcanic islands that constitute the Tofua arc. The arc has been
active historically, with many submarine eruptions having been reported within the last 50
years. Rocks of the Tofua arc constitute an arc tholeiitic series that ranges in composition
from basalt to low-K rhyolite; the main magma types are basalt and basaltic andesite (Bryan
et at., 1972; Ewart and Bryan, 1972; Ewart et aI., 1977; Bryan, 1979a; Ewart, 1979; Gill,
1981; Ewart and Hawkesworth, 1987). The age of intitiation of the Tofua arc is poorly
constrained and may have varied along strike with oldest edifices occurring at the north
end. For example, a 3-Ma age has been reported for Niuatopatapu Island by Tappin et al.
(1994), but there is no evidence for similar ages at the south end. It is not likely that the
present distribution of arc volcanoes existed prior to about 1 Ma (see discussion in
Hawkins, 1994). There is a broad compositional similarity to the main phases of volcanism
on the Lau Ridge, but, as we will discuss, the Tofua arc appears to have a more significant
contribution of a subduction component than is seen on the Lau Ridge.
The eastern belt, or Tonga platform, comprises uplifted blocks of Tertiary platform
carbonates that overlie a crystalline basement formed of middle Eocene to late Miocene
arc-composition volcanic and plutonic rocks. Quaternary reef limestones form a cap on the
platform (Scholl et aI., 1985). Depths increase from south to north on the Tonga platform.
Numerous cross faults with differential uplift give rise to three major morphologic seg-
ments. The northern segment is deeper (e.g., north of Vava'u depths range from 1000 to
1500 m), whereas the central and southern segments are from 500 to 1000 m deep, with
many emergent banks and islands. The chain of active volcanoes of the Tofua arc is aligned
along the crest of the northern segment, whereas it is offset to the west in the region south
of Vava'u.
The oldest rocks on the platform are exposed on the island of 'Eua, where there are
beach boulders of arc-tholeiitic-composition hypersthene gabbro which have been dated as
46 to 40 Ma (Duncan et at., 1985). These are overlain by late-middle Eocene calcareous
conglomerates and breccias (Cunningham and Anscombe, 1985). 'Eua also has arc tholei-
itic andesitic lava flows that give ages of 33-31 Ma and andesitic to silicic andesite dikes
with ages of 19-17 Ma (Duncan et aI., 1985; Hawkins and Falvey, 1985). These crystalline
rocks are interpreted as part of the Lau Ridge that was rifted away as the Lau Basin opened.
Drill core material (e.g., exploration wells, Exon et at., 1985; and ODP Site 840, Parson
et at. 1992), dredged samples (Exon et at., 1985; Stevenson, 1985), and island exposures
(Cunningham and Anscombe, 1985) show that the upper levels of the platform consist
largely of gravity flow deposits formed of volcaniclastic turbidites and debris flows of sand
and gravel. These were derived from a Miocene age volcanic arc; probably it was the Lau
Ridge, but Cawood (1985) points out that volcanic rocks exposed on the Lau Ridge are not
compositionally equivalent to the clastic rocks of equivalent age found on the Tonga Ridge.
Along-strike compositional zonation may be an explanation for this, or the sediment
sources may be submerged in the Lau Basin. The clastic rocks are interbedded with thin
intervals of pelagic, or hemipelagic, nannofossil chalks and marlstones (Clift and Dixon,
1994; Ledbetter and Haggerty, 1994). The clastic rocks show an overall fining upward in the
late Miocene prior to beginning of Lau Basin rifting. The late Pliocene-Pleistocene beds
are a carbonate-rich sequence with minor interbeds of ash and volcanic sands (Cunningham
and Anscombe, 1985; Exon et at., 1985; Clift and Dixon, 1994).
Drill core data from ODP Site 841 on the Tonga Ridge (Fig. 3.1) show that the Miocene
sediments are faulted against volcanic rocks comprising a low-K rhyolitic edifice formed of
flows, welded and nonwelded tuffs, and tuff breccias that must have been erupted in
shallow water or subaerially (Bloomer et at., 1994). Radiometric dates (44 ± 2 Ma) and
paleontologic ages for overlying sediments indicate a late-middle Eocene age (McDougall,
1994). Also found at Site 841 are arc tholeiitic basalt and basaltic andesite dikes or sills
which have intruded distal facies of upper Miocene (foraminifer Zone N16 to Subzone
N17a) volcaniclastic turbidites (Parson et at., 1992). These intrusives, and the turbidites,
were formed in the forearc to the Lau Ridge volcanic arc prior to opening of the Lau Basin.
2.2. Tonga Trench

The deepest levels of the Tonga Ridge are exposed on the wall of the Tonga Trench,
where there is a cross section of oceanic crust that is largely formed of arc-related mafic and
ultramafic rocks. Rocks dredged from the inner slope of the trench include highly depleted
serpentinite, dunite, harzburgite, clinopyroxenite, gabbro, diabase, arc tholeiitic basalts,
andesite, boninite, quartz diorite, tuffs, and volcanic breccia. Petrologic studies of this rock
assemblage indicate that it is largely derived from a volcanic arc (Fisher and Engel, 1969;
Hawkins et at., 1972; Vallier et at., 1985; Bloomer and Fisher, 1987; Hawkins, 1988). Some
rocks indicative of the accretion of seafloor material have been dredged as well, but the bulk
of the material suggests that the inner slope exposes the lower levels of an island arc as
originally proposed by Fisher and Engel (1969). Bloomer and Fisher (1987) proposed a
4000-m stratigraphic reconstruction, based on depths of dredge hauls, that establishes an
arclike crustal section capped by mafic and intermediate composition volcanic rocks.
2.3. Lau Ridge

The Lau Ridge remnant arc forms the western boundary of the Lau Basin. It comprises
islands and atoll reefs as well as submerged conical features assumed to be volcanic
seamounts; the emergent Lau Island Group forms a chain about 500 km long and up to
80 km wide. Barrier reefs, fringing reefs, or both, commonly surround the islands. The Lau
Ridge and the submarine extension to the south, the Colville Ridge, form a 2400-km-long
remnant arc. Lau Ridge volcanism began at least by mid-Miocene time, 14 Ma, and was
active until early Pliocene time, 1.5-2.5 Ma (Gill, 1976; Cole et at., 1985; Woodhall, 1985).
Basaltic andesite and andesite, including both tholeiitic and calc-alkaline low- to high-K
magma series, are the main rock types preserved on the Lau Ridge. Lesser amounts of
basalt, dacite, and rhyolite are also present as well as intrusive rocks compositionally
similar to the volcanic rocks (Woodhall, 1985). The Lau Ridge magma series differs from
broadly similar rocks of the Tofua arc in having lower ratios of 87Sr/86Sr (e.g., 0.7030-
0.7033 for the Lau Ridge (Gill, 1976), and 0.70361-0.70399 for the Tofua arc (Ewart and
Hawkesworth, 1987)) that are interpreted as representing less of a subduction component
in the source for Lau Ridge arc magmas.
2.4. Lau Basin-An Overview

The Lau Basin (Fig. 3.1) is a trapezoidal-shaped backarc basin that separates the
inactive Lau Ridge remnant volcanic arc from the Tonga Ridge (Karig, 1970; Hawkins,
1974). The shape of the basin suggests that it has opened more widely at its north end,
perhaps because opening began there earlier and has progressed southward in the basin. At
present, the southern end apparently is hinged where the Louisville seamount chain
intersects the Tonga Trench. Opening of the Lau Basin, as a consequence of crustal
extension above the Tonga Trench subduction system, is the most recent (i.e., <6 Ma) event
in a long sequence of crustal extension episodes in the southwestern Pacific that may be
traced back to late Cretaceous time and the initial breakup of the eastern Australian
continental margin (Kroenke, 1984; Hawkins, 1994).
The Lau Basin has many geologic characteristics that are typical of other western
Pacific backarc basins, but it also has several features that may be unique. Like most
backarc basins, the Lau Basin is situated above a well-defined Wadati-Benioff zone that
marks the location of the subducted Pacific plate. The inclined seismic zone reaches depths
on the order of 700 km to the west under the North Fiji Basin (Fig. 3.3; Isacks and
Barazangi, 1977; Billington, 1980; Giardini and Woodhouse, 1984, 1986; Pelletier and
Louat, 1989). Seafloor magnetic lineations and drill core data indicate that Cretaceous-age
Pacific plate lithosphere is being subducted into the Tonga Trench beneath the Indo-
Australian plate at rates estimated from 10.5 cm/yr (Minster and Jordan, 1978) to as much
as 17.8 cm/yr (Pelletier and Louat, 1989), depending on the latitude. The seismic zone lies
at a depth of about 140 km below the active volcanoes of the Tofua (Tonga) arc and is about
250 km under the central Lau spreading center (CLSC) and eastern Lau spreading center
70 JAMES IN. HAWKINS, JR.
100 0 Kilometers 100
FIGURE 3.3. Composite cross section of earth-

quake hypocenters for the Tonga Trench subduction
zone. The "0 kilometer" mark on the horizontal
axis corresponds to the location of the Tofua arc.
Locations of projections of the trench axis are
shown by the vertical lines. Scale is in kilometers
(Isacks and Barazangi, 1977).
(ELSC) (lsacks and Barazangi, 1977). The upper surface of the seismic zone has been
modeled as an irregular and strongly curved surface (Billington, 1980). It terminates at its
northern end where the Tonga Trench curves sharply to the west and the plate boundary
becomes a transform fault boundary between the Tonga and New Hebrides-Vanuatu
trenches. An inflection in the seismic zone at depths of 525 to 575 km, first identified by
Billington (1980), was interpreted as due to imbrication of the subducted lithosphere by
Louat and Dupont (1982). They estimated that the lithosphere at the depth of this inflection
was subducted about 7 to 8 m.y. ago. This is close to the 6-Ma age estimated for beginning
of opening of the Lau Basin (Parson et aI., 1992); the two events may be related. Interpreta-
tion of the configuration of the Tonga Trench seismic zone suggest a half-spoon shape
sharply curved to the west at its north end and abruptly terminated along the west-trending
part of the Tonga Trench (Fig. 3.4; Billington, 1980). A broad zone of strong seismic wave
attenuation was recognized west of the Tonga Ridge beneath the Lau Basin by Barazangi
and Isacks (1971), who attributed it to high temperature or partial melting in the mantle. The
mantle under the basin has low Q (strong seismic-wave attenuation) and is inferred to be
hotter than the surrounding mantle (Fig. 3.5; Barazangi and Isacks, 1971). The ridge axes of
the modem spreading centers rise to depths as shallow as 2200 m, and the average basin
depth is about 2500 to 3000 m (Hawkins, 1974); this is anomalous because it is the depth
range for most of the ridge crest of mid-ocean ridge system as originally noted by Sclater
et al.(1972). They interpreted this depth anomaly as being related to the relative youth of the
basin and the likelihood of a broad region of upwelling hot mantle beneath it. The Lau
Basin depths are in striking contrast to the Mariana Trough, where depths as great as
4500 m are common near the axial ridge and the crest of this ridge is about 3500 m deep
(e.g., Hawkins et al., 1990). A possible corollary of this is that the Mariana Trough lacks
a similar large region of upwelling hot mantle.
Karig's proposal that the Lau Basin was floored with basalt was substantiated by the
first geologic studies based on dredged samples (Hawkins et aI., 1970; Sclater et al., 1972).
These papers, and subsequent work by Hawkins (1974, 1976), established that there was
young basaltic crust in the Lau Basin and that it showed considerable similarity to MORB.
A range in basaltic types was found, but all were distinctly different from arc tholeiites. The
modem spreading ridge system of the Lau Basin, comprising the central and eastern Lau
100
I
1 I
I
,o0r-
I
>-
FIGURE 3.4. View, looking south, of a grid representing the upper surface of the Wadati-Benioff zone of the
Tonga-Kermadec subduction zone. The projection is slightly distorted in that it does not take into account the
Earth's sphericity (from Billington, 1980).
spreading centers, was discovered, and sampled, on the Scripps Institution of Oceanogra-
phy 7-TOW Expedition in 1970 (Hawkins et al., 1970). Our single-beam profiler survey
lines helped to delineate part of the ridge (Hawkins, 1974) and to recognize symmetric
magnetic anomalies, but the anomalies proved difficult to trace throughout the basin
(Lawver et al., 1976; Lawver and Hawkins, 1978). Initially we all assumed that nearly the
entire width of the Lau Basin had formed by seafloor spreading. Neither the overlapping
spreading centers nor the wedge shape of the magnetic pattern and spreading system was
recognized. Only the presence of fresh pillow lavas and the general trend of the ridge could
be established. Additional data did not help solve the problem. The western part of the basin
lacks a continuous magnetic fabric, and symmetric anomalies cannot be traced beyond the
region near the axial ridges (e.g., Lawver et aI., 1976; Weissel, 1977; Lawver and Hawkins,
Fiji
Islands
0C:::~~~~
FIGURE 3.5 Cross section of the Fiji

Islands to Tonga Trench region show-
ing inferred and extrapolated extent of
the region of high and low attenuation
in the uppermost mantle. Note that the
Lau Basin is underlain by mantle with
extremely low Q (high attenuation)
(from Barazangi and Isacks, 1971).
1978). Interpretations of the age of the earliest backarc crust in the Lau Basin varied from
5 to 10 Ma (Sclater et al., 1972) to 2.5 to 3 Ma (Malahoff et al. 1982). Lawver et al. (1976)
and Lawver and Hawkins (1978) concluded that true seafloor spreading had not been an
important process throughout the entire history of the basin and proposed that some form of
diffuse spreading on short ridge segments may have operated. Thus, it was clear early on
that it was difficult to explain the entire evolution of the basin by seafloor spreading,
although several attempts to do so (e.g., Weissel, 1977) appeared to give satisfactory models
for some of the data.
Our understanding of the petrology of the basin's crust has undergone great change
from an originally simple comparison to "normal" oceanic crust to the much more
complex view now held. The earliest petrologic studies of Lau Basin crust focused on high-
standing features that form the modem axial ridge system and the Peggy Ridge. Petrologic
data for these ridges indicated a close similarity to MORB and led to the interpretation that
petrogenetic process for backarc basins were similar to those forming oceanic crust
(Hawkins et aI., 1970; Sclater et aI., 1972; Hawkins, 1974, 1976; Gill, 1976; Pineau et aI.,
1976; Carlson et aI., 1978). More extensive sampling, including seamounts and high-
standing older crust on the margins of the basin, showed that there were rocks having
compositions transitional to arc chemistry and that the basin crust was zoned with borders
more like arc compositions than MORB (Hawkins and Melchior, 1985). The compositional
variability includes a wide range in rock types from basaltic to andesitic and multiple
mantle source types have been inferred from a number of subsequent studies (e.g., Poreda,
1985; Jenner et aI., 1987; Volpe et aI., 1987, 1988; Hawkins 1989; Hawkins et aI., 1989;
Nilsson et aI., 1989; Boespflug et aI., 1990; Davis et aI., 1990; Frenzel et al., 1990; Loock
et aI., 1990; Sunkel, 1990; Vallier et aI., 1991; Ernewein et aI., 1994). The most recent
insights to the geology of the Lau Basin and its evolution have come from GLORIA
imagery of much of the northern part ofthe basin (Parson et aI., 1990) and drilling on ODP
Leg 135. The drill cores have given samples of hitherto inaccessible crust from the older,
sediment-covered, western part of the basin and from sediment-covered crust near the
modem axial ridges (Parson et aI., 1992; Hawkins et al., 1994).
2.5. The New View of Lau Basin Evolution

Extensive bathymetric surveys with single-beam and multibeam profiling systems,
seismic reflection profiling, shipboard and aeromagnetic surveys, and GLORIA imagery
have given us a good understanding of both the regional bathymetric fabric and the
relations between basin morphology and the magnetic patterns. Additional insight has
come from drilling on ODP Leg 135. Together, these data have helped develop a new view
of Lau Basin evolution that also has implications to other arc-backarc systems. The reason
why the magnetic data have been so difficult to interpret is because they are not all part of a
continuous symmetric system developed by seafloor spreading (Parson et aI., 1989, 1990,
1992; Parson and Hawkins, 1994).
Our new view of Lau Basin evolution is that it has opened in two stages each with its
own tectonic style. Initially, the forearc to the Lau Ridge was stretched and attenuated by
crustal extension and rifting. This formed a basin-range-type structure. Lavas ponded in the
basins gave rise to magnetic anomalies, but the magnetic patterns do not form a continuous
symmetric record because the lavas did not generate new crust at a "fixed" spreading
center. Data from ODP Site 834 demonstrate that rifting started prior to 5.6 Ma; we estimate
that it may have started by about 6 Ma (Hawkins, 1994; Hawkins and Allan, 1994). The
early backarc crustal extension was contemporaneous with volcanism on the Lau Ridge
(e.g., 4.5- to 2.5-Ma basalts of the Korobasaga Group). Whelan et al. (1985) correlate the
Korobasaga Group with an "early rifting stage" that occurred late in the evolution of the
Lau Ridge. The arc rifting and eruption of the Korobasaga Group, which was dominated by
tholeiitic basalt, overlaps with beginning of crustal extension and basaltic volcanism in the
western Lau Basin. This predates the time inferred for the beginning of seafloor spreading.
Geometrically this process occurred in theforearc, and it is only because the modem Tofua
arc subsequently formed to the seaward that the term backarc basin has any meaning.
The second stage of opening involved seafloor spreading that may have started at
about 5.5 to 5 Ma. Parson and Hawkins (1994) propose that this spreading was initiated by
southward propagation of a rift that started on the trace of the Peggy Ridge; this propagator
formed the present ELSe. Thus, the age of crust formed by seafloor spreading is pro-
gressively younger to the south culminating at the Valu Fa Ridge at the south end of the
ELSe. The wedge-shaped age pattern of new seafloor formed by the ELSC is further
complicated by the development of a second propagator that cut through the older (ELSC)
seafloor and formed the present CLSC (Parson et al., 1989, 1990, 1992; Parson and
Hawkins, 1994). The two spreading centers overlap and are separated by complex seafloor
with traces of pseudofaults and an abandoned intermediate ridge or relay zone. Presently,
the Lau Basin appears to be opening in an east-west direction by symmetric seafloor
spreading on the well-defined, but segmented and offset, axial ridges aligned along longi-
tude 176°30' W.
The evolution of the Lau Basin has involved crustal extension and the partial dismem-
bering of the Lau Ridge. The evidence suggests that the initial rift was on the outboard
edge of the former arc or in the forearc (Hawkins, 1994). Other arc-backarc systems (e.g.,
Mariana) may have formed in a similar way although this evolutionary style is not
necesarily true for all arcs. For example, the Sumisu rift has formed on the inboard side
of the Izu arc (Taylor, 1992). There has been a long-standing controversy as to whether or
not backarc and arc magmatism are synchronous. At least for the Lau Basin we have good
evidence that they were synchronous although we have no good constraints as to relative
volumetric importance. The 4.5-Ma to 2.5-Ma basalts of the Korobasaga Group erupted
during the "early rifting stage" in the evolution of the Lau Ridge (Whelan etal.,1985). The
timing of this rifting and associated volcanism, which was dominated by tholeiitic basalt,
broadly overlaps the beginning of crustal extension in the western Lau Basin (Hawkins,
1994). We also have evidence within the basin that a wave of arc composition volcanism
migrated across it as the basin opened (Bednarz and Schmincke, 1994; Clift and Dixon,
1994). Presently, Tofua arc volcanism and backarc spreading are proceeding concurrently.
It is also evident that arc and backarc magmatism operate contemporaneously in the
modem Mariana arc and backarc.
3. LAU BASIN MORPHOLOGIC PROVINCES
3.1. Introduction
The Lau Basin may be subdivided into several morphologic-tectonic provinces that
together constitute the extensional backarc basin (Fig. 3.1). The central part of the basin,
where the ODP Leg 135 drill transect was located, comprises a western extensional basin-
ridge province (here informally called WEB) and a triangular region of young crust that has
formed by seafloor spreading. The latter area includes the two actively spreading ridges-
the CLSC and the ELSC-and the relay zone (RZ) between them that collectively extend
for about 600 kIn. These ridges are not centered in the Lau Basin but are offset toward the
eastern side; much of the southern part of the ELSC lies within 50 kIn of the active Tofua
arc. The Peggy Ridge (Fig. 3.1) separates the northern and southern parts of the Lau Basin.
Peggy Ridge appears to be the location of a transform fault with right-slip displacement.
The northern basin is not well sampled, but GLORIA imagery has helped define a probable
spreading center, the northwestern Lau Spreading Center (NLSC), that trends northeasterly
from the Peggy Ridge. A well-defined ridge-ridge-ridge triple junction has been recog-
nized in the northeastern basin. Originally named Mangatolu triple junction (MTJ) by the
discoverers (Hawkins, 1989; Hawkins et aI., 1989; Nilsson et at., 1989), it is also called the
Kings triple junction. Much of the northern basin is covered by seamounts. One robust
feature forms the island of Niuafo'ou. Only three submerged features have been surveyed
and sampled: Donna Seamount, Rochambeau Bank, and Zephyr Shoal. Each is petrologi-
cally distinct and has a different geologic history.
3.2. Eastern and Central Lau Spreading Centers

At present, new seafloor is being generated at three well-mapped areas of spreading
(Fig. 3.1). These are the ELSC, CLSC, and MTJ. A fourth spreading center, NLSC, has
been postulated but data for it are limited. The main site of seafloor spreading is on the well-
defined axial ridge system that extends for more than 600 kIn in a north-south direction near
176°30'W. This ridge system is segmented in a morphotectonic sense, as well as having a
petrologic segmentation, that allows further subdivision into secondary and tertiary seg-
ments on the scale of 10 s of kIn. It comprises the CLSC, ELSC and a RZ in the overlap
region. Detailed charts and discussions of the morphology of parts of each ridge system are
presented in Weidicke and Habler (1993) and Weidicke and Collier (1993).
3.2.1. Central and Eastern Lau Spreading Centers

The CLSC is shallower than 2300 m throughout its 190-km length, and it is flanked by
2800- to 2900-m-deep basins that parallel the ridge trend. The ridge crest morphology of
the CLSC varies along axis. Parts of it have a narrow axial rift basin with flanking walls that
rise 200 to 500 m above a low relief floor. Some areas are capped with small mounds and
pinnacles. The southern end near 19°22'S is split by a narrow rift, and the ridge narrows and
deepens and ends in an area of rough topography. The ELSC extends for about 400 kIn. Its
crest has inward-facing scarps that define an axial rift valley. Small edifices are found
within the valley, and at least one intrarift seamount has been split into two segments
now on opposite sides of the axis. The floor of this rift shoals from about 3000 m at the
dying northern end to about 2300 m near the southern end of the rifted part of the ridge at
21°S. Farther to the south, the ELSC axial crest lacks a rift and is capped by small elevated
areas. The ELSC ends near 22°50' S on a segment of the ELSC known as the Valu Fa Ridge
(VFR; Scholl et aI., 1985; Morton and Sleep, 1985). The morphology and segmentation of
the VFR have been described by Weidicke and Collier (1993). The VFR lies only 30 to 40
kIn west of the trace of the Tofua arc. The VFR is petrologically distinct from the northern
ELSC and is here treated as a separate major segment.
The ELSC began as a propagating rift at about 5.5 to 5 Ma and was followed by the
CLSC at about 1.5 to 1.2 Ma (Parson and Hawkins, 1994). Both have propagated southward
from the Peggy Ridge. The two ridges form an overlapping system near 19°22'S with the
CLSC propagating southward at the expense of the ELSC. The southern end of the ELSC,
the Valu Fa spreading center (Jenner et al., 1987; Vallier et aI., 1991), appears to be
propagating southward into older crust, but essentially nothing is known about the age or
composition of that older crust. The rift tip of the propagator is near 23°S. Parson and
Hawkins (1994) estimate that the rift tip of the ELSC has propagated southward at about
120 mm/yr.
Presently the northern Lau Basin may be opening at a rate of 120-145 mm/yr on a line
between Lakemba, Fiji, and Vava'u as determined by global positioning system (GPS)
measurements (Bevis et al., 1993; F. W. Taylor, personal commun., 1994). The opening rate
decreases southward toward the VFR. The complex magnetic fabric and the effects of
propagating ridges make it difficult to estimate long-term average spreading rates. How-
ever, it is likely that similar rates of opening, 100-170 mm/yr, may have been in effect south
of the Peggy Ridge since the initiation of spreading on the ELSC (Parson and Hawkins,
1994).
The most extensively studied part of the Lau Basin is along the axial ridge systems.
The ridges comprise a range in rock types that is mainly basaltic with lesser amounts of
fractionated types, including Fe-Ti-enriched basalt, "oceanic andesite," and low-K rhyo-
lite. These are discussed in more detail below.
3.2.2. Mangatolu Triple Junction

The Mangatolu triple junction, in the northeastern part of the Lau Basin (Fig. 3.1) near
15°30'S, 175°45'W, comprises three limbs of a ridge-ridge-ridge triple junction (Hawkins
et al., 1989; Nilsson et al., 1989; Nilsson, 1993). The MTJ is also known as the Kings triple
junction (Falloon et al., 1992). Three well-defined spreading axes have been defined for the
MTJ by bathymetry, GLORIA imagery, magnetic anomalies, and dredged samples. The
west-trending limb breaks up into at least three separate traces that have a poorly developed
ridgelike cross section. The northeasterly trending limb intersects the line of volcanic
edifices that constitutes the Tofua arc. A deep channel, or trough, cuts the Tonga Trench
wall on the projection of the trace of this limb, but we have no data to support a direct
relationship between the two. The southern axial limb has an overlap zone at about 15°53' S.
The southern end is poorly defined by the available bathymetric data but can be traced as
far south as 16°05'S.
The MTJ is in the widest part of the Lau Basin, more than 250 km wide, and has
replaced older crust of uncertain age. This crust may include some of the oldest backarc
basin crust, remnants of the forearc to the Lau Ridge, or trapped old ocean crust. Data
for the seafloor are very limited, but, as discussed in a separate section, both arclike and
MORB-like crust have been found.
3.2.3. Northwestern Lau Spreading Center

A northeasterly trending region of high sonar backscattering, near 177°W on the
Peggy Ridge (Parson et al., 1990; Parson and Tiffin, 1993; Parson and Hawkins, 1994) has
been interpreted as a spreading center. The NLSC appears to trend northeasterly and may
link the Peggy Ridge transform with the major structural discontinuity at the west trending
segment of the Tonga Trench. Parson and Tiffin (1993) note that it is about 40 kIn wide
where it intersects the Peggy Ridge and has the finely lineated GLORIA image typical of
medium and slowly spreading axial ridges. The NLSe has been superposed on the crustal
fabric of older sedimented rift basins. The trend of the NLSe is toward Rochambeau Bank,
a prominent seamount described later. The interpretation of evidence from GLORIA
imagery is supported by fresh basaltic samples from a few dredge sites near the Peggy
Ridge and from seamounts and scarps on the probable trace of the spreading center
(Hawkins, 1988). The NLSe cuts older crust that had a different origin.
3.3. Western Extensional Basin

The broad area south of Peggy Ridge, between the Lau Ridge and the active eLse and
ELSe, is characterized by narrow, partly sedimented, subbasins, elongated in a north to
northeasterly direction. This part of the Lau Basin has rugged topography with up to 1500 m
of relief in contrast to the smoother topography that flanks the two spreading centers. We
described it as having basin and range topography to differentiate it from crust formed at the
spreading centers (Parson et at., 1992) and used the name western extensional basin (WEB)
province. The basins are about 1-15 kIn wide and 10-85 kIn long. The subbasins are
separated by ridges of thinly sedimented crust. Limited data from dredged samples show
that these are mainly basalt, some of it moderately differentiated, having compositions
transitional between MORB and arc material (Hawkins and Melchior, 1985). The age of the
crust forming these ridges is not known, but several of the basins were drilled on ODP Leg
135 and we have good age data for these. There is a general age progression from about
6 Ma on the west to 0.6-0.8 Ma near the present axial ridges (Hawkins, et aI., 1994). The
high-standing blocks between the subbasins are assumed to be older crust. The interpreta-
tion of Leg 135 data suggests that the WEB formed by crustal extension and rifting, not by
seafloor spreading.
3.4. Peggy Ridge and the Northern Basin

The Peggy Ridge (Fig. 3.1) is a high-standing feature (1250 m) of uncertain origin.
Parson and Tiffin (1993) noted that its upper surface is ribbed, irregular, characterized by
clusters of linear peaks and troughs, and appears to be dominated by superimposed
seamounts. The northeast side drops sharply into a 2500-m-deep valley. They interpret the
steep northeast flank of the ridge as a fault. The northeast side of the deep valley rises
abruptly about 1800 m to form a ridge subparallel to the Peggy Ridge. Presently Peggy
Ridge marks the location of many shallow earthquakes that have right-lateral first motions
(Eguchi, 1984; Hamburger and Isacks, 1988). The morphology of the ridge suggests to
Parson and Tiffin (1993) that it may be a leaky transform fault. The ridge may have had a
complex history, and all we can say with any degree of certainty is that its composition
precludes it having been part of an island arc, and regional magnetic patterns make it
unlikely that it was a spreading center (Sclater et aI., 1972; Hawkins, 1974, 1976). At
present it marks the location of a transform fault accommodating differential motion
between the eLSe and the postulated NLse north of Peggy Ridge (Parson et aI., 1990;
Parson and Tiffin, 1993; Parson and Hawkins, 1994). As discussed in a subsequent section,
the Peggy Ridge may have played a major role in the development of the ELSe and eLse
rift propagators. Rocks from the Peggy Ridge are MORB-like basalts that resemble the
rocks of the active ridges.
At least four distinct provinces may be recognized north of the Peggy Ridge. In
addition to the MTJ and the NLSC, there is a broad region north of Peggy Ridge that Parson
and Tiffin (1993) call a "sedimented block terrain." This includes an old ridge partially
buried in sediment and a region of elevated topography that has several large seamounts
and the island of Niuafo'ou. There are several areas of high heat flow (e.g., up to 4.22 HFU;
Sclater et ai., 1972), and the recently active volcanic island of Niuafo'ou (Reay et aI., 1974;
Taylor, 1991). Samples from the seamounts include a wide range of rock types such as
N-MORB, E-MORB, OIB, and dacite.
South of the MTJ and west of the Tonga Ridge, the seafloor is characterized by north-
south-trending ridges that separate sedimented basins. Parson and Tiffin (1993) call this the
"ridge/interridge terrain." The narrow basins have relatively thick sediment fill similar to
those in the WEB (e.g., more than 0.2-s two-way travel time, Hawkins, 1974). Parson and
Tiffin (1993) suggest that this part of the basin may have had an origin by crustal extension
and rifting like that which formed the western part of the Lau Basin.
4. PETROLOGIC DISCUSSION OF THE NEOVOLCANIC ZONES
4.1. SSZ Mantle Influence

Before discussing the petrogenesis of magmas that erupt on the axial ridges, a
digression to comment on the tectonic setting is in order. Magmas of the Lau Basin
neovolcanic zones form by partial melting of upwelling mantle in the SSZ tectonic
environment (Fig. 3.2). The crust flooring subbasins of the WEB also came from this
source. Current ideas about the nature of SSZ mantle are that it has been depleted by
previous melting and selectively reenriched with a subduction component (e.g., Pearce
et aI., 1984; Hawkins, 1994). It is not surprising that many backarc magmas carry the
distinctive compositional signatures of this complex mantle source; that is, they show some
affinity to arc magmas. This is expressed in the major element chemistry of some backarc
basin basalt glasses in their relatively higher Al 20 3 and NazO, but lower FeO* (all iron
determined as FeO) and Ti0 2, when compared to MORB glasses with equivalent levels of
MgO. These characteristics were first pointed out by Fryer et al. (1981) for Mariana Trough
basalts. They are found in most other backarc basins as well, but are by no means typical of
all backarc basalts. Forearc magmas as well as those of island arcs also carry the SSZ
signature. In addition to the different abundances of some major elements, there are several
major distinctions in trace and minor constituents, differing from true MORB, that are
typical of SSZ magma systems (e.g., Fryer et aI., 1981; Briqueu et aI., 1984; Hawkins and
Melchior, 1985; Sinton and Fryer, 1987; Woodhead et ai., 1993). These are (1) a relative
depletion in high-field-strength elements (HFSE), especially Nb and Ta. (2) The relatively
higher Na is paralleled by a tendency for some, but not all, samples to be accompanied by
variably higher K, Rb, Ba, and Sr relative to MORB; Ba in particular is relatively enriched.
(3) The rare-earth element (REE) patterns vary from flat to MORB-like in form although
for some of the least fractionated samples the low abundances indicate a depleted source.
La is variable, but enrichment relative to MORB is common in many SSZ magmas (e.g.,
Gill, 1981). (4) SSZ magmas commonly are depleted relative to MORB in Cr, Ni, and Co.
(5) Backarc samples are generally more hydrous and more oxidized than otherwise compa-
rable MORB. The distinctive differences from MORB in major elements, as well as some
trace elements, are attributed to the more hydrous mantle source of the SSZ environment.
Elevated PH20 delays the onset of plagioclase crystallization and leads to higher Al content
of the melts. Higher oxygen fugacity promotes the crystallization of Fe-Ti oxides as well as
Cr-spinel and causes the relative depletions in Fe, Ti, Cr of SSZ magmas. The field of
olivine stability is expanded, and this causes the relative depletion of Ni in SSZ magmas.
Clinopyroxene is an important liquidus phase and contributes to Co depletion.
Many, but not all, backarc basin basalts show varied effects of the SSZ influence
described. The Lau Basin is no exception, as both MORB-like and arclike compositional
varieties are found. The Mariana Trough axial ridge also displays both MORB-like and arc-
like rock types (Hawkins et aI., 1990). This requires that the SSZ mantle must receive an
input of MORB-source mantle as shown schematically on Fig. 3.2. In the Lau Basin there is
a general pattern of MORB-like basalt being more common on the axial ridges and the
transitional to arc compositions being more common on the eastern, western, and southern
edges of the basin (Fig. 3.6; Hawkins and Melchior, 1985). The data for the edges of the
basin are very limited, and this proposal still needs to be tested with many more samples
than we have. However, the idea appears to be fairly well supported with the limited data
we have. As we discussed, there are some deviations from this general pattern that probably
are related to the extent or longevity of melt production in different areas as well as to
source heterogeneity. The general inference for the Lau Basin is that, in spite of the SSZ
setting, a MORB-source mantle has been the major contributor to the backarc crust, and
there has been restricted mixing with arclike SSZ mantle sources. Some Lau Basin axial
ridge lavas show an imprint of SSZ ("arc") compositional features but many do not.
Compositions very close to MORB are found on all of the active ridges as well as in the drill
core samples. Some of the earliest lavas to form new crust as the Lau Basin opened (ODP
Site 834) are among the most MORB-like (Hergt and Farley, 1994; Hawkins and Allan,
1994). This requires massive upwelling and melting of MORB-source mantle that has
overwhelmed the SSZ signature. This is discussed further in the section presenting a model
for evolution.
4.2. Eastern and Central Lau Spreading Centers

4.2.1. Introduction
The combined ELSC and CLSC extend for about 600 km through the Lau Basin in a
generally north-south direction culminating at the Valu Fa Ridge segment of the ELSe.
There is an extensive data base for the ELSC and CLSC that offers insight to their evolution
and the mantle sources for the melts that formed them (e.g., Hawkins, 1974, 1976, 1977,
1988,1989; Hawkins and Melchior, 1985; Jenner et aI., 1987; Sunkel, 1990; Boespflug et aI.,
1990; Vallier et aI., 1991; Falloon et aI., 1992; Ernewein et aI., 1994). Each spreading ridge
carries its own distinctive isotopic and chemical signature, but all have a compositional
stamp that indicates a predominantly MORB-like character with an SSZ overprint. A
summary of representative data is in Table I and sample locations are in Fig. 3.6. The
modem spreading axes all are situated above the Wadati Benioff zone which lies at about
250 km beneath the ELSC, CLSC, and MTJ, and about 300 to 350 km below the NLSe. As
179°W
15°S
ITS
19°5
21"S
FIGURE 3.6. Locations of dredge sites referred to in text and on tables. Dredge site acronyms are A = ANT,
L = 123 (both from Hawkins and Melchior, 1985); P = PPTU (Hawkins, 1988); R = RNDB (Hawkins, 1989);
E =Emewein et al. (1994); F =Frenzel et al. (\990); S =Sunkel (\990); V =Vallier et al. (\99\). Numbers shown
are the dredge site numbers.
a consequence, a major source of the melts must be in the SSZ mantle, but, as outlined,
there is a general MORB-like character to all of the basaltic samples. Local variations in
chemistry that point to variable affinity to arc-composition magmas are attributed to
variable mixing with a subduction component. Proximity to the trace of the Tofua arc is
suggested as an explanation for the compositional variation seen from north to south along
the ELSe (Emewein et al., 1994).
ODP Site 836 was drilled on some of the oldest ELSe crust formed at this latitude, and
the cores give data for early stages of ELSe magmagenesis on the propagating rift. Valu Fa
Ridge samples also are early ELSe crust formed at the rift tip. They are discussed in a
separate section because of their many distinctive features.
~
TABLE I
Representative Analyses, ELSe, eLse, and Relay Zone
Location ELSe
Site L-74 A-225 R-5 R-6 R-7 R-II E-25 S-54 S-84
Sample 123 ANT RNDB RNDB RNDB RNDB Darwin-3 Sonne Sonne
74-1 GLa 225-1 GLa 5-3 GU 6-1 GLa 7-2 GLa II-I GLa 25-5-1 GC 54-1 KD 84-1
Si0 2 50.24 53.59 51.44 48.37 52.58 50.62 51.3 52.4 65.5
Ti02 0.76 1.05 0.9 0.85 2.11 1.02 0.96 0.7 0.64
Alz03 15.8 14.56 15.44 15.38 13.43 15.18 15.93 15.77 13.4
FeO* 9.13 11.55 9.39 7.95 14.76 9.54 9.28 8.07 6.69
MnO 0.21 0.21 0.19 0.17 0.21 0.18 0.17 0.16 0.16
MgO 7.39 5.27 8.02 6.21 3.99 7.51 6.5 7.78 1.34
CaO 13.37 10.22 12.87 11.01 8.55 12.83 11.29 12.5 4.54
Na20 2.1 2.34 1.91 2.09 3.24 1.9 2.11 1.79 4.2
KzO 0.11 0.17 0.03 0.16 0.14 0.13 0.17 0.2 0.73
P205 0.09 0.11 0.06 0.11 0.2 0.12 0.08 0.08 0.14
Sum 99.2 99.07 100.26 99.21 99.21 99.04 97.79 99.45 97.34
Mg# 62.9 48.3 62.9 61.6 35.6 61.8 58.9 66.4 31.3
CaO/AI20 3 0.85 0.702 0.83 0.72 0.64 0.85 0.71 0.79 0.34
Trace elements (ppm)
Cr 300 50 337 342 116 216 16
Ni 94 33 110 62 8 103 55 69 4
Co 90 43 43 35 38 12
V 282 359 307 277 502 338 376 258 54 $:
Zf 40 45 53 50 127 57 47 36 141
Y 21 26 27 31 55 26 23 21 58 ~
Nb <3 1 0.69 <5 5 ~
Hf 1.16 1.55 1.41 ~
Ta < .05 0.08 0.04
Rb 2 0.5 1 5 4 6 2.8 <3 12 ~
Ba 10.8 32.5 4 18 27 8 40.7 44 120 ffi
Sf 166 98 69 118 95 III 119 116 128
~
(j)
Location ELSC
Site V-I ODP 836 A ODP 836 B ODP 836 B ELSC ELSC ELSC ~0
Sample Lee Unit 3 RK Unit 4 RK Unit 5 RK Mg# > 60 Mg# 50-59 Mg# < 50 (j)
1-5 3H7/54-55 3R2/32-38 7R2/56-62 (Mean 20) (Mean 7) (Mean 2) -<:
Si02 56.69 47.13 49.94 48.08 50.95 52.79 52.79

~
Ti02 1.38 0.69 0.79 0.72 0.95 0.92 1.85 :i!III
A1 20 3 14.57 19.18 16.11 16.21 15.41 15.13 14.08
FeO* 11.51 8.08 8.27 8.13 9.36 10.47 13.62
~
c:
MnO 0.22 0.16 0.12 0.14 0.18 0.19 0.21
7.68 9.07
~
C/)
MgO 3.73 8.88 7.52 6.52 4.28
CaO 7.65 13.03 14.85 13.78 12.64 11.47 8.9 ~
Nap 3.36 1.9 1.96 2.33 2.03 1.92 3.03
~O 0.52 0.01 0.15 0.01 0.08 0.12 0.16
P20 S 0.16 0.14 0.06 0.14 0.08 0.08 0.19
Sum 99.2 99.2 99.93 98.59 99.22 99.61 99.09
Mg# 39.9 69.3 65.6 69.6 62.2 56.1 39.2
CaO/A1 20 3 0.52 0.68 0.92 0.85 0.82 0.758 0.632
Trace elements
Cr 4 318
Ni 25 III 139 120
Co
V 350 216 157
Zr 81 47 53 44
Y 36 19 18 20
Nb 2 2
Hf 2.3
Ta
Rb 8.6 0.1 3 2
Ba 96 17 14 19
Sr 168 140 288 319
(continued)
CO
-
~
TABLE I
( Continued)
Location CLSC
Site A-229 L-79 P-29 P-31 P-33 R-13 R-14 R-25
Sample ANT 123 PPTU PPTU PPTU RNDB RNDB RNDB
229 GL 79-1 GL 29-3 GL 31-1 GL 33-I-GL 13-2 GL 14-1 GL 25-1 GL
Si02 48.11 49.33 72.96 49.51 50.45 60.41 51.28 50.96

Ti02 0.88 1.94 0.34 1.64 1.93 1.62 2.39 1.05
AIP3 17.35 13.93 12.08 16 14.11 12.93 12.63 14.65
FeO* 9.02 13.27 4.56 9.76 13.29 12.16 16.1 10.6
MnO 0.17 0.28 0.13 0.18 0.25 0.22 0.25 0.2
MgO 8.73 6.64 0.12 7.56 5.41 1.99 4.64 7.6
CaO 12.56 11.74 2.21 11.33 10.17 6.15 9.12 12.76
Nap 2.59 2.59 2.23 2.66 2.85 3.91 3.12 2.26
Kp 0.03 0.09 1.18 0.42 0.08 0.42 0.14 0.03
PP5 0.05 0.14 0.02 0.15 0.13 0.32 0.2 0.07
Sum 99.49 99.95 95.83 99.21 98.67 100.13 99.87 100.17
Mg# 67 51.2 5.4 61.4 45.5 25.2 37.1 59.5
CaO/Alp3 0.72 0.84 0.18 0.71 0.72 0.48 0.72 0.87
Cr 198 20 303
Ni 3 68 37 7 18 116
Co 46 44
V 13 345 455 147 469 332 ~
Zr 201 68 128 441 165 45
Y 82 31 59 138 64 30 ~
Nb 2 :::E
Hf 1.84 3.57 1.89 $;
Ta 0.081 0.16 0.04
Rb 31 I 10 3 0.1 ~
Ba 143 7 2 75 31 2 ~
Sr 113 86 114 86 91 49
~
(j)
Location CLSC Relay Zone
Site CLSC CLSC CLSC CLSC CLSC R-12 R-31 R-31
~
r-
Sample RNDB Mg# > 60 Mg# 50-59 Mg# 40-49 Mg# < 40 RNDB RNDB RNDB a(j)
26-2 GL (Mean 5) (Mean 17) (Mean 10) (Mean 7) 12-6 GL 31-1 GL 31-4 GL -<:
a
Si02 50.25 49.15 52.04 50.81 53.76 53 51.44 49.25
Ti0 2 0.94 0.99 1.24 1.94 2.07 0.57 1.21 0.79
"~
IT!
AlP3 15.86 16.97 14.46 13.41 13.01 16.2 15.31 17.36
FeO* 9.79 9.19 10.7 14.1 14.44 7.95 10.61
~
8.65 c:
MnO 0.19 0.18 0.2 0.26 0.24 0.14 0.22 0.17
6.71 5.63 3.82
~
C/)
MgO 8.25 8.48 6.89 6.28 9.1
CaO 13.06 12.14 11.37 10.21 8.34 12.28 11.6 12.82 ~
Nap 2.05 2.57 2.53 2.88 3.27 1.54 2.54 2.2
0.07 0.11 0.14 0.12 0.26 0.32 0.08 0.03
Kz° 0.07 0.1 0.16 0.28 0.1
P205 0.08 0.07 0.03
Sum 100.53 99.88 99.67 99.55 99.56 99 99.37 100.41
Mg# 63.3 65.4 56.2 45 35.2 64 54.8 68.3
CaO/Alp3 0.82 0.715 0.786 0.761 0.641 0.758 0.758 0.738

Cr 325 334
Ni 149 67 55 163
Co 43 44
V 323 271 326 212
Zr 51 66 83 68
Y 2929 17 31 24
Nb 2 3 3
Hf 1.14
Ta <0.05
Rb 0.3 6 1
Ba 4 29 7 3
Sr 63 197 122 93
"GL = glass data. All other data are for aphyric rocks.
"Trace elements from aphyric rock.
~
4.2.2. Petrography-ElSe and elSe

Rocks from the ELSC-CLSC ridge system are representative of nearly all of the rocks
from the Lau Basin in terms of their petrography. They exhibit textures and mineralogy
similar to rocks of mid-ocean ridges as summarized here.
4.2.2.1. Textures. Texturally, the dredged and cored samples from the ELSC-CLSC
ridge system are mainly aphyric to sparsely phyric basalts. Highly phyric samples are
relatively uncommon; those found are largely plagioclase phyric. Less common are olivine-
plagioclase phyric rocks; olivine-phyric and clinopyroxene-plagioclase phyric samples are
rare. Glass-rich vitrophyric chill margins are common as rinds on pillow fragments and
probable sheet flows. Samples with chilled glass margins are common in the ODP Site 836
drill core samples and were useful in identifying five separate petrologic units and cooling
units (Parson et at., 1992). Essentially unaltered volcanic glass has been found on samples
representing a wide range in composition and from rocks formed during the earliest stages
of spreading on the ELSC. Some samples dredged from the VFR, discussed later, are
almost entirely silica-rich glass (e.g., up to 57.5% Si02) with only a few microlites of
plagioclase (Vallier et aI., 1991).
An important textural feature of all Lau Basin rocks is their highly vesicular nature.
This is the only textural feature different from typical MORB samples. Vesicle contents
ranging from 5% to 50% were observed in ODP drill core samples, and similar vesicularity
occurs in samples dredged from ridge depths as great as 2600 m. Emewein et al. (1994)
contend that the ELSC samples are more vesicular (more enriched in volatiles) than CLSC
rocks, but other data for the CLSC (Hawkins, 1976) do not support this. The high-silica
glasses from the VFR have vesicle contents typically ranging from 10% to 25%, with some
samples having up to 35%. Some vesicles are as large as 3 cm in length and spherical
vesicles up to 1 cm in diameter are common (Vallier et at., 1991). Glasses of comparable
composition from the MTJ are equally vesicular; some samples have as much as 50%
vesicles. The high vesicularity of all of the Lau Basin samples is indicative of the high
volatile content of their parental magmas and may be typical of many backarc magma
systems. For example, many Mariana Trough samples also are highly vesicular (Hawkins
and Melchior, 1985; Hawkins et aI., 1990). High-water contents and high HPIC0 2 mea-
sured on backarc glasses (Garcia et aI., 1979; Muenow et aI., 1980; Newman, 1989) suggest
that water was the main constituent forming the vesicles, but CO 2 and oxides of sulfur must
have been important constituents as well (Nilsson, 1993; Nilsson and Peach, 1993; Farley,
1994).
4.2.2.2. Mineralogy. The mineralogy of the ELSC-CLSC axial ridge samples is typi-
cal ofMORB assemblages. That is, they consist ofCr-spinel (SP), olivine (OL), plagioclase
(PL), clinopyroxene (CPX), Fe and Ti oxides (OX), and sulfides of Fe and Cu-Fe (SU). This
is the probable crystallization sequence as interpreted from petrographic study. Low-K
rhyolitic samples, now off-axis from the CLSC, have minor quartz but lack alkali feldspar.
These rocks may have formed close to the propagating rift tip of the CLSC.
Cr-spinel is common as an early-crystallized phase in relatively primitive basaltic
magmas (i.e., least fractionated melts close in composition to the primary melts). Cr-spinel
is useful in evaluating magma chemistry and evolution because the composition reflects
parental melt chemistry, oxygen fugacity, pressure, and temperature during crystallization
(e.g., Irvine, 1965, 1967; Fisk and Bence, 1980; Dick and Bullen, 1984; Allan, et al. 1988;
Roeder and Reynolds, 1991; Allan, 1994). There is a good correlation between Mg# (where
Mg# =(Mg/(Mg + Fe2+» and cr# (where cr# =(Cr/(Cr + AI» of Cr-spinel with Mg# of
the melt and the forsterite (Fo) content of coexisting olivine (Hawkins and Allan, 1994).
Hawkins and Melchior (1985) reported data for Cr-spinels that range from cr# = 0.464 and
Mg# = 0.735 to 0.295 and 0.706 respectively. Coexisting olivine ranges from FOS9.3 to
FOS7 .7 . Sunkel (1990) reported Cr-spinel with cr# =0.54 and Mg# =0.755 coexisting with
olivine (F091 ) in relatively primitive basalts. She also described basaltic andesites in which
Cr-spinel with cr# = 0.534 and Mg# = 0.613 coexisted with FoS6 . The abundance of Cr-
spinel in Lau Basin samples may be a reflection of their relatively high initial f0 2 (e.g.,
Nilsson and Peach, 1993).
Plagioclase is the most abundant mineral in all samples and is present as phenocrysts,
microphenocrysts, and as quench-textured microlites in vitrophyre. It also forms crystal
aggregates with olivine, clinopyroxene, or both, and xenocrystic grains are common in
some samples. The compositional range is typical of MORB (e.g., An60 to Ans5 ). Some
samples are extremely calcic and more typical of arc-composition magmas (e.g., up to
AIlw). In nearly all instances, these high-Ca grains are in rocks with compositions transi-
tional between MORB and arc. It is important to note that plagioclase is present and
abundant in all samples; thus, the elevated PH 0 may have delayed onset of crystallization
but did not completely suppress it. 2
Clinopyroxene is present in nearly all samples with Mg# <60. It occurs as a pheno-
cryst phase as well as a groundmass constituent. Compositions range from diopside-salite
to augite. The Fe/Mg ratio increases from least- to most-evolved rock compositions. Oxides
are very minor constituents of all but the FeTi-enriched basalts. Both ilmenite and Ti-
magnetite are present.
4.2.3. Petrology and Geochemistry-ElSe and elSe

The general MORB-like character of crust formed at the active spreading centers is
seen both in their mineralogy and chemistry. However, minor but distinctive chemical
compositional features set some of the samples apart from N-MORB sensu stricto, and each
ridge also has its own chemical characteristics. A summary of representative rock types
dredged from the ELSC axial ridge, areas of ELSC crust sampled by drilling at ODP Site
836, and from the CLSC is in Table I and Figs. 3.7a-7d. Similar data for the Leg 135 drill
sites are in Table II and Figs. 3.8a-8d for comparison. Trace element ranges, normalized to
N-MORB, are in Fig. 3.9. Sample locations are in Fig. 3.6.
The chemical distinctions between MORB and the axial ridges samples are mainly due
to subtle large-ionic-radius lithophile elements (LILE) enrichments and HFSE depletions
that give variable similarity to arc tholeiitic magmas. This has been seen in other backarc
basins (e.g., Mariana Trough; Hawkins and Melchior, 1985), and I consider it to be a
reflection of the SSZ environment in which the magmas formed. The axial zone of the
ELSC varies in composition along strike from N-MORB-like in the north to more arclike
rocks in the south. This along-strike variation may in part be controlled by the increasing
proximity to the trend of the Tofua arc and its SSZ mantle source as proposed by Ernewein
et al. (1994). Sunkel (1990) describes samples from the axial ridge south of 21°50'S as
being low-K tholeiitic basalts, depleted in HFSE, but enriched in LILE relative to
N-MORB (i.e., showing more of the SSZ component). The southernmost part of the ELSC,
TABLE 1/ 8l
Representative Analyses, ODP Leg 135 Core Samples
Location WEB
Site ODP 834 A ODP 834 B ODP 834 B ODP 834 B ODP 834 B ODP 835 ODP 835
Sample Unit 1 Unit 7 Unit 7 GL Unit 12 Unit 13 GLa Unit 1 GLa
12XCC/13-15 mean (17) 29RlI07-12 56RlI07-15 59R2I16-24 mean (27) 6R1I32-37
Si02 49.78 50.2 50.17 52.2 51.18 52.05 50.35

Ti02 1.4 1.32 1.42 2.11 1.33 1.08 1.09
A120 3 17.51 16.33 16.33 15.71 16.14 15.52 16.47
FeO* 8.47 8.78 8.67 12.41 9.08 9.99 8.46
MnO 0.15 0.17 0.2 0.2 0.19 0.19 0.16
MgO 5.06 7.73 7.73 3.4 6.1 6.12 7.67
CaO 12.71 11.88 11.87 8.58 11.41 11.14 12.9
N~O 3.08 3 3.08 3.87 2.98 2.35 1.82
K20 0.26 0.1 0.09 0.73 0.14 0.21 0.28
P20 S 0.15 0.12 0.11 0.21 0.1 0.12 0.11
Sum 99.48 99.63 99.67 100.78 98.66 98.77 100.23
Mg# 55 64.1 64.6 36 57.9 55.7 65
CaO/A1 20 3 0.73 0.73 0.73 0.55 0.71 0.72 0.78

b
Cr 267 0 179
Ni 53 102 12 78
Co
V 227 ~
'Ix 120 102 146 79 44 s:
y 28 29 40 27 22 m
Nb 2 3 0.95 ~
Hf $;
Ta
Rb 5 3 13 1.46 5 ~
~
Ba 41 22 60 29 42 S1l
Sr 245 176 195 169 130 ;n
G)
Location WEB
Site ODP 837 ODP 837 ODP 838 ODP 839 A ODP 839 B ODP 839 B ~
r-
Sample Unit I Unit I Unit I GL Unit I Unit 2 Unit 3 GL 0
G)
mean (5) 4$1136-41 9H2I95-117 24XllOO-06 18R1I54-60 22R1I19-23 -<
0
Si0 2 55.16 55.23 52.95 53.03 54.92 53.21
Ti02 1.34 1.33 0.92 0.65 0.9 0.7 ":i!
111
AI 20 3 15.34 15.34 15 15.06 16.08 15.9
FeO* 11.7 11.67 11.19 9.04 10.1 8.71 ~
c::
MnO 0.2 0.2 0.23 0.17 0.18 0.17
MgO 3.5 3.56 5.64 9 4.64 6.43 ~
CaO 8.08 8.12 10.72 11.66 10.08 12.24 ~
N~O 2.87 3.06 1.93 1.35 1.9 1.8
Kp 0.69 0.75 0.35 0.23 0.56 0.32
pps 0.21 0.23 0.09 0.07 0.12 0.09
Sum 99.09 99.5 99.01 101.26 100.62 99.57
Mg# 38 38.5 50.8 67.1 48.5 60.2
CaO/AI20 3 0.53 0.53 0.96 0.77 0.63 0.77
Cr 618 7
Ni 5 123
Co
V 314
Zr 79 32 49
Y 35 16 21
Nb 5 0
Hf
Ta
Rb 14 5 8
Ba 80 78 85
Sr 165 148 191
"GL = glass data. All other data are for aphyric rocks.
hMajor elements on glass trace elements, aphyric rock.
(\)
.,...
20 .7,-------:----:---------,
- ....... ELSe
16
---- ClSC
11
- MTJ
I0 .&
i ••
12 ... " .... ElSe
---- ClSC
_ MTJ
.0
0 '0
A MgO(wI%) B MgO(wI%)
s.o
"-"
I
s~
• .0
" ..._" ElSe

- ClSC
_ MTJ
0.0
0 .0 .0
C MgO(wI%) D MgO (wI%)
FIGURE 3.7. Variation diagrams showing major elements (A) AIP3' (B) FeO*, (e) Ti02, (D) Nap versus
MgO. Fields shown are mid-ocean ridge basalt (MORB) (after Hochstaedter et aI., 1990a) and Mariana Trough
glasses (MARA) (Hawkins et al., 1990). Other fields are eastern Lau spreading center (ELSe), central Lau
spreading center (eLSe), Mangatolu triple junction (MTJ). Data sources are discussed in the text.
the Valu Ridge, includes "oceanic andesite" (e.g., 55-60% Si02) having chemical and
isotopic signatures resembling the nearby Tofua arc (Vallier et aI., 1991).
The ELSe began to form at about 5.5-5 Ma by rift propagation into older (pre-late
Miocene) crust that may have been part of the forearc to the Lau Ridge or part of the arc
(Fig. 3.1). Most ofthe ELSe data are from samples dredged on the neovolcanic zone ofthe
ridge axis. Other data are from some of the older ELSe crust (about 0.6-0.8 Ma) that was
drilled at ODP Site 836 and from dredge samples. Magmas that formed the northern part
of the ELSe are mainly MORB-like tholeiitic basalt in terms of major and trace elements
and element ratios (e.g., Bryan et al., 1976; Bryan, 1979b; Viereck et aI., 1989). They range
from relatively unfractionated (glasses have up to 8% MgO, Mg# 64,0.9% Ti0 2 , <0.15%
K20) to moderately fractionated (4-5% MgO and Mg# 43-50). There are lesser amounts
of more strongly fractionated types, including Fe-Ti-enriched basalt (up to 2.1 % Ti0 2 and
14.8% FeO*). The range in composition of ELSe lavas is easily explained by low-pressure
fractional crystallization of observed phenocryst assemblages.
ODP Site 836 is located close to the boundary between the WEB and probable ELSe
crust. Five petrologic units were recognized. Units 1 and 2 (14-16 mbsf) are andesitic
gravels probably derived from a nearby arc-composition seamount. The other three are
MORB-like basaltic units which lie in the data fields for ELSe on Fig. 3.8; this supports the
interpretation that they are genetically related to the ELSe and are some of the earliest
ELse crust formed at that latitude. Their trace element and REE abundances also resemble
~.--------------------------,
II
~
1.
e;
i ..
.•
1~
,
A MgO(wt%) MgO(wt%)
u
2.0
~
~H
1.0
... 0
,.
C MgO(wt%) MgO(wt%)
FIGURE 3.8. Variation diagrams showing major elements (A) AIP3' (B) FeO*, (C) Ti0 2, (0) Nap versus
MgO. Numbered fields are for OOP Leg 135 drill sites (Hawkins, 1994). Other fields are as in Fig. 3.7.
the ELSe (Figs. 3.9a and 3.l0a). Site 836 trace element data are discussed together with the
ELSe data in a subsequent section.
Representative ELSe data are in Table I, and plots of major element composition
versus MgO are shown in Fig. 3.7. ELSe samples with more than 7% MgO overlap the
MORB and Mariana Trough fields for A1 20 3 . This may reflect higher k 2 , which in tum
may be due to higher water content in parental melts. Samples with lower MgO define fields
that are shifted toward the Mariana Trough field for Al 20 3 and lie parallel to the MORB
field, but at a lower level, for FeO*. The Ti02 field is subparallel to Mariana Trough but
shows depletion. The Fe and Ti behavior also reflects the relatively higher water content
than MORB. The behavior of AI, Fe, and Ti in fractionated samples resembles the (SSZ)
trends seen in the Mariana Trough (e.g., Fryer et aI., 1981; Hawkins and Melchior, 1985;
Hawkins, 1989; Hawkins et aI., 1990). The Nap field is MORB-like, and ELSe glasses
lack the relatively higher-K content that distinguishes Mariana Trough (transitional to arc-
type) basalts from MORB.
The ELSe magma chemistry appears to have remained fairly constant over at least
2 m.y., as shown by the similarity between axial ridge samples and dredge samples from the
older part of the eastern limb of the ELSe (e.g., 123-74-1, Table I). An approximate age of
about 2 Ma for this dredge site can be estimated from the model of Parson and Hawkins
(1994). This relatively unfractionated sample (Mg#=62.9) has the AI, Fe, Ti characteristics
of the Mariana Trough transitional basalts, although it has low KzO. It is similar in most
respects to the modem lavas of the axial zone. The record of ELSe magmatism from about
2 Ma to the present suggests a fairly uniform long-term composition for the magma
100 SITE834 835

E
-e-StTE 834-13
....... SITE 834·98
-e-SITE 834·10-' --z:r.SITE 83$-1
CD -o-SITE 834-12
!5 10
:::t
Z
J!
..
a.
E
If)
O.l RbBa K Sr TaNb laCe P ZI HfSmTi Tb Y Yb
100 l00~F-------~S=IT=E~83~7~8~~~83~9~--'
B CLSCANLSC
...... PPTU31-1 ....... SITE 837-1 ____ SITE 839-3
~PPTU 33·2 Gl
CD -e-SITE 838 -o-SITE 839-9
CD
II: -o-PPTU 19-1GL !5 --0- sITe 839-1
~ 10 -e-PPTU 2O-SGL ~ 10
Z
Z
.,
..
D-
E
If)
SrTaNbLaCe P ZI HfSmTi Tb Y Yb
100 ,------------,:-=-:-:-:=---, 100 SEAMOUNTS

G
........ ZEPHYR ____ IX)NNA
CD --e-ROCH -trNF
II: -o-ROCH
~ 10
Z
!E
..
If)
O.l RbBa K SrTaNbLaCe P ZI HfSmTi Tb Y Yb
100
0 ~ 100
CD
l....slTE 8"'~2-:;:'SiTE .....71
--e-SITE 834·5 -o-SITE 834-8
H OLD CRUST
...... PPl\J ... 1 ____ PPTV7-'
II: CD -e-PPTU6-1 --o-TWD74·1

010 II:
~
-o-SITE 834-6
- - - - - ~ 10
-o-PPTU 6·3 ....... ANT 225-6
z
., Z
f.. .9!
If) ..
E
a.
If)
O.l RbBa K SrTaNbLaCe P ZI HfSmTI Tb Y Yb
FIGURE 3.9. Trace elements normalized to N-MORB values of Sun and McDonough (1989). Note the relative
depletion in Nb and Ta and enrichments in LILE that characterize SSZ magmas. (A) eastern Lau spreading center
(ELSC) and Valu Fa ridge (VF). (B) central Lau spreading center (eLSC) and northwestern Lau spreading center
(NLSC). (e) Mangatolu triple junction (MTJ) and relay zone (RZ) between ELSe and eLSe. (D,E,F) ODP drill
sites (Hawkins et aI., Allan, 1994). (G) seamounts in northern basin, RoeH is Rochambeau Bank, and NF is
Niuafo' ou Island. (H) PPTU samples are from nortbeastern Lau Basin (locations in Hawkins, 1988). TWD 74-1 is
from eastern side of the eLSe and ANT 225-6 is from west side of the ELSe (locations in Hawkins, 1976;
Hawkins and Melchior, 1985).
60r---------------------~A~ 60 r----------------------::D~
.~
~~ 10
___ RNOB 5-3 (P) -.-VAlU FA !en. ___ SITE 834-9B
-e-SITE 834·IOA
-e-RNOB&4(F) ....... PPTU31·1 (P) ~SITE 834-12
-o-SITE 838 (F) -<>-PPTU 33-2 (F) -.-SITE834-13
-.-SITE 838 (P) -.-SITE 835-1
ta Ce Pr Nd 8m Eu Gd Tb Oy Ho Er Tm Vb Lu l La Ce Pr Nd 8m Eu Gd Tb Oy Ho Er TmVb Lu
60 r----------------------:B~ 6Or------------------------,
E
-e-RNOB 17-()2 (F)

-.-SlTE 837-1 -.-srre IJ».I
-o-PPTU 19-1 NLSC
-e-SITE 838 .......TAFAHI
-.-PPTU 23-2 8MT
-a-SITE 8»1 -<>-FONUALEI
-.-ANOB 12·' RZ
-.-SITE 839-3
l La Ce Pr Nd 8m Eu Gd Tb Oy Ho Er Tm Vb Lu \a Ce Pr Nd 8m Eu Gd Tb Oy Ho Er Tm Vb Lu
60 60
C F
.n-
.....
..0.
~ ~
~
~ .---:::::
___ SITE 834·2
-e-SITE 834·5
~
-o-SITE 834-8
........ SITE 834·7
--.- .......TAFAHI -.-KAO
-e-FONUAlEI .......TOFUA
-.-SITE 834-8 -o-LATE -<>-HUNGAH
-.-METIS
1La Ce Pr Nd 8m Eu Gd Tb Oy Ho Er Tm Vb Lu \a Ce Pr Nd 8m Eu Gd Tb Oy Ho Er Tm Vb Lu
FIGURE 3.10. Chondrite normalized REE data for representative rock types. Normalizing values from Masuda
eta/. (1973). (A) ELSC (RNOB 5-3, 6-4 and Site 836); CLSC (PPTU 31-1, 33-2); and Valu Fa Ridge. P=relatively
primitive, F = fractionated. (B) MTJ (RNB 19-1, 17-02); NLSC (PPTU 19-1); Rochambeau Bank (PPTU 23-2) and
relay zone (RNOB 12-1). (C) OOP Site 834 data; suffix is petrologic unit. (0) OOP Sites 834 and 835. Labels as for
Fig. lOco (E) Arc-like OOP Leg 135 data and representative Tofua arc data. Labels as for Fig. lOco (F) Representa-
tive Tofua arc data (Ewart et a/., 1994a,b).
sources. The main variability in ridge chemistry seems to be established near the rift tip
or as the rift approaches the Tofua arc (e.g., Emewein et al., 1994).
The eLSe also formed by a propagating rift that penetrated crust formed by the ELSe
(Fig. 3.1). It was initiated at about 1.5-1 Ma. In general, it comprises rock types like those
for the ELse and exhibits a similar range in composition. eLSe rocks have the chemical
characteristics of MORB and are even closer to MORB in composition than are the ELSe
rocks. Representative data are in Table I. eLSe rocks range from relatively unfractionated
basalts, represented by glasses with up to 8.8% MgO and Mg# = 67.5, to more evolved Fe-
Ti-enriched basalt glasses with 4.8% MgO, Mg# = 39.9, TiO = 2.31%, and FeO* = 14.7%.
Rocks from near the propagating rift tip include some low-K, high-silica andesites and
low-K rhyolite (RNDB-13-2; PPTU 29-3, Table I) that probably are extreme fractionates
formed as the rift penetrated older cold crust.
The CLSC magma chemistry appears to represent a marked change in the source
chemistry from that which fed the ELSe. Sample ANT-229 (Table I) was collected from a
prominent scarp near the inferred boundary between older crust generated by the ELSC and
younger crust formed by the CLSC (Fig. 3.1). We can estimate the propagation rate for the
CLSC from its length and the proposed age of 1.2 to 1.5 Ma for its initiation (Parson and
Hawkins, 1994). The estimated crustal age at sample site ANT-229 is between 0.75 and 0.94
Ma. The dredged material could be either the youngest remnant of crust formed by the
ELSC at this latitude or the oldest formed by the CLSC propagator. The glass data indicate a
"primitive" Chr-Ol-Plag basalt with 8.73% MgO, Mg# = 67, and 0.88% Ti02; ~O is
0.03%. When plotted on the variation diagrams for oxides of AI, Fe, Ti, and Na, it is clear
that it is part of the CLSC magma series and distinctly different from ELSC. Compare these
data with those for sample 123-74-1, which was collected at essentially the same longitude
but on crust that formed at the ELSe.
Representative CLSC data are in Table I, and plots of major element composition
versus MgO are shown in Fig. 3.7. CLSC samples with more than 7% MgO plot in the
MORB field for Al20 3 and thus are similar to the ELSC; the CLSC data follow the MORB
trend at lower levels of MgO and define a distinctly different trend than for the ELSC. The
field for Na20 is displaced away from MORB and ELSC toward the Mariana Trough
glasses. Both FeO* and Ti0 2 define fields very close to those for MORB and suggest that
102 was lower than for the ELSC, and presumably the source was less hydrous. Samples
with lower MgO define fields that are shifted toward the Mariana Trough field for Al20 3
and run parallel to the MORB field, but at a lower level, for FeO*. The Ti0 2 field is
subparallel to Mariana Trough but shows depletion, whereas the Nap field is MORB-like.
The compositional range for the ELSC and CLSC may be explained in terms of low-
pressure fractional crystallization of observed phenocryst assemblages. The spectrum of
glass compositions for both the ELSC and CLSC lie close to the presumed low-pressure,
anhydrous, multiply saturated cotectic as projected from PL in the OL-CPX-QTZ diagram
(Hawkins and Allan, 1994), supporting the inference that fractionation may explain the
compositional range.
Samples from the relay zone between the CLSC propagator and the dying north end of
the ELSC have both transitional and MORB-like chemistry (Table I and Fig. 3.9c). Two
samples with Mg#=68 are compared. Sample RNDB 12-6 shows the SSZ signature of
having depletion in HFSE and elevated LILE (e.g., 0.3% ~O, 24 ppm Ba), whereas RNDB
31-4 has HFSE somewhat depleted relative to MORB but lacks the elevated LILE (~O
= 0.03%, 7 ppm Ba).
High volatile contents, mainly water, are a distinctive feature of the Mariana Trough
and many Lau Basin samples (Garcia et aI., 1979; Hawkins and Melchior, 1985; Newman,
1989). This high level of volatiles is reflected in the high vesicularity of Lau Basin rocks.
There are limited data for H20, CO2 , or S measured in ELSC or CLSC basaltic glasses; the
range is 0.42-1.55%, 0-91 ppm, and 310-1880 ppm respectively (Nilsson, 1993; Nilsson
and Peach, 1993). High volatile levels also are distinctive of samples from other backarc
basins, for example, East Scotia Sea (Muenow et al., 1980) and Sumisu rift, (Ikeda and
Yuasa, 1989; Hochstaedter et al., 1990a,b). Dick (1982) inferred high volatile contents as an
explanation for the high vesicularity of Shikoku Basin rocks.
4.3. Valu Fa Ridge

4.3.1. Introduction
The southern end of the ELSC extends to at least 22°50' S, where it loses its mor-
phologic identity in a zone of rough topography. The southernmost part of the ELSC, south
of 2l 0 20'S, is known as the Valu Fa Ridge (Morton and Sleep, 1985). The VFR is a well-
defined, linear feature that extends for at least 165 km more or less parallel to the Tonga
Ridge and approaches to about 40 km of the volcanic island of Ata in the Tofua arc. The
VFR is propagating southward into older crust of uncertain composition. This crust may be
equivalent to the rifted forearc terrane that forms the basin-range seafloor of the western
Lau Basin.
The VFR rises about 800 m above the surrounding seafloor shoaling to depths of 1500
m. The ridge is narrow, about 10 km wide or less, lacks a crestal axial rift valley, and is
capped by many small domelike edifices (Collier and Sinha, 1992; Wiedicke and Collier,
1993). It is offset about 3 km, in a left-lateral sense, near 22°l2'S and there are several other
minor offsets of the axis. Magnetic anomaly data indicate that the ridge has been spreading
for at least 0.7 to 0.9 Ma, and the spreading rate is estimated to be about 6-7 cm/yr (Morton
and Pohl, 1990). Multichannel seismic reflection data and seismic refraction lines give
evidence for a strong reflection at about 3.5 km below seafloor (Morton and Sleep, 1985).
Additional data indicate the depth as 3.2 ± 0.2 km below the seafloor (Collier and Sinha,
1992). The polarity of the reflection phase is shifted 180° as expected if this were the top of a
low-velocity zone. It has been interpreted as the top of a flat-topped magma chamber about
2-3 km wide (Morton and Sleep, 1985). The magma chamber (lens?) probably extends
along the ridge crest for at least 20 km between 22°1O'S and 22°30'S (Collier and Sinha,
1992). These authors see a correlation between the width of the VFR crest and the width of
the magma chamber relector. The offset in the ridge near 22°l2'S is interpreted as an
overlapping spreading center with a magma chamber under the overlap basin.
Hydrothermal circulation and metallogenesis on the VFR ridge crest is evidenced by
high concentrations of particulate and dissolved Mn, by elevated 3He, ferromanganese
crusts up to 1 cm thick on apparently young fresh samples of lava dredged from the ridge
crest, by sulfide mineralization, and by active hydrothermal vents. These features are
dicussed in a separate section.
4.3.2. Petrography-VFR
The main rock types recovered from the VFR are highly vesicular, glass-rich blocky,
tabular or bulbous fragments having an andesitic composition. Some samples have ropy
surfaces and flow banding. Other types include andesitic breccias, ferromanganese crusts,
and breccias of smectite and ferromanganese minerals. Photographs and imagery of the
seafloor show "viscous domes," rare large pillows, thin sheets, and irregular rubble piles of
lava fragment; some of the rubble piles are cemented by ferromanganese minerals and
nontronite (Vallier et aI., 1991).
The samples are mainly vitrophyric or aphyric to sparsely phyric fine-grained rocks.
Plagioclase is the main crystalline phase, but it is rare. It forms both microphenocrysts
(Ans7 _77 ) and microlites (An62-An3s ). These have low K content (e.g., OrO.l_O.4). Clinopyrox-
ene is rare. It occurs as microphenocrysts; some are compositionally augite, others are
subcalcic or pigeonitic. All are compositionally distinct from pyroxenes of the Tofua arc,
including samples from Ata. The VFR pyroxenes all are Fe-rich with from 21 to 30% mol
proportion Fs. Very minor amounts of Fe-Ti oxides form grains less than 0.1 mm in size.
4.3.3. Geochemistry-VFR
The glass samples analyzed are mainly low-K andesites ranging in composition from
54% to 60% Si02, 0.36-0.57% K20 (Davis et aI., 1990; Vallier et al., 1991). Representative
analyses are in Table I. Sunkel (1990) reported the presence of three samples of high-silica
dacitic glasses with from 65.5% to 69.8% Si02, 0.44% to 0.64% Ti0 2 , and 0.62% to 0.87%
K20. Davis et al. (1990) included analyses of two basalt samples from the VFR that are
moderately fractionated (MgO 5.66-6.07%). A plot of total alkalis versus silica (not
shown) indicates that the suite is calcic. Ti0 2, N~O, and FeO* all are relatively high when
compared to Tofua arc data at comparable levels of MgO (Vallier et aI., 1991).
Trace element data show that the VFR samples are typical of other Lau Basin samples
(Jenner et al., 1987; Vallier et aI., 1991). That is, they have LILE enrichment and HFSE
depletion, especially Ta and Nb, when compared to MORB, as seen in Fig. 3.9. However,
these relative depletions and enrichments, while following the arc or SSZ pattern, are not as
extreme. The chondrite normalized REE patterns (Fig. 3.10a) have a slightly convex
upward (MORB-like) shape similar to the ELSe and eLse rather than the flat pattern
considered typical of arc tholeiites. The overall enrichment can be explained as the result of
olivine and plagioclase fractionation. The shapes of the VFR patterns are similar to most
of the Mariana Trough glasses and many of the Lau Basin samples (Hawkins, 1994;
Hawkins and Melchior, 1985). However, note that the pattern also is similar to that for the
Tofua arc volcano of Kao (Fig. 3.10f). Ratios of LILE to light rare-earth elements (LREE)
for VFR, such as KlLa, are enriched and resemble ratios for arc volcanics (Jenner et aI.,
1987). The VFR data lie in the field defined by the Tofua arc on the Hf/3-Ta-Th diagram
(Fig. 3.11).
4.3.4. Petrogenesis-VFR
The origin of this assemblage of intermediate composition rocks that forms the
carapace of a very young segment of the ELSe is problematic; mixing of melts or sources,
fractionation, and assimilation all may be involved. VFR rocks have many similarities to
the Tofua arc, yet they also have isotopic and elemental characteristics of the ELSe and
eLSe. Their major element chemistry resembles "oceanic andesites," like those found
on the Galapagos Ridge (Sinton et aI., 1983) or on the CLSe, that are explained as the result
of low-pressure crystal fractionation.
Vallier et al. (1991) summarize the elemental data as follows: A similarity to Tofua arc
lavas is seen in the high-alkali metal and alkaline-earth values and ratios such as Rb/Cs and
BalLa. The depletions in HFSE are arclike, and the depletion in Nb gives high arclike ratios
of Ba/Nb. An origin by differentiation of backarc basin basalt is suggested by the similarity
to backarc signature shown in values for SrIREE, RblZr, and BalZr lower than for most
arcs. Higher values for Fe, Ti, and V than seen for arcs, as well as MORB-like TiIV, are
typical of backarc lavas. The arclike, SSZ signature would be an expected consequence of
mixing of sources, or assimilation of older altered crust, at the tip of a propagating backarc
rift into previously undisturbed SSZ lithosphere. This helps to explain the trace element and
GEe 95
eLse
'" MORO-Sourco
o VFR
'I t/.TJ
* "ARC'
~ RZ
A o O!d, N
::.: PH
)~ RochftmbilaU
Th __ ._ _ _ _ _ _ _ _·•_ _ (
L~"
+ tIILSG
B
Tn L._".____ ._.._.__.. _ ........____/
c WPB
Th L_L-_... ________ .. _._. _____..._________··_·_·.._____·_' "fa
FIGURE 3.11. Discriminant diagram based on HFSE after Wood (1980) and Wood et al. (1981). WPB = within
plate basalt. Other acronyms and labels as defined in the text. (A) Axial ridge system data. Shaded field is for Site
834 data (Hawkins, 1994). T = Tofua arc (Ewart eta!., 1994a,b). (B) Valu Fa Ridge and Mangato1u triple junction.
Shaded field L=axia1 ridge from Fig. lla, T = Tofua arc. (C) Old crust. ARC = seafloor southeast ofMTJ; Old, N =
crust northwest of MTJ; RZ = relay zone; PR = Peggy Ridge; NLSC = northwestern Lau spreading center.
REE patterns shown in Figs. 3.9a and lOa. Their similarity in Nd-Sr-isotope ratios to the
Lau Basin axial ridge samples, and the Pb-isotopic similarity to Tofua arc lavas, suggest
either mixing of sources or mixing of magmas. The isotope data are discussed in a separate
section.
4.4. Mangatolu Triple Junction

4.4.1. Introduction
A northeasterly trending spreading axis was discovered in the northeastern Lau Basin,
and samples were collected, on Scripps Institution of Oceanography PAPATUA Expedi-
tion, 1986. At one site we collected high-silica rocks from a small dome and recovered
fragments of zinc sulfide-enriched hydrothermal vent chimneys-the first to be found in
the western Pacific (Hawkins, 1986; Hawkins and Helu, 1986). Subsequently we discov-
ered, charted, and sampled the other two limbs of this spreading system on SIO ROUND-
ABOUT expedition (Hawkins et al., 1989; Nilsson et aI., 1989; Nilsson, 1993).
4.4.2. Petrography and Petrology

The samples include basaltic pillow fragments with glass-rich chill margins and blocky
fragments of silica-rich vitrophyric material of broadly andesitic composition. The basaltic
rocks are sparsely to moderately phyric comprising euhedral to subhedral phenocrysts of
plagioclase, olivine, and clinopyroxene surrounded by glass, multi phase assemblages of
spherulitic textured or finely crystalline material. Fe-Ti oxides are minor constituents and
minute globules of Fe or Cu-Fe sulfides, 10 to 50 microns in diameter, are rare. The rocks
are commonly vesicular, with vesicle contents as high as 20% and some vesicles as large as
5 mm in maximum dimension. As for other Lau Basin samples, a high content of volatiles is
implied. Data for a limited number of samples have verified this (Nilsson and Peach, 1993).
The more evolved rocks have basaltic andesite and andesite bulk compositions.
Nearly all of the samples are vitrophyres, with varied abundances of vesicles, and are
sparsely phyric to nearly aphyric. Typically the dredged fragments are angular or block
shaped, many have flow structures delineated by aligned microphenocrysts and trains of
millimeter-sized vesicles. The sulfide-enriched chimney fragments comprise wurzite and
sphalerite, with lesser amounts of chalcopyrite and pyrite, in a matrix of amorphous silica
and barite (Hawkins, 1986).
Representative data for the range in rock types are in Table III and are summarized in
Figs. 3.8-14. They share many of the characteristics of the VFR samples for major and
trace elements, but in general are much more like the ELSC-CLSC lavas than like those of
the Tofua arc. For example, on the Harker diagrams (Fig. 3.7), the trends for oxides of AI,
Fe, Ti, and Na are MORB-like, although the field for Ti is intermediate to ELSC and CLSC;
it coincides with the Mariana Trough field. The REE and the extended REE plots (Figs. 3.9
and 10) also resemble MORB as do some of the data shown on the Hf/3-Ta-Th diagram
(Fig. 3.11). Samples with the highest Th levels lie close to, but not within, the SSZ field.
Thus, even though lying close to the Tofua arc, and with one limb projecting into it, the MTJ
samples are like those of the NLSC and CLSC. Nilsson (1993) concluded that major
element variations can be explained by extensive low-pressure fractionation of plagioclase,
olivine, clinopyroxene, and titanomagnetite. Despite the proximity to the Tofua arc, the
MTJ lavas have been derived from a different mantle source.
5. WESTERN EXTENSIONAL BASIN
5.1. Introduction
The presently active spreading centers create new crust from mantle-derived melts
with little evidence for a subduction component. The spreading centers have replaced older
crust that formerly was part of the forearc to the Lau Ridge. The WEB has horsts of older
crust separated by grabens partly filled with younger basalt. Some of the older crust of the
WEB may be remnants of the Lau Ridge volcanic arc, some of it may be formed of basalts
erupted in the early stages of extension of what had been the forearc of the Lau Ridge, and,
possibly, there may be trapped fragments of old Pacific lithosphere. None of the latter has
been identified. Part of the present Tonga Ridge is underlain by low-K rhyolites, dated at
about 44 Ma, which formed in an arc setting, presumably an ancestral Lau Ridge. This was
drilled at ODP Site 841 (Bloomer et at., 1994). Other remnants of Eocene age arc crust, as
well as Miocene-age volcanic units that probably were part of the Lau Ridge, are exposed
TABLE 11/
Representative Analyses, Peggy Ridge, NLSC and MTJ
Location Peggy Ridge NLSC MTJ

Site L-95 L-86 L-103 P-19 P-20 R-17 R-19
Sample 123 123 123 PPTU PPTU RNDB RNDB
95-12 GL 86-1 103-1 GL 19-1 GL 20-5 GL 17-2 GL 19-1 GL
Si0 2 50.42 49.66 52.84 49.17 48.19 59.98 50.89

Ti0 2 0.34 2 1.75 0.82 0.79 1.37 0.94
Al 20 3 15.75 14.65 15.37 17.15 17.41 15.16 15.93
FeO* 8.59 11.88 9.46 8.72 9.11 10.5 8.53
MoO 0.16 0.29 0.19 0.18 0.18 0.23 0.17
MgO 9.78 6.77 4.72 9.51 8.98 1.91 7.52
CaO 14.67 11.54 9.35 12.75 12.42 5.62 13.1
NazO 0.99 3.01 3.35 2.11 2.34 3.89 2.14
Kz° 0.01 0.16 1.01 0.03 0.04 0.97 0.21
PP5 0.Q2 0.18 0.5 0.04 0.04 0.42 0.11
Sum 100.73 100.14 98.54 100.62 99.69 100.05 99.54
Mg# 70 53.9 50.5 69.1 66.9 27.2 64.4
CaO/AI 20 3 0.93 0.79 0.61 0.74 0.71 0.37 0.82
Cr 670 700 741 470 39 309
Ni 271 380 312 221 6 90
Co 75 56 53 15 41
V 252 212 215 51 226
Zr 7 44 48 238 64
Y 11 20 21 78 24
Nb 3 9
Hf l.15 1.08 6.09 1.59
Ta 0.045 0.045 0.65 0.133
Rb 0.153 2.9 0.3 0.1 27 7
Ba l.16 30 1 147 27
Sr 12 109 79 100 135 133
v (ppmi v (ppm)
A T1 (ppm) B Ti jppm)
(thoiJSsnds) (thousands)
FIGURE 3.12. Ti-V plot. ARC and MORB fields from Shervais (1982). Field labels as for Figs. 3.10 and 3.11.
ELSC field includes Site 836 data.
200,-)-.--________________.____- - ,
MTJ
Sa (ppm) ,
sOo 5too 200
A Zr (ppm) B Zr (ppm)
FIGURE 3.13. Ba-Zr plot. ELSe field includes data for Site 836.
on 'Eua in the present forearc (Cunningham and Anscombe, 1985). We know very little
about the seafloor north and northeast of the Peggy Ridge, which may be comparable to the
WEB. Dredged samples from several ridges and scarps in the WEB are compositionally
transitional between MORB and arc (Hawkins and Melchior, 1985). Similar rock types
were dredged on the eastern edge of the Lau Basin from similar morphologic features. This
suggests that they were formerly part of the same petrologic/tectonic province.
The backarc crust of the WEB was sampled at six drill sites on ODP Leg 135. All but
one of these sites were in subbasins that had opened during the early extension and rifting
history of the basin. Three sites sampled MORB-like crust, and three were on arclike or
transitional crust. One site (836) was on crust formed very early in the spreading history
of the ELSe. A comprehensive discussion of the petrology of these drill sites is in Parson
et al. (1992) and Hawkins et al. (1994). A brief summary of the drill site data and the
dredged samples is presented in the following section.
5.2. Old Crust

There are large areas of seafloor that form ridges between the subbasins drilled in the
WEB. For want of a better name I will call this "old crust." The name is also used for
seafloor on the eastern side of the axial ridge system, in the northeastern part of the basin
near the Mangatolu triple junction and for much of the region north of Peggy Ridge.
Hawkins and Melchior (1985) proposed that the crust of the Lau Basin was compositionally
12-r-------------------.
10
MTJ
K (ppm) 6
(!.hm.iNJf'I<1s)
A B
FIGURE 3.14. K-Ba plot. M is field for fresh N-MORB (Viereck et ai., 1989).
zoned from a MORB-like central area on the axial ridge systems to more arclike, or
transitional to arc, composition crust on the outer margins. The "old crust" constitutes the
outer margin. It is likely that this term includes a range in crustal ages; for most of it only a
relative age can be assigned. It should be emphasized that the zoned pattern was based on
analyses of glasses and fresh aphyric rocks and is not a reflection of alteration. The general
pattern of the zonation, and locations of sample used to infer it, are in Fig. 3.6. Some
representative analyses are in Table IV.
Examples of dredged samples from ridges in the WEB include RNDB 1-1 from a ridge
crest west of Site 836, and RNDB 2-7, 2-8 (Table IV); the latter are from east of Site 836 but
resemble older ridges to the west. DSDP Site 203 was drilled in a subbasin; it did not reach
basement but recovered a small basalt cobble deep in the core. Glass in the vitrophyric rim
has transitional chemistry (e.g., 52% Si02, 0.7% Ti0 2, 0.28% Kp in a sample with
Mg#=66.l; Fryer et al., 1981). It is likely that it came from the adjacent intrabasin ridge.
There is but one sample from the southern end of the WEB, F-1l4 on Fig. 3.6, but it also
shows the transitional chemistry of the postulated zonation. Rocks from dredge sites
123-81, PPTU 34 and ANT 223 (Table IV), are from crust which Hawkins and Melchior
(1985) interpreted as transitional between MORB and arc. ANT-223 is from close to the
western side of the ELSC crust about 100 km off the present axis and southwest of ODP Site
836. Sample 123-81 is from the WEB northwest of Site 837. Sample 123-101-4 is from the
lower slopes of the Lau Ridge and represents either a part of the arc or some of the early
formed backarc crust. Examples of old crust from the northeastern part of the basin,
adjacent to the Mangatolu triple junction, are PPTU 4-1, 6-1 and 7 -1. PPTU 4-1 and 7 -1 have
compositions transitional to arc chemistry, whereas PPTU 6-1 is more MORB-like.
Collectively these samples of "old crust" indicate that much of the earlier-formed
crust in the opening of the Lau Basin was transitional in composition between MORB and
arc, as originally proposed by Hawkins and Melchior (1985). However, there are problems
with this simple model as evidenced by data discussed in the following section.
5.3. MORB-like Drill Sites

ODP Sites 834, 835, and 836 drilled rocks having MORB-like chemistry; representa-
tive data are in Table II. Sites 834 and 835 are in extensional basins, partly filled with
sediments, and Site 836, discussed previously, is part of the earliest crust formed at the
ELSC. It is in an extensional basin like Sites 834 and 835 that may represent a tectonic
setting transitional from rifting to spreading. Age estimates from the paleontologic data
indicate that Site 834 basalts range in age from about 5.5 Ma, or older, to 3.8 Ma. We
proposed that magmatism may have started at about 6 Ma (Hawkins, 1994). It is important
to note that the narrow subbasin at Site 834 was the locus of basalt eruption for more than
1.7 m.y. as it opened. The basaltic filling is interbedded with thin sedimentary layers. At Site
835, the age of the end of basaltic infilling is late Pliocene (3.5 Ma) as determined from the
paleontologic age of overlying sediments. As for Site 834, we do not know when volcanism
began. Site 836 basalts are intercalated with Pleistocene sediments. The basaltic crust is
estimated to be from 0.8 to 0.64 m.y. old. The correlation of Site 836 basalts to the ELSC is
based on the probable age of rift propagation past Site 836 (Parson and Hawkins, 1994) and
the chemistry of Site 836 basalts.
Thirteen petrologic units were iqentified at Site 834 that range from a relatively
"primitive" massive unit (unit 7) having Mg#=6l-64 to fractionated units (e.g., unit 12)
having Mg#=34-36 and Si02=54-56%. Identification of petrologic units was based on a

combination of petrography, mineralogy, and chemistry as well as the presence of chilled
margins and sedimentary interbeds. The sequence drilled shows cyclic eruptions of vari-
ably fractionated magmas and large variations in thickness of the units. The variations in
thickness probably reflect variable volumes of production of parental melts. There is wide
range in melt composition with depth and no correlation between depth and magma
chemistry. For example, "primitive" unit 7 was encountered from 214-287.6 below
seafloor (mbsf) and is early Pliocene or older, whereas "evolved" unit 12 was encountered
from 363.5-407.9 mbsf. The deepest, oldest, unit sampled (unit 13) at 423 mbsf is a
fractionated unit with Mg#=57-58 and about 51% Si02 • One of the youngest units (unit
2), at 112 mbsf, has Mg#=52-55. It is clear that we did not drill to the bottom of the lava
sequence into the earliest melts erupted at Site 834, but had encountered some relatively
evolved rocks at the point where drilling had to be stopped. Fields for element abundances
are in variation diagrams (Figs. 3.8-14). Trace element and isotope data and their bearing
on petrogenesis are discussed in separate sections.
The cyclicity in magma chemistry and thickness of petrologic units indicates periodic
eruption, refilling, and melt storage in near-surface magma chambers which experienced
variable amounts of fractionation and mixing of melts. The variable thicknesses (volumes)
of each petrologic unit may reflect variations in the volume of parental melt production.
This in tum may be a response to nonuniform mantle upwelling. Several factors may have
been involved, such as episodic increases in the flux of thermal energy to the mantle source,
episodic pulses of trench rollback or rate of subduction, or perturbations in the convective
overturn or mixing of SSZ mantlelMORB-source mantle.
Previous work (Hawkins and Melchior, 1985) suggested that the basin is composi-
tionally zoned with rocks transitional to arc-composition forming zones on the east and
west around the central part of the basin that has crust with more MORB-like composition.
Figure 3.6 shows the location of samples that were used in making this interpretation. The
work of Ernewein et al. (1994) adds to this picture by showing that the ELSC becomes more
arclike from north to south, culminating at Valu Fa Ridge, which has a strong arclike
signature. Site 834 was selected to sample some of the oldest crust in the basin and indeed
the cores include the oldest Lau Basin crust yet identified. We expected that Site 834 lavas
would be arclike because of their location and relative age. However, they include some of
the most MORB-like rocks in the basin. The hypothesis was tested, and it appears to have
failed. A similar relation between early rifting and MORB composition lavas has been
documented at the Sumisu rift (Hochstaedter et aI., 1990a,b). What went wrong with this
intuitively appealing idea?
Lavas cored at Site 834 formed over a time span of more than 1.8 m.y. by continued
eruptions into a slowly opening basin. Relatively unfractionated basalts form thick petro-
logic units that have Na, Ti, and low-partition-coefficient trace element abundances,
indicating a relatively small extent of source melting in comparison to other Lau Basin drill
sites (Hawkins, 1994). The deepest samples recovered are fractionated rocks, which imply
that we did not reach the base of new crust at Site 834. We have no idea of the chemistry of
the first erupted lavas, but intricately zoned cores of plagioclase are interpreted as the result
of mixing between arc and MORB-type melts (Bryan and Pearce, 1994). Strongly zoned Cr-
spinels also support the idea of magma mixing (Allan, 1994). The Site 834 data do not
necessarily refute the general idea of compositional zoning in the basin because rocks with
transitional composition exist at the dredge sites on the intrabasin ridges (Fig. 3.6). The data
for unit 13, the deepest petrologic unit sampled, also are instructive. The glass samples
carry the depleted SSZ signature both in terms of LILE and HFSE. For example, sample
59R2/52-55 has Rb=1.46 ppm, Ba=29.4 ppm, Nb=0.95 ppm, Zr=79 ppm, and Hf=2.14 ppm
(Hergt and Farley, 1994; Table II). The lack of K, Rb, Ba enrichment relative to MORB
differs from the values we would have expected if these were the transitional basalts
mapped elsewhere in the western Lau Basin. However, the AI, Fe, Ti, and Na values relative
to MgO all would lie in the MARA (Mariana Trough basalt) field (Figs. 3.7 and 3.8) and are
definitely not comparable to MORB but are like other backarc basin basalts. These samples
must have come from depleted SSZ mantle that lacked the LILE enrichment. Was it a
depleted MORB-source mantle? As discussed later, Hergt and Hawkesworth's (1994)
isotopic data establish that the source had the Pacific MORB-source mantle isotopic
composition.
An explanation for the lack of LILE enrichment may be that the SSZ mantle source for
Site 834 escaped infiltration of subduction-derived fluids. A more likely answer is that the
large volume of melt erupted to fill the basin at Site 834 may have extracted all of the LILE
enriched SSZ component available and continued to melt depleted MORB-like mantle
source material. For purposes of comparison with the other drill sites and the axial ridges,
we need to consider that Site 834 lavas were erupted about 2 to 3 m.y. before the ELSC
propagator tip passed by on the east, and that they are on a "flowline" relative to the present
spreading direction of the CLSC (Parson and Hawkins, 1994). Site 834 samples are even
more like the lavas from the modem CLSC than are some of the ELSC glasses. This is
supported by the trace element abundances and ratios as well as the isotope chemistry. Site
834 glasses are, however, distinctive in having relatively high Na content when compared
to either ELSC and CLSC glasses or to MORB. They resemble Mariana Trough glasses in
having this relatively high NazO. Although they have a strong MORB element signature,
data for least-fractionated samples (>6.5% MgO) overlap with the Mariana Trough fields
and with the CLSC fields except for AIP3 (Fig. 3.8). Differentiated samples «6.5% MgO)
define different trends relative to glasses from the axial ridge of the Mariana Trough. They
lack the high A1 20 3 and low FeO* and Ti0 2 considered by some authors to be the backarc
basin basalt signature.
The sodium anomaly presents a paradox. The thickness of some of the petrologic units
indicates extensive melting of the source yet the high Na content implies a small extent of
source melting. Site 834 magmas may have come from a source like the MORB source but
relatively enriched in Na, or they may represent relatively small amounts of melting
compared to typical MORB, or they may be the result of extensive interaction between
magma and the conduits, or combinations of these processes may have occurred. Repeated
eruption of melt batches, each formed by a small extent of melting of a source continually
replenished with MORB-source mantle, could help explain the Na relations and would be
consistent with the mantle counterflow and diapirism envisoned in Fig. 3.2.
Site 835 data are limited to one moderately fractionated basaltic petrologic unit. Like
the Site 834 rocks, it is MORB-like in most respects, as seen in Figs. 3.8-3.10, and the data
lie within the fields for ELSC rocks. An affinity to SSZ magmas is shown by the moderate
LILE enrichment and the depletion in Nb and Ta (Fig. 3.ge). Comparison with the ELSC
data is relevant because the ELSC propagator tip passed by to the east at about the time
(3.6 Ma) that Site 835 lavas were erupted. The NazO and Ti0 2 relative to MgO are lower
than for Site 834, which may indicate that the source experienced a relatively larger extent
of melting, or the source was more depleted in Na and Ti, or both. The A1 20 3 and FeO* for
Site 835 lavas resemble the ELSe and show even less of the Mariana backarc basin
signature. This displacement may reflect a less hydrous, less oxidizing, melt source than
that for the Mariana Trough.
Data for Site 836 were presented in the discussion of ELSe crust. The two youngest
units of the five cored at this site appear to have been derived from a nearby arc-
composition seamount. Rocks dredged from a nearby ridge (e.g., RNDB 2-7, Table IV)
may be representative of this source. The presence of this arc-composition material in close
proximity to the spreading ridge points out the extreme small-scale heterogeneity of the
backarc mantle melt sources.
Rocks from WEB drill Sites 834 and 835 resemble those from the active ridges in
terms of major and trace element chemistry, which suggests a similar source. It is inferred
that they are melting products from a MORB-source mantle with little or no chemical
influence from a subduction component in spite of the SSZ setting in which they formed.
5.4. Arclike Drill Sites

Rocks at Sites 837, 838 and 839 are more arclike (Table II and Figs. 3.8-3.14) than
those already described; some are closely similar to modern Tofua arc tholeiites (Ewart et
aI., 1973, 1977, 1994a,b; Ewart and Hawkesworth, 1987; Hawkins and Allan, 1994). All
three sites were drilled in small, narrow, partly sedimented basins, elongated in a north-
easterly direction. The sites are all less than 90 km from the present axial zone of the ELSe
and on what had been assumed to be backarc crust. Their crustal ages are estimated to be
about 2 Ma, which corresponds closely to the time of nearest approach of the ELSe
propagating tip. Their arclike composition was unexpected in view of their location,
although some similar rocks had been dredged from nearby high-standing ridges (e.g.,
RNDB 1-1, 2-7, 2-8, Table IV).
Site 837 and 838 igneous rock samples are moderately differentiated arclike basalts
and basaltic andesites. They are gravel clasts probably derived from a nearby seamount. No
glasses were recovered at Site 837; bulk rock analyses have 54-56% Si02, and 3-3.4%
MgO. These are markedly similar to VFR samples (Jenner et al., 1987; Vallier et al., 1991)
and to island-arc tholeiites in general. FeO* and Al20 3 are like ELSe lavas, but N~O and
Ti02 resemble Mariana Trough data. Glasses from Site 838 comprise a bimodal suite with
52-53% Si02 and about 6% MgO as one mode and 69-73% Si02 and 0.2% MgO as the
other. All are like Mariana Trough glasses for Al 20 3 but more like the ELse for FeO*,
Ti0 2 , and N~O (Fig. 3.8). They resemble rocks from the relay zone between the ELse and
eLse and other Lau Basin crust transitional between MORB and arc chemistry.
Site 839 lavas are among the more unusual samples collected by drilling on Leg 135.
Some samples resemble those from Sites 837 and 838, whereas others are even more
arclike. The samples include two-pyroxene basalts and basaltic andesite. Their Si02 ranges
from 50-55% and MgO from 4.4% to 15%. The samples with more than 9% MgO are not
liquid compositions as they have abundant phenocrysts of olivine and clinopyroxene as
well as phenocrysts of chrome-spinel. Allan (1994) showed that the strongly zoned olivine
and chrome-spinel phenocrysts support magma mixing as an explanation for some of the
Site 839 lavas. Site 839 is considered to be an arc-composition edifice that formed within
the Lau Basin at about 1.9 Ma. It was cut off and abandoned on the west side of the ELSe
as the rift tip passed by.
The melts erupted at Sites 837, 838, and 839 were formed from mantle with a robust
arc-source signature. If Site 839 lavas had been collected on the Tonga Ridge, they would
have been considered to be part of the Tofua arc magma series. However, even though the
arc signature is strong, the Site 839 samples lack the high A1 20 3 seen in the Mariana Trough
or in typical arc lavas. The Site 839 source may have been less hydrous than the Tofua arc
SSZ source. Even though erupted synchronously with crust being formed at the nearby
ELSC, rocks from these three drill sites are much less like MORB than is the ELSC crust.
The mantle source must have remained heterogeneous, retaining arc-source SSZ domains,
even while the MORB-source mantle influx was feeding melts to the ELSC propagating
rift. Although unusual, this is not unique, as we found a young arc-composition volcanic
edifice, aligned with other edifices having typical "backarc basin basalt" chemistry, on the
crest of the axial ridge in the Mariana Trough (Hawkins et al., 1990).
6. THE PEGGY RIDGE AND SEAMOUNTS OF THE NORTHERN BASIN
6.1. Introduction
The geology of the area north of Peggy Ridge is poorly known, although the morphol-
ogy is well illustrated by sonar side-scan imagery made with GLORIA (Parson and Tiffin,
1993). The NLSC extends northeasterly from Peggy Ridge into older crust having nu-
merous seamounts (Fig. 3.1). The morphology ofthis area suggests that there are numerous
partly sedimented basins as well as some high-relief areas in which seamounts rise more
than 3000 from nearby narrow basins (Hawkins, 1974). The region north of Peggy Ridge is
poorly sampled and the data are mainly from seamounts. Each of the three major seamounts
we have dredged and Niuafo'ou Island constitute distinctive magma types. This section
presents a brief discussion of the area, although the data are few and it is not possible to
develop an integrated picture of the regional geology.
6.2. Geology of Peggy Ridge

Peggy Ridge is a major feature that coincides with the trace of a fault zone having
right-lateral displacement (Eguchi, 1984). This fault probably acts as a (leaky?) transform
fault. It is not clear that the fault is related to the origin of the ridge, although a prior (?)
phase of compression across the fault trace could be responsible for the ridge morphology.
Although the shape of the ridge suggests that it may have formed as a spreading center, the
magnetic anomaly patterns do not support this (Sclater et aI., 1972).
Samples from several places along its length (Fig. 3.6) indicate that it is formed of
several basalt types (Sclater et al., 1972; Hawkins, 1976, 1988, 1989). The westernmost end,
near the Lau Ridge, has transitional basalt (e.g., 123-103-1, Table III) similar to that which
forms much of the high areas of the western basin. These have element abundances and
ratios suggesting an affinity to backarc basin basalts (BABB) or Mariana Trough-type lavas
(Hawkins and Melchior, 1985). One site on the western part of the ridge yielded very
primitive, highly depleted basalt that has many of the characteristics of a basaltic komatiite
or a variety of boninite. Rocks and glasses from this dredge site (123-95-12, Table III;
Hawkins and Melchior, 1985) represent extensive melting, up to 30%, of a depleted mantle
source. Although having many of the chemical aspects of high Ca-boninites, the samples
TABLE IV
Representative Analyses, Old Crust and Seamounts
Location Old Crust
Site A-223 F-1I4 L-8l L-IOI P-4 P-6 P-7 P-34

Sample ANT Sonne 123 123 PPTU PPTU PPTU PPTU
223-1 GL 114-GL 81-GL 101-4 4-1 GL 6-1 GL 7-1 GL 34-1 GL
Si02 50.89 50.88 51.57 55.97 49.72 48.22 50.08 49.18

Ti02 1.32 0.83 1.11 1.04 0.96 0.79 1.28 0.79
A120 3 17.1 18.84 16.03 14.45 16.17 17.46 15.48 15.58
FeO* 8.01 9.34 8.38 10.35 9.13 9.29 10.38 10.91
MnO 0.17 0.21 0.18 0.21 0.19 0.19 0.18 0.2
MgO 6.35 4.59 6.23 4.4 6.64 9.02 6.41 6.97
CaO 11.06 11.39 11.77 8.87 12.66 12.52 12.22 13.38
N~O 3.23 1.89 3.03 2.81 2.31 2.28 2.71 1.9
K 20 0.38 0.46 0.43 0.61 0.66 0.02 0.3 0.17
P 20 5 0.23 0.14 0.2 0.13 0.2 0.04 0.13 0.08
Sum 98.79 98.59 98.92 98.84 98.64 99.83 99.17 99.08
Mg# 61.9 50.1 60.9 46.6 59.8 66.6 55.9 56.7
CaO/Al 20 3 0.647 0.6 0.73 0.61 0.78 0.72 0.79 0.86

Cr 640 74 510 681 177 222
Ni 332 34 193 257 114 52 133
Co 225 31 85 52 50 32
V 262 212 297 249 250 244 279
Zr 76 64 38 62 95 81 34
Y 26 29 25 24 26 21
Nb 2 2
Hf 1.49 2.28 1.27
Ta 0.175 0.11 0.078
Rb 6 6 18 1 6 3
Ba 46.3 12 34.9 69 17 61 12
Sr 186 281 174 224 191 222 244
have relatively high Al 20 3 (15-16%), which argues against that classification. However,
CaO/Alp3 is 0.9-1.0 which is comparable to many high-Ca boninites (Crawford et at.,
1989). The trace element data, especially REE, Cr, and Ni, offer a strong argument for a
boninitic character (Hawkins and Melchior, 1985). Regardless of the name assigned, it is a
rock type similar to early-formed melts in many SSZ settings and may represent some of
first magma types erupted at the beginning of opening of the Lau Basin. I consider it
unlikely that it formed by magma leakage on the transform fault because of the large
amount of melting and unfractionated composition, implied by its chemistry.
The middle and eastern parts of the ridge are composed of MORB-like basalt (e.g.,
123-86-1, Table III). These resemble some of the fractionated basalts from Site 834 (e.g.,
unit 13) and the axial ridges. Although they have relatively high Ti0 2 (e.g., 2%), FeO* (e.g.,
11-12%), and Nap (e.g., 3%), their calculated abundances at 8% MgO (after Klein and
Langmuir, 1987) are about 1.6%, 9.7%, and 2.6% respectively. They may represent a
relatively smaller extent of melting of a source similar to the source for Site 834 lavas.
TABLE IV
(Continued)
Location Old Crust Donna Smt. Rochambeau Bank Zephyr Sh.
Site R-l R-2 R-2 L-89 L-106 P-23 L-98

Sample RNDB RNDB RNDB 123 123 PPTU 123
1-1 RK 2-7 GL 2-8 GL 89-4 GL 106-1 GL 23-2 98-4RK
Si0 2 49.33 55.86 56.43 50.53 48.77 48.88 62.99

Ti0 2 0.51 1.2 1.25 1.26 1.6 1.89 0.47
AIP3 11.89 14.95 14.14 14.3 16.09 14.94 14.34
FeO* 9.07 11.32 12.19 11.97 10.34 11.35 5.32
MnO 0.16 0.22 0.22 0.25 0.2 0.19 0.09
MgO 12.91 3.87 3.72 6.73 7.22 6.63 3.62
CaO 12.29 8.32 8.14 11.72 12.36 11.79 5.32
Nap 1.83 2.91 2.41 2.88 2.85 3.1 3.65
Kp 0.35 0.24 0.29 0.08 0.32 0.26 0.88
PP5 0.13 0.15 0.14 0.12 0.17 0.17 0.Q7
Sum 98.47 99.2 99.11 99.84 99.92 99.21 96.75
Mg# 72.8 41.2 38.5 53.5 58.9 54.5 58.2
CaO/Alp3 1.03 0.557 0.576 0.82 0.77 0.79 0.37

Cr 150 255 238
Ni 306 17 10 45 122 72
Co 50 106 47
V 197 435 404 334 287 355
Zr 33 98 93 77 105 127
Y 13 38 37 32 61 31
Nb 3 2 3 5 7
Hf 3.27
Ta 0.375
Rb 9 6 5 16 5.6 5
Ba 55 55 11.1 66.6 40
Sr 298 131 137 90 230 215
aFrom aphyric sample 123-101-2.
bFrom aphyric rock samples. Other data for aphyric rock
The postulated NLSC abuts the Peggy Ridge and samples from the intersection are
fresh glass-rich basalts also having MORB-like chemistry. We consider these to be part
of the NLSC (e.g., PPTU-19-1, PPTU 20-5, Table III).
The ridge may be a submarine range of crustal blocks forced up by transpression on
the northwest-trending fault. Partial support for this suggestion comes from the west-to-
east variation in chemistry from arclike, to boninitic to MORB-like. The disposition of
magma types collected from Peggy Ridge mimics the regional zonation seen in the western
Lau Basin. The newly formed NLSC puts an overprint of young MORB composition lavas
on the older crust.
6.3. Geology of Donna Seamount

Donna seamount (Fig. 3.1) is a very prominent high-standing feature that rises more
than 2300 m above its surroundings to about 888 m deep. It is somewhat elongated in an
east-west direction. The summit is highly irregular, and it may never have reached to wave
base. The south flank drops into a deep basin from which diabasic/microgabbroic textured
rocks were dredged (Hawkins, 1976). The morphology suggests that the seamount may be a
very large uplifted, uptilted crustal block rather than a point-source magma leak. Rocks
dredged from the southern and eastern flanks are mainly glass-rimmed pillow fragments
comprising olivine-augite-labradorite basalt. Other dredged material includes minor
amounts of breccia and tuffaceous sediment. Most of the breccias and sediment samples
have Fe-Mn oxide-hydroxide rinds up to 1 cm thick. The age of this seamount is not known,
but all indications are that it is one of the older features in the basin.
Rock mineralogy and chemistry (e.g., 123-89-4, Table IV) are typical of the MORB-
like basalts found widely throughout the basin. The calculated Na and Fe at 8% MgO (2.4%
and 9.9% respectively) are typical of other MORB-like basalts from this part of the basin.
The REE pattern is typical for MORB (Hawkins and Melchior, 1985), and other trace
elements also indicate derivation from a MORB-source mantle. There is no evidence that
Donna seamount formed at the same time as the eruptions at Site 834, but the source
composition and petrogenesis for the two sites were the same.
6.4. Geology of Rochambeau Bank

Rochambeau Bank is elongated in a northwesterly direction; it sits atop a broad 2000-
m-deep platform and rises nearly 1500 above it. Parson and Tiffin (1993) noted that it lies
along strike of the NLSC and suggest that the NLSC may extend at least to it. Rochambeau
Bank offers several contrasts with Donna Seamount and with other samples from the
postulated trace of the NLSC. Its form and the nature of the dredged samples suggest a
point-source volcanic origin. Fresh glass-rinded plagioclase-rich pillow fragments or sheet
flows have been collected. These are relatively more enriched in alkalis and HFS elements
than most other Lau Basin rocks and are quite different from presumed NLSC samples
(123-106-1, PPTU 23-2, Table IV). In some respects the major element chemistry is arclike,
but the enriched HFSE suggest an ocean island basalt (OIB) affinity. Ti, Zr, Hf, Ta, Nb all
are enriched relative to other Lau Basin rocks. Volpe et al. (1988) designated Rochambeau
Bank samples as E-MORB on the basis of Nd- and Sr-isotope ratios. Poreda (1985) and
Poreda and Craig (1993) reported 3He/4He ofll.O and 14.1 x RA (RA = atmospheric ratio) for
Rochambeau Bank samples and a ratio of 22.0 x RA for a sample from the nearby seafloor.
They propose that these ratios represent contributions from a Samoan plume component
that has been entrained under the northern Lau Basin.
6.5. Geology of Niuafo'ou Island

The island of Niuafo' ou must represent a melting anomaly. It rises more than 1300 m
above the surrounding seafloor, which itself is part of a broad pedestal, elongated in a
northwesterly direction, and rising about 500 m above its surrounding seafloor. Its maxi-
mum elevation above sea level is 213 m. Although a Tongan island, it is not part of the
Tofua arc; it is formed of MORB-like basalt similar to the nearby.
Niuafo'ou is a shield volcano, about 8 krn in diameter at sea level and has a central
collapse caldera about 5 krn across. There have been 12 recorded eruptions in the last 170
years, the most recent being in the mid-1940s (Reay et aI., 1974; Taylor, 1991). The 1886
eruption was reported to have sent tephra to an estimated height of 2.3 krn.
Surface exposures on the island include both lavas and tephra. Some tephra layers are
as much as 6 m thick; both airfall and base surge deposits have been recognized (Taylor,
1991). Lavas range from glassy to porphyritic; plagioclase forms phenocrysts, and both
olivine and plagioclase occur as microphenocrysts. The data for these samples (Reay et aI.,
1974) indicate that the samples are moderately fractionated basalts: Si02=49-50%,
MgO=5.l-7.2%, and Ti0 2=1.4-1.75%. Reay et al. (1974) observed that the Ni content is
low and resembles arc lavas and the Ba is high. However, the Ba enrichment is not unusual
for Lau Basin lavas. The ratio TiN is 47 (from Reay etal., 1974), which is at the high end of
MORB-ratios (Fig. 3.8), and would plot close to a trend defined by seamounts (Shervais,
1982).
Isotope data (Sr, Nd, Pb) all show that Niuafo'ou lavas are distinctly different from
Tofua arc samples. Ewart and Hawkesworth (1987) conclude that they are from a distinct
magma source that has an OIB signature in spite of the general chemical similarity to
MORB. Poreda and Craig (1993) presented 3He/4He data for gases collected in the crater
lake Vai Kona on Niuafo'ou that have 7.8 x RA • This MORB-1ike ratio is typical of the
surrounding seafloor and contrasts with ratios up to 22 X RA (the Samoan "plume"
component proposed by Poreda, 1985) found at Rochambeau Bank some 100 km to the
northwest. The "plume" component and the OIB component must be decoupled.
6.6. Geology of Zephyr Shoal

Zephyr Shoal is a prominent feature with more than 1500 m relief that rises to water
depths of less than 700 m. The top is capped with corals and reefal limestone (Sclater et aI.,
1972; Hawkins, 1974, 1976, 1988). This seamount is anomalous in that it has arc composi-
tion rocks but is out in the middle of the northern Lau Basin. The rocks are hypersthene
(Fs 34)-augite-labradorite (An 7o _55 ) phyric dacite (Hawkins, 1976). The groundmass is
brown glass having low-K rhyolite chemistry (Hawkins, 1985). This glass composition is
remarkably similar to glass from the 1967-1968 eruptions of Metis Shoal (Melson et aI.,
1970). The bulk composition of the vitrophyres is dacitic with about 63% Si02 , 0.5% Ti02 ,
and 0.9% KzO (sample 123-98-4, Table IV; Hawkins, 1976)_
The phenocryst assemblage could not have formed in equilibrium with the enclosing
glass (melt); they probably are xenocrysts. The dacite chemistry is best explained as a
hybrid formed by mixing between a highly fractionated liquid and "exotic" crystalline
material. The extremely low REE patterns and general depletion in HFSE (Gill, 1976)
suggest that Zephyr Shoal lavas may have been derived from a fractionated boninitic
parental magma that mixed with crystal residue from other melts.
Although Zephyr Shoal samples are broadly similar in chemistry and mineralogy to
Tofua arc lavas, it is unlikely that the two are related. Zephyr Shoal samples are much more
sodic (N~O/K20 about 4) than Tofua arc samples, which typically have a ratio of 2-3.
Also, Zephyr Shoal has lower Ti0 2 than otherwise comparable Tofua arc samples (Ewart
et aI., 1973).
7. LAU BASIN PETROLOGIC EVOLUTION
7.1. Introduction
Mid-ocean ridge basalts form a distinct class of rocks that are relatively depleted in
magmaphilic elements such as LILE and HFSE. It is inferred that they are derived from a
previously melted, and relatively depleted, mantle source that gives them distinctive trace
element abundances and ratios. These inferred depletions are further supported by isotopic
signatures such as for the Sm-Nd, Rb-Sr, and U-Pb systems. The source for SSZ magmas
has been depleted even more than the MORB source and is considered to be like the residue
left after extraction of MORB (i.e., it is a multiply depleted mantle source). Another
characteristic of many SSZ magmas is a relative enrichment in LILE and water. The
subduction zone setting permits the reenrichment of the depleted mantle by fluids derived
from the subducted ocean crust, by sediment involvement, either directly or through an
intervening fluid phase, or by the addition of partial melts (e.g., plagiogranite) derived from
the subducted amphibolitized ocean crust (Pearce et al., 1992).
The early studies of backarc basin magma types emphasized their similarity to MORB
(Hawkins et at., 1970; Hart et al., 1972; Sclater et al., 1972), although the volatile and alkali-
metal contents of some were recognized as being anomalously high (Fryer et al., 1981).
For the latter the term backarc basin basalt (BABB) was proposed. Some workers (e.g.,
Gill, 1976), pointed out that some samples were similar to arc magmas. The recently
completed studies on ODP Leg 135 samples (Hawkins et at., 1994) demonstrate the range in
element and isotopic chemistry of Lau Basin rocks that span the range from MORB to arc
composition.
A discussion of the petrologic evolution of the Lau Basin, and other backarc basins,
needs to consider the nature of the mUltiple mantle and subduction-derived source materials
present in the SSZ tectonic setting. In the section that follows, I discuss various element-
element discriminants and isotope data that help delineate the nature of the Lau Basin crust
and give insight to the magma sources.
7.2. Nature of the Magma Source

The SSZ mantle source for arc and backarc magmas has a complex chemical character
that is the result of previous melting, addition of sediment, hydrous fluids, or melts from the
subducted slab, and counterflow of "new" mantle into the source region. This new mantle
may carry MORB-source or OIB-source chemical and isotopic signatures. It is well known
that there is an extreme decoupling between the LILE and HFSE in SSZ magmas (e.g., Gill,
1981). In this section I discuss several of the chemical and isotopic tracers that help in
constraining end members. The HFSE and REE perhaps offer the best constraints on
depletion of the original mantle source relative to the MORB-source. Compatible elements
(e.g., Cr, Ni, and V), give insight to melt/conduit reactions and to oxygen fugacity. The
LILE are a possible tracer of slab-derived elements transported in fluid or assimilated with
sediments. The isotope systems, especially Sr and the U-Pb series, give insight to mixing
between different sources.
7.3. The Immobile Elements-HFSE and REE

The HFSE (e.g., Ti, Zr, Hf, Nb, Ta, Th) including the heavy rare-earth elements
(HREE, Sm-Yb) are generally considered to be immobile under hydrous conditions up to
the greenschist facies of metamorphism (Ludden and Thompson, 1979; Pearce and Norry,
1979). Inasmuch as the HFSE are not likely to be added to the mantle wedge by subduction-
derived fluids, their abundances in IAV, and backarc basin crust, are determined by
abundances in the SSZ mantle and the melting process. Moderate levels of alteration seem
not to modify their relative abundances; thus, they are useful as tracers of the mantle source
composition when least-fractionated samples are evaluated.
Island arc volcanic (IAV) magma series typically show depletions in HFSE relative to
N-MORB (e.g., Arculus and Powell, 1986). Many IAVs show an overall relative depletion
in HREE (Gill, 1981) but do not show progressively increasing depletion with atomic
number as expected from garnet fractionation. A source depleted in all HFSE including the
HREE, or a mechanism to retain all HREE uniformly in the mantle residue is required. The
IAV are characterized by a marked relative depletion in Nb and Ta (Dupuy et aI., 1982;
Briqueu et aI., 1984; Kelemen et aI., 1990) as seen on plots of trace element normalized to
MORB (Fig. 3.9). The other HFSE show variable depletions; for some IAV series there is
lesser depletion in Zr than in other arcs.
The general tendency for intraoceanic IAV series to be depleted in the HFSE has been
explained by some authors as being due to previous depletion of their mantle source; that is,
the mantle source is multiply depleted MORB-source mantle (Hawkins et aI., 1984;
Woodhead et aI., 1993). Other explanations include retention of HFSE in a mantle mineral
(e.g., a titanate) phase (Briqueu et aI., 1984; Ryerson and Watson, 1987), more extensive
melting of MORB-source mantle (Dupuy et al., 1982), and exchange reactions between
melt and mantle enroute to the surface (Kelemen, 1986; Kelemen et aI., 1990). Kelemen
et al. (1990) showed that melts derived from a relatively more fertile mantle may react with
overlying depleted mantle and develop melts depleted in HFSE relative to MORB. This
may be especially significant because long transit of small volumes of melts through large
volumes of previously depleted mantle (e.g., through SSZ mantle) will develop strong
depletions in HFSE found in lAY. The bulk distribution coefficients (Do) for HFSE in
depleted mantle will be controlled largely by the amount of residual pyroxene. Do for
clinopyroxene-poor mantle increases for HFSE in the sequence Ta and Nb < Zr < Hf < Ti.
Subduction-derived "dehydration fluids" would cause enrichments in the LILE but
would have no effect on HFSE or HREE abundances in the mantle wedge because of
immobility of these elements in aqueous fluids. However, partial melts of the subducted
(amphibolitized) basaltic crust would be able to impregnate the SSZ mantle with the
"immobile" elements (Pearce et al., 1992). Consideration of distribution coefficients (e.g.,
Kelemen et aI., 1990) suggests that small amounts of melt would be enriched selectively in
the relative order of LREE >Ta and Nb>Zr>Ti>Hf>HREE.
The discussion of HFSE partitioning in the genesis of IAV magmas applies to backarc
magmagenesis as well. Plots of trace element abundances for Lau Basin basalts having
Mg#>60, normalized to N-MORB (Fig. 3.9), show the characteristics ofIAV. Niobium, in
particular, is generally at or below x-ray fluorescence (XRF) detection limits of about 1-2
ppm. Nearly all Nb data obtained by ICP-MS or by emission spectrography are in the range
of 0.4-1 ppm; Ta values are in the range 0.01-0.2 ppm (Ewart et aI., 1994a,b; Hergt and
Hawkesworth, 1994). Tofua arc data show Nb = 0.5-1.4 ppm and Ta = 0.03-0.09 ppm
(Ewart et al., 1994a,b). Collectively, the data shown in Fig. 3.9 support a multiply depleted
MORB-souice mantle for the source of Lau Basin basalts. There is no indication of
enrichment in Ti or Zr that might be due to impregnation of the SSZ mantle with melts from
the subducted basaltic crust. We cannot rule out a possible role for these melts, but if they
were a factor they did not impart a recognizable signal.
The REE data for the Lau Basin samples (Fig. 3.10) include both MORB-like patterns,
with LalSmef <1 and some with LalSmef ~1. The variable LREE enrichment is typical of
SSZ settings. The HREE are unremarkable in that they resemble MORB or show slight
relative depletion. These depleted samples thus resemble low-K tholeiites of the Tofua arc
(Gill, 1981; Ewart and Hawkesworth, 1987).
Thorium is generally considered as an HFSE, but it could be mobilized in subduction
fluids. Data on the Hf/3-Ta-Th diagram (Fig. 3.11) indicate that nearly all of the Lau Basin
samples have MORB-like relative abundances for these elements. Data for ELSC, CLSC,
and Sites 834, 835, and 836 all lie in the MORB field. Many of the data are shifted toward
Hf on these diagrams mainly because of the low levels for Ta and Th. Thorium is in very
low concentrations, generally <0.2 ppm) and argues for a minimal contribution from
sediment or sediment-derived fluids. This plot has been used to help evaluate the sources
for ophiolites and is intended to help discriminate between SSZ sources and MORB
sources. We see here that this is not a conclusive discriminant diagram inasmuch as the Lau
Basin clearly is in an SSZ setting but the crust largely exhibits a MORB signature. The few
data that plot in the SSZ field are either from old crust in the northeastern part of the basin,
close to the Tofua arc; from the Valu Fa Ridge, where the ELSC is propagating into older
(forearc?) crust; or from the relay zone separating the ELSC and CLSe. In each place the
mantle source may retain fragments of the SSZ mantle that formed arclike melts. In
contrast, the Mangatolu triple junction extends from the backarc basin into the Tofua arc.
At the latitude of the MTJ, the Tofua arc may be as old as 3 Ma (Tappin et al., 1994), and
it has erupted depleted melts as seen on Tafahi Island (Ewart and Hawkesworth, 1987). Data
for the MTJ trend across the MORB field toward the SSZ field in Fig. 3.11 but do not reach
it. Possibly, the reason for the minimal arc signal in MTJ lavas, in spite of their location, is
because the original "arc component" in the SSZ mantle wedge has long since been
stripped out and has been replaced by MORB-source mantle.
Collectively, the HFSE data demonstrate that the Lau Basin mantle source is domi-
nated by multiply depleted MORB-source mantle with local remnant domains of less
depleted mantle or zones that have been replenished by counterflow with more fertile
MORB-source mantle. There is little or no evidence for enrichment with partial melts of
subducted crust. The HFSE shed no light on the role of infiltration of subduction-derived
aqueous fluids.
7.4. Titanium and Vanadium

The covariation of Ti with V, and TiIV ratios have been proposed as useful discrimi-
nants for recognizing the original tectonic setting of magma series now preserved in
ophiolites (Shervais, 1982). The crystallliquid partition coefficients for V (KD ) are sensitive
to oxygen fugacity and range from> 1 to <1; higher 102 causes KD for V to decrease.
Inasmuch as the bulk KD for Ti is almost always <1, the relative enrichment or depletion
of V relative to Ti is strongly influenced by 102 , Other factors include the original composi-
tion of the source, extent of melting, and effects of fractional crystallization. Ti-V data for
the Lau Basin are shown in Fig. 3.12.
MORB-series magmas are characterized by TiN ranging from 20 to 50 and reflect
derivation under conditions of relatively low .fJ2 , whereas arc-series magmas have TiIV
ranging from 10 to 20, reflecting more oxidizing environments (Shervais, 1982). As noted
by Shervais (1982), backarc basins display a range in TiIV that includes both arclike and
MORB-like ratios. This is well illustrated in the data for the Lau Basin (Fig. 3.12). For
clarity, the arrays of data for each of the morphotectonic provinces are shown on separate
plots.
Nearly all of the CLSC data lie within the MORB field. Some samples have TiN as
low as 15, whereas Ti-rich samples (> 12,000 ppm) have ratios between 30 and 50. The
ELSC data lie on the plot between ratio lines of 15 and 30. The lower ratios straddle the
boundary designated to separate MORB from arc. Thus it could be argued that the Ti-V
relations indicate a transitional character for some CLSC and most of the ELSC samples.
Data from VFR lie in the MORB field, even though many HFSE data give them an arc
affinity (Figs. 3.9-3.11). Mangatolu triple junction data suggest that there may be two
classes of magma types. Many of the data lie in the MORB field and overlap with Valu Fa
data. There is also a broadly scattered field for high-Ti, low-V samples that extends into the
field for alkali basalts (Shervais, 1982). Many of the low-V samples (e.g., <100 ppm) are
from highly fractionated samples that likely lost both Ti and V as a result of magnetite
fractionation, but additional V was lost to another phase. Fractionation of either clinopyrox-
ene or amphibole or both would preferentially remove V relative to Ti.
Site 836 data lie within the ELSC field and are not shown separately. Sites 834 and 835
each define separate fields, although both are MORB-like.
7.5. Alkalis and Alkaline Earth Elements

A distinctive feature of N-MORB is their very low levels of K, Rb, Cs, Ba, and LREE
(Kay et at., 1970; Bryan et at., 1976; Viereck et aI., 1989). N-MORB also have distinctive
low ratios for 87Srf8 6Sr (Hedge and Peterman, 1970; Hart et at., 1974). Abundances ofthese
trace elements (LILE), and isotope ratios, increase with low-temperature alteration. The
K, Rb, and Cs increase much more readily than Ba, Sr, or LREE (Hart, 1969; Hart and
Nawalk, 1970; Staudigel and Hart, 1983; Hart and Staudigel, 1989).
Unlike isotope ratios, abundances and ratios of trace elements can be modified by
fractional crystallization. Inasmuch as the LILE are highly incompatible elements, one
expects to find positive correlations between K, Rb, Cs, Ba and La, and Si0 2 or with HFS
elements such as Zr. However, the difference in abundances of alkalis and alkaline earth
elements between least-fractionated samples of IAV and MORB indicates that the IAV are
relatively enriched in these elements. Equally important are the high ratios of BalLa,
alkalislLa, and an overabundance of Cs in IAV; these cannot be explained by crystal
fractionation. They are considered to be signatures of IAV sources or signatures that arise
during the melting process (White and Patchett, 1984). Similar chemical signatures are seen
in many backarc basin basalts, especially those erupted on relatively youthful features such
as the VFR (Jenner et aI., 1987; Vallier et aI., 1991) and the ELSe. Most BABB typically
have abundances of K and Sr not much different from MORB. However, they are charac-
terized by relatively high levels of Ba, and variable, but commonly high, Rb and La relative
to MORB with comparable abundances of Mg. Many of the axial ridge glasses from the
Mariana Trough and Lau Basin also are relatively enriched in Na (Fig. 3.7). These
observations suggest that the backarc mantle source has been enriched selectively in a
"subduction component."
Fractionation should cause a linear correlation between the LILE and elements such
as Zr. Conversely, they should show scatter if they are the result of fluid transfer into the
rock system. The LILE are highly mobile under hydrous low-temperature conditions of
alteration and are likely constituents of fluids derived from dehydration of subducted crust
(Tatsumi et aI., 1986; Tatsumi, 1989). Their presence in otherwise apparently unaltered
glass in SSZ settings is strong evidence that they reflect a subduction fluid component in the
melt source. For example, if Ba abundance correlates well with Zr, it can be argued that the
Ba is a primary feature of the melt rather than an effect of alteration. Lau Basin data are
shown in Fig. 3.12. We might then look for correlation between Ba and other cogeners such
as K, Rb, and La to evaluate an enriched melt source rather than alteration to explain
enrichments in this group of elements. The Lau Basin data for K and Ba are in Fig. 3.14.
Ewart and Hawkesworth (1987) note that the average KlBa for Tonga (Tofua) arc lavas
is about 30 and contrast this with values of 90 for fresh MORB and about 60 for altered
MORB. They proposed that a subduction fluid component has been added to the source of
the arc lavas. The range for all of the Lau Basin basalts (30-90) encompasses both MORB-
source ratios and those for a subduction component. The most arclike ratios are from the
VFR and parts of the ELSC and CLSC. This may be a reflection of their proximity to the
Tofua arc. However, the Mangatolu triple junction actually cuts into the arc trend, yet it has
ratios ranging mainly from 50 to 70. The highest KlBa ratios (as high as MORB values of
90) are from the CLSC, from part of the Mangatolu triple junction, and areas of "old" crust.
For comparison, Mariana Trough axial ridge glass samples have KlBa ranging from 50
to 90 (Hawkins et aI., 1990).
Cesium is an especially sensitive tracer of alteration and is considered to be highly
mobile in aqueous fluids (Tatsumi et aI., 1986). This mobility makes it potentially useful as
an indication of a contribution to the SSZ mantle from subduction fluids. Cesium is
enriched in IAV relative to MORB together with the other alkali metals. The limited data
for Cs in Lau Basin samples require caution in attempting to determine how important
subduction fluids have been in petrogenesis. Fresh, unaltered MORB has about 0.0086 ppm
Cs, and Rb/Cs ranges from about 43-140 (e.g., Hart and Staudigel, 1989), whereas Aleutian
IAV have 0.7-2 ppm Cs and Rb/Cs ranging from 24 to 30 (Morris and Hart, 1983). There
are few data for the Tofua arc, but a Cs range of 0.04-0.8 ppm and Rb/Cs ranging from
10-95 is reported by Ewart et al. (1994a). The decrease in Rb/Cs of arc magmas can be
attributed to input of sediment or sediment-derived fluids (e.g., Ben Othmman et aI., 1989).
The Lau Basin data are limited, but most samples have MORB-like ratios. ODP Site 834
glasses have Cs abundances ranging from 0.01 to 0.36 ppm and Rb/Cs is about 50 (Ewart
et aI., 1994a; Hergt and Hawkesworth, 1994). ODP Site 836 also has MORB-like values
with Cs = 0.01-0.05 ppm and Rb/Cs = 90-180. The CLSC has <0.08-0.15 ppm Cs and
Rb/Cs = 37-77. The VFR has some ofthe most arclike rocks; Cs data are very limited, but
those available are MORB-like (Cs =0.17-0.19 ppm and Rb/Cs =39-45). More data may
change the picture, but the data available indicate that introduction of Cs by subduction
fluids or sediments has been very limited in importance.
7.6. Evidence for a Signature from Subducted Sediment

It is highly probable that at least some of the seafloor sediments delivered to the
subduction zone are introduced to the mantle and influence the SSZ magma chemistry.
Most of the studies that have addressed this problem have focused on island-arc volcanism,
but the results are applicable to the entire SSZ system. The most useful tracers of sediment
input to magmagenesis are abundances and ratios of the LILE, especially Ba and Sr, Cs
(discussed above), the LREE, Th, Pb, B, Be, and isotope ratios such as 180/160, 87Srf8 6Sr,
lOBef/Be, and the Pb isotopes. Negative Ce anomalies, expressed as the Ce/Ce* ratio, are
considered by some to be indicative of sediment input (e.g., Hole et at., 1984). With the
exception of Ce anomalies, nearly all of these element signatures may be explained equally
well by transport in a fluid phase rather than by direct incorporation of sediment (see
discussions in Ewart and Hawkesworth, 1987; Tatsumi, 1989; Woodhead, 1989). The
argument for extensive sediment involvement seems weak and unnecessary, although it
seems clear that dehydration fluids may play an important role. An effective way to
evaluate this is with isotope data.
Meijer (1976) studied Mariana arc volcanoes and concluded that the Pb-isotope data
permitted no more than 1-2% sediment input. He cautioned that the constraints depend
critically on the assumption that the sediment data used are appropriate to what is actually
incorporated. White and Patchett (1984) pointed out that other assumptions, such as using
depleted MORB-source mantle or altered MORB as end members, are critical in the
modeling. They showed that mixes of sediment and depleted MORB-source mantle were
adequate to explain the observed Nd and Pb data for the Mariana arc, but the Sr isotope
ratios derived from the modeling were lower than observed values.
Ewart and Hawkesworth (1987) showed that the Tonga (Tofu a) arc and Kermadec arc
volcanic series have Sr- and Nd-isotope data that lie in the mantle array. They noted that the
Pb-isotope data for the Tonga-Kermadec arcs do not give unambiguous evidence for the
incorporation of a subduction component, although the data allow it. Southern Tofua arc
volcanoes may represent mixtures between MORB and about 2% sediment, whereas OIB
sources could give rise to the Pb-isotope ratios observed throughout the region. They
propose that the subduction component could have been dehydration fluids derived from
subducted sediment.
B. DISCUSSION OF THE ISOTOPE DATA
B.1. Introduction
Isotope data provide important insights to the origin of magma systems; in particular,
for studies of backarc basins they are important in identifying potential end members and
for constraining possible mixing between them. There are extensive data for Sr and Nd
isotopes for Lau Basin rocks; these are complemented by a somewhat lesser array of Pb-
isotope data and some data for He and O. Here I will summarize some of the petrogenetic
constraints inferred by other workers. Collectively the data show that multiple magma
sources have been involved in magmagenesis. These include Pacific and Indian MORB-
sources, "plume" or OIB sources, and a subduction component. Data for the Sr, Nd, and
Pd isotopes are shown in Fig. 3.15-17.
The problems of magmagenesis in backarc basins are similar to those for IAV series.
One of the distinctive signatures ofIAV relative to MORB is their slightly more radiogenic
207PbJ204Pb_ and Sr-isotope ratios, an overabundance of Ba and Sr as well as alkalis, an
enrichment in La, and a common, but not ubiquitous, depletion in Ceo Many of these
features are seen in backarc basin basalts. Recognition of lOBe in some arc magmas also has
important implications for the role of subducted sediment in magmagenesis. To date there
are no data indicating lOBe in backarc lavas, but they seem to be an unlikely place to find
it in view of the short half-life for lOBe and the depth (transit time) to which the Be carrying
phase would need to be subducted. IAV lavas tend to have more radiogenic Sr at a given
level of 143Nd/l44Nd than comparable MORB (e.g., DePaolo and Wasserburg, 1977),
although White and Patchett (1984) argue that the field of MORB is broad enough to
I
A B
1 J
0.1041
FIGURE 3.15. Nd-Sr-isotope data. Tofua arc field from Ewart and Hawkesworth (1987). Pacific and Indian
MORB fields and ODP data from Hergt and Hawkesworth (1994). eLSe and ELSe fields from Volpe et al. (1988),
Boespftug et al. (1990), and Loock et al. (1990).
include most of the IAV data. The IAV are, nonetheless, shifted to the right and at the lower
Nd end of the MORB field. This characteristic is shared by many backarc basalts and is
observed in Lau Basin samples (Volpe et aI., 1987, 1988, 1990; Loock et al., 1990). Data
from the ODP Leg 135 transect show that there are also less radiogenic, more MORB-like,
basalts in the western Lau Basin (Hergt and Hawkesworth, 1994).
B.2. Isotope Data

It has long been recognized that the Lau Basin has anomalously high 87Sr/86Sr relative
to MORB and to other western Pacific arc and backarc systems (Gill, 1976; Hawkins, 1976;
Stem, 1982; Hawkins and Melchior, 1985; Volpe et aI., 1988; Loock et al., 1990). Although
most of the Sr-isotope ratios are higher than typical MORB values, the Nd-isotope ratios lie
in the MORB field (Carlson et al., 1978; Volpe et aI., 1988; Fig. 3.15). The isotope data,
together with incompatible trace element abundances, have been used to suggest that several
distinct mantle sources are required to explain the observed isotopic range of the active
ridges and seamounts (e.g., Volpe et aI., 1988; Boespflug et aI., 1990; Loock et aI., 1990).
Volpe et al. (1988) showed that Nd- and Sr-isotope data for basalt glasses from the
15.7-,----------------, 38.9-,---------------,
N
15.5
208pbJ 204Pb
38.0
ta.1 15.3 18.5 18.7 18.9 19.1 17.9 1S.i ~8.3 18,5 18.7 18.9 19.1
A 208Pb{204Pb B 20Spb I 2t)I;Pb
FIGURE 3.16. Pb-isotope data. N = Niuafo'ou. Sources as for Fig. 3.15.

FIGURE 3.17. Ph-isotope data. Sources as for Fig. 3.15.
eLse and seamounts in the northern Lau Basin require multiple mantle sources. They
proposed (1) a MORB-like component depleted in incompatible elements, (2) a component
similar to an E-MORB or OIB source that is enriched in incompatible elements, and (3) a
distinct OIB-like component having high 87Srf8 6Sr and low 143Ndll44 Nd. The Nd-Sr rela-
tions are more like MORB from the Indian Ocean than to the East Pacific Rise, although all
data are shifted to Sr-isotope ratios slightly higher than Indian MORB.
The Valu Fa Ridge (Jenner et aI., 1987; Vallier et aI., 1991) is the youngest segment
of the ELSe rift zone where it is propagating into older oceanic crust at the southern end of
the Lau Basin. The isotopic data give a mixed signal. The Sr- and Nd-isotopic data are
similar to Lau Basin axial ridge samples, whereas the Pb-isotopic data trend toward arc
values. Jenner et al. (1987) proposed that the trace element and isotopic data indicate a
minor, but significant, contribution from the subducted slab. Vallier et al. (1991) conclude
that VFR lavas are isotopically similar to the (Tofu a arc) volcanic island of Ata that is
40 km to the east of the VFR. However, Ata is anomalous in the Tofua arc in having a
MORB-like character. The VFR has 87Sr/86Sr ratios lower than most of the other Tofua arc
volcanoes and thus resembles the backarc data. The VFR ratios for 143Ndll44 Nd isotope data
are higher than those of the Tofua are, and they too are more like the backarc data. The high
UlTh for VFR is arclike, as are the high ratios for 238U,J23O'fh, 23DThP32Th, and 226Raf23O'fh.
Loock et al. (1990) compared isotopic data for the eLSe, the Mangatolu triple junc-
tion in the northeastern Lau Basin, and Valu Fa Ridge. Their data further support the
interpretation that glasses from much of the northern Lau Basin and eLSe have the Indian
Ocean signature (87Sr/86Sr = 0.70309 to 0.70318 and 143Ndll44Nd = 0.51309 to 0.51312). The
Mangatolu triple junction has ratios of 0.7040 and 0.51283, respectively, and the VFR has
ratios of 0.70320 to 0.70339 and 0.51303 to 0.51306, respectively. Loock et al. (1990) and
Boespflug et al. (1990) suggested that the arclike isotope data for the VFR could be
explained by mixing between Pacific MORB and either a Tongan arc component or a flux
of trace elements from the subducted slab. Hergt and Hawkesworth (1994) note that the
VFR is the only part of the modem spreading ridge system in the Lau Basin that lacks an
Indian MORB mantle source signature. Inasmuch as it is the southernmost and youngest
part of the axial rift system, the influx of Indian-source mantle has not yet reached that
far south.
Prior to drilling on ODP Leg 135, the best interpretation of the nature of Lau Basin
mantle magma sources was that the axial ridges were dominated by Indian Ocean MORB,
propagating rifts in the southern and northeastern parts of the basin have a strong signature
of an arc component mixed with Pacific MORB, and the northern Lau Basin has an OIB
component mixed with Indian MORB. The best evidence for each of these components is
the Pb-isotope data, but it also is implicit in the Sr and Nd isotopes (Fig. 3.15; Hergt and
Hawkesworth, 1994). These interpretations, based largely on samples from the modem
spreading centers and other sites less than 1 m.y. old did not account for the sources
important in the early history of the basin.
Drilling in the western part of the Lau Basin has given us a new view of the
heterogeneity of Lau Basin mantle. One of the most significant results of Leg 135 is the
isotope studies presented by Hergt and Hawkesworth (1994) that bear on mixing of mUltiple
mantle components in the evolution of Lau Basin crust. They proposed that magmas
erupted at the modem spreading centers are derived from a recent influx of mantle having
the isotopic signature of the Indian plate mantle. The Indian source has displaced an older
source having the isotopic signature of Pacific plate asthenosphere. The Pb data suggest two
different mixing trends. These require an "arc" component and a component from either a
Pacific MORB or an Indian MORB source. Hergt and Hawkesworth (1994) evaluate
several possible explanations for the trace of the arc component. These are (1) magma
mixing between upwelling mantle and arc magmas during the early stages of extension,
(2) low but significant infiltration of the backarc magma source by slab-derived flux, and
(3) contamination of the source by arc lithosphere during extension and upwelling. They
propose that the "arclike" component has low Ce/Pb, low Nb/U, and relatively radiogenic
Sr and Pb. The Pb composition of the "arc" end member is more radiogenic than most of
the data from the Tofua arc but overlaps with Tafahi and Niuatoputapu, two of the
northernmost Tofua arc volcanoes, and with old Tonga arc crust such as the Eocene rocks of
'Eua. No data are available from the Lau Ridge to make a comparison for Pb isotopes, but,
arguably, 'Eua may be equivalent. The N-MORB component is defined by unit 7 at Site
834, which has relatively unradiogenic Pb and Sr like the Pacific N-MORB mantle sources.
The trace element data for glasses from this unit have a smooth MORB-like normalized
pattern (Hawkins, 1994; Hergt and Farley, 1994)
At Site 834, we cored basalts erupted in the earliest stage of Lau Basin extension.
These basalts probably represent the isotopic composition of the "original" SSZ mantle.
The Nd-isotope data are remarkably uniform (0.513109 to 0.513125) and are similar both to
MORB and to the modem Lau spreading centers. The Sr data range from 0.702544 to
0.702958 (i.e., lower than most of the Lau Basin basalt). Although several samples are more
radiogenic than MORB, they do not reach values typical of the modem Lau spreading
centers (e.g., 0.7033). The Site 834 samples have the most MORB-like Nd and Sr values
reported to date for the Lau Basin. Good correlations exist between Sr and Pb isotope
compositions. The Pb-isotope data define linear trends indicating mixing that project into
the field for Pacific MORB. This is quite remarkable in view of the widespread evidence
elsewhere in the backarc basin for an Indian mantle source. Unit 7, at mid-depth in the hole,
is the most MORB-like unit and has the least radiogenic Pb. The Pb-isotope ratios become
progressively more radiogenic upward to the youngest flow studied (unit 2). Units 13
(deepest) to 8 also show a similar upward increase from less to more radiogenic; unit 8 and
unit 2 are very similar. Units 7 and 8 offer the greatest contrast in Pb-isotope ratios. The
variations with depth, and the apparent cyclicity of the variations, are interpreted as the
result of mixing of different proportions of MORB-like and arclike melts (Hergt and Farley,
1994; Hergt and Hawkesworth, 1994). For example, these authors propose that the variation
between unit 7, the most MORB-like sample, and units 2 and 8, the most radiogenic
samples, may be explained by mixing between "essentially pristine MORB" and up
to 50% of the "arc" endmember. The timing and nature of the mixing process are fairly
well constrained and require mixing of mafic melts rather than melting of a heterogeneous
source (Hergt and Farley, 1994).
Lavas from Sites 837 and 839 have a strong arc signature in terms of mineralogy and
chemistry (e.g., they have two-pyroxene basaltic andesite and andesite). They also have arc
isotopic characteristics (Figs. 3.15-17). Site 837 samples are similar to the modem Tofua
arc, and Pb-isotope ratios lie on a trend between Indian mantle and arc. Site 839 Pb-isotope
ratios are confined to a relatively small field, suggesting that they come from a well-mixed
and homogeneous source. The Pb-isotope ratios lie on a trend between arc and the Pacific
mantle and thus resemble Site 834 samples. The isotopic signatures for Site 839 resemble
those for the modem Tofua arc volcanic island Ata as well as those for arc-derived intrusive
dikes at forearc Site 841 (Bloomer et aI., 1994). The arc rocks of Site 839 probably formed
at an arc edifice that was cut off and isolated from the Tonga Ridge by the southward
propagation of the ELSC (Ewart et aI., 1994b; Hawkins, 1994).
Site 836 sampled crust formed early in the history of the ELSC. The propagating rift
tip would have passed close to the site at about the time the Site 836 crust was formed
(0.6 to 0.8 Ma; Parson and Hawkins, 1994). Probable igneous basement at Site 836 has a
strong Indian mantle isotopic signature and resembles rocks of the present axial ridges.
Samples from volcanic breccias or gravels are more radiogenic than the presumed base-
ment and are similar to arclike Site 837 samples. The breccias probably came from a nearby
seamount.
Collectively, the data of Hergt and Hawkesworth (1994), together with data of Loock
et al. (1990), suggest that a remnant of "old Pacific" asthenosphere has been available
under the easternmost part of the Lau-Tonga system for at least the past 6 m.y. This
inference helps explain a Pacific MORB source as evidenced in the oldest drill site samples
(Site 834). This depleted source mixed with an Indian MORB source and an arc component
to form magmas of the modem spreading centers (e.g., CLSC, ELSC, and VFR). The
remnant of Pacific mantle must have migrated eastward since about 5.5 Ma, perhaps in
response to the eastward rollback of the Tonga Trench. Hergt and Hawkesworth (1994)
suggest that the influx oflndian MORB-source mantle under the Lau Basin replaced Pacific
MORB-source mantle within the past 5.5 m.y. They postulate that the influx was coupled
with the southward propagation of the ELSe. The rift propagation may have been the result
or the cause of clockwise rotation of the Tonga Ridge and its underlying mantle. They note
that the complex mixture of Indian, Pacific, and arc sources at the Valu Fa rift tip may be
explained as a consequence of this process. Another possibility is that the "Pacific"
signature is derived from the subducted slab.
Mixing between MORB-source mantle and other components is not unique to the Lau
Basin. For example the Mariana Trough may have both a MORB and an arc component
(Volpe et aI., 1987, 1990). They interpreted the data as supporting a heterogeneous source
rather than mixing of melts as proposed for the Lau Basin. The northern Mariana Trough
narrows to a propagating rift tip penetrating the Mariana arc; mixing between MORB and
either an arc component or a mantle metasomatized by fluid from the subduction zone
(Stem et al. 1990). Nascent backarc basins behind the Shichito Ridge, in the Izu-
Ogasawara arc, are interpreted as showing evidence for mixing between E-MORB and
island arc magmas rather than derivation from heterogeneous magma sources (Ikeda and
Yuasa, 1989). However, Hochstaedter et al. (1990a,b) invoke mixing relations between
sources having E-MORB and a subduction component for the Sumisu Rift of the Izu-
Ogasawara arc system.
The question of whether the isotopic heterogeneity of backarc magmas is caused by

selective melting of separate domains comprising a heterogeneous source, or by mixing
with injections of different batches of melt, may be closely tied to the dynamics of backarc
opening. Injection of new batches of MORB melt to mix with another type of mantle or
melt (e.g., SSZ mantle) could be a consequence of mantle counterflow, upwelling, and
melting in the extensional region above the subduction zone. Magmas should evolve
toward more MORB-like systems as the "other" component is melted, mixed with, and
overwhelmed by the influx of MORB-source mantle. Alternatively, a passive dynamic
response may be visualized in which heterogeneous mantle is progressively melted with the
more fusible (enriched) domains being consumed first. This too would lead to crust
dominated by less enriched rocks. The second sequence of events would lead to a linear
progression in composition rather than cyclicity and would not necessarily have N-MORB
crust as the evolutionary end point. The Lau Basin data suggest that the first model is a more
likely explanation for Lau Basin evolution.
Helium-isotope data offer one of the more intriguing insights to mantle sources
(Poreda, 1985; Poreda and Craig, 1993). The axial ridge system has MORB-like 3He/4He
ratios of about 8-10 x atmospheric ratio (RA ). Seamounts and seafloor rocks in the northern
part of the basin have RA of up to 22. These are taken as evidence for a "Samoan plume"
component in the mantle source that may be entering the SSZ region from the north or
northwest. Other isotope data suggest an OIB component in these rocks. Oxygen-isotope
data are limited to a collection of samples from the Peggy Ridge and the region around the
CLSC (Pineau et aI., 1976). They all are similar to MORB values (e.g., 8 180 = +5.7 to +6)
and indicate equilibration with mantle source material.
9. HYDROTHERMAL ACTIVITY
9.1. Introduction
We have long known that the Lau Basin has many sites of high heat flow presumed to
be related to hydrothermal activity and other areas with very low heat flow presumed to be
in recharge zones (Sclater et at., 1972). Considerable effort has been spent in sampling and
mapping surface sediments in the basin to look for evidence of hydrothermal minerals and
to understand their type, abundance, and distribution pattern. This has culminated in the
discovery of several fields of active "black smoker" hydrothermal vents on the southern
part of the ELSC. The presence of these vents and metalliferous deposits, on crust that is a
potential progenitor to an ophiolite assemblage, has important implications for the origin of
many economic deposits in continental settings.
9.2. Hydrothermal Sediments

The first reported occurrence of hydrothermal mineral deposits in the Lau Basin were
of ferromanganese oxides in sediments (Berthine, 1974) and barite and opal in volcaniclas-
tic sediments (Berthine and Keene, 1975) dredged from Peggy Ridge (dredge site 123-86,
Fig. 3.1). Subsequently, there has been an intensive effort to map hydrothermal ferro-
manganese oxide (HTFM) mineral distribution in Lau Basin sediments (e.g., Cronan et at.,
1984; Hodkinson et at., 1986; Hodkinson and Cronan, 1991, 1994). Sediment cores from
ODP Leg 135 have given additional information about the regional and temporal distribu-
tion of HTFM minerals in nondetrital and carbonate-free sediments (Hodkinson and

Cronan, 1994).
Hodkinson and Cronan (1991) recognized two hydrothermal phases in Lau Basin
surface sediments: (1) an "oxide phase" with Mn, Fe, Co, Ni, and Cu, and (2) a "su1fide/
weathered sulfide phase" with Fe, Cu, and Zn. The sulfide/weathered sulfide phase is
restricted to sediments at the neovolcanic zone of actively spreading ridges, whereas the
oxide phase is more widely dispersed throughout the basin. Using multivariate statistical
analysis to evaluate the trace and minor element chemistry of the ODP drill core sediments,
they were able to discriminate between a hydrothermal component and a weathered detrital
component (Hodkinson and Cronan, 1994). The data were used to evaluate the relative
abundances of HTFM minerals with depth in cores from the backarc basin drill sites. Only
the oxide phase was found in the six backarc drill cores. The HTFM mineral concentrations
give an integrated record of dispersal and fallout from hydrothermal plumes. In five of the
six backarc drill cores the concentrations increase upward to a maximum value and then
decrease. This pattern is different from that seen in sediments near mid-ocean ridges, where
there is a concentration maximum at the interface between sediment and basalt and an
upward decrease. Hodkinson and Cronan (1994) interpret these relations as indicating
minimal input of HTFM minerals from the igneous crust of the backarc subbasins. They
conclude that the HTFM concentration maximum at high levels in the core records
hydrothermal activity on the ELSC or CLSC distant from the drill sites. They find a fairly
good correlation with the stratigraphic age of the HTFM maximum in the cores and the
probable time of passage of the propagating rift tip of the axial ridge. They note a time lag
of from 0.4 to 0.7 Ma between passage of the rift tip and the age of the maximum HTFM
signal in the cores. They explain the uphole decrease in HTFM as the result of increasing
distance with time as the drill site location migrated away from the spreading ridge axis.
The record at Site 836 differed from that described above. Site 836 is on some of the
early crust formed by the ELSC. There is a major HTFM mineral peak near the base of the
sediment column at about 0.6 Ma. This is the approximate age of the igneous basement.
There are several lesser spikes in the HTFM mineral concentrations at age increments of
about 0.1 Ma. Even though this was drilled to igneous basement close to the former rise
crest, no sulfide/weathered sulfide phases were found. All of the drill core data show many
minor peaks in HTFM concentration in addition to the maximum. These are interpreted as
being the reflection of variable hydrothermal activity at the ridge axis. The variable
hydrothermal activity may be the result of repeated minor reorganization of the tectonic
fabric of the ridge crest, which would lead to variations in crustal fractures and per-
meability.
9.3. Hydrothermal Vents, Crusts, and Massive Sulfides

Active hydrothermal vent systems and polymetallic sulfide deposits are known from
many places on the global ocean ridge system and from several other backarc basins (e.g.,
Mariana Trough; Hawkins et al., 1990). Evidence of hydrothermal circulation in the Lau
Basin in the form of stained pillow surfaces and thin Mn-oxide crusts was photographed on
the CLSC (von Stackelberg et al., 1985); occurrences of hydrothermally formed nontronite
and manganese oxide crusts (birnessite), and metallic sulfides (pyrite and chalcopyrite)
impregnating volcanic rocks were dredged from the VFR (von Stackelberg et at., 1985).
Three types of hydrothermal metal-rich crusts were identified on the VFR (von
Stackelberg and von Rad, 1990): (1) ochre-colored Fe-rich crusts, (2) olive-colored Fe-rich
crusts, and (3) silver-gray-colored Mn-rich crusts. The Fe-rich crusts are well-crystallized
nontronite, whereas the Mn-rich crusts are birnessite. The authors interpret the growth of
the sponge-like crusts as having been influenced by microorganisms.
The first evidence for hydrothermal vents impregnated with metallic sulfides in the
Lau Basin was discovered on the east limb of the Mangatolu triple junction (Hawkins,
1986; Hawkins and Helu, 1986). Broken, cylindrical fragments of a sulfide-impregnated
mass of barite and amorphous silica were dredged along with angular blocks of fresh, high-
silica andesite vitrophyre. The sulfides are primarily wurtzite (ZnS) with minor amounts
of sphalerite, chalcopyrite, and pyrite. The chimney fragments appeared to be unaltered,
and it was assumed that they were from either an active vent system or one that had recently
become inactive.
An extensive hydrothermal vent system, extending for about 100 km, has been found
on the Valu Fa Ridge; four separate vent fields have been studied (von Stackelberg and von
Rad, 1990). This field was explored and sampled with the deep-submersible Nautile
(Fouquet et aI., 1991a,b). The sulfides occur mainly with basaltic andesite and rhyodacite.
Four types of deposits were recognized, which the authors relate to the tectonic style of
different segments of the VFR. These are (1) low-temperature hydrothermal Fe-oxides
without visible hydrothermal activity; (2) low-temperature Mn-oxide crusts that cover
higher-temperature sulfides, the low-temperature crusts being related to extensive hydro-
thermal discharges (40°C) through permeable andesitic rocks; (3) barite chimneys and
masses of barite mixed with sulfides that were formed by medium- to high-temperature
discharges and found in an inactive vent field; and (4) very high temperature black (320-
400°C) and white (250-320°C) smokers. The chimneys are formed of Cu and Zn sulfides
and barite. Fragments of altered andesite beneath the chimneys have stockwork mineraliza-
tion with cm-thick veins of chalcopyrite.
The mineralogy of the VFR deposits consists of abundant barite, sphalerite, pyrite,
chalcopyrite, and marcasite. Minor minerals include tennantite, galena, Pb-As sulfosalts
(e.g., gratonite), and primary grains of native gold. The vent effluent has the lowest pH (2)
ever measured in the ocean and the highest measured contents of Zn, Cd, As, and Pb
(Fouquet et ai., 1991a).
The Valu Fa field is considered to be one of the most active on the seafloor and is
distinctive both in the chemistry of the vent effluent and the mineralogy of the sulfide
deposits (Fouquet et aI., 1991b). They describe it as being intermediate between typical
mid-ocean ridge mineralization and the massive sulfide deposits of the Kuroko type. Von
Stackelberg and von Rad (1990) speculate that the seafloor below the VFR may be
impregnated with high-temperature sulfide minerals that would make the VFR comparable
to a Cyprus-type ore deposit.
10. BACKARC BASIN SEDIMENTATION
Comprehensive discussions of Lau Basin sediment studies and their tectonic signifi-
cance are presented in the scientific results volume for ODP Leg 135 (Hawkins et ai., 1994).
Here I present a brief synthesis of data for volcanogenic sediments that is relevant to the
magmatic evolution of the Lau Basin and the adjacent island arcs. One of the more
important discoveries of ODP Leg 135 is the evidence for ephemeral arc-composition
volcanism at numerous centers within the Lau Basin as it opened. Previously unrecognized,
evidence for the existence of these intrabasinal volcanoes comes from the abundance of
arc-composition detritus recovered in drill cores. Textures indicate minimum transport, and
nearby intrabasinal arc constructs are inferred to be the sources. They were a major source
of the volcaniclastic sediment deposited in the subbasins and were active synchronously
with the nearby eruption of MORB-like lavas that formed much of the backarc basin crust
(Bednarz and Schmincke, 1994; Clift and Dixon, 1994; Parson et aI., 1994).
The backarc basin drill sites, as well as the two drilled on the Tonga Ridge, display
similar patterns of sedimentation. The oldest sediments recovered are late Miocene and the
youngest are Holocene. The age of basal sediments differs at each site with a general
pattern of younging eastward from Site 834, alongside the Lau Ridge, to Site 836 that was
drilled on some of the oldest crust formed by the ELSe. Except for Site 836, the sediments
primarily comprise a lower sequence of volcaniclastic turbidites interbedded with hemi-
pelagic clayey nannofossil mixed sediments and nannofossil clays. These are overlain by
an upper sequence of hemipelagic clayey nannofossil oozes, locally containing calcareous
turbidites (Parson et aI., 1994; Rothwell et aI., 1994). These authors interpret the volca-
niclastic turbidites as being derived from nearby arc fragments or constructs within the
basin rather than from adjacent island arcs. Pliocene proximal volcaniclastic sediments are
interpreted as having been derived from intrabasinal seamounts. The Pleistocene sediments
are mainly hemipelagic. It is important to note that they view intrabasinal "arc" volcanism
as a major source of the volcaniclastic material. Site 836, drilled on old ELSC crust, is
distinctive in having only 20 m of sediment overlying presumed igneous basement. The
lower half of this sediment sequence comprises more than 50% volcaniclastic material. The
glass shards, representing a limited compositional range from 54.3% to 58.1 % Si02, are
best interpreted as the result of fractional crystallization of the MORB-like basaltic magmas
that formed the basement at Site 836 and the nearby ELSe. Similar basaltic andesite has
been dredged from the high-standing ridges near Site 836 (e.g., RNDB-02, Table IV).
The interpretation that the Pliocene age volcaniclastic beds were derived from intra-
basinal seamount volcanoes is based on sedimentologic data. Rothwell et al. (1994) note
that volcaniclastic rocks of the subbasin depocenters have delicate needlelike shards that
could not have survived long-distance transport, and the glass shards are fresh. Neither the
Lau Ridge nor the Tofua arc is considered a likely source. The chemistry of the glass shards
indicates that nearly all have low-K arc tholeiitic compositions ranging from basaltic to
rhyolitic. Some shards from Sites 834, 835, and 837 have medium-K arc chemistry. Glasses
from Sites 834 and 835 span the full range from basaltic to rhyolitic, whereas Site 836 has a
fairly unimodal composition (basaltic andesite, clustered around 55% Si02). Site 838
glasses also have a unimodal concentration of values ranging from 50% to 55% Si02 •
Bimodal suites are found at Sites 837 and 839 where basaltic andesites and rhyolites
represent maxima. The Site 836 glasses are distinct in having trace element ratios similar to
those of the nearby ELSC (e.g., MORB-like ratios for BalZr of 0.9 to 1.4), in contrast to the
much higher, arclike values (3.9 to 3.10), found at the other sites. In general, all of the trace
element data for glasses from the backarc sites have broadly similar patterns when nor-
malized to N-MORB. These patterns show the SSZ magma signature, namely moderate to
strong enrichment in LILE, strong depletion in Nb, and minor depletion in other HFSE
(Clift and Dixon, 1994; Rothwell et at., 1994).
The provenance of the widespread volcaniclastic units is important to understanding
the magmatic evolution of the Lau-Tonga system. The only epiclastic unit that can be tied
to the Lau Ridge is a 3.3-Ma bed at Site 834 that closely matches the chemistry of the
Korobasaga volcanic group. Rothwell et al.(1994) conclude that it was unlikely that there
was widespread or voluminous volcanism on the Lau Ridge after extension and rifting
started in the Lau Basin. This idea is contradicted by the evidence that the Korobasaga
Group was laid down on the Lau Ridge between 4.5 and 2.5 Ma (Whelan et aI., 1985) and
was a potential source of sediment. Differences in eruptive style, sediment entrapment in
the basin west of Site 834, a different pattern of dispersal, or a more limited locus of
volcanism, all may have played a role in limiting volcaniclastic sedimentation in the sub-
basins sampled on Leg 135. Cole et al. (1985) and Woodhall (1985) note that the Koroba-
saga Group is limited to six islands, forming a 35-km-wide belt, in the northern 200 km of
the Lau Group. This may account for its relative unimportance as a source of volcanic
detritus for infilling the subbasins. The Tofua arc probably was not a robust system until
about 1 Ma. It may have evolved over 3 m.y. by southward propagation of volcanic centers
(Hawkins, 1994; Parson and Hawkins, 1994). There is an extensive record of Miocene
volcanism on the Tonga Ridge. The early Miocene beds probably had a source on the Lau
Ridge, but in latest Miocene time the most likely sources were small and ephemeral (?)
intrabasinal eruptive sites (Clift and Dixon, 1994; Rothwell et ai., 1994). Zephyr Shoal
(Table IV, Fig. 3.9g) is an example of such a feature. The low-K arc-series chemistry of
many of the Tonga Ridge volcaniclastic units matches the compositions found in the
sedimentary units at the backarc drill sites; this supports the idea of derivation from
intrabasin volcanoes.
The age of opening and sedimentation of the basins appears to young eastward. The
time of transition from deposition of volcaniclastic sediments to sediments dominated by
clayey nannofossil ooze varies from site to site but also tends to young eastward. The
eastward age progression in timing of the locally derived voluminous eruptions of arclike
material suggests that there was a wave of arc volcanic activity that "tracked" the eastward
retreat of the Tonga forearc as it followed the rollback of the trench. This is a previously
unrecognized aspect of backarc basin evolution. Intrabasinal arc edifices erupted alongside
magma leaks that form MORB-like backarc basin crust. The arc volcanic front, or zone, has
been stabilized for about 1 Ma or less and constitutes the present Tofua arc. Well-dated
volcaniclastic units help constrain the longevity of the intrabasinal arclike eruptive centers.
For example, Site 835 records nearby activity from 3.5 to 2.9 or 3.0 Ma. At Site 837, proximal
sources laid down deposits from 2.1 to 0.34 Ma; at Site 838, proximal sources formed vol-
caniclastic beds from 1.92 to 0.50 Ma; at Site 839, the range is from 1.73 to 0.37 Ma (Bed-
narz and Schmincke, 1994). I propose that the "intrabasinal arc" constructs had eruptive
life spans around 1 to 2 m.y. before becoming inactive as the arc front migrated eastward.
11. REGIONAL GEOLOGIC HISTORY
Before presenting a synthesis of Lau Basin evolution, I will summarize the geologic
history as we know it. Citations to appropriate references are given elsewhere in the text.
1. The Lau Ridge was an active volcanic island arc from at least 14 Ma to 2.5 Ma.
The volcanism was in response to subduction at an ancestral Tonga Trench that may have
had the same general orientation as the present one, but the trench axis probably lay 100 km
or more to the west. Volcanic activity younger than 2.5 Ma appears to have been related to
extensional stresses to the rear of the arc and backarc rather than to frontal arc volcanism.
2. The geologic history of the Lau Ridge may extend to late Eocene time, as
suggested by the occurrence of Eocene arc volcanic rocks on Viti Levu, Fiji Islands, and
Eocene arc plutonic rocks on 'Eua, in the Tonga forearc. Leg 135 drilling at Site 841 also
revealed pre-late Eocene and late Miocene arc series volcanic rocks on the Tonga Ridge.
All of these rocks are interpreted as remnants of a former "Melanesian island arc" that
included what are now parts of the Lau Ridge, Tonga Ridge, Fiji Platform, and New
Hebrides-Vanuatu Ridge. The orientation of the trench related to this ancestral arc is not
known, but it probably faced to the east or the northeast.
3. The Lau Basin began to open in latest Miocene time (about 6 Ma) as the eastern
part of the Lau Ridge and its forearc were first extended and then rifted away to the east.
Fragments of the Lau Ridge and its older basement are found on the Tonga Ridge.
4. There has been more or less continuous arc volcanism in the Lau-Tonga region
since late Eocene time. Lau Ridge volcanism overlapped with the time of initial crustal
extension that formed the Lau Basin, but during the last 2.5 Ma it has diminished and
seafloor spreading in the Lau Basin has supplanted it. Within the last 1 m.y. the Tofua arc
has become a major feature on the Tonga Ridge. The Tofua arc may have developed
sequentially from north to south in consort with the rotational opening of the Lau Basin.
5. Lau Basin crust is largely floored by basaltic rocks having MORB mineralogy and
many MORB-like chemical and isotopic characteristics. The central Lau and eastern Lau
spreading centers (Fig. 3.1) are well-defined axial ridge systems formed mainly of
N-MORB type basalts.
6. Although dominated by MORB-like chemistry, the crust of the Lau Basin, like
other backarc basins, is heterogeneous. On a regional scale, the Lau Basin is composi-
tionally zoned from older rocks having boninitic and "transitional to arc" compositions to
younger MORB-like rocks on the spreading centers. However, this general pattern is
contradicted by the rocks from Site 834. The basin drilled at Site 834, the oldest crust
known in the basin, has MORB-like chemistry. Seamounts in the northern Lau Basin
include rocks having an ocean island basalt composition component, and dacitic rocks of
Zephyr Shoal may be silicic differentiates of boninite magma series. Rocks of the Valu Fa
Ridge and the Mangatolu triple junction range from basaltic to basaltic andesite; the VFR
has a strong arc signature and probably comes from a mixed source, whereas the MTJ has
few arc characteristics. Sites 834, 835, and 836 (Fig. 3.1), although not on a true "flowline"
in the sense of progressive accretion of new crust, sampled the evolution of magma types
over a time span of about 5 m.y. The sites are located, respectively, on the oldest crust
drilled (about 5.6 to 6 Ma, latest Miocene), intermediate age crust (about 3.4 Ma, late
Pliocene), and the youngest crust drilled (about 0.64 to 0.8 Ma, middle Pleistocene). Site
836 lies about 50 km west ofthe axis of the ELSe, which is formed of MORB-like tholeiite.
The ELSe becomes transitional from MORB to arc tholeiite toward the VFR. Sites 837,
838, and 839 form a cluster, about 125 km west of the active ELSe, on crust estimated to
be 1.9, 2.0, and 2.1 Ma, respectively. Site 839 samples are arclike in chemistry in spite of
their proximity to the ELSe.
12. SUMMARY AND A MODEL FOR LAU BASIN EVOLUTION
The evolution of the Lau Basin, and other backarc basins, must represent the kine-
matic response of the SSZ crust and upper mantle to oceanic plate subduction, mantle
counterflow, and upwelling. This response is expressed as an extension of the suprasubduc-
tion zone lithosphere and the formation of new crust by emplacement of melts derived from
the SSZ mantle. The magmatism derives from both the SSZ mantle and "new" mantle
brought in by convective overturn. The tectonic setting and magmatic response are similar
to that of extensional provinces in continental settings such as the Basin and Range
Province of western North America, where crustal extension and thinning have been
accompanied by voluminous volcanism. Backarc basin evolution may be a small-scale
replica of the early stages of formation of the major ocean basins.
The opening of the Lau Basin has been accomplished by two modes of extension.
Initially, rifting and extension formed basins separated by high-standing blocks; subse-
quently, seafloor spreading began and continues today. The rifting and extension phase
resulted in formation of new crust within the subbasins. We do not know the deep structure
of these subbasins, but it is likely that it is different from "true" deep-sea floor in that
the new basaltic crust may overlie attenuated remnants of older crust (e.g., fragments of old
arc or forearc). Extension was accompanied by extrusion of both arclike and mid-ocean-
ridge-type magmas that have formed new crust. The proximity to active island arcs, and the
similarity in many respects between arc and backarc magmas, suggests that there may be
similarities in petrogenesis. Extreme rifting and extension ultimately separated all remnants
of the older crust, and the potential void was replaced with new ocean crust that directly
overlay its mantle source. Seafloor spreading began. The seafloor spreading was promoted
by propagating rifts. Interpretations of the results of studies of the western Pacific backarc
basins have given rise to a common model in which extension followed by rift propagation
has played a key role in their evolution (Stem et aI., 1984; Parson et aI., 1990; Tamaki et aI.,
1992; Fryer and Pearce, 1992; Parson et aI., 1992; Taylor, 1992; Parson and Hawkins, 1994).
A successful model for Lau Basin evolution has many problems to explain. These
include an explanation for the occurrence of a spectrum of crustal rocks that range from
MORB-like to arclike; the occurrence of both Pacific and Indian MORB-source mantle
isotopic signatures as well as arc-source isotopic ratios; the cause for widespread extension
at a convergent plate margin; the mechanism causing extension of the crust and its eventual
rupture after which seafloor spreading becomes operative; and the apparent random loca-
tion of the part of the remnant arc or forearc that becomes the site of extension and rifting.
The model presented in Fig. 3.18 is an updated version of earlier models by the author
(Hawkins et aI., 1984; Hawkins and Melchior, 1985) and draws on more recent models of
Stem et at. (1990), Parson et al. (1990), and Parson and Hawkins (1994). Stem and Bloomer
(1992) proposed a similar interpretation of the early history of the lzu-Bonin-Mariana
system and used it to explain the origin of ophiolite remnants in the California Coast Range.
Interpretations of data from ODP drilling in the Sea of Japan, the lzu-Bonin system, and
the Mariana arc all have led to broadly similar models (e.g., Fryer and Pearce, 1992; Pearce
et aI., 1992; Tamaki et aI., 1992; Taylor, 1992). All data such as ages, rock and sediment
chemistry, geophysical data, and the regional geology used in this Lau Basin model are
from the numerous contributions cited elsewhere in the text in addition to extensive data
collected by the author.
A discussion of the initial conditions in the southwestern Pacific at the beginning of
Cenozoic time is beyond the scope of this chapter, but there is evidence for a vast zone of
extension and upwelling along the 4400-km trace of the West Melanesian-Solomon-Vitiaz
trench and arc system (see Kroenke, 1984, for an extensive discussion and Hawkins, 1994,
for a summary). A similar tectonic/petrologic feature is required to explain the evolution of
the Palau-Kyushu Ridge (e.g., Mattey et at., 1980) and was used by Stem and Bloomer
(1992) for the early stages of the lzu-Bonin-Mariana system. A dramatic change in
rheology of the upper mantle, perhaps accompanied by a change in relative plate motion,
was important in the genesis of these major geologic features. Mantle upwelling and crustal
extension promoted the development of boninitic magma systems early in arc history and at
relatively shallow mantle levels. This has been invoked for the Izu-Bonin arc (Pearce et at.,
1992; Stem and Bloomer, 1992) and for the Mariana forearc (Bloomer and Hawkins, 1987;
Stem et at., 1991). Boninite magmas were followed by, and perhaps overlapped with,
eruption of arc tholeiitic series melts. There is a major problem in explaining the mismatch
in the age of initiation of the western Pacific Eocene arc systems and the timing of the
relative change in motion of the Pacific plate (about 43 Ma) as inferred from the age of the
bend of the Emperor-Hawaii seamount chain. Evidence for volcanism at least 5 to 6 m.y.
earlier is found on the Palau-Kyushu Ridge and in the Bonin Islands (Taylor, 1992, and
references therein). Thus, it is difficult to relate initiation of arc volcanism to the change in
motion even though plate convergence has been important subsequently. The boninitic and
early arc tholeiitic stages of an arc require melting of a suitable mantle source at relatively
shallow depths (e.g., Stem et at., 1991; Pearce et at., 1992; Stem and Bloomer; 1992); plate
convergence may not be required, and, in fact, lithosphere extension could well have been
the control on early boninitic magmatism. Admittedly this is speCUlative, but an early phase
of extension could initiate the boninitic magmatism and set up crustal buoyancy differences
that could be exploited by subsequent plate motion changes to cause subduction. A similar
model was proposed by Stem and Bloomer (1992), and a less refined version was presented
by Hawkins et at. (1984). Different, and equally plausible, models presented by Pearce et at.
(1992) include subduction of a spreading ridge beneath young lithosphere. Whatever the
mechanism for its initiation, the earliest stages of Eocene arc development in the north-
western Pacific resulted in an arc that was up to 300 km wide and 3000 km long formed of
boninite and arc tholeiite (Taylor, 1992). A comparable arc developed concurrently in the
southwestern Pacific.
Considering the present Papua-Vanuatu-Fiji-Lau-Tonga region, there is evidence
for an Eocene Vitiaz arc or West Melanesian arc on the northeastern margin of the Indian
plate (Colley and Hindle, 1984; Kroenke, 1984; Colley et at., 1986, 1993; Gill, 1987;
Wharton et at., 1992) that may have been as much as 200 km wide and nearly 2000 km long.
This ancestral arc is shown schematically in Fig. 3.l8a; the 40- to 46-Ma rocks include arc
tholeiitic gabbro and gabbro-norite, pillow basalts, low-K rhyolite, and boninite. Although
these ages are close to the inferred age for the Emperor-Hawaii bend, they are hard to
reconcile with a cause-effect relationship. We would not expect an instant response of
magmatism to the beginning of subduction. It is probable that there was a delay of a few
million years between the onset of magma separation and upwelling from a mantle source
and the beginning of eruption. The silicic rocks are voluminous (?), highly differentiated
melts, that must have had mafic precursors. A further delay is likely to allow the extreme
fractionation. Whether these rocks are related to extension or are the result of subduction,
they constitute some of the oldest crust of the West Melanesian or Vitiaz arc. Evidence for
Oligocene arc magmatism is widespread and includes rocks on Fiji, 'Eua, and Vanuatu, as
well as New Britain and New Ireland (Gill, 1987). These were no doubt a response to
subduction after the change in Pacific plate motion. The Eocene-Oligocene arc is shown
in Fig. 3.l8b and in the following figures is labeled "E-O." Opening of the South Fiji Basin
as a backarc system appears to have been a response to convergence at the Vitiaz Trench in
Oligocene time, but for simplicity it is not shown on the figures.
The Miocene arc (Lau Ridge) shown in Fig. 3.l8c probably faced the present Tonga
Trench; although we have no control on its actual orientation, it is likely that there may have
been some rotation. It is uncertain whether there was a break in arc volcanism in the early
Miocene; neither outcrops nor drill site data offer an indication for continuity. The Eocene-
-
A. Eocene and Oligocene
SSZ
rrnSOURCE
--
FA = Forearc
A Arc Tholeiite
-
B Boninite
R Low - K rhyolite
n - Arc Magmas
B. Miocene (16 10 8 Ma)

0(
W 840 E841
--~;j :";'_ ~=====: c~:~;;

SSZ
SOURCE '$
FA - Forearc
E-O = Eocene-Oligocene Arc
E - ' Eua
n- Arc Tholeiite
[ill] - Subduction Fluid
D = Arc-derived Volcaniclastics
~= Reefs
C. 8106 Ma
FIGURE 3.18.
D. 6t02 Me
837
B34 835 g Else 840 841
.................
11\ - BABB magma
,t _ MOAB·like
magma
ELSC - Eastern Lau

Spreading Center
E. 1 Me to Recent
«
837
838 E
834 835 839 836 840 841
-----
T - Tolua Arc
E - • Eua
,t- MOAB· like

magma
O- Arc Tholeiite
Series
FIGURE 3.18. A model for the evolution of the Lau Basin. (A) An early phase of the Vitiaz or west Melanesian
arc system forming arc rocks now found in the Fiji Islands and the Tonga platform. The rhyolitic rocks of Site 841
are postulated to be part of this system. Arc polarity is not known but is assumed to have faced north or
northeasterly. (B) The Lau Ridge volcanic arc and possible locations of exposures on 'Eua and Sites 840 and 841.
This view and subsequent views are projections onto latitude 200 S. (C) The beginning of crustal extension to form
the WEB and rifting of the Lau arc. Magmatic activity has not yet filled the extensional basin, but mantle
upwelling is setting the stage for it. Note the insertion of the flexure or imbrication in the subducted lithosphere.
(D) Effects of extension, mantle upwelling, and magmatism in the backarc basin. The ELSe has been propagated
southward and is fed by the Indian MORB source. Not shown are the ephemeral intrabasin arc-composition
volcanoes that formed progressively from west to east. (E) In the present configuration the Tofua arc and the ELSe
are both active. Not shown is the eLSe that is propagating southward. The vertical dimension in all sketches has
been exaggerated to show details of crustal geology.
Oligocene arc probably formed part of the forearc to the Lau Ridge, as evidence from 'Eua
and other Tongan islands indicates that Miocene dikes and volcaniclastic material were
emplaced into, or deposited on, the older forearc rocks. Nearly all of the rocks of the forearc
have an affinity to arc magmas, and this suggests that the underlying crust is largely arc
material. Some evidence exists for older MORB-like crust under the forearc in the occur-
rence of basaltic clasts in the low-K rhyolite at Site 841 (Bloomer et aI., 1994). These are not
unequivocal evidence for trapped older deep-sea floor inasmuch as MORB-like rocks are
among the earliest magma types formed in arc rifting (e.g., Taylor et aI., 1992). MORB-like
lavas in the Mariana forearc were reported by Johnson and Fryer (1990), who suggested that
they may represent injection of magma during forearc rifting. I conclude that it is likely that
much, or all, of the forearc is underlain by crust formed in an island arc system.
The SSZ source for the Lau Ridge volcanic arc is considered to be depleted harz-
burgite that had been metasomatized by fluid derived from subducted, altered ocean crust
(e.g., Pearce et aI., 1984; Tatsumi et aI., 1986; Tatsumi, 1989) and possibly from subducted
sediments. The initial depleted mantle source carried the isotope signature of the source of
Pacific MORB. Late in the evolution of the Lau Ridge, at about 6 Ma, it began to be rifted,
and at about the same time crustal extension began to form rift basins in the forearc to the
Lau Ridge. An imbrication or a flexure of the subducted plate may have been formed at the
trench at about 7 to 8 Ma, as shown schematically in Fig. 3.l8d (see Billington, 1980, for a
discussion of this important feature). This change in the subducted plate may have been
either a cause or an effect of extension. The progressive deepening of the location of the
flexure, presently at about 535 km deep, is shown in Fig. 3.l8d and 18e. The opening of the
Lau Basin included an extension and rifting phase followed by seafloor spreading. Exten-
sion and rifting formed the basins drilled at all of the backarc sites except for Site 836,
which is related to the seafloor-spreading phase. The magmas at Site 834 appear to
contradict the compositional zonation of the Lau Basin proposed by Hawkins and Melchior
(1985) and shown by the distribution of sample types in Fig. 3.6. They should be transi-
tional in composition like the samples dredged from nearby ridges .. Some Site 834 petro-
logic units resemble Mariana Trough basalts in major elements and HFSE, but they lack
LILE enrichment. A possible explanation may be that the long-term melting that partly
filled the subbasin at Site 834 thoroughly depleted the source in the LILE-enriched SSZ
component. Continued melting of the mantle tapped the depleted MORB-source compo-
nent which dominated the later stages of melting. The rocks we drilled recorded only the
products of the later melting.
Site 834 was fed mainly by a "Pacific" MORB source, as shown by Hergt and
Hawkesworth (1994), whereas Sites 837 to 839 had an SSZ source with some mixture of the
Pacific source. Site 836 magmas, and the modem spreading centers, were derived from an
"Indian" MORB source, as originally proposed by Volpe et al. (1988). The two-
dimensional sketch does not show the complexity of the probable mixing pattern. The
influx of the Indian MORB source would have swept down the axis of the basin from the
north in the wake of the propagating rift tip of the ELSC. The trace of a "Samoan plume
component" in the northern part of the basin (Poreda, 1985) offers further support for this
idea. The Pacific MORB source either was consumed, bypassed, or overwhelmed by the
Indian source as the basin evolved. The final phase of evolution included the initiation and
southward propagation of a second rift system at about 1 to 2 Ma that formed the CLSC.
Mixing of MORB-like and arc magmas has been important throughout the history of
opening of the basin (Hergt and Farley, 1994).
The development of a new volcanic arc (Tofua arc) on the outboard edge of the Lau
Basin probably occurred within the past 3 m.y. and has propagated from north to south
down the axis of the Tofua arc. A complexity not shown because of limitations of the scale
of the sketches is the seaward march of the small, intrabasinal, arc composition volcanoes
that contributed to the infilling of the rift basins. These may have been no less significant at
any time than the scattering of arc volcanoes that constitute the present Tofua arc. Thus, the
long-standing question of whether there was synchronous backarc and arc magmatism has
an indefinite answer in the Lau Basin. Both phenomena have been operative more or less
continuously within the basin as it opened. However, a well-defined arc on the Tonga Ridge
probably did not appear until after about 6 m.y. of backarc spreading.
The Lau-Tonga system serves as a model for understanding relict convergent margin
rocks found as terranes accreted to continental margins and as ophiolite series assemblages.
It seems likely that eventually plate reorganization will lead to the amalgamation of the arc
and backarc system of the western Pacific and their eventual accretion to a continent. There
is ample evidence in areas of composite terranes such as the Philippines for amalgamation
of arc-backarc-forearc systems like the Lau-Tonga system (e.g., Hawkins and Evans, 1983;
Hawkins et aI., 1985). There is also general acceptance of the idea that many ophiolites are
remnants of SSZ magma systems. In many ways the Leg 135 transect, together with the
other western Pacific ODP studies, provides important new data and insight to understand-
ing continental evolution through terrane accretion.
Acknowledgments
This synthesis summarizes results of many years of study of the Lau Basin and its
surroundings by the author as well as the work of many others cited in the text. Within the
last five years there have been a number of intensive studies with wide-beam echo sounding
(SeaBeam), sonar imaging (GLORIA), submersibles, and drilling by the Ocean Drilling
Program. Shipmates on ODP Leg 135 are acknowledged for their contributions toward a
better understanding of Lau Basin geology. The ODP work in particular has brought
together a wide range of studies that have given us new understanding of the complex
history of the Lau Basin. This summary has drawn extensively on the collected works of the
ODP and especially from the Scientific Results Volume for ODP Leg 135. The author
acknowledges support received from the National Science Foundation, in particular from
awards OCE 84-15472, OCE 87-16660, JOI JSC 5-89, and Texas A&M 20556. I thank Jo
Griffith, Evelyn Hegemier, and Fred Florendo for making the illustrations. Fred Hochstaed-
ter and Kevin Johnson are thanked for many helpful suggestions that have improved the
original manuscript.
REFERENCES
Allan, 1. E 1994. Cr-spinel in depleted basalts from the Lau Basin backarc, ODP Leg 135; petrogenetic history
from Mg-Pe crystal-liquid exchange, in Proc. ODp, Sci. Results, 135 (1. W. Hawkins, L. M. Parson, and J. E
Allan, et aI., eds.), pp. 565-584, Ocean Drilling Program, College Station, TIC
Allan, J. E, Sack, R. 0., and Batiza, R. 1988. Cr-rich spinels as petrogenetic indicators: MORB-type lavas from the
Lamont seamount chain, Eastern Pacific, Am. Mineral. 73:741-753.
Arculus, R. J., and Powell, R. 1986. Source component mixing in the regions of arc magma generation, 1. Geophys.
Res. 91:5913-5926.
Barazangi, M., and Isacks, B. 1971. Lateral variations of seismic-wave attenuation in the upper mantle above the
inclined earthquake zone of the Tonga Island Arc: Deep anomaly in the upper mantle, 1. Geophys. Res.
76:8493-8516.
Bednarz, U., and Schmincke, H. U. 1994. Composition and origin of volcaniclastic sediments in the Lau Basin
(southwest Pacific), Leg 135, (Sites 834-839), in Proc. ODp, Sci. Results, 135 (J. W. Hawkins, L. M. Parson,
and J. F. Allan, et aI., eds.), pp. 51-74, Ocean Drilling Program, College Station, TX.
Ben Othman, D., White, W. M., and Patchett, J. 1989. The geochemistry of marine sediments, island arc magma
genesis, and crust-mantle recycling, Earth Planet. Sci. Lett. 94:1-21.
Berthine, K. K. 1974. Origin of Lau Basin Rise sediment, Geochim. Cosmochim. Acta 38:629-640.
Berthine, K. K., and Keene, J. B.1975. Submarine barite-opal rocks of hydrothermal origin, Science 188:150-152.
Bevis, M., Taylor, F. W., Schutz, B. E., and Calmant, S. 1993. Geodetic observations of convergence and back-arc
spreading in the SW Pacific (1988-1992), Eos 74:60.
Billington, S. 1980. The morphology and tectonics of the subducted lithosphere in the Tonga-Kermadec-Fiji
region from seismicity and focal mechanism solutions, Ph. D. dissertation, Cornell University, Ithaca, NY.
Bloomer, S. H., Ewart, A., Hergt, J. and Bryan, W. B. 1994. Geochemistry and origin of igneous rocks from the
outer Tonga forearc, Site 841, in Proc. ODp, Sci. Results, 135, (J. W. Hawkins, L. M. Parson, and J. F. Allan
et aI., eds.), pp. 625-646, Ocean Drilling Program, College Station, TX.
Bloomer, S. H., and Fisher, R. L. 1987. Petrology and geochemistry of igneous rocks from the Tonga Trench-a
non-accreting plate boundary, 1. Geol. 95:469-495.
Bloomer, S. H., and Hawkins, J. W. 1987. Petrology and geochemistry ofboninite series volcanic rocks from the
Mariana Trench, Contrib. Mineral. Petrol. 97:361-377.
Boespfiug, X., Dosso, L., Bougault, H., and Joron, J.-L. 1990. Trace element and isotopic (Sr, Nd) geochemistry of
volcanic rocks from the Lau Basin, Geol. lahrb. 92:503-516.
Briqueu, L., Bougault, H., and Joron, J.-L. 1984. Quantification of Nb, Ta, Ti, and V anomalies in magmas
associated with subduction zones: Petrogenetic implications, Earth Planet. Sci. Lett. 68. 297-308.
Bryan, W. B. 1979a. Low K 20 dacite from the Tonga-Kermadec island arc: Petrography, chemistry, and
petrogenesis, in Trondhjemites, Dacites, and Related Rocks (F. Barker, ed.), pp. 581-600, Elsevier, Am-
sterdam.
Bryan, W. B. 1979b. Regional variation and petrogenesis of basalt glasses from the FAMOUS area, Mid-Atlantic
Ridge, 1. Petrol. 20:293-325.
Bryan, W. B., and Pearce, T. H. 1994. Plagioclase zoning in selected lavas from Holes 834B, 839B, and 8418, in
Proc. ODp, Sci. Results, 135 (J. W. Hawkins, L. M. Parson, J. F. Allan, et aI., eds.), pp. 543-556, Ocean
Bryan, W. B., Stice, G. D., and Ewart, A. E. 1972. Geology, petrography, and geochemistry of the volcanic islands
of Tonga, 1. Geophys. Res. 77:1566-1585.
Bryan, W. B., Thompson, G., Frey, F. A, and Dickey, J. S. 1976. Inferred geologic settings and differentiation in
basalts from the Deep Sea Drilling Project, 1. Geophys. Res. 81:4285-4304.
Carlson, R. W., Macdougall, J. D. and Lugmair, G. W. 1978. Differential SmlNd evolution in oceanic basalts,
Geophys. Res. Lett. 5:229-232.
Carlson, R. L., and Melia, P. J. 1984. Subduction hinge migration, in Geodynamics of Back-Arc Regions (R. L.
Carlson and K. Kobayashi, eds.), Tectonophysics, 102:399-411.
Cawood, P. A 1985. Petrography, phase chemistry, and provenance of volcanogenic debris from the southern
Tonga Ridge: Implications for arc history and magmatism, in Geology and Offshore Resources of Pacific
Island Arcs-Tonga Region (D. Scholl and T. Vallier, eds.), pp. 149-170, Circum-Pacific Council for Energy
and Mineral Resources, Houston, TX.
Chase, C. G. 1978. Extension behind island arcs and motions relative to hot spots, 1. Geophys. Res. 83:5385-5387.
Clift, P., and Dixon, J. E. 1994. Variations in arc volcanism and sedimentation related to rifting of the Lau Basin
(Southwest Pacific), in Proc. ODP, Sci. Results, 135 (J. W. Hawkins, L. M. Parson, J. F. Allan, et aI., eds.), pp.
23-50, Ocean Drilling Program, College Station, TX.
Cole, J. w., Gill, J. B., and Woodhall, D. 1985. Petrologic history of the Lau Ridge, Fiji, in Geology and Offshore
Resources of Pacific Island Arcs-Tonga Region (D. Scholl and T. Vallier, eds.), pp. 379-414, Circum-
Pacific Council for Energy and Mineral Resources, Houston, TX.
Colley, H., and Hindle, W. H. 1984. Volcano-tectonic evolution of Fiji and adjoining marginal seas, in Marginal
Basin Geology (B. P. Kokelaar and M. F. Howells, eds.), pp. 151-162, Blackwell, Oxford.
Colley, H., Hindle, H. w., and Hathaway, B. 1986. Early arc and basinal igneous activity on Viti Levu, Fiji,
IAVCEI Congress, New Zealand, p. 140 (Abstract).
Colley, H., Wharton, M. R., and Hathaway, B. 1993. Magmatism and initial arc formation in Fiji, in Ancient
Volcanism and Modem Analogues, IAVCEI General Assembly, Canberra, Australia, p. 22 (Abstract).
Collier,1. S., and Sinha, M. C. 1992. Seismic mapping of a magma chamber beneath the Valu Fa Ridge, Lau Basin,
1. Geophys. Res. 97:14,031-14,053.
Crawford, A 1., Falloon, T. J., and Green, D. H. 1989. Classification, petrogenesis and tectonic setting of
boninites, in Boninites and Related Rocks (A. 1. Crawford, ed.), pp. 2-49, Unwin Hyman, London.
Cronan, D. S., Moorby, S. S., Glasby, G. P., Knedler, L., Thomson, J., and Hodkinson, R. A 1984. Hydrothermal
and volcaniclastic sedimentation on the Tonga-Kermadec Ridge and its adjacent marginal basins, Geol. Soc.
London Spec. Publ. 16:137-149.
Cunningham, J. K, and Anscombe, K. J. 1985. Geology of 'Eua and other islands, Kingdom of Tonga, in Geology
and Offshore Resources of Pacificlsland Arcs-Tonga Region (D. Scholl and T. Vallier, eds.), pp. 221-258,
Circum-Pacific Council for Energy and Mineral Resources, Houston, TX.
Davis, A. S., Clague, D. A., and Morton, 1. L. 1990. Volcanic glass compositions from two spreading centers in
Lau Basin, southwest Pacific Ocean, Geol. lahrb. 92:481-501.
DePaolo, D. 1., and Wasserburg, G. 1. 1977. The sources of island arcs as indicated by Nd and Sr isotope studies,
Geophys. Res. Lett. 4:465-468.
Dick, H. 1. B. 1982. The petrology of two back arc basins of the northern Philippine Sea, Am. 1. Sci. 282:644-700.
Dick, H. 1. B., and Bullen, T. 1984. Chromian spinel as a petrogenetic indicator in abyssal and alpine-type
peridotites and spatially associated lavas, Contrib. Mineral. Petrol. 86:54-76.
Duncan, R. A, Vallier, T., and Falvey, D. A. 1985. Volcanic episodes on 'Eua, Tonga, in Geology and Offshore
Pacific Council for Energy and Mineral Resources, Houston TX.
Dupuy, c., Dostal, J., Marcelot, G., Bougault, H., Joron, J. L., and Treuil, M. 1982. Geochemistry of basalts from
central and southern New Hebrides arc: Implications for their source rock composition, Earth Planet. Sci.
Lett. 60:207-225.
Eguchi, T. 1984. Seismotectonics of the Fiji Plateau and Lau Basin, Tectonophysics 102:17-32.
Elsasser, W. M. 1971. Sea-floor spreading as thermal convection, 1. Geophys. Res. 76:1101-1112.
Ernewein, M., Pearce, J. A., Bloomer, S. H., Parson, L. M., Murton, B. J., and Johnson, L. E. 1994. Geochemistry
of Lau Basin volcanic rocks: Influence of ridge segmentation and arc proximity, in Volcanism Associated
with Extension at Consuming Plate Margins (1. Smellie, ed.), Geo!. Soc. Spec. Pub!., London.
Ewart, A. 1979. A review of the mineralogy and chemistry of Tertiary-Recent dacitic, latitic, rhyolitic, and
related salic volcanic rocks, in Trondhjemites, Dacites, and Related Rocks (F. Barker, ed.), pp. 13-122,
Amsterdam Elsevier.
Ewart, A., Brothers, R. N., and Mateen, A. 1977. An outline of the geology and geochemistry, and the possible
petrogenetic evolution of the volcanic rocks of the Tonga-Kermadec-New Zealand island arc, 1. Volcanol.
Ewart, A., and Bryan, W. B. 1972. Petrography and geochemistry of the igneous rocks from 'Eua, Tongan Islands,
Geol. Soc. Am. Bull. 83:3281-3298.
Ewart, A., Bryan, W. B., Chappell, B. w., and Rudnick, R. L. 1994a. Regional geochemistry of the Lau-Tonga arc
and back-arc systems, in Proc. ODp, Sci. Results, 135 (J. W. Hawkins, L. M. Parson, J. F. Allan et al., eds.),
pp. 385-426. Ocean Drilling Program, College Station, TX.
Ewart, A., Bryan, W. B., and Gill, J. B. 1973. Mineralogy and geochemistry of the younger Tongan Islands, S. W.
Pacific, 1. Petrol. 14:429-466.
Ewart, A., Hergt, J., and Hawkins, J. W. 1994b. Major, trace element and Pb, Sr, and Nd isotope geochemistry of
Site 839 basalts and basaltic andesites: Implications for arc volcanism, in Proc. ODp, Sci. Results, 135 (J. W.
Hawkins, L. M. Parson, J. F. Allan et aI., eds.), pp. 519-532. Ocean Drilling Program, College Station, TX.
Ewart, A E., and Hawkesworth, C. 1. 1987. Pleistocene to Recent Tonga- Kermadec arc lavas: Interpretation of
new isotopic and rare earth data in terms of a depleted mantle source model, 1. Petrol. 28:495-530.
Exon, N. F., Herzer, R. H., and Cole, J. W. 1985. Mixed volcaniclastic and pelagic sedimentary rocks from the
Cenozoic southern Tonga Platform and their implications for petroleum potential, in Geology and Offshore
Resources of Pacific Island Arcs-Tonga Region (D. Scholl and T. Vallier, eds.), pp. 75-108, Circum-Pacific
Council for Energy and Mineral Resources, Houston, TX.
Falloon, T. J., Malahoff, A., Zonenshain, L. P., and Bogdanov, Y. 1992. Petrology and geochemistry of back-arc
basin basalts from Lau Basin spreading ridges at 15, 18, and 19°5, Mineral. Petrol. 47:1-36.
Farley, K. N. 1994. Oxidation state and sulfur concentrations in Lau Basin basalts, in Proc. ODp, Sci. Results, 135
(I. W. Hawkins, L. M. Parson, 1. F. Allan et al., eds.), pp. 603-614, Ocean Drilling Program, College Sta-
tion, TX.
Fisher, R. L., and Engel, C. G. 1969. Ultramafic and basaltic rocks dredged from the nearshore flank of the Tonga
Trench, Geol. Soc. Am. Bull. 80:1373-1378.
Fisk, M. R., and Bence, A. E. 1980. Experimental crystallization of chrome spinel in FAMOUS basalt 527-1-1,
Earth Planet. Sci. Lett. 48:113-121.
Fouquet, Y., von Stackelberg, U., Charlou, I. L., Donval, 1. P., Foucher, I. P., Erzinger, I., Herzig, P., Muhe, R.,
Weidicke, M., Soakai, S., and Whitechurch, H. 1991a. Hydrothermal activity in the Lau back-arc basin:
Sulfides and water chemistry, Geology 19:303-306.
Fouquet, Y., von Stackelberg, U., Charlou, 1. L., Donval, I. P., Erzinger, I., Foucher, I. P., Herzig, P., Muhe, R.,
Soakai, S., Weidicke, M., and Whitechurch, H. 1991b. Hydrothermal activity and metallogenesis in the Lau
back-arc basin, Nature 348:778-781.
Frenzel, G., Muhe, R., and Stoffers, P. 1990. Petrology of the volcanic rocks from the Lau Basin, southwest Pacific,
Geol. Jahrb. 92:395-479.
Fryer, P., and Pearce, I. A. 1992. Introduction to the scientific results of Leg 125, in Proc. ODp, Sci. Results, 125 (P.
Fryer, I. A. Pearce, L. B. Stokking et al. eds.), pp. 3-14, Ocean Drilling Program, College Station, TX.
Fryer, P., Sinton, I., and Philpotts, I. A. 1981. Basaltic glasses from the Mariana Trough, in Init. Repts. DSDP, 60
(D. M. Hussong, S. Uyeda et aI., eds.), pp. 601-610, U. S. Govt. Printing Office, Washington, DC.
Furlong, K. P., Chapman, D. S., and Alfeld, P. W. 1982. Thermal modeling of the geometry of subduction with
implications for the tectonics of the overriding plate, J. Geophys. Res. 87:1786-1802.
Garcia, M., Liu, N. W. K., and Muenow, D. W. 1979. Volatiles from submarine volcanic rocks from the Mariana
Island Arc and Trough, Geochim. Cosmochim. Acta 43:305-312.
Giardini, D., and Woodhouse, I. H. 1984. Deep seismicity and modes of deformation in Tonga subduction zone,
Nature 307:505-509.
Giardini, D., and Woodhouse, I. H. 1986. Horizontal shear flow in the mantle beneath the Tonga arc, Nature
319:551-555.
Gill, I. B. 1976. Composition and age of Lau Basin and ridge volcanic rocks: Implications for evolution of an
interarc basin and remnant arc, Geol. Soc. Am. Bull. 87:1384-1395.
Gill, I. B. 1981. Orogenic Andesites, Springer-Verlag, Berlin.
Gill, I. B. 1987. Early geochemical evolution of an oceanic island arc and backarc: Fiji and the South Fiji Basin, J.
Geol. 95:589-615.
Hamburger, M. W., and Isacks, B. L. 1988. Diffuse back-arc deformation in the southwestern Pacific, Nature
332:599-604.
Hart, S. R. 1969. K, Rb, Cs contents and K1Rb, KlCs ratios of fresh and altered submarine basalts, Earth Planet.
Sci. Lett. 6:295-303.
Hart, S. R., Erlank, A. I., and Kable, E. I. 1974. Sea floor basalt alteration: Some chemical and strontium isotope
effects, Contrib. Mineral. Petrol. 44:219-230.
Hart, S. R., Glassley, W. A., and Karig, D. E. 1972. Basalts and seafloor spreading behind the Mariana Island arc,
Earth Planet. Sci. Lett. 15: 12-18.
Hart, S. R., and Nawalk, A. 1. 1970. K, Rb, Cs, and Sr relationships in submarine basalts from the Puerto Rico
Trench, Geochim. Cosmochim. Acta 34:145-155.
Hart, S. R. and Staudigel, H. 1989. Isotopic characterization and identification of recycled components, in Crust!
Mantle Recycling at Convergence Zones (S. R. Hart and L. Gulen, eds.), pp. 15-28, Kluwer, Boston.
Hawkins, I. W. 1974. Geology of the Lau Basin, a marginal sea behind the Tonga Arc, in The Geology of
Continental Margins (c. A. Burk and C. L. Drake, eds.), pp. 505-520, New York, Springer-Verlag.
Hawkins, 1. W. 1976. Petrology and geochemistry of basaltic rocks of the Lau Basin, Earth Planet. Sci. Lett.
28:283-297.
Hawkins, I. W. 1977. Petrological and geochemical characteristics of marginal basin basalts, in Island Arcs, Deep
Sea Trenches, and Back Arc Basins (M. Talwani and W. C. Pitman III, eds.), Vol. 1, pp. 355-365, Maurice
Ewing Series, American Geophysical Union, Washington, DC.
Hawkins, I. W. 1985. Low-K rhyolitic pumice from the Tonga Ridge, in Geology and Offshore Resources of
Pacific Island Arcs-Tonga Region (D. Scholl and T. Vallier, eds.), pp. 171-178, Circum-Pacific Council for
Energy and Mineral Resources, Houston, TX.
Hawkins, I. W. 1986. "Black smoker" vent chimneys, Lau Basin, Eos 67:430.
Hawkins,1. W. 1988. Cruise Report-PAPATUA expedition. Leg 04, RN Thomas Washington, SIO Ref Ser. 88-14.
Hawkins, J. W. 1989. Cruise Report-ROUNDABOUT expedition. Legs 14,15, RIV Thomas Washington, SIO
Ref Ser. 89-13.
Hawkins, J. W 1994. Petrologic synthesis, ODP Leg 135-Lau Basin transect, in Proc. ODp, Sci. Results, 135 (J.
W Hawkins, L. M. Parson, J. E Allan, et aI., eds.), pp. 879-908, Ocean Drilling Program, College Station,
TX.
Hawkins, 1. W, and Allan, 1. E 1994. Petrologic evolution of the Lau Basin, Sites 834-839, in Proc. ODp, Sci.
Results, 135 (1. W Hawkins, L. M. Parson, J. E Allan, et aI., eds.), pp. 426-470, Ocean Drilling Program,
College Station, TX.
Hawkins, J. W, and Evans, C. A. 1983. Geology of the Zambales Range, Luzon, Philippines: Ophiolite derived
from an island-arc backarc basin pair, in The Tectonic Evolution of Southeast Asian Seas and Islands, Part 2
(D. Hayes, ed.), Vol. 27, pp. 124-138. Geophys. Monogr. Ser., American Geophysical Union, Washing-
ton, DC.
Hawkins, 1. W., Bloomer, S. H., Evans, C. A, and Melchior, J. T. 1984. Evolution of intra-oceanic arc-trench
systems, Tectonophysics 102:174-205.
Hawkins, 1. W, and Falvey, D. A. 1985. Petrology of andesitic dikes and flows from 'Eua, Tonga, in Geology and
Offshore Resources of Pacific Island Arcs-Tonga Region (D. Scholl and T. Vallier), pp. 269-280, Circum-
Hawkins, J. W, Fisher, R L., and Engel, C. G. 1972. Ultramafic and mafic rock suites exposed on the deep flanks
of the Tonga Trench, Geol. Soc. Am. Abs. Prog. 4:167-168 (Abstract).
Hawkins,1. W, and Helu, S. 1986. Polymetallic sulphide deposit from "black-smoker" chimney: Lau Basin, Eos
67:378 (Abstract).
Hawkins, J. W, Lonsdale, P. E, Macdougall, J. D., and Volpe, A. M. 1990. Petrology of the axial ridge of the
Mariana Trough backarc spreading center, Earth Planet. Sci. Lett. 100:226-250.
Hawkins,1. W., and Melchior, J. T. 1985. Petrology of Mariana Trough and Lau Basin basalts, J. Geophys. Res.
90:11431-11468.
Hawkins, 1. W, Melchior, 1. T., Florendo, E E, and Nilsson, K. N. 1989. Evolution of backarc basin magmas and
their mantle sources-examples from the Lau Basin and Mariana Trough, Eos 70:1389.
Hawkins, J. W., Moore, G. E, Villamor, R, Evans, S., and Wright, E. 1985. Geology of the composite terranes of
east and central Mindanao, in Tectonostratigraphic Terranes of the Circum-Pacific Region (D. G. Howell,
ed.), pp. 437-463, Circum-Pacific Council for Energy and Mineral Resources, Earth Sci. Ser., Vol. I,
Houston, TX.
Hawkins, 1. W, Parson, L. M., Allan, J. E, and Leg 135 Scientific Party. 1991. New insight to the evolution of arc-
backarc systems, results of Ocean Drilling Program (ODP) Leg 135, Lau-Tonga transect, Eos 72:541.
Hawkins, 1. W, Parson, L. M., and Allan, J. E et al. 1994. Proc. ODP, Sci. Results, 135, Ocean Drilling Program,
Hawkins, J. W., Sclater, J. G., and Hohnhaus, G. 1970. Petrologic and geophysical characteristics of the Lau Basin
Ridge: A spreading center behind the Tonga arc, Geol. Soc. Am. Bull. 2:71, (Abstract).
Hedge, C. E., and Peterman, Z. E. 1970. The strontium isotopic composition of basalts from the Gorda and Juan de
Fuca Rises, northeastern Pacific Ocean, Contrib. Mineral. Petrol. 27:114-120.
Hergt, J. M. and Farley, K. N. 1994. Major, trace element, and isotope (Pb, Sr, and Nd) variations in Site 834
basalts: Implications for the initiation of backarc opening, in Proc. ODp, Sci. Results, 135 (J. W. Hawkins, L.
M. Parson, J. E Allan et al. eds.), pp. 471-486, Ocean Drilling Program, College Station, TX.
Hergt, J. M., and Hawkesworth, C. 1. 1994. The Pb, Sr, and Nd isotopic evolution of the Lau Basin: Implications
for mantle dynamics during the backarc opening, in Proc. ODP, Sci. Results, 135 (1. W. Hawkins, L. M.
Parson, J. E Allan et aI., eds.), pp. 505-518, Ocean Drilling Program, College Station, TX.
Hochstaedter, A. G., Gill, J. B., Kusakabe, M., Newman, S., Pringle, M., Taylor, B., and Fryer, P.1990a. Volcanism
in the Sumisu Rift. I: Major element, volatile, and stable isotope geochemistry, Earth Planet Sci. Lett.
100:179-194.
Hochstaedter, A. G., Gill, J. B., and Morris, J. D. 1990b. Volcanism in the Sumisu Rift, II: Subduction and non-
Hodkinson, R. A., and Cronan, D. S. 1991. Geochemistry of recent hydrothermal sediments in relation to tectonic
environment in the Lau Basin, southwest Pacific, Mar. Geol. 98:353-366.
Hodkinson, R A., and Cronan, D. S. 1994. Variability in the hydrothermal component of the sedimentary sequence
in the Lau back-arc basin: ODP Leg 135 Sites 834A, 835A, 836A, 837 A, 838A and 839A, in Proc. ODp, Sci.
Results, 135 (1. W Hawkins, L. M. Parson, J. E Allan, et al., eds.), pp. 75-86, Ocean Drilling Program,
Hodkinson, R A, Cronan, D. S., Glasby, G. P., and Moorby, S. A 1986. Geochemistry of marine sediments from
the Lau Basin, N. Z. J. Geol. Geophys. 29:335-344.
Hole, M. J., Saunders, A. D., Marriner, G. E, and Tarney, J. 1984. Subduction of pelagic sediments: Implications
for the origin of Ce-anomalous basalts from the Mariana Islands, J. Geol. Soc. London 141:453-472.
Hynes, A, and Mott, 1. 1985. On the causes of back-arc spreading, Geology, 13:387-389.
Ikeda, Y., and Yuasa, M. 1989. Volcanism in nascent back-arc basins behind the Shichito Ridge and adjacent areas
in the Izu-Ogasawara arc, northwest Pacific: Evidence for mixing between E-type MORB and island arc
magmas at the initiation of back-arc rifting, Contrib. Mineral. Petrol. 101:377-393.
Irvine, T. N. 1965. Chromian spinel as a petrogenetic indicator. I: Theory. Can. J. Earth Sci. 2:648-672.
Irvine, T. N. 1967. Chromian spinel as a petrogenetic indicator, 2, Petrological applications, Can. J. Earth Sci.
4:71-103.
Isacks, B. L., and Barazangi, M. 1977. Geometry of Benioff zones: Lateral segmentation and downwards bending
of the subducted lithosphere, in Island Arcs, Deep Sea Trenches, and Back Arc Basins (M. Talwani and W. C.
Pitman III, eds.), Vol. I, pp. 94-114, Maurice Ewing Series, American Geophysical Union, Washington, DC.
Jenner, G. A, Cawood, P. A, Rautenschlein, M., and White, W. M. 1987. Composition of back-arc basin
volcanics, Valu Fa Ridge, Lau Basin: Evidence for a slab-derived component in their mantle source, J.
Volcanol. Geotherm. Res. 32:209-222.
Johnson, L. E., and Fryer, P. 1990. The first evidence for MORB-like lavas from the outer Mariana forearc:
Geochemistry, petrography and tectonic implications, Earth Planet. Sci. Lett. 100:304-316.
Karig, D. E. 1970. Ridges and basins of the Tonga-Kermadec Island arc system, J. Geophys. Res. 75:239-254.
Karig, D. E. 1971. Origin and development of marginal basins in the western Pacific, J. Geophys. Res. 76:2542-
2561.
Kay, R W., Hubbard, N., and Gast, P. 1970. Chemical characteristics and origins of oceanic ridge volcanic rocks,
J. Geophys. Res. 75:238-254.
Kelemen, P. B. 1986. Assimilation of ultramafic rocks in subduction related magmatic arcs, J. Geoi. 94:829-843.
Kelemen, P. B., Johnson, K. T. M., Kinzler, R J., and Irving, A. J. 1990. High-field strength element depletions in
arc basalts due to mantle-magma interaction, Nature 345:521-524.
Klein, E. M., and Langmuir, C. H. 1987. Global correlations of ocean ridge basalt chemistry with axial depth and
crustal thickness, J. Geophys. Res. 92:8089-8115.
Kroenke, L. 1984. Cenozoic Development of the Southwest Pacific CCOP/SOPAC, Tech. Bull., No.6. U.N. Econ.
Soc. Comm. Asia Pac., Suva, Fiji.
Lawver, L. A., and Hawkins, J. W. 1978. Diffuse magnetic anomalies in marginal basins: Their possible tectonic
and petrologic significance, Tectonophysics 45:323-339.
Lawver, L., Hawkins, J. W., and Sclater, J. G. 1976. Magnetic anomalies and crustal dilation in the Lau Basin,
Ledbetter, 1. K., and Haggerty, J. A. 1994. Late Miocene sedimentation history of the Tonga forearc at Site 840, in
Proc. ODp, Sci. Results, 135 (1. W. Hawkins, L. M. Parson, J. E Allan et ai., eds.), pp. 163-172, Ocean
Loock, G., McDonough, W. E, Goldstein, S. L. and Hoffman, A W. 1990. Isotopic composition of volcanic
glasses from the Lau Basin, Mar. Mining 9:235-245.
Louat, R, and Dupont, J. 1982. Seismicite'de l'arc des Tonga-Kermadec, in Equipe de Geologie-Geophysique du
Centre ORSTOM de Noumea, contribution a' I 'etude geodynamique du Sud-Ouest Pacifique, pp. 299-317,
Trav. Doc. 147, ORSTOM (Nouvelle Caledonie').
Ludden, 1. N., and Thompson, G. 1979. An evaluation of the behavior of the rare earth elements during the
weathering of sea-floor basalt, Earth Planet. Sci. Lett. 43:85-92.
Malahoff, A., Feden, R. H., and Fleming, H. S. 1982. Magnetic anomalies and tectonic fabric of marginal basins
north of New Zealand, J. Geophys. Res. 87:4109-4125.
Masuda, A., Nakamura, N., and Tamaka, T. 1973. Fine structure of mutually normalized rare-earth element
patterns of chondrites, Geochim. Cosmochim. Acta. 37:239-248.
Mattey, D. P., Marsh, N. G. and Tarney. J. 1980. The geochemistry, mineralogy. and petrology of basalts from the
west Philippine and Parece Vela Basins and from the Palau-Kyushu and West Mariana Ridges, DSDP Leg 59.
in Init. Repts. DSDP. 59 (L. Kroenke and R Scott, et al., eds.), pp. 753-800, U. S. Govt. Printing Office,
Washington, DC.
McDougall, I. 1994. Dating of rhyolitic glass in the Tonga forearc (Hole 841 B), in Proc. ODP Sci. Results. 135
(1. W. Hawkins, L. M. Parson, and J. E Allan, eds.), pp. 923-924, Ocean Drilling Program, College Station, TX.
McKenzie, D., and Bickle, M. J. 1988. The volume and composition of melt generated by extension of the
lithosphere, 1. Petrol. 29:625-679.
Meijer, A 1976. Pb and Sr isotopic data bearing on the origin of volcanic rocks from the Mariana island-arc
system, Bull. Geol. Soc. Am. 87:1358-1369.
Melson, W. G., Jarosewich, E., and Lundquist, C. A 1970. Volcanic eruption at Metis Shoal, Tonga, 1967-1968:
Description and petrology, Smithsonian Contrib. Eanh. Sci. 4:1-18.
Minster, J. B., and Jordan, T. H. 1978. Present day plate motions, 1. Geophys. Res. 83:5331-5354.
Moberly, R. 1972. Origin of lithosphere behind island arcs with reference to the western Pacific, Geol. Soc. Am.
Mem. 132:35-55.
Morris, 1. D., and Hart, S. R. 1983. Isotopic and incompatible element constraints on the genesis of island arc
volcanics from Cold Bay and Amak Island, Aleutians, amd implications for mantle structure, Geochim.
Cosmochim. Acta. 47:2015-2-30.
Morton,1. and Sleep, N. H. 1985. Seismic reflections from a Lau Basin magma chamber, in Geology and Offshore
Morton, J., and Pohl, W. 1990. Magnetic anomaly identification in the Lau Basin and North Fiji Basin, southwest
Pacific Ocean, Geol. lahrb. D. 92:93-108.
Muenow, D. W., Liu, N. W. K., Garcia, M. 0., and Saunders, A. D. 1980. Volatiles in submarine volcanic rocks
from the spreading axis of the East Scotia Sea backarc basin, Earth Planet. Sci. Lett. 47:272-278.
Newman, S. 1989. Water and carbon dioxide contents in the basaltic glasses from the Mariana Trough, Eos
70:1387.
Nilsson, K 1993. Oxidation state, sulfur speciation, and sulfur concentration in basaltic magmas: Examples from
Hess Deep and the Lau Basin, Ph. D. dissertation, Scripps Institution of Oceanography, University of
California-San Diego, La Jolla.
Nilsson, K, Florendo, F., and Hawkins, J. W. 1989. Petrology of a nascent triple junction, northeastern Lau Basin,
Eos 73:1389.
Nilsson, K, and Peach, C. L. 1993. Sulfur speciation, oxidation state, and sulfur concentration in backarc magmas,
Geochim. Cosmochim. Acta 57:3807-3813.
Packham, G. H., and Falvey, D. A 1971. An hypothesis for the formation of marginal seas in the western Pacific,
Parson, L. M., and Hawkins, 1. W. 1994. Two-stage ridge propagation and the geological history of the Lau backarc
basin, in Proc. ODP, Sci. Results, 135 (1. W. Hawkins, L. M. Parson, J. F. Allan et al., eds.), pp. 819-828,
Ocean Drilling Program, College Station, TX.
Parson, L. M., Hawkins, J. w., Allan, J. F., et al. 1992. Proc. ODP, Init. Repts., 135, Ocean Drilling Program,
Parson, L. M., Pearce, 1. A., Murton, B. J., and RRS Charles Darwin Scientific Party. 1990. Role of ridge jumps
and ridge propagation in the tectonic evolution of the Lau back-arc basin, SW Pacific, Geology 18:470-473.
Parson, L. M., Rothwell, R. G, and MacLeod, C. J. 1994. Tectonics and sedimentation in the Lau Basin, southwest
Pacific, in Proc. ODp, Sci. Results, 135 (1. W. Hawkins, L. M. Parson, J. F. Allan, et al., eds.), pp. 9-22, Ocean
Parson, L. M., and Tiffin, D. L. 1993. Northern Lau Basin: backarc extension at the leading edge of the Indo-
Australian plate, Geo-Marine Lett. 13:107-115.
Parson, L. M., et al. 1989. RRS Charles Darwin Cruise 33/88, 5 May-l June, 1988, Geophysical and geological
investigations of the Lau back-arc basin, SW Pacific. Inst. Oceanogr. Sci. Deacon Lab. Cruise Rep., No. 206.
Pearce, J. A., Lippard, S. J., and Roberts, S. 1984. Characteristics and tectonic significance of supra-subduction
zone ophiolites, in Geology ofMarginal Basins (P. Kokelaar and M. Howells, eds.), Geo!. Soc. London. Spec.
Pub!. 16:77-94.
Pearce,1. A., and Norry, M. J. 1979. Petrogenetic implications ofTi, Zr, Y, and Nb variations in volcanic rocks,
Contrib. Mineral. Petrol. 69:33-47.
Pearce, J. A, van der Laan, S. R., Arculus, R. J., Murton, B. 1., Ishii, T., Peate, D. w., and Parkinson, I. 1. 1992.
Boninite and harzburgite from Leg 125 (Bonin-Mariana forearc): A case study of magma genesis during the
initial stages of subduction, in Proc. ODP, Sci. Results, 125 (P. Fryer,1. A. Pearce, L. B. Stokking et aI., eds.),
pp. 623-659, Ocean Drilling Program, College Station, TX.
Pelletier, B., and Louat, R. 1989. Seismotectonics and present day relative plate motions in the Tonga-Lau and
Kermadec-Havre region, Tectonophysics 165:237-250.
Pineau, E, Javoy, M., Hawkins, 1., and Craig, H. 1976. Oxygen isotope variations in marginal basin and ocean
ridge basalts, Earth Planet. Sci. Lett. 28:299-307.
Poreda, R. 1985. Helium-3 and deuterium in backarc basin basalts: Lau Basin and Mariana Trough, Earth Planet.
Sci. Lett. 76:244-254.
Poreda, R., and Craig. H. 1993. He and Sr isotopes in the Lau Basin mantle: Depleted and primitive mantle
components, Earth Planet. Sci. Lett. 119:319-329.
Reay, A., Rooke, J. M., Wallace, R. C., and Whelan, P. 1974. Lavas from Niuafo'ou Island, Tonga, resemble
ocean-floor basalts, Geology 2:605-606.
Ribe, N. M. 1989. Mantle flow induced by back arc spreading, Geophys. J. Int. 98:85-91.
Roeder, P. L., and Reynolds, I. 1991. Crystallization of chromite and chromium solubility in basaltic liquid, J.
Petrol. 32:909-934.
Rothwell, R. G., Bednarz, U., Boe, R., Clift, P., Hodkinson, R. A., Ledbetter, J. K., Pratt, C. E., and Soakai, S.
1994. Sedimentation and sedimentary processes in the Lau Basin: Results from Leg 135 of the Ocean Drilling
Program, in Proc. ODp, Sci. Results, 135 (J. W. Hawkins, L. M. Parson, J. E Allan, et aI., eds.), pp. 101-130,
Ocean Drilling Program, College Station, TX.
Ryerson, E J., and Watson, E. B. 1987. Rutile saturation in magmas; implications for Ti-Nb-Ta depletion in island
arc basalts, Earth Planet. Sci. Lett. 86:225-239.
Scholl, D. W., Vallier, T. L., and Packham, G. H. 1985. Framework geology and resource potential of southern
Tonga Platform and adjacent terranes-a synthesis, in Geology and Offshore Resources of Pacific Island
Arcs-Tonga Region (D. Scholl and T. Vallier, eds.), pp. 457-488, Circum-Pacific Council for Energy and
Mineral Resources, Houston, TX.
Sclater, J. G., Hawkins, J. W., Mammerickx, J., and Chase, C. G. 1972. Crustal extension between the Tonga and
Lau Ridges; petrologic and geophysical evidence, Geol. Soc. Am. Bull. 83:505-518.
Shervais, 1. W. 1982. Ti-V plots and the petrogenesis of modem and ophiolitic lavas, Earth Planet. Sci. Lett.
59:101-118.
Sinton, J., and Fryer, P. 1987. Mariana Trough lavas from 18°N: Implications for the origin of back arc basin
basalts, J. Geophys. Res. 92:12782-12802.
Sinton, J. M., Wilson, D. S., Christie, D. M., Hey, R., and Delaney, J. R. 1983. Petrologic consequences of rift
propagation on oceanic spreading ridges, Earth Planet. Sci. Lett. 62:193-207.
Sleep, N., and Toksoz, M. N. 1971. Evolution of marginal basins, Nature 233:548-550.
Staudigel, H., and Hart, S. R. 1983. Alteration of basaltic glass; mechanisms and significance for the ocean's crust-
seawater budget, Geochim. Cosmochim. Acta 47:337-350.
Stem, R. J. 1982. Strontium isotopes from circum-Pacific intraoceanic island arcs and marginal basins: Regional
variations and implications for magmagenesis, Geol. Soc. Am. Bull. 93:477-486.
Stem, R. 1., and Bloomer, S. H. 1992. Subduction zone infancy: Examples from the Izu-Bonin-Mariana and
Jurassic California arcs, Geol. Soc. Am. Bull. 104:1621-1636.
Stem, R. J., Lin, P. N., Morris, J. D., Jackson, M. c., Fryer, P., Bloomer, S. H, and Ito, E. 1990. Enriched back-arc
basin basalts from the northern Mariana Trough: Implications for the magmatic evolution of back-arc basins,
Stem, R. J., Morris, 1., Bloomer, S. H., and Hawkins, J. W. 1991. The source of the subduction component in
convergent margin magmas: Trace element and radiogenic isotope evidence from Eocene boninites, Mariana
forearc, Geochim. Cosmochim. Acta 55:1467-1481.
Stem, R. J., Smoot, N. C., and Rubin, M. 1984. Unzipping of the Volcano Arc, Japan, Tectonophysics 102:153-174.
Stevenson, A. J. 1985. Dredging and coring; southern Tonga platform, in Geology and Offshore Resources of
Pacific Island Arcs-Tonga Region (D. Scholl and T. Vallier, eds.), pp. 31-38, Circum-Pacific Council for
Energy and Mineral Resources, Houston, TX.
Sun, S.-S., and McDonough, W. E 1989. Chemical and isotopic systematics of oceanic basalts: Implications for
mantle composition and processes, in Magmatism in the Ocean Basins (A. D. Saunders and M. J. Norry,
eds.), Geo!. Soc. Spec. Pub!., 42:313-346, Blackwell, London.
Sunkel, G. 1990. Origin of petrological and geochemical variation of Lau Basin lavas (SW Pacific), Mar. Mining
9:205-234.
Tamaki, K., Suyehiro, K., Allan, J., Ingle, J. C., and Pisciotto, K. A. 1992. Tectonic synthesis and implications of
Japan Sea ODP drilling, in Proc. ODP, Sci. Results, 127/128, Pt. 2 (K. Tamaki, K. Suyehiro, J. Allan, M.
McWilliams et aI., eds.), pp. 1333-1348, Ocean Drilling Program, College Station, TX.
Tappin, D. R., Bruns, T. R., and Geist, E. L. 1994. Rifting ofthe Tonga/Lau Ridge and formation of the Lau backarc
basin: Evidence from Site 840 on the Tonga Ridge, in Proc. ODp, Sci. Results, 135 (J. W. Hawkins, L. M.
Parson,1. F. Allan et al., eds.), pp. 367-372, Ocean Drilling Program, College Station, TX.
Tatsumi, Y 1989. Migration of fluid phases and genesis of basalt magmas in subduction zones, 1. Geophys. Res.
94:4697-4707.
Tatsumi, Y, Hamilton, D. L., and Nesbitt, R. W. 1986. Chemical characteristics of fluid phase released from a
subducted lithosphere and origin of arc magmas: Evidence from high pressure experiments and natural rocks,
1. Volcanol. Geotherm. Res. 29:293-309.
Taylor, B. 1992. Rifting and the volcanic-tectonic evolution of the Izu-Bonin-Mariana Arc, in Proc. ODP, Sci.
Results, 126 (B. Taylor and K. Fujioka, et aI., eds.), pp. 627-651, Ocean Drilling Program, College Sta-
tion, TX.
Taylor, P. W.1991. The geology and petrology of Niuafo'ou Island, Tonga: Subaerial volcanism in an active back-
arc basin, M.Sc. thesis, Macquarie University.
Taylor, B., Fujioka, K, et al. 1992. Proc. ODP, Sci. Results, 126, Ocean Drilling Program, College Station, TX.
Taylor, B., and Karner, G. D. 1983. On the evolution of marginal basins, Rev. Geophys. Space Phys. 21:1727-
1741.
Takazawa, E., Frey, F. A, Shimizu, N., Obata, M., and Bodinier, 1. L. 1992. Geochemical evidence for melt
migration and reaction in the upper mantle, Nature 359:55-58.
Uyeda, S., and Kanamori, H. 1979. Back-arc opening and the mode of subduction, 1. Geophys. Res. 84:1049-1061.
Vallier, T. L., Jenner, G. A., Frey, F. A, Gill, 1. B., Volpe, A. M., Hawkins, 1. w., Morris, J. D., Cawood, P. A,
Morton, J. L., Scholl, D. w., Rautenschlein, M., White, W. M., Williams, R. w., Stevenson, A. J., and White,
L. D. 1991. Subalkaline andesite from Valu Fa Ridge, a back-arc spreading center in the southern Lau Basin;
petrogenesis, comparative chemistry, and tectonic implications, Chern. Geol. 91:227-256.
Vallier, T. L., O'Conner, R. M., Scholl, D. w., Stevenson, A. 1., and Quinterno, P. 1985. Petrology of rocks dredged
from the landward slope of the Tonga Trench: Implications for middle Miocene volcanism and subsidence of
the Tonga Ridge forearc, in Geology and Offshore Resources of Pacific Island Arcs-Tonga Region (D.
Scholl and T. Vallier, eds.), pp., 109-120, Circum-Pacific Council for Energy and Mineral Resources,
Houston, TX.
Viereck, L. G., Rower, M. F. J., Hertogen, J., Schmincke, H. U., and Jenner, G. A 1989. The genesis and
significance of N-MORB sub-types, Contrib. Mineral. Petrol. 102:112-126.
Volpe, A. M., Macdougall, J. D., and Hawkins, J. W. 1987. Mariana Trough basalts (MTB): Trace element and Sr-
Nd isotopic evidence for mixing between MORB-like and arc-like melts, Earth Planet. Sci. Lett. 82:21-254.
Volpe, A M., Macdougall, J. D., and Hawkins, J. W. 1988. Lau Basin basalts (LBB): Trace element and Sr-Nd
isotopic evidence for heterogeneity in back-arc basin mantle, Earth Planet. Sci. Lett. 90:174-186.
Volpe, A M., Macdougall, J. D., Lugmair, G. w., Hawkins, J. w., and Lonsdale, P. F. 1990. Fine-scale isotopic
variation in Mariana Trough basalts: Evidence for heterogeneity and a recycled component in backarc basin
mantle, Earth Planet. Sci. Lett. 100:251-264.
von Stackelberg, U., and von Rad, U. 1990. Geological evolution and hydrothermal activity in the Lau and North
Fiji basins (Sonne cruise SO-35), Geol. lahrb. D92:629-660.
von Stackelberg, U., and Shipboard Party. 1985. Hydrothermal sulfide deposits in back-arc spreading centers in the
southwest Pacific, Bundes. Geowiss. Rohstoffe Circ. 2:2-14.
Wegener, A 1929. Die Entstehung der Kontinente und Ozeane (translated by 1. Biram, 1969, as The Origin of
Continents and Oceans), Dover, New York.
Weidicke, M., and Collier, J. 1993. Morphology of the Valu Fa spreading ridge in the southern Lau Basin, 1.
Geophys. Res. 98:11769-11782.
Weidicke, M., and Habler, W. 1993. Morphotectonic characteristics of a propagating spreading system in the
northern Lau Basin, 1. Geophys. Res. 98:11,783-797.
Weissel, J. K 1977. Evolution of the Lau Basin by the growth of small plates, in Island Arcs, Deep Sea Trenches,
and Back Arc Basins (M. Talwani and W. C. Pitman III, eds.), Vol. I, pp. 429-436. Maurice Ewing Series,
American Geophysical Union, Washington, DC.
Wharton, M. R., Hathaway, B., and Colley, H.1992. Volcanism associated with extension in the Vitiaz island arc,
Fiji, Symposium on Volcanism Associated with Extension at Consuming Plate Margins, Geological Society
of London (Abstract).
Whelan, P. M., Gill, J. B., Kollman, E., Duncan, R., and Drake, R. E. 1985. Radiometric dating of magmatic stages
in Fiji, in Geology and Offshore Resources of Pacific Island Arcs-Tonga Region, (D. Scholl and T. Vallier,
eds.), pp. 415-440, Circum-Pacific Council for Energy and Mineral Resources, Houston, TX.
White, W. M., and Patchett, J. 1984. Hf-Nd-Sr isotopes and incompatible element abundances in island arcs:
Implications for magma origins and crust-mantle evolution, Earth Planet. Sci. Lett. 67:167-185.
Wood, D. A. 1980. The application of a Th-Hf-Ta diagram to problems of tectonomagmatic classification and to
establishing the nature of crustal contamination of basaltic lavas of the British Tertiary volcanic province,
Wood, D. A., Marsh, N. G., Tamey, J., Joron, J. L., Cotten, J., and Treuil, M. 1981. Geochemistry of igneous rocks
recovered from a transect across the Mariana Trough, Are, Fore-arc, and Trench, Sites 453 through 461, Deep
Sea Drilling Project Leg 60, in Init. Repts. DSDP, 60 (D. M. Hussong and S. Uyeda, et al., eds.), pp. 611-646,
U. S. Gov!. Printing Office, Washington, DC.
Woodhall, D., 1985, Geology of the Lau Ridge, in Geology and Offshore Resources of Pacific Island Arcs-Tonga
Region (D. Scholl and T. Vallier, eds.), pp. 351-378, Circum-Pacific Council for Energy and Mineral
Resources, Houston, TX.
Woodhead, J. D. 1989. Geochemistry of the Mariana arc (western Pacific) source compositions and processes,
Chern. Geol. 76:1-24.
Woodhead, J., Eggins, S., and Gamble, J. 1993. High field strength and transition element systematics in island arc
and back-arc basin basalts: Evidence for multi-phase melt extraction and a depleted mantle wedge, Earth
Planet. Sci. Lett. 114:491-504.
4
The North Fiji Basin

Geology, Structure,
and Geodynamic Evolution
Jean-Marie Auzende, Bernard Pelletier, and Jean-Philippe Eissen
ABSTRACT
As the result of intensive studies conducted by U.S., French, and Japanese scientific teams,
the North Fiji Basin ridge, poorly known 10 years ago, is one of the most exhaustively
investigated ridge axes of the world's oceans. Today, a ridge segment more than 800 km
long and 100 km wide has been fully mapped with the Sea Beam and Furono echo sounders.
This ridge axis shows four main segments characterized by the same morpho structural
aspect and limits that characterize mid-oceanic ridges. Along the whole length of the axis,
a water-column sample has been taken every 20 km and rock samples every 10 km.
Different types of hydrothermal activity have been discovered and explored either during
the Nautile cruise in 1989 or during the Shinkai 6500 cruise in 1992. The most famous site is
the "White Lady," located around 17°S; it is characterized by 285°C shimmering hot water,
which is very poor in metallic elements, expelled through an anhydrite chimney. This water
probably represents the low salinity end-member resulting from phase separation in the
deep levels of the oceanic crust. Other active sites have been observed all along the axis
showing different characteristics such as low-temperature diffusion zones. Even though
some parts of the North Fiji Basin remain poorly investigated, the newly acquired data from
the ridge axis and from the eastern and northwestern parts allow us to develop a new
tectonic model of basin evolution since its creation 12 m.y. ago.
1. INTRODUCTION
The study of the North Fiji Basin (NFB) (Fig. 4.1) started between 20 and 25 years ago
with focused efforts on selected areas and large scale profiling across the basin mainly
conducted during U.S. cruises (Scripps Institution of Oceanography, Hawaii Institute of
Jean-Marie Auzende • IFREMERlCB, 29280 Plouzane, France. Present address: ORSTOM, BPA5,
Noumea, New Caledonia. Bernard Pelletier • ORSTOM, Noumea, New Caledonia. Jean-Phillippe
Eissen • Antenne ORSTOM-IFREMERlCB, 29280 Plouzane, France.
139
140 J-M. AUZENDE et aJ.
10·
S
20·
PACIFIC
OCEAN
30·
160·E 170·E 180· 175·W

FIGURE 4.1. General geodynamic setting of the North Fiji Basin (NFB) in the SW Pacific.
Geophysics, and Woods Hole Oceanographic Institution) and French ORSTOM cruises.
Since 1976, ORSTOM (lnstitut Fran9ais de Recherche Scientifique pour Ie Developpement
en Cooperation) has conducted the EVA (Evolution des Arcs insulaires) program, which
has been partly devoted to the study of the Lau and North Fiji basins. Fifteen cruises have
been conducted in different areas, especially in the New Hebrides arc domain, in the south-
ernmost part of the basin around the Matthew-Hunter zone, and more recently in the
northwestern part of the basin.
The SEAPSO (Sea Beam Pacifique Sud Ouest) program in 1985 represented the first
coordinated approach by French teams for backarc basin studies. As a part of this program,
IFREMER (Institut Fran9ais de Recherche pour l'Exploitation de la Mer), ORSTOM, and
INSU (Institut National des Sciences de l'Univers) jointly explored the NFB where the
existence of an axial ridge was assumed but not demonstrated. One of the major results
of the SEAPSO-Leg 3 cruise of the R.Y. Jean Charcot was the partial mapping of an ac-
tive accreting ridge in this marginal basin.
On the basis of the SEAPSO cruise results, Japanese and French scientists decided to
undertake a joint project to study the rift systems of the western Pacific. This project,
STARMER, coordinated by IFREMER in France and the Science and Technology Agency
(STA) in Japan, was initiated in 1987. Its objective was a five-year interdisciplinary
(geology, geochemistry, and geophysics) study of the NFB ridge. Since this date, seven
cruises have been conducted, representing more than nine months at sea. Four cruises were
dedicated to surface-ship surveys, including swath bathymetric mapping, geophysical
profiling, water sampling, gravity coring, and dredging. The three others were diving
cruises using the Nautile (June-July 1989), the new Japanese submersible Shinkai 6500
(September-November 1991), and Cyana (December-January 1992).
THE NORTH FIJI BASIN 141
Simultaneously, U.S., German, and SOPAC (South Pacific Applied Geoscience Com-
mission) teams conducted SeaMARC, Sea Beam, and GLORIA (geologic long range
inclined asdic) surveys on selected areas, especially in the northern and northeastern parts
of the NFB. In total, more than 20 months of ship time have been spent in the basin in the
past 10 years. The result of these efforts is an extensive multibeam bathymetric coverage
of the central spreading axis, detailed rock and water sampling along the axis, and in situ
observations by deep-tows and submersibles.
The NFB is one of the marginal basins located at the converging boundary between the
Pacific and Australian major plates (Fig. 4.1). It lies between the New Hebrides arc to the
west, the Fiji Platform to the east, the Vitiaz fossil subduction zone to the north, and the
arcuate Matthew-Hunter zone to the south. Different models for opening of the NFB have
been proposed (Chase, 1971; Gill and Gorton, 1973; Falvey, 1975; Dubois et al., 1977;
Malahoff et al., 1982a; Auzende et al., 1988b). These authors suggest that the opening of the
basin started 10 m.y. ago after the locking of the Vitiaz subduction by the Ontong-Java
Plateau and the reversal of its subduction polarity. This change of polarity involved the
clockwise rotation of the New Hebrides arc and by secondary effect the anticlockwise
rotation of the Fiji Platform. This first phase implies a NW-SE-trending spreading axis and
N45° to N55° flowlines for the opening motion. The second phase resulted in the beginning
of the collision of the New Hebrides arc with the d'Entrecasteaux and Loyalty Islands
ridges (Daniel, 1982; Louat, 1982; Monzier et al., 1984,1990), the change of the traction
stresses to an E-W direction, and the emplacement of the N-S-trending spreading center
in the central part of the basin. During this time N-S spreading occurred in the northern
part of the basin and north of the Fiji Islands (Auzende et al., 1988a,b).
1.1. Bathymetry and Structure

Bathymetric data concerning the NFB were rare before the SEAPSO III cruise of the
R. V. Jean Charcot (December 1985) and the five surface cruises of the STARMER project.
The only existing data were general and imprecise maps (Chase, 1971; Mammerickx et al.,
1971) or more detailed local surveys (Halunen, 1979; Kroenke et al., 1987, 1991). The NFB
was then called the North Fiji Plateau, and the main feature mapped was a 3000-m-deep flat
area occupying the whole central part of the basin.
The bathymetric map shown in Fig. 4.2 (Maze et al., 1992, unpublished map) results
from the compilation of all the available bathymetric data; it combines good-quality
classical bathymetric surveys and recent multibeam surveys performed mainly within the
SEAPSO and STARMER projects. This map especially uses the data from the maps of
Monzier et al. (1984, 1991), Auzende et al. (1990b, 1992b) and Urabe et al. (1992). It shows
the different physiographic zones characterizing the NFB. We have divided the basin into
five physiographic and structural provinces: the western basin, the central spreading ridge,
the eastern basin, the northwestern basin, and the northeastern basin.
1.1.1. The Western Basin

This domain is located south of the Hazel Holme Ridge (14°S) between the New
Hebrides arc and 173°E (Fig. 4.2). It is still poorly known and contains the largest portion of
old crust of the basin, including the initial NW-SE spreading center (Auzende et al., 1988b;
Pelletier et al., 1993a). Although some topographic features exist, especially in its southern
142 J-M. AUZENDE et al.
FIGURE 4.2. Bathymetric map of the NFB. This map results from a compilation (J.P. Maze) of all multibeam
data existing on the NFB (SEAPSO, Proligo, STARMER, Multipso cruises). The contour interval is 200 m. CSR:
central spreading ridge; WB: western basin; E: eastern basin; NWB: northwestern basin; NEB: northeastern basin.
Locations of Figs. 4.3, 4.5, 4.7, and 4.11 are shown.
part (south of 17°S), the western basin is mainly characterized by gently undulating and flat
areas 3200-3300 m deep. The primitive oceanic crust topography is smoothed or obliter-
ated by sedimentary cover which thickens toward the west from 0.1 s to more than O.S s of
two-way travel time (Luyendyck et ai., 1974; Halunen, 1979; Kroenke et ai., 1994). The
thickest sediments occur in the westernmost part of the basin along the New Hebrides arc,
especially east of the central New Hebrides islands where the thickness reaches l-s two-
way travel time in seismic reflection profiles (Luyendyck et ai., 1974; Pelletier et ai., 1988).
From 17°20'S to 20 0 S0'S, the westernmost part of the western basin is characterized
by the N1S0° trending southern New Hebrides backarc troughs (Fig. 4.2) (Monzier et al.,
1984, 1991; Recy et ai., 1990; Price et ai., 1993). These troughs deepen southward from
2800 to 3600 m and are flanked eastward by a N1SO° volcanic ridge with an average depth
of 1000 m and rising to 6S0 m above sea level at Futuna Island. A secondary N1S0° -170°
trending volcanic ridge (600 to 1800 m deep) lies 30 km east of the Futuna Ridge. Isolated
volcanoes exist farther east; one of them, the Constantine Bank, shoals to a depth of 104 m
(Monzier et al., 1984).
The morphology of the southern part of the western basin is still largely unknown but
appears to be complex and characterized by NE-SW-trending structural features (Fig. 4.2)
(Monzier et aI., 1984,1991). Around 21°S, 172°E, two N40-45° trending ridges rising to a
depth of 2000-2200 m and enclosing a 3600-m depression are the most prominent features
and likely a major tectonic element of this southern area. Because NW-SE trending
magnetic anomaly lineations exist immediately to the southeast, this paired trough-ridge is
interpreted as a fracture zone.
Other important tectonic features are located at 17°30'S, 172°E in the central eastern
part of the western basin, where NE-SW-trending, steep-sided, linear ridges rise to 1300 m
and are commonly interrupted by 3000-3300-m depressions (Halunen, 1979). The ridge
area, also characterized by absence of sediments, high heat flow, and presence of shallow
earthquakes, has been interpreted as an active NE-SW spreading center (Halunen, 1979).
1.1.2. The Central Spreading Ridge

Locally recognized from bathymetric and magnetic profiles (Chase, 1971; Malahoff
et aI., 1982a; Maillet et aI., 1986), the central spreading ridge (CSR) was for the first time
partially multibeam-mapped during the SEAPSO III cruise of the R.Y. Jean Charcot
(December 1985) (Auzende et aI., 1986a,b, 1988a). Up to now within the French-Japanese
joint STARMER project it has been mapped between 14°30'S and 21°40'S on more than 1°
width by multibeam bathymetric full coverage (Auzende et aI., 1990b; Urabe et aI., 1992;
Fig. 4.3). The ridge axis can be divided into four major segments, with lengths varying from
120 to more than 200 km, that are described below from south to north.
1.1.2.1. The Southernmost Segment. The southernmost segment is about 120 km long
between 21°40'S and 20030'S. It is characterized by a N-S trend and a complicated
morphology of alternating ridges reaching 2500-m depth and depressions reaching below
3000 m. The atypical morphology and the lack of in situ observation make it difficult to
locate precisely the present-day spreading axis on this segment. Only magnetic lineation
analysis (Maillet et aI., 1989; Ruellan et aI., 1989) confirms the existence of an active
spreading axis centered on 174°05'E. The transverse bathymetric profile of Fig. 4.4
illustrates this atypical morphology intermediate between fast and slow spreading ridges
(Macdonald, 1983; Karson et at., 1987).
1.1.2.2. The 3JO-km-Long North-South Segment. The 3tO-km-Iong north-south segment,

between 210S and 18°tO'S, is apparently the simplest segment. Its linear ridge axis crosses a
2800-m-deep flat-topped dome cut in its center by a graben that is a few hundred meters
wide and a few tens of meters deep. The width of the axial ridge, limited by the 3000-m
isobath, is about 20 km. On both sides, the dome is flanked by symmetrical grabens more
than 3000 m deep and trending N-S. The southern and northern tips of this segment exhibit
V-shaped features interpreted as inner and outer pseudofaults of propagating rifts (de
Alteris et aI., 1993; Ruellan et aI., 1994). In the middle part of the segment, around 19-
19°40'S, symmetrical oblique volcanic lines crosscut the N-S oceanic structural grain. On
the northwestern side of the area west of the inner pseudofault, the oceanic bottom shows
successive arcuate ridges abutting the pseudofault. These peculiar features can be either
interpreted as fossil overlapping spreading centers (OSC) (Ruellan et at., 1994) or as failed
rifts of a propagating rift system (de Alteris et aI., 1993).
1.1.2.3. The N15° Segment. The N15° segment is about 120 km long and shows spec-
tacular along-strike morphological variation. In its southern part from 18°tO'S to 17°55'S,
144 J-M. AUZENDE at al.
FIGURE 4.3. Simplified bathymetry (contour interval = 1000

m) and magnetic lineations (heavy lines) along the central
spreading ridge from Auzende et al. (199Oc) and de Alteris et al.
(1993) (after Auzende et aI., 1994b). The present-day axis is
shown with medium lines. In gray: anomaly I; J: Jaramillo
anomaly (I Ma); 2: anomaly 2 (1.98 Ma); 2A: anomaly 2A (2.6-
3.5 Ma). The ages of anomalies are after Cande and Kent (1992).
the present-day axis is not well defined and the accretion is distributed over numerous small
volcanoes scattered over a wide area. This zone corresponds to the change of direction of
the central spreading ridge from N-S to NI5°. North of 17°55'S, the spreading axis is
represented by a double ridge less than 2500 m high bounding a 2-3-km-wide, 200-300-
m-deep graben. At the northern tip of the N15° segment, close to the 16°50' S triple junction,
the ridge axis is located on the top of a shallow massif culminating at less than 1900 m deep
and cut by a graben 500 m to 2 km wide and 200 m deep. This area is the site of the active
hydrothermal vents explored by Nautile and Shinkai 6500 (see following). The N15° axial
ridge is flanked by curved grabens more than 3000 m deep interpreted by Ruellan et al.
(1994) as fossil oses.
wsw ~ ENE
~OOO~
SW 4000~
~OOO~
~ W ~
~ IS"27'S
4000~ ___________________________
____________
~ --=
;: ~IS"~~
~OO~ ~
200~NW 17"04'5 ESE
3000~
200~~W 17~~O~~E
3000E..~
__- +_ _----"=-"
WNW 'S"34'S ESE
3000~~
N,W~ S,E
3000",~=---+--=""
3000~E
30:~~",,__ ;re
~
2S00~
3000~
SW 1 10'5 N
3000 ""'--'====------l----=""
~E
3~0 """'--'-9"-20-'S--''''''''E=-'
3000~
NW
~
__=------,"-
3000",,~=----+
~S'sS.E
'0 20Km
300~
__~==~_-._~~
I I
E.V.S
_
W~-=~E
_
3000~E...
~_=----'"
WSW 19"40'S ENE
~~gg~j
~~ggi~i
W I E
~~ggi~1
FIGURE 4.4, Transverse bathymetric profiles across the different segments of the central spreading ridge
(16°17'5 to 21°15'5), A: ridge axis,
1.1.2.4. The N160° Segment. The N160° segment is about 200 km long and can be
described as three parts (Auzende et at., 1991a, 1994a; Gracia-Mont, 1991, 1992; Jarvis et
aI., 1993; Figs. 4.5 and 4.6): (1) From 16°50S to 15°30S the spreading axis is located in a
4000-4S00-m-deep graben located between two subvertical walls. The average width of
the graben is 8 km along the whole segment. In its axial part, the graben is cut by a 2-3-km-
wide, 400-S00-m-high ridge (Fig. 4.4). This morphology is very similar to those described
at slow spreading (less than 40 mm/yr) ridges like the Mid-Atlantic Ridge (Kappel and
Ryan, 1986; Karson et aI., 1987; Karson, 1990; Gente et aI., 1991). Around 16°IOS, the
remarkable linearity of the N160° axis is interrupted by a slight curve toward the north,
offsetting the graben by about 4 km. On both sides the active domain is flanked by a large
volcanic massif that shallows to less than 1700 m deep. The width of the volcanic massif
FIGURE 4.5. Bathymetry of the NI60 segment and of the 14°50'S triple junction (enlargement of Fig. 4.2). The
contour interval is 200 m.
decreases from 100 km in the south to a few kilometers in the north until it disappears north
of 15°30' S. The magnetic data analysis allows an estimate of the beginning of the volcanic
construction and the massif uplift at about 1 Ma. This uplift affects an older oceanic crust as
demonstrated by the age of the sediments sampled from the northern edge of the N55°
trending graben located east of the 16°50'S triple junction (Lagabrielle et al., 1994). (2)
Between 15°30' Sand 15°00' S the accretion is distributed over a wide domain of two 6O-km-
long, 4OOO-m-deep en echelon grabens offsetting the axis by about 40 km to the northeast.
Each of these grabens is made up of a succession of to-km-Iong en echelon segments (Fig.
4.6). In this area, the magmatic supply appears to be concentrated along a narrow ridge that
separates the grabens. For the past 1 m.y. the accretion seems to have been mainly
amagmatic (Gracia-Mont, 1991, 1992). (3) North of 15°OOS the spreading axis becomes
more complex and is located within two distinct branches. The western one, trending
NI20°, is characterized by a 4-km-wide, 4000-m-deep graben that crosscuts an older
oceanic crust showing different trends from N160° to NI20°. The northern branch consists
172°30 174° 174°30

c:::> ,
14°30 .......... 2
14°30
-- 3
--- 4
15° 15°
15°30 15°30
."
16° 16°
16°30 16°30
17" 17"
172"30 173° 173°30 174° 174°30
FIGURE 4.6. Structural map of the Nl60 segment and of the l4°50'S triple junction (after Auzende et ai., 1994a).
I: axial domain deduced from bathymetry and magnetic anomaly, 2: normal fault and scarps, 3: crests, 4: depres-
sions, 5: isolated volcanoes, 6: ridge axis.
of a 2400-m-deep, Nl400-trending ridge that connects with the N160° graben system at a
4000-m-deep depression at 14°50'S. This depression can be interpreted as a triple junction
between the N160° and N120° grabens and the N140° ridge (Auzende et aI., 1994a).
1.1.3. The Eastern Basin

The eastern basin occupies the area between the Fiji Islands and the central spreading
ridge. On the basis of different arguments, a spreading ridge was postulated west of Fiji
(Sclater and Menard, 1967; Chase, 1971; Brocher and Holmes, 1985). Few recent data exist
for this area, except a full-coverage, one-square-degree bathymetric survey located imme-
diately west of Fiji (Fig. 4.7). This area is underlain by important shallow seismicity with
strike-slip (Hamburger and Isacks, 1988, 1994) and normal-fault-type focal mechanism
solutions (Louat and Pelletier 1989; Pelletier and Louat, 1989). Previously interpreted as a
strike-slip deformation zone (Auzende et aI., 1986b), this area was recently reinterpreted in
-'-L-'-"~-'-'-.J....L-'-'-J..J..J 18°00
S
FIGURE 4.7. Sea Beam bathymetry of the West Fiji area (after Auzende et aI., J988a). The contour interval is 50 m.
tenns of a propagating rift (Auzende et aI., 1993, 1994b). The main morphostructural units
of the area are represented in Fig. 4.8.
1.1.3.1. The Propagating Ridge (PR) and Its Tip (PT). The PR is located in the western
part and comprises a central ridge bounded by two grabens. The eastern graben is well
developed with a constant width of 10 km and depth of 4000 m. The western graben is
narrower (2-3 km) and shallower (3000 m). The central ridge is 7 to 8 km wide and 2750 m
deep. Its N-S to N05° trend changes to N155° at its southern tip. The width ofthe PR varies
from 40 km in the north to 8 km in the south. South of 17°44'S, the ridge disappears and is
replaced by a 3000 m deep flat area limited by converging faults. This flat area is similar to
the "propagating tip" described by Hey et al. (1986) in the case ofthe 95.5°W propagating
rift on the East Pacific Rise close to the Galapagos.
1.1.3.2. The Pseudo/aults. The outer pseudofault (OPF) corresponds to a NI55°-

trending fault bounding the propagating system on the west. This fault separates the
propagating from the "nonnal" oceanic bottom grain trending NI5-20°. The inner pseudo-
fault (IPF) is repeated north of 17°35' S (Fig. 4.8, dashed line, top center), with a N15° fault
,-----.------.-----.---------,------,---------,17°10·
S
8°00'
176°40'
FIGURE 4.8. Structural sketch of the West Fiji Ridge (after Auzende et al.. 1993). PR axis: propagating axis; PT:
propagating tip; OPF: outer pseudofault; IPF: inner pseudofault; SR: southern rift; TL: transferred lithosphere;
TZ: transform zone. Main positive and negative magnetic anomaly lineations are underlined in gray and light gray
respectively. Anomalies I and J are labeled.
converging toward the OPF with a 40° angle to the west and a more diffuse lineament
trending N30° and converging with the OPF with a 55° angle to the east. South of 17°35'S,
there is a unique fault converging with the OPF at a 20° angle.
1.1.3.3. The Southern Ridge (SR). This SR is located in the southeastern part of our
survey (Fig. 4.8) and is characterized by a 3000-m-deep, few-kilometer-wide axial graben
trending NlO°. In the eastern part of the domain, the failed rift is represented by a
succession of NlO-NI5° trending ridges abutting a N150° fault. These ridges are 2 to 3 km
wide and 2750 m deep. They probably represent the old spreading axis abandoned during
the southward propagation of the western active rift.
1.l.3.4. The Transform Zone (17) and the Transferred Lithosphere (TL). The junction be-
tween the propagating rift and the southern ridge is a wide zone of arcuate small ridges
representing a transform zone as defined by Hey et al. (1986) on the East Pacific Rise. The
TL is located north of the TZ and comprises fan-shaped structures deepening to 3000 m in
the northern part.
The close structural similarities between this area and the Galapagos propagating rift
confirm the interpretation of an active spreading axis. The spreading rate was estimated
to be 30 mmlyr by Louat and Pelletier (1989) from seismicity analysis and kinematic
reconstruction of the Lau and North Fiji basins. The width of the propagating rift area is
consistent with this rate estimate (see following).
1.1.4. The Northwestern Basin

The northwestern basin, located north of 14 oS and west of 174°E between the northern
part of the New Hebrides island arc to the west and the Vitiaz Trench to the east, has been
the subject of different speculations and interpretations, caused mainly by the lack of data.
It was interpreted as (1) the result of a late Miocene fan-shaped opening along a median
NW-SE-trending axis (Falvey, 1975; Malahoff et aI., 1982a; Auzende et al., 1988b), (2) an
old piece of Pacific plate (Chase, 1971; Luyendyck et aI., 1974), or (3) Australian plate
(Halunen, 1979) trapped behind the Vitiaz paleotrench. Recently, the bathymetry and main
morpho structural units of the whole northwestern part of the NFB have been recognized
during the EVA 14 (1987) and Santa Cruz (1991) cruises of R.Y. Coriolis and Le Noroft,
respectively, and are described here (Pelletier et al., 1988, 1993a,b) (Figs. 4.9 and 4.10).
1.1.4.1. The N-S-Trending Northern New Hebrides Backarc Troughs. The N-S-trending
northern New Hebrides backarc troughs were previously known south of 12°S (Charvis and
165° 168° 169° 170° 171° 172° 173° 174°E

----L---~----~----~8°S
9°
10°
11°
12°
14°
15°
16°
17"
c:::J 0 - 1500 2500 - 3000 4000 - 4500
_ 1500-2000 3000 - 3500 4500 - 6000
2000 -2500 _ 3500-4000
>6000
FIGURE 4.9. Bathymetric map of the northwestern basin.
166°E
a
b
c
d
o e
FIGURE 4.10. Structural map of the northwestern basin (after Pelletier et al., 1993a). The studied area is shown
in insert. AUS: Australia; NC: New Caledonia; FJ: Fiji; NZ: New Zealand. NHT: New Hebrides Trench;
NHA: New Hebrides arc; VT: Vitiaz Trench; BAT: backarc troughs domain; 9°30 R: 9°30'S Ridge; DR: Duff
Ridge; WTR: West Tikopia Ridge; TR: Tikopia Ridge; HHR: Hazel Holme Ridge; SPR: South Pandora Ridge;
Nl60 R: Nl60 ridge of the central spreading ridge; TT: Tikopia Trough; SCT: Santa Cruz Trough; WTP: West
Torres Plateau; ER: d'Entrecasteaux ridge. Islands and reefs are in black. D: Duff Islands; R: Reef Islands;
Tn: Tinakula Island; N: Ndende Island; V: Vanikoro Island; Tk: Tikopia Island; A: Anuta Island; F: Fatutaka
Island; P: Pandora Bank. (a) trough and depression; (b) structural trend and fracture; (c) structural high;
(d) structural low; (e) volcanic high.
Pelletier, 1989; Recy et aI., 1990; Johnson et al., 1993) and extend from 13°30'S to 9°30'S,
east of the New Hebrides arc platform in the westernmost part of the basin. The troughs are
formed by volcanic cones and ridges bounding N-S trending, 3000-3500-m-deep depres-
sions, between a N-S- to NNW-SSE-trending major scarp to the west and Duff ridge to the
east. They are filled by large volcanic highs at 12°S and lO o 30'S and are 60 kIn wide south
of 12°S and widen northward up to 100 kIn at lO o 45'S. At lO o 30'S the western part of the
troughs abuts the N95° trending Santa Cruz trough, which cuts the arc platform.
1.1.4.2. The Duff Ridge. The Duff Ridge separates the backarc troughs from the central
part of the northwestern basin. It is a continuous volcanic ridge more than 400 kIn long and
30 kIn wide, which strikes N-S and rises to around 1500 m south of lO o 45'S. It trends NW-
SE and shoals to 20 m farther north. The volcanic Duff Islands are located in the north-
ernmost tip of the ridge. Duff Ridge is interpreted as a fossil volcanic line related to the New
Hebrides subduction.
152 J-M. AUZ£ND£ et al.
1.1.4.3. The TIkopia, West TIkopia, and 9°30'S Ridges. The Tikopia, West Tikopia, and
9°30'S ridges are a succession of orthogonal ridges, interpreted as spreading ridges or
transform zones, lying in the axial part of the northwestern basin, which generally deepens
to the north from 3200 to 4200 m. Tikopia ridge trends E-Wand is located around 12°30' S
between the Tikopia volcano and island at 168°45'E and the volcanoes located south of
Fatutaka Island at 170°45 'E. East of the large Tikopia volcano (3500 m high, 35 km wide at
its base), the ridge is a 50-km-wide, N900-1oo° trending elongated dome cut along strike
by a lO-km-wide, 65-km-Iong trough (4200 m deep) bounded by looO-1500-m-high scarps.
Fresh basalts have been dredged on the northern wall of the trough. West Tikopia Ridge is a
N-S-trending, discontinuous ridge composed of aligned seamounts rising to 800 m west of
the Tikopia volcano along 168°30'E from 12°20'S to 10045'S. The 9°30 ridge is an almost
continuous, Nl00° trending volcanic ridge from 165°45'E to 168°15'E north of the New
Hebrides arc platform and the Duff Islands. The 9°30 ridge ends to the east at a large
volcanic massif shoaling to 10 m in the northern prolongation of the West Tikopia Ridge.
Numerous volcanic highs, including Anuta and Fatutaka Islands and Pandora Bank,
are located in the eastern part of the area between 10030'S to 13°S and 169°E to 174°E ,
south of the Vitiaz paleotrench. These volcanic highs, previously considered as a part of the
inactive Vitiaz arc, do not constitute a continuous chain but are isolated and occur as far as
240 km from the Vitiaz paleotrench. They constitute a series of massifs aligned on N-S
lineaments. No volcanic arc lies immediately south of the Vitiaz paleotrench. However, a
narrow ridge rising to about 2000 m between 12° and 11°S and a narrow and discontinuous
swell (3200-3400 m deep) north of 10030'S parallel the Vitiaz Trench (4500-6000 m
deep), which shows a succession of NW-SE and E-W trending troughs.
The northwestern basin is separated from the rest of the basin by a complex series of
E-W-trending ridges and troughs called the Hazel Holme fracture zone by Chase (1971).
Right-lateral or left-lateral strike slip or even compressional motions have been proposed
along this enigmatic feature (Luyendyck et al., 1974; Halunen, 1979; Eguchi, 1984; Ham-
burger and !sacks, 1988,1994), which can be described as two parts.
1.1.4.4. The Hazel Holme Ridge, which corresponds to the western part of the Hazel
Holme fracture zone of Chase (west of 171 °E), was partly surveyed during the EVA 14 and
Santa Cruz cruises and interpreted as an active extensional zone (Pelletier et al., 1988,
1993a; Charvis and Pelletier, 1989; Louat and Pelletier, 1989). This seismically active
feature is composed of several parallel narrow ridges and deep troughs trending N80-100°
and extending over a maximum width of 120 km. The troughs are, on average, 3500-4000
m deep and are bounded by 500-1000-m-high scarps. The deepest trough (4500 m deep at
about 169°E) is located in the axial part of the feature and appears to be right-laterally
offset. West of 168°30'E the width of the ridge decreases, lateral troughs disappear, and a
2500-3500-m-deep E-W trough is bounded by symmetrical ridges rising to 1700 m. This
trough ends at 168°IO'E where it connects with the southern tip of the northern New
Hebrides backarc troughs. The bathymetric profiles across the area exhibit a slow spread-
ing ridge morphology.
1.1.4.5. The South Pandora Ridge, which corresponds to the eastern part of the Hazel
Holme fracture zone of Chase (1971), was explored between 171°E and 173°30'E during a
R.Y. Kana Keoki reconnaissance cruise (1982) and near 174°E by a SeaMARC survey
during the R.Y. Moana Wave cruise (1987); it is interpreted as an E-W trending active slow
spreading ridge within a broad transform domain (Kroenke et al., 1991; Price and Kroenke,
1991). The ridge is 110 km wide and is cut by numerous transform offsets. It is composed
of a series of E-W trending segments cut along strike by a central trough that is 10 to 20 km
wide and 500 to 4500 m deep. The axial trough is flanked by 1000-2000-m-high scarps and
is locally filled by large volcanoes rising in some places to sea level. Fresh pillow basalts
recovered on the South Pandora Ridge have geochemical characteristics intermediate
between mid-ocean ridge basalts (MORBs) and ocean island basalts (OIBs) (Price et at.,
1990; Price and Kroenke, 1991; Sinton et at., 1994).
1.1.5. The Northeastern Basin

The northeastern basin is located east of 174°E and west of the northern Lau Basin
(Fig. 4.2), between the Fiji Platform and the North Fiji fracture zone in the south and an
imprecisely located lineament in the north that separates the young NFB oceanic crust from
the old Pacific crust. This lineament is marked by a succession of deep troughs (Alexa
Trough, Rotuma Trough, Hom Trough; Brocher, 1985) and connects the Vitiaz Trough to
the west and the northern tip of the Tonga Trench to the east. Although some restricted areas
have been recognized (see below), the overall structure of the northeastern basin is largely
unknown. The northeastern basin is now the least-known part of the NFB.
On the basis of an aeromagnetic survey over the entire area, Cherkis (1980) reported
E-W-trending magnetic lineations and proposed an active spreading center located near
14°30'S from 175° to 178°E and near 14°S from 178°E to 180°. The only area explored by
surface ship is located north of Viti Levu Island between 176°30'E and 178°E and 13°30'S
and 15°S (Halunen, 1979; Brocher, 1985). Halunen (1979) described a prominent NW-SE
trending paired ridge-trough in the northwest part, more subdued WSW-ENE trending
topography in the southwest, and uniform sediment cover about 0.08 s double-travel-time
thick. On the basis of WNW-ESE magnetic lineations and bathymetric trends offset by an
oblique NNW-SSE trough at 177°30'E and interpreted as a pseudofault, Brocher (1985)
proposed a WNW-ESE inactive spreading center at 14°30'S and 14°S located respectively
west and east of a fossil propagating rift at 177°30'E.
More recently, a multibeam bathymetric survey of a N-S elongated area at 177°-
177°30'E from 13°50'S to 16°40'S was conducted during the R.Y. Sonne cruise SO-35 (von
Stackelberg et at., 1985; von Stackelberg and von Rad, 1990). The different morphostruc-
tural units are described below from north to south (Fig. 4.11).
1. A 3500-m-deep, lO-km-wide steep-sided trough trends N120° at 14°05'S and is
bounded by linear ridges rising to 2200 m. This trough, also described by Halunen (1979),
corresponds to the northwestern extension of the pseudofaults of Brocher (1985) and is
surprisingly interpreted by von Stackelberg and von Rad (1990) as a short narrow basin
possibly belonging to the Vitiaz Trench lineament.
2. Two N120° trending narrow ridges, rising to 2400-2700 m and enclosing a 2900-
m-deep depression, are observed at 14°30'S and are interpreted, following Brocher (1985),
as an extinct spreading center. Although magnetic lineations are associated with these
ridges, the unidentified anomalies do not allow an estimate of the age for the spreading.
Altered ferrobasalts derived from a parental MORB magma and covered with hydrothermal
Mn crusts were recovered from these ridges (Sinton et at., 1985; Johnson and Sinton, 1990).
3. Volcanoes are present between 14°50'S and 15°15'S.
4. Two N70° trending prominent and parallel ridges exist at 15°40'S and 15°50'S. The
northern one, called the Braemar Ridge, rises to 1600-1900 m deep and is bounded to the
north by a 3600-m-deep depression. The southern ridge is 2000-2200 m deep and is
r -_ _ _ _ _ _ _ _ _ _ _--..:-:--_ _ _ _ _ _ _ _ _ _ _ _ _i I4•S
I~_~
2 __ _
]~
51111/
IS'
FIGURE 4. 11. Structural sketch of a part of the northeastern basin based partly on the bathymetric map obtained
during the R.Y. Sonne cruise SO-35 (von Stackelberg et ai., 1985; von Stackelberg and von Rad, 1990), the
GLORIA image (Jarvis etai., 1994) and the map of Fig. 4.2.1: structural high; 2: structural low; 3: scarp; 4: main
trough; 5: main ridge. FP: Fiji Platform; NFFZ: north Fiji fracture zone; BrR: Braemar Ridge; BaR: Balmoral
Ridge.
narrower than and separated from the Braemar Ridge by a 2900-m-deep depression. These
two ridges are interpreted as uplifted faulted blocks of Pliocene volcanics and volcaniclas-
tics derived from the fossil Fiji volcanic arc and deposited in a forearc basin between the
Vanuatu-Fiji-Lau-Tonga arc and the Vitiaz Trench lineament (von Stackelberg and
von Rad, 1990).
5. An E-W trending ridge lies at 15°55'S and appears to be made up of aligned
volcanoes.
6. A 30-40-km-wide pull-apart basin with N-S structural trends, centered at 177°25'E
between 16°lO'S and 16°35'S, is developed between an E-W trending 2000-m-deep ridge
and 3400-4000-m-deep troughs (von Stackelberg and von Rad, 1990). Fresh MORB and
backarc basin basalts (BABB) lavas with hydrothermal sulfide mineralization have been
recovered in the pull-part basin.
The topography of the area immediately north of the Fiji Platform is controlled by the
active left-lateral north Fiji fracture zone. The zone extends from the northern tip of the
Tonga Trench to the central part of the NFB and marks the present-day boundary between
the Pacific and Australian plates. The fracture zone is mainly composed of E-W structural
trends with some pull-apart basins (Louat and Pelletier, 1989; Pelletier and Louat, 1989;
Johnson and Sinton, 1990; Hughes Clarke et aI., 1993; Jarvis et al. 1994). The fracture zone
has been partly imaged by SeaMARC (Kroenke et aI., 1991) and GLORIA during the
SOPAC cruise in 1990 (Hughes Clarke et aI., 1993; Jarvis et aI., 1994). A second pull-apart
basin trending N45° was found at 15°30'S, 178°40'E. The ridges located in the northeastern
basin north of the Fiji Platform, like the Braemar Ridge and the Balmoral Ridge and
Balmoral Reef farther west, are interpreted by Jarvis et al. (1994) to be pieces of the Fiji
Platform rifted away by successive spreading segments (pull-apart basins) during changes
in the location of the north Fiji fracture zone.
2. MAGNETISM AND PALEOMAGNETISM
The first published data concerning the magnetism (Sclater and Menard, 1967; Chase,
1971; Luyendyck et al., 1974; Halunen, 1979) emphasized the complexity of the basin.
These authors developed various interpretations of the magnetic lineations with different
locations of spreading centers being proposed in the NFB south of the Hazel Holme Ridge.
The area located to the north of the Hazel Holme Ridge was considered to be part of the old
Pacific or Australian plates.
Paleomagnetic results from Vanuatu and Viti Levu indicate a clockwise rotation of the
New Hebrides arc of about 28° since 6 Ma (Falvey, 1978) and a counterclockwise rotation
of 21° since 4 Ma (James and Falvey, 1978) and of 90° since 7 Ma (Malahoff et aI., 1982b)
of the Fiji Platform. These rotations and the polarity reversal from the Vitiaz to the New
Hebrides subduction since 10-8 Ma are the basis for the hypothesis of a NW-SE trending
spreading center northwest of Fiji and an E-W spreading axis north of Fiji (Falvey, 1978;
Malahoff et al., 1982a,b).
An aeromagnetic survey carried out by the National Oceanic and Atmospheric Ad-
ministration (NOAA) and the U.S. Naval Research Project in 1979 (Cherkis, 1980; Larue
et aI., 1982; Malahoff et aI., 1982a, 1994) defined magnetic lineations in the NFB which
have been differently interpreted by different authors. After reprocessing these aeromag-
netic data, Auzende et al. (1988b) proposed a new interpretation of the magnetic lineations.
In the central basin a very clear N-S axial lineation is bounded by two parallel lineations
identified as the J (Jaramillo) (1 Ma) and 2 (1.9 Ma) anomalies (Cande and Kent, 1992). On
the eastern limb of the axis, a possible 2A (2.6-3.5 Ma) anomaly can be distinguished but is
not continuously identified on the western limb. In the western basin the magnetic pattern is
dominated by NW-SE lineations crosscut by transverse N45° trending features interpreted
as transform faults. The NW-SE lineations represent the trends of the initial stage of
opening of the NFB between 10 and 3 Ma (Auzende et aI., 1988b) or 12? to 7 Ma (Pelletier
et aI., 1993a). The magnetic pattern of the eastern basin is less well defined and shows
indistinctly NW-SE trends mixed with N-S directions which are particularly dominant
immediately west of the Fiji Islands. The northern and northeastern parts of the NFB are
mainly characterized on the aeromagnetic survey by roughly E-W trending lineations that
have not been dated.
In the northwestern basin recent data acquired during the EVA 14 and Santa Cruz
cruises permit precise definition ofthe magnetic pattern (Pelletier et aI., 1988, 1993a,b). The
magnetic map (Fig. 4.12) mainly shows two trends of lineations. NW-SE trending linea-
tions disrupted by NE-SW transform faults occur both in the northern edge of the area along
166·E 170· 172·
a
b
c
-d
~e
FIGURE 4.12. Map of magnetic anomaly lineations in the northwestern NFB (after Pelletier et aI., 1993b).
(a) positive magnetic anomaly lineation; (b) negative magnetic anomaly lineation; (c) fault zone; (d) contour of the
axial part of the spreading ridges; (e) contour of the main morphostructural features.
the Vitiaz Trench up to 8°30'S and in the southwestern part of the area along the New
Hebrides arc, north and south of the western end of the Hazel Holme Ridge. Some NW-SE
lineations are also present in the western side of the basin, north of 11 oS from the eastern
edge of the New Hebrides arc platform to the Duff Islands. The NW-SE lineations identified
as anomalies 5A ?-5 to 4 (12?-11 to 7 Ma) resulted from the initial opening of the basin in a
NE-SW direction. E-W trending magnetic lineations disrupted by N-S transform faults are
distributed over the entire central part of the northwestern basin from 9 0 S to 15°S and are
flanked by the NW-SE lineations. The E-W trending lineations identified as anomalies
3A to 2A (7 to 2.5 Ma) resulted from a second stage of opening in a N-S direction and are
associated with the South Pandora, Tikopia, and 9°30' S ridges. The latter is interpreted as a
spreading center. At 168°30'E, the E-W lineations abut the N-S trending West Tikopia
Ridge, which is interpreted as a major transform fault that offsets northward the spreading
center from the Tikopia Ridge to the 9°30' Ridge. The Hazel Holme Ridge is interpreted as
an active and very young extensional zone or slow spreading ridge which crosscuts older
oceanic crust and connects the N160° trending segment of the central spreading ridge with
the southern tip of the northern New Hebrides backarc troughs. E-W trending magnetic
lineations seem to be associated with the Hazel Holme Ridge but are not datable.
During the SEAPSO and STARMER French-Japanese joint projects, a detailed
magnetic survey (Fig. 4.3) of the central part of the NFB was conducted. It allows us to
refine identification of the magnetic anomalies and calculation of the spreading rate on the
different segments during the past 3 m.y. (Auzende et aI., 1990c; Huchon et aI., 1994). The
important features are the following.
1. An axial anomaly is identified all along the structural ridge axis. The width of this
anomaly varies, giving spreading rates ranging from 50 to 82 mm/yr.
2. The J (Jaramillo) event is only clearly represented on the N-S segment. Its exis-
tence along the N15° and the southern part of the N160° segments is debatable. Based on the
interpretation of the axial and J (Jaramillo) anomalies, the estimated spreading rate ofthe
N160° segment has varied between 40 and 50 mm/yr, which are rates representative of
intermediate rate spreading ridges. Along the entire N160° segment the present-day ridge
axis is characterized by a morphology usually encountered at slow spreading ridges
(Macdonald et aI., 1984). It is made up of a succession of deep grabens reaching more than
4000 m deep and bounded by steep lOOO-m-high walls.
3. The anomalies 2 and 2A are clearly identified on both sides of the ridge axis, at
least along the N-S and N15° segments. Along the southernmost N-S segment only anomaly
2 has been identified.
In the eastern part of the NFB, west of the Fiji Islands, the identification of magnetic
anomalies (Fig. 4.8) also favors the existence of an active spreading center (Auzende et al.,
1994b). In the western part of the survey, corresponding to the propagating rift system, a
well-defined anomaly is interpreted as anomaly 1. The width of this anomaly decreases
from north to south. On both sides of the area, a magnetic lineation could be anomaly J
(Jaramillo-1 Ma). The spreading rates calculated from these anomaly identifications are
close to 40 mm/yr. The magnetic pattern exactly superimposes the structure. Between both
axes, a large transverse E-W trending negative anomaly can be interpreted as the transform
zone between the propagating rift and the southern rift (Hey et aI., 1986). This anomaly has
no morphological expression. From the calculated spreading rate and the angles between
the pseudofaults, the southward propagation velocity can be calculated. For an angle of 55°
to 40°, as is observed north of 17°35'S, the propagation velocity is respectively 25 or
45 mm/yr. The tip of the propagator is characterized by a 20° angle with the pseudofault,
implying an increase of the propagation velocity up to 100 mm/yr in the present phase.
The age of this propagation could be related to the emplacement or the reactivation of the
north Fiji fracture zone 1 to 1.5 m.y. ago (Lafoy et aI., 1990), resulting in the formation of the
16°50'S triple junction.
3. SEISMICITY
Gutenberg and Richter (1954) described isolated seismicity between the Tonga and
New Hebrides arc major seismic zones. Sykes (1966) later succeeded in identifying a NE-
SW trending seismic zone running from the southern tip of the New Hebrides arc to the
center of the NFB. On the basis of seismic evidence, Sykes et al. (1969) suggested that the
two triangular areas between the Tonga and New Hebrides arcs should be described as two
different basins, the Lau and North Fiji basins. Chase (1971) extended this analysis and
proposed a model including five microplates for the same area. The deep structure of the
Lau and North Fiji basins was studied by Dubois (1971), Aggarwal et al. (1972), and Dubois
et al. (1973). They showed the existence of a low-seismic-velocity zone in the upper
mantle. Barazangi et al. (1974) demonstrated that this low-velocity zone was related to
seismic-wave attenuation due to active spreading.
Crustal deformation in the NFB was addressed by Eguchi (1984), Hamburger and
Isacks (1988), and Louat and Pelletier (1989), using a large number of shallow earthquake
data and focal mechanism solutions from international catalogues. Despite discrepancies in
their interpretations concerning the sense of active crustal motion, they provide a descrip-
tion of the distribution of shallow seismicity in the NFB and the major seismic lineaments
(see following). Hamburger and !sacks (1988) claimed that there are no steady-state
spreading centers in the NFB but only diffuse extensional zones in a wide strike-slip
boundary between the Pacific and Australian major plates. In contrast, Louat and Pelletier
(1989) tried to quantify the crustal motions along discrete spreading centers or transform
faults, using the seismological and marine geological data.
Figure 4.13 shows the distribution of the shallow seismicity in the NFB (Louat and
Pelletier, 1989). Earthquakes concentrate within certain areas delineating linear zones.
Some of the seismic lines are associated with known structural features.
1. The north Fiji fracture zone (NFFZ) is clearly marked by a WSW-ENE trending
seismic belt between 16° and 17°S from the central triple junction to the north of the Fiji
Platform. Numerous focal mechanism solutions attest to E-W left-lateral strike-slip motion
FIGURE 4.13. Shallow seismicity (0-70 km) in the NFB (after Louat and Pelletier, 1989). CSR: central
spreading ridge; NFFZ: north Fiji fracture zone; WFR: West Fiji Ridge; HHR: Hazel Holme Ridge; SPR: South
Pandora Ridge; CSL: central seismic lineament; ESL: eastern seismic lineament; K: Kandavu Island. See Louat
and Pelletier (1989) for the data source.
along the NFFZ. However, events with normal fault solutions have occurred east of the
177°25'E pull-apart basin at 178°E and at the western end of the north Fiji fracture zone.
2. The West Fiji Ridge (WFR) coincides with a NNE-SSW trending seismic line
located near 176°E, immediately west of the Fiji Platform between 17° and 200S. Events
with strike-slip or normal fault solutions occur along the west Fiji Ridge.
3. A WNW-ESE seismic belt correlates with the Hazel Holme Ridge between the
New Hebrides backarc area and the northern tip of the Nl600-trending segment of the
central spreading ridge. Focal mechanism solutions indicate strike-slip faulting.
In contrast, the central spreading ridge, which is the major active spreading center
of the basin, is almost unidentifiable with shallow seismicity. However, a few events cluster
at particular areas: around 21°S at the tip of the southward propagating rift and at the offset
of the axis, at 19°40'S where transverse volcanic alignments exist, around 18°lO'S at the tip
of the northward propagating rift, and between 16° and 17°S along the N160° trending
segment. Events with strike-slip faulting solutions occur all along the central spreading
ridge, but are especially concentrated at the 21°S ridge axis offset. However, one event with
normal fault solution is located on the N160° segment.
Two seismic lineaments are observed in the basin where the bathymetry largely
remains unknown (Lou at and Pelletier, 1989): a NNE-SSW line (CSL) west of the central
spreading ridge and characterized by pure normal faulting solutions, and a NE-SW line
(ESL) between the Fiji Platform and Kandavu Island. Moreover, diffuse areas of seismicity
are located in the southernmost part of the NFB and in the northeastern basin north of the
NFFZ. All these seismic features indicate active crustal deformation which is still not
understood.
4. HEAT FLOW DATA
Few heat flow measurements exist for the NFB. They all indicate high heat flow values
(Sclater and Menard, 1967; Macdonald et at., 1973; Halunen, 1979) over the entire basin,
which is consistent with it being interpreted as a recent oceanic basin. At a smaller scale,
two different provinces are distinguished. North of the Hazel Holme Ridge, the average
heat flow value is 2.29 HFU; however, Halunen (1979) suggests that the heat flow is very
low (0.88 HFU) if the New Hebrides arc volcanism effect is subtracted. These low values
can be related to hydrothermal circulation or to the older age of this part of the NFB. South
ofthe Hazel Holme Ridge, the values are higher with an average of 4 HFU. The relatively
high heat flow values have been interpreted as indicative of the abnormally shallow depth
of the NFB, called for this reason the North Fiji Plateau. In fact, the depths measured in the
NFB are "normal" in terms of vertical evolution related to the age of the oceanic crust.
Except in some peculiar areas such as the 16°50'S triple junction, the ridge crest depth
varies from 2500 to 2800 m, which is the standard for mid-oceanic ridges (Sclater and
Francheteau, 1970).
During the STARMER project a few heat flow measurements have been performed
within the active rift valley by the submersibles Nautile and Shinkai 6500, close to the
16°50'S triple junction and in the area of the White Lady active hydrothermal site (see
following). They all give low values (less than 1 HFU; Joshima, pers. comm.) that are
probably related to water circulation.
5. ROCK GEOCHEMISTRY
The compilation of the geochemical data collected along the North Fiji Basin spread-
ing system between latitude BOS and 22°S during the STARMER project (Eissen et aI.,
1994; Lagabrielle et aI., 1994; Nohara et al., 1994), in addition to all published data (Sinton
et aI., 1985; Aggrey et aI., 1988; Auzende et al., 1990a; Boespflug, 1990; Eissen et aI., 1990,
1991; Johnson and Sinton, 1990; Price et aI., 1990; Monjaret et aI., 1991; Sinton etal., 1994),
shows systematic geochemical variation along the different spreading segments. Crystal
fractionation-controlled almost exclusively by olivine for the less fractionated lavas,
or by plagioclase + olivine ± clinopyroxene for the most fractionated lavas-can explain
part of the observed major element geochemical variations. However, the large ion
lithophile elements (LILE), high field strength elements (HFSE), and rare earth elements
(REE) contents of the basalts of the NFB as well as their isotopic compositions are strongly
heterogeneous. Mixing of mantle sources must play a major role in order to explain the
observed LILE, HFSE, REE, and isotopic variations. Thus, three different mantle sources
have been identified in the NFB (Eissen et aI., 1994):
1. A typical depleted N-MORB source, producing basalts with flat patterns on
N-MORB-normalized extended element variation diagrams (Fig. 4.14d), 1 < LalNb < 2
(Fig. 5.15; N-MORB field of Gill, 1981), and Sr- and Nd-isotopic ratios close to the Pacific
MORB field (Fig. 4.14i).
2. An E-MORB- or OIB-related source giving basalts with a strong enrichment in
HFSE, LILE, and LREE, on N-MORB-normalized extended element variation diagrams
(Fig. 4.14a), LalNb < 1 (Fig. 4.15; E-MORB field of Gill, 1981), and Sr- and Nd-isotopic
ratios trending toward the intra-plate field (e.g. Samoan field in Fig. 4.14f).
3. A subduction-modified (or -related) mantle source, with basalts having generally a
higher volatile content (Aggrey et aI., 1988), and a clear enrichment in HFSE, LILE, and
LREE, on N-MORB-normalized extended element variation diagrams with a negative Nb
anomaly indicating some subduction-related contamination (Fig. 4.14d), LalNb > 2 (Fig.
4.14; orogenic field of Gill, 1981), and Sr- and Nd-isotopic ratios ranging from the Pacific
MORB field to the New Hebrides arc field (Fig. 4.14h). These basalts derived from this
source might also have been previously called BABB (Saunders and Tamey, 1984;
Hawkins and Melchior, 1985; Sinton and Fryer, 1987).
The observed variations, reported as a function of their respective spreading segments,
are as follows. Along the N-S segment, which also represents morphologically the most
regular spreading ridge of the NFB, the magma source produces only basalts with
N-MORB characteristics (Figs. 4.14d, 4.15). However, their isotopic compositions extend
slightly over a range wider than the Pacific MORB field, with one trend toward the Samoan
field and a weak trend toward the New Hebrides field (Fig. 4.14i). Along the N15° segment
the three sources coexist, giving a wide range of geochemical signatures for the basalts
collected (Fig. 4.15). Their mixing results in a large number of transitional compositions
giving patterns of E-MORB types (Fig. 4.14c). Their isotopic compositions are essentially
similar to the more Sr radiogenic of the Pacific MORB with two additional groups having,
respectively, an OIB- and a subduction-related trend (Fig. 4.14h). The source mixing
transitional toward OIB (hot-spot-related) increases northward from 18°20'S to 12°S, as
the Rotuma-Samoan hot-spot lineament is approached. Thus, along the N160° segment,
the three sources still coexist, but the influence of the OIB source increases, whereas the
influence ofthe subduction-related source decreases (Figs. 4.14b and 4.15). It is particularly
~lO
i
?
.. 140 508
.....\ T.J.?
... 15°5
\\~
\~
".-
.. :1'~:508
~
~l
, :'
6 j T.J.
o
:'
""""" 'J'::::)
" NS 1MIIgm.m
c
~~
Za 20·S
I ...
-
" 0 Sl2S
w: ,_,
I---"'-<:--~--'~
" ,.....
0..5131
w southern~ w
most:
I (::::~ segment! Ie "-
"-
"
"-
"- 0 ..5128
e "-
0.1025 0.7030 0.7035 "0:'1040

"S,I"'Sr
FIGURE 4.14. Incompatible elements and isotopic variations of the North Fiji Basin basalts as a function of a
schematic spreading segmentation (central sketch); (a-e) N-MORB-normalized extended element variation diagrams
(normalization values after Sun and McDonough, 1989); (f-j) variations of 87Sr/86Sr with 143Ndll44 Nd (Pacific
MORB field after Boespftug, 1990; Samoan field after Wright and White, 1987; NH: New Hebrides I field after
Briqueu et ai., 1994). Light gray field: OIB-related source; dark gray field: N-MORB source; thin lines field: tran-
sitional toward OIB (or E-MORB) source; heavy lines field (with negative Nb anomaly): subduction related source.
Rotuma
S
+
South jfi N 168
4
3
Pandora
..
•
·c
••:;:: t.• ~:.:j;

I- . : : :
1
·tE~l ..
~ ¥:~..:
!
~ .11:
[ 1 •
8L--1~3~~~1~S~~~1~7~~~1~9--~~2~1--~
Latitude South
FIGURE 4. 15. La/Nb along-strike variations of the North Fiji Basin basalts as a function of the latitude of the
sampling site.
well marked on the isotopic composition diagram (Fig. 4.14g) where all the compositions
spread between the Pacific MORB and the Samoan fields. Along the Pandora-Rotuma
Ridge, only the OIB-derived lava type is present, the eventual contribution from other
sources being completely diluted (Figs. 4.14a, 4.14f, 4.1S). Mixing calculations based on
incompatible elements involving OIB and N-MORB sources show that the OIB source
contribution is between SO% and 80% for the Rotuma-Pandora Ridge lavas and is lower
than 2S% for the N160° and N1So segments (Eissen et aI., 1994). On the southernmost
segment (N174°E), N-MORB are present, but the dying subduction along the Hunter Ridge
still influences several basalt compositions, which show a significant subduction-related
contamination with negative Nb anomalies and high volatiles content. A weak E-MORB
source contribution is also present in several others and may be related to subducted-OIB
seamounts from the South Fiji Basin (Fig. 4.14e, 4.14j). However, no significant isotopic
variations are observed along this segment (Fig. 4.1S).
Thus, the geochemistry of the NFB magma sources is directly influenced by the
surrounding active or dead subduction zones and the regional OIB (or hot-spot)-related
sources. The magma genesis is dominantly controlled by magma production along the
spreading system, resulting in the dominance of N-MORBs. However, this source mixes
to a various extent with the two sources mentioned previously. In the entire northern NFB
many basalts result from the mixing of an N-MORB and a OIB source, similar to transi-
tional alkalic lavas from oceanic intraplate magmatism. This source is related either to the
Samoan lineament (Wright and White, 1987) or to the Fiji Platform (Gill, 1984) or even to
the Wallis lineament (Price et aI., 1991). Similarly, if the subduction-related contamination
along the 174°E segment is directly linked to the dying subduction at the southern end of the
New Hebrides arc, the presence of perceptible subduction-related geochemical characteris-
tics in basalts from the central NFB, SOO km away from the active subduction zone, is more
questionable. This influence results in fact from the partial melting of an upper mantle
source that was affected by subduction contamination during the clockwise rotation of the
New Hebrides arc leading to the opening of the NFB during the past 10 m.y. (Ma1ahoff
et aI., 1982a; Auzende et ai., 1988a) and is no longer directly linked to the presently active
subduction.
6. SULFIDE DEPOSITS
Since the beginning of the SEAPSO and STARMER projects, two main hydrothermal
sulfide deposit fields have been explored by Nautile in 1989 (Auzende et aI., 1991b) and
Shinkai 6500 in 1991 (Auzende et aI., 1992a) in the vicinity of the 16°S0'S triple junction
(Lafoy et aI., 1987, 1990). The first (named "White Lady") at 16°S9'S and the second
(named "Pere Lachaise") at 16°S8'S were both discovered during the Nautile dives
in 1989.
The main structural features, which characterize the 2-km-wide, l00-IS0-m-deep rift
valley of the NFB in the White Lady area, are steep fault scarps associated with open
fissures. Fissures a few meters long and a few centimeters to I-m wide are commonly
observed on the seafloor and on the terraces adjacent to the fault scarps. The fault trend is
parallel to the main trend of the central spreading ridge, NIO° to N20° (Fig. 4.16). The
White Lady chimney (Figs. 4.16 and 4.17) is located on the top of a 6-7 -m-high,
SO-m-diameter mound composed of sulfides and oxides. The chimney bifurcates and is
[2]1
~2
1»13.
!t::t:~tI4
II1II5
FIGURE 4.16. Detailed structural map of the ax-

ial graben close to 17°S. Location of the active
hydrothermal sites (White Lady and STARMER)
and major sulfides deposit (pere Lachaise) (after
Bendel et ai., 1993). I: normal faults; 2: crest; 3: 01'
main axial graben; 4: intersection zone; 5: second-
ary graben; 6: active hydrothermal site; 7: fossil lKm
chimneys; 8: temperature anomalies; 9: lava lakes;
fl6 010
10: hydrothermal oxydes staining; 11: mussels and , 7
gastropods colonies. The small arrows indicate the (!)8 A 11
sampling areas out of the White Lady site. <e9
!I Shimmering water
- Shimmering water
;IIl!IE"lIlFIrIf;'~ Anhydrite chimney
Animal colonies
1m
FIGURE 4.17. Artistic view of the White Lady evolution from (a) 1989 and (b) 1991 (after Auzende etai., 1992a).
composed exclusively of anhydrite. It expels shimmering 285°C water, which is poor in

particulates. At the foot of the main chimney, six secondary vents expelling the same
transparent water are colonized by gastropods, mussels, crabs, and cirripeds. In 1991, six
Shinkai 6500 dives were devoted to the study of the White Lady and of the surrounding
active hydrothermal sites. The present-day condition of the White Lady can be summarized
as follows: the anhydrite chimney currently comprises a massive 2-m-high, 2-m-diameter
main conduit and a 1.5-m-high, lO-cm-diameter secondary chimney. These vents expel the
same shimmering water, but the flux is about twice that of two years ago. The measured
temperature is 265°C, 20°C less than the measured temperature in 1989. Finally, significant
changes (Fig. 4.17) have been observed for the colonies of living fauna. They are in all cases
more numerous, and their territories are expanding toward the top of the mound. The
sulfates, oxides, and sulfides have been studied in great detail by Bendel (1993) and Bendel
et al. (1993), who deduced from geological observations and mineralogical study four
stages in the evolution of the White Lady site. The sulfide spires are formed in the first stage
by mixing of hydrothermal fluid with seawater. The sulfide mound and the maturation of
sulfides reflect the second stage. Copper-rich massive sulfides are precipitated at the core
of the mound. During the third stage the sulfides are tectonically brecciated, and the
resulting fissures are filled with amorphous silica, talc, and barite. The fourth stage
corresponds to anhydrite precipitation related to the present-day activity at the top of the
sulfide mound.
The Pere Lachaise site is characterized by the disappearance of the axial graben and
the increasing width of the domain. This area, close to the 16°50' S triple junction, exhibits
three sets of normal fault and fissures trending NI5°, NI40-150°, and N60°. The explored
fossil hydrothermal field consists of several tens of individual spires growing directly on
basalt scattered over a 2-km x 2-km surface. The deposits sampled from the Pere Lachaise
site indicate the same compositions as are observed on the White Lady mound. The absence
of a massive mound in this area is interpreted as the result of the conjunction of the three
directions of faulting, which creates a zone of high porosity where hydrothermal discharge
is dispersed over a large area (Bendel et ai., 1993).
7. WATER SAMPLING
The first significant water samples on the ridge axis were taken in 1985 during the
SEAPSO III cruise (Auzende et aI., 1988a) and the Papatua cruise in 1986 (Craig, 1986).
Other water samplings have been conducted on the northern part of the NFB ridge during
the Moana Wave cruise in 1987. They indicate hydrothermal activity around the South
Pandora Ridge (Kroenke et aI., 1987).
During the SEAPSO III cruise, seven hydrocasts were taken on the axis between 16°S
and 20 0 S. Four of these indicate significant methane and manganese anomalies close to the
seafloor; two hydrocasts are located at the junction of the N-S and N15° segments and the
other two are on the N160° segment close to the 16°50'S triple junction. These last
hydrocasts show anomalies 10 times higher than background level for the deep Pacific
water. During the STARMER project, hydrocasts were taken about every 20 miles all along
the ridge axis. They demonstrate the quite constant distribution of the hydrothermal activity
along the spreading ridge axis.
In situ sampling was conducted during the STARMER Nautile cruise (1989) and the
Shinkai 6500 cruise (1991) on the White Lady active site at 17°S in the triple junction area
(Auzende et at., 1991b). The main results can be summarized as follows (Grimaud et at.,
1991; Ishibashi et at., 1994): the samples were collected with conventional titanium syringes
by Nautile or with a pumping system by Shinkai 6500. In total, 16 samples were taken on
the White Lady site. Their compositions range from a few percent hydrothermal water to
nearly pure 285°C temperature fluids. The elemental concentrations show linear trends on
element versus Mg diagrams. All the element concentrations of the parent end member are
significantly lower than all other hydrothermal systems. Chloride (255 mmol/kg) and
sodium (210 mmol/kg) have especially low concentrations compared with fluids collected
on the East Pacific Rise (EPR), Juan de Fuca Ridge, Galapagos Ridge (von Damm and
Bischoff, 1987) and Mid-Atlantic Ridge (Campbell et at., 1988). The only similar fluids
have been sampled in the Ashes vent field on the Juan de Fuca Ridge (Massoth et at., 1989;
Butterfield et at., 1990). In spite of these low concentrations the characteristic elemental
ratios KINa and Ca/Sr are close to the values usually found in other hydrothermal systems
(von Damm and Bischoff, 1987).
B. DISCUSSION AND CONCLUSION
The evolution of the NFB, located between the opposed New Hebrides and Tonga
subduction zones, results from variation in the relative motions of the Australian and
Pacific plates. At a large scale the NFB can be interpreted as a deformation zone at the
boundary of the two plates. Depending on the geometry of the boundary, the shortening
between the plates will be accommodated by compressive stresses such as along New
Hebrides and Tonga trenches or by strike-slip motion such as in the northern part of the
basin. The geometry of the ridge axis, its segmentation, and its changes of direction are
controlled by this relative motion.
B.1. Hectokilometric, Decakilometric, and Kilometric Ridge Segmentation

The precise mapping of the central spreading ridge illustrates different scales of
segmentation. At a large scale, five main segments can be considered. Their length varies
from about 140 km for the southernmost N-S segment to more than 280 km for the N-S
segment. Although the spreading rate is intermediate all along the central spreading ridge,
ranging from 50 mmlyr on the N160° and N15° segments to 80 mmlyr on the N-S segment,
the segments show very different morphologies. The N-S and N15° segments' morphology
is typical ofEPR fast-spreading-type morphology (Fig. 4.4) with a central dome 8 km wide
cut in its axial part by a graben 50 m to several hundred meters wide and 50 m deep. In
contrast, the N160° segment is characterized by an 8-km-wide, 1000-m-deep graben
occupied in its axial part by a neovolcanic zone (Fig. 4.4). This morphology is typical of
slow-spreading MAR-type morphology. The conclusion drawn from these observations is
that there is not a unique correlation between the ridge morphology and the spreading rate.
It is also necessary to take into account the magmatic budget of each area and its
geodynamic context. The drastic change of morphology north and south of the 16°50'S
triple junction suggests that the triple junction area constitutes a magmatic supply boundary
between the N15° segment, characterized by significant volcanic construction, and the
amagmatic extension of the N160° segment.
Within each major segment, segmentation at the scale of 10 s of kilometers is

observed. These subsegments are limited by offsets, OSCs, and propagating rifts as
described on the EPR (Macdonald, 1983) and as illustrated in Fig. 4.3. The average lengths
of these intermediate subsegments varies from 40 Ian on the N-S segment to more than
60 Ian on the N160° segment.
Segmentation at a scale of a kilometer has also been documented on the N-S segment
around 19°5 by the in situ exploration with Nautile and Shinkai 6500 (Ondreas et al., 1993;
Gracia-Mont et aI., 1994). The active accretion is located in tectonic grabens a kilometer
long and tens of meters wide separated by kilometer-wide magmatic saddles between the
grabens. The active hydrothermal sites are located in the grabens, whereas very fresh mag-
matic edifices are observed on the saddles.
B.2. Unusual Tectonic Features

One of the characteristics of spreading in the NFB is the existence all along the ridge
axis of unusual tectonic features such as propagating rifts, OSC, offsets, and others,
representative of the complex geodynamic environment of the basin. At a global scale the
NFB can be considered as a megashear zone at the boundary between the Australian and
Pacific plates (Hamburger and !sacks, 1988). The effect of these permanent shear stresses
applied on the whole basin is the deformation of the spreading ridge. This deformation is
accommodated by changes of morphology and migration of the ridge axis with initiation of
triple junctions (examples of 16°50'S and 14°50'S triple junctions) and ridge propagation
(northward propagation of the N-S segment at 18°lO'S, southward propagation of the N-S
segment at 20 0 30'S) and activation of major transform faults (north Fiji fracture zone,
central fracture zone, and south Fiji fracture zone) (Fig. 4.18).
Also, probably as a result of the geodynamic environment, the accretion south of the
north Fiji fracture zone is distributed on two parallel active spreading ridges, the cen-
tral spreading ridge and the West Fiji Ridge, propagating at a large scale, respectively to the
north and to the south (Auzende et aI., 1993, 1994b). Since at least 1.5 to 1 Ma, this twin-
spreading-ridge system isolates an intermediate West Fiji microplate between the central
spreading ridge and the West Fiji Ridge (Fig. 4.18). The same phenomenon is observed for
the Easter (Francheteau et aI., 1987; Larson et al., 1992) and Juan Fernandez microplates.
The spreading rates on both ridges vary, increasing from south to north on the West Fiji
Ridge and from north to south on the central spreading ridge. The sum of spreading rates
is close to 100 mmlyr all along the N-S oceanic crust created during the last 1 to
1.5 m.y.
B.3. Petrology and Geochemistry

The compilation of all the geochemical data available for the basalts of the NFB shows
that low-pressure crystal fractionation controls the major element geochemistry, and man-
tle source heterogeneities and their mixing dominantly control the LILE, HFSE, and REE
patterns and isotopic compositions of these basalts. The dominant magma of this backarc
basin is an N-MORB type, producing basalts very similar to those emplaced along the most
classical intermediate-rate mid-oceanic spreading centers. But other geochemical signa-
tures, representative of the local upper mantle geochemistry, are observed: (1) an OIB
source, similar to transitional alkalic lavas from oceanic intraplate magmatism; its influ-
ence increases northward from 18°20'S to 12°S, this lava type being the only one present
FIGURE 4.18. Kinematic sketch of the twin ridges (central spreading ridge and the West Fiji Ridge) with
simplified bathymetry. NFFZ: north Fiji fracture zone; CFZ: central Fiji fracture zone; SFFZ: south Fiji fracture
zone; N-S, Nl5 and N160: the different segments of the main central spreading ridge. 5-6: spreading rate
calculated from magnetic data. 0-2?: inferred spreading rate in crn/yr. The arrows at the tip of ridge segments
indicate the direction of propagation (after Auzende et aI., 1994b).
along the Rotuma-Pandora Ridge; and (2) a source that was affected by subduction
contamination from the New Hebrides subduction, with a clear enrichment in incompatible
elements, a negative Nb anomaly, and a high loss on ignition (LOI). This contamination
with variable arc affinity, which affected the entire NFB upper mantle, was left behind after
the clockwise rotation of this arc, as a result of the opening of the NFB. This subduction-
related material was then recycled during partial melting of the NFB upper mantle beneath
the central spreading axis.
8.4. Hydrothermal Activity: Water Chemistry and Sulfides

The explored hydrothermal active sites of the NFB central spreading ridge are
probably only representative of a few percent of the hydrothermal activity. The hydrocasts
taken all along the ridge axis suggest a widespread activity with the indication in some
places of a very strong but unstable activity. For example, the 19°5 "megaplume" (Nojiri
et aI., 1989) discovered during the Kaiyo cruise in 1987 was extinct two years later when
attempts were made to sample it again.
One of the main characteristics of the NFB ridge axis hydrothermal waters sampled by
Nautile and Shinkai 6500 on the White Lady site is their low salinity, which suggests that
fluids have undergone phase separation (Grimaud et al., 1991; Ishibashi et aI., 1994). Such a
phase separation would result in the formation of one vaporlike phase with low salinity and
one liquid-like phase with high salinity (brine). Dissolved species will be depleted in the
vapor, but phase separation will not significantly modify the elemental ratios as observed
in the White Lady waters. To conclude, the NFB hydrothermal waters could result from
three-component mixing of normal hydrothermal water, low-salinity fluid from condensed
vapor, and brine.
Concerning the sulfides, the analysis of the White Lady site deposits succession
implies three different hydrothermal episodes (Bendel, 1993; Bendel et aI., 1993): (1) The
first episode is characterized by deposition of Cu, Fe, and Zn-rich sulfides and implies that a
hot (more than 300°C) water was being expelled. This episode is close to the mid-oceanic
ridge type. (2) The second hydrothermal episode produced opal, barite, and talc precipi-
tates. It certainly occurs after a tectonic event which was responsible for the sulfides mound
brecciation. This tectonic event caused the fluids to become more oxidized with tempera-
tures ranging from 150°C to 300°C, depending on subsurface mixing. (3) The third
hydrothermal episode only produced anhydrite and minor opal and talc. This deposit was
associated with low elemental abundances in the expelled water and could be, as suggested
by Grimaud et al. (1991) and Ishibashi et al. (1994), the result of a phase separation after
boiling near the surface. The succession of these three hydrothermal phases results in a
great instability of the hydrothermal system over a short period of time (less than 1 m.y.).
This instability could be related to the tectonic events affecting the area, especially the
intense tectonic fracturing associated with the 16°50'S triple junction.
B.S. Evolution of the Basin

The recent data acquired on the northern and the northwestern parts of the NFB
(especially those of Pelletier et aI., 1988, 1993a,b) allow construction of a previously
proposed geodynamical evolution model (Auzende et aI., 1988b) to be completed. The
evolution of the NFB is illustrated by the six cartoons of Fig. 4.19.
l. At 12 Ma the Vitiaz-New Hebrides-Fiji-Lau-Tonga arc split and NFB rifting started with
the change of subduction polarity from the Vitiaz to the New Hebrides system. The main part
of the initial arc remained on the southern side of the spreading axis.
2. Spreading along a NW-SE axis was synchronous with the clockwise rotation of the New
Hebrides arc and the anticlockwise rotation of the Fiji Platform. The initial axis located north
of the Fiji Platform jumped southward between the New Hebrides and Fiji. The crust created
during this stage is mainly located in the western basin.
3. At 7 Ma the NW-SE spreading axis stopped and was replaced by an E-W trending spreading
center from the northwestern tip of the basin to the north of the Fiji Platform. This center
corresponds to the 9°30'S, Tikopia, and proto-south Pandora ridges. This opening induced a
new subduction zone along the southern limit of the NFB.
4. A triple junction was active around 3 Ma (anomaly 2A) between the former E-W axis and the
newly created N-S-trending spreading center. At this stage there was a major change in the
stress direction from N-S to E-W. This period was synchronous with the beginning of
spreading in the Lau Basin.
5. Around 1.5 Ma the opening along the E-W axis was reorganized by the development of the
north Fiji fracture zone along the Fiji Platform up to the N-S spreading axis, creating the
16°50'S triple junction. The left-lateral strike-slip motion along the north Fiji fracture zone
induced the change of the direction of the N-S axis at the 16°50'S triple junction. The West
Fiji and Hazel Holme ridges developed to accommodate strike-slip motion along the north
Fiji fracture zone.
, ..... - ..., 1.5Ma

Pre •• nt day 10.S
" ...I't
......
' .... _-- ,--
......
20'
170' 180'E
FIGURE 4.19. Geodynamic evolution of the North Fiji Basin. See text for explanation. I: active ridge axis;
2: incipient ridge axis; 3: transform fault; 4: flowlines; 5: active subduction zone; 6: incipient subduction zone. VT:
Vitiaz Trench; NH: New' Hebrides arc; F: Fiji Platform; NC: Nouvelle-Caledonie.
6. At present, the spreading south of the north Fiji fracture zone is distributed along the parallel
central spreading ridge and West Fiji Ridge isolating an intennediate microplate. To the south,
the southernmost spreading axis is connected to the New Hebrides Trench by a large left-
lateral strike-slip fault. North of the 16°50' S triple junction, opening occurs along the N160°
segment and the Hazel Holme Ridge. The Santa Cruz Trough, which crosscuts the north-
ernmost part of the New Hebrides arc, could be linked to the Hazel Holme spreading system.
Crustal motion probably occurs in the north and the northeastern part of the NFB along the
South Pandora Ridge and its eastern prolongation, interpreted either as a slow spreading ridge
or a strike-slip fault. This lineament is connected with the N1600-Hazel Holme system by the
14°50'S triple junction.
Finally the opening of the NFB can be divided into three major stages: an opening in a
NE-SW direction from 12 to 7 Ma, an opening in a N-S direction from 7 to 3 Ma, and an
opening in an E-W direction from 3 Ma to the present day. The triangular shape of the basin
results from these three successive spreading phases. Since the beginning of the creation of
the NFB, the location of the successive spreading centers has migrated southward to
accompany the migration of the New Hebrides arc.
Acknowledgments
We thank all the scientific parties and the ships and submersibles crews involved in the
studies of NFB during the SEAPSO, EVA, and STARMER projects. We thank Thomas
Brocher, Brian Taylor, and an anonymous reviewer for their constructive comments and for
their help in improving our English. J. Butscher and J. Perrier are thanked for the prepara-
tion of the illustrations.
REFERENCES
Aggarwal, Y. P., Barazangi, B., and Isacks, B. 1972. P and S travel times in the Tonga-Fiji region: A zone of low
velocity in the uppermost mantle behind the Tonga Island arc, J. Geophys. Res. 77:6427-6434.
Aggrey, K. E., Muenow, D. W., and Sinton, J. M. 1988. Volatile abundances in submarine glasses from the North
Fiji and Lau backarc basins, Geochim. Cosmochim. Acta 52:25401-2506.
Alteris (de), G., Ruellan E., Auzende J. M., Ondreas H., Bendel v., Gracia-Mont E., Lagabrielle Y., Huchon P., and
Tanahashi M. 1993. Propagating rifts in the North Fiji Basin (Southwest Pacific), Geology 21:583-586,
Auzende, J. M., Bendel, V., Fujikura, K., Geistdoerfer, P., Gracia-Mont, E., Joshima, M., Kisimoto, K., Mitsu-
zawa, K., Murai, M., Nojiri, Y., Ondreas, H., Pratt, C., and Ruellan, E. 1992a. Resultats preliminaires des
plongees du "Shinkai 6500" sur la dorsale du Bassin Nord-Fidjien (SW Pacifique)-STARMER, CR. Acad.
Sci. Paris 314(11):491-498.
Auzende, J. M., Boespflug, X., Bougault, H., Dosso, L., Foucher, J. P., Joron, J. L., Ruellan, E., and Sibuet, J. C.
1990a. From intraoceanic extension to mature spreading in back arc basins: Examples from the Okinawa, Lau
and North Fiji Basin, Actes du Colloque Tour du Monde Jean Charcot, Oceano/. Acta 10:153-163.
Auzende, J. M., Eissen, J. P., Caprais, M. P., Gente, P., Gueneley, S., Harmegnies, E, Lagabrielle, Y., Lapouille, A.,
Lefevre, c., Maillet, P., Maze, J.P., Ondreas, H., Schaaf, A., and Singh, R. 1986a. Accretion oceanique dans la
partie meridionale du bassin Nord-Fidjien: Resultats preliminaires de la campagne oceanographique
SEAPSO III du N.O. Jean Charcot (decembre 1985), CR. Acad. Sci. Paris 303:93-98.
Auzende, J. M., Eissen, J. P., Lafoy, Y., Gente, P., and Charlou, J. L. 1988a. Seafloor spreading in the North Fiji
Basin (Southwest Pacific), Tectonophysics 146:317-351.
Auzende, J. M., Gracia-Mont, E., Bendel, v., Huchon, P., Lafoy, Y., Lagabrielle, Y., de Alteris, G., and Tanahashi,
M. 1994a. A possible triple junction at 14°50'S on the North Fiji Basin Ridge (Southwest Pacific), Mar. Geo/.
116:25-36.
Auzende, J. M., Gracia-Mont, E., Bendel, v., Lafoy, Y., Lagabrielle, Y., Okuda, Y., and Ruellan, E. in press.
Amagmatic extension at intermediate spreading ridge (North Fiji Basin), in Oceanic Lithosphere, Special
Issue (P. Kapezinskhas, ed.), Wiley, New York.
Auzende, J. M., Hey, R. N., Pelletier, B., Lafoy, Y., and Lagabrielle, Y. 1993. Propagation d'une zone d'accretion a
l'est de la dorsale du bassin Nord Fidjien (SW Pacifique), CR. Acad. Sci. Paris 317(11):671-678.
Auzende, J. M., Honza, E., Maze, J. P., and the STARMER Group. 1992b. Comments on the Seabeam map of the
North Fiji Basin Ridge between 16°IO'S and 21°40'S, Ofioliti 17(1):43-53.
Auzende, J. M., Honza, E., and the STARMER Group. 1990b. Bathymetric map of the North Fiji Basin Ridge be-
tween 16°IO'S and 21°40'S, published by IFREMER and STA Japan, six colored sheets edited by Beicip, Paris.
Auzende, J. M., Lafoy, Y., and Marsset, B. 1988b. Recent geodynamic evolution of the North Fiji Basin (SW
Pacific), Geology 16:925-929.
Auzende, J. M., Lagabrielle, Y., Schaaf, A., Gente, P., and Eissen, J. P. 1986b. Tectonique intra-oc6anique
d6crochante al'ouest des iles Fidji (Bassin Nord Fidjien). Campagne SEAPSO III du N.O. Jean Charcot, C.R.
Acad. Sci. Paris 303:241-246.
Auzende, J. M., Okuda, Y., Bendel, v., Ciabrini, J. P., Eissen, J. P., Gracia, E., Hirose, K., Iwabushi, Y., Kisimoto,
K., Lafoy, Y., Lagabrielle, Y., Marumo, K., Matsumoto, T., Mitsusawa, K., Momma, H., Mukai, H., Nojiri,
Y., Okuda, Y., Ortega-Osorio, A., Ruellan, E., Tanahashi, M., Tupua, E., and Yamaguchi, K. 1991a.
Propagation "en 6chelon" de la dorsale du Bassin Nord Fidjien entre 16°40 et 14°50S (Yokosuka 90-
STARMER), C.R. Acad. Sci. Paris 3U(II):1531-1538.
Auzende, J. M., Pelletier, B., and Lafoy, Y. 1994b. Twin active spreading ridges in the North Fiji Basin (S.w.
Auzende, J. M., and the STARMER Group. 199Oc. Active spreading and hydrothermalism in North Fiji Basin (SW
Pacific). Results of Japanese-French cruise Kaiyo 87, Mar. Geophys. Res. U:269-283.
Auzende, J. M., Urabe, T., Bendel, V., Deplus, c., Eissen, J. P., Grimaud, D., Huchon, P., Ishibashi, J., Joshima,
M., Lagabrielle, Y., Mevel, c., Naka, J., Ruellan, E., Tanaka, T., and Tanahashi, M. 1991b. In situ geological
and geochemical study of an active hydrothermal site on the North Fiji Basin Ridge, Mar. Geol. 98:259-269.
Barazangi, M., Isacks, B. L., Dubois, J., and Pascal, G. 1974. Seismic wave attenuation in the upper mantle
beneath the Southwest Pacific, J. Geophys. Res. 76:8493-8516.
Bendel, V. 1993. Cadre geologique et composition des min6ralisations hydrothermales en contexte arriere-arc:
exemple de la dorsale du Bassin Nord Fidjien, These Universit6 Brest, France.
Bendel, v., Fouquet, Y., Auzende, J. M., Lagabrielle, Y., Grimaud, D., and Urabe, T. 1993. Metallogenesis at a
Triple Junction system: The White Lady hydrothermal field (North Fiji Back-Arc Basin, SW Pacific),
Economical Geol. 88:2237-2249.
Boespflug, X. 1990. Evolution g60dynamique et g60chirnique des bassins arriere-arcs. Exemples des bassins
d'Okinawa, de Lau et Nord-Fidjien, These, Universit6 Brest, France.
Briqueu, L., Laporte, C., Crawford, A., Hasenaka, T., Baker, P., and Coltorti, M. 1994. Temporal magmatic
evolution of the Aoba basin-central New Hebrides Island Arc: Pb, Sr and Nd isotopic evidence for the
coexistence of two mantle components beneath the arc, in Proc. OD?, Sci. Results, 134 (H. G. Greene, J.- Y.
Collet, L. B. Stokking et al., eds.), pp. 393-401, Ocean Drilling Program, College Station, TX.
Brocher, T. M. 1985, On the formation of the Vitiaz Trench lineament and North Fiji Basin, in Investigations of the
Northern Melanesian Borderland (T. M. Brocher, ed.), pp. 13-34, Circum-Pacific Council for Energy and
Mineral Resources, Houston, TX, Earth Science Series, Vol. 3.
Brocher, T. M., and Holmes, R. 1985. The marine geology of sedimentary basins south of Viti Levu, Fiji, in
Investigations of the Northern Melanesian Borderland (T. M. Brocher, ed.), pp. 123-138, Circum-Pacific
Council for Energy and Mineral Resources, Houston, TX, Earth Science Series, Vol. 3.
Butterfield, D. A., Massoth, G. J., McDuff, R. E., Lupton, J. E., and Lilley, D. 1990. Geochemistry of hydrothermal
fluids from Axial Seamount Hydrothermal Emissions Study vent field, Juan de Fuca Ridge: Subseafloor
boiling and subsequent fluid-rock interaction, J. Geophys. Res. 95:12,895-12,921.
Campbell, A. C., et al. 1988. Chemistry of hot springs on the Mid-Atlantic Ridge, Nature 335:514-518.
Cande, S. c., and Kent, D. V. 1992. A new geomagnetic polarity time scale for the late Cretaceous and Cenozoic, J.
Geophys. Res. 97(BIO):13,917-13,951.
Charvis, P., and Pelletier, B. 1989. The northern New Hebrides backarc troughs: History and relation with the
North Fiji Basin, Tectonophysics 170:259-277.
Chase, C. G. 1971. Tectonic history of the Fiji Plateau, Geol. Soc. Am. Bull. 82:3087-3110.
Cherkis, N. Z. 1980. Aeromagnetic investigations and sea floor spreading history in the Lau basin and the northern
Fiji Plateau, in Symposium on Petroleum Potential in Island Arcs, Small Basins, Submerged Margins and
Related Areas (W. 1. Clark, ed.), United Nations, ESCAP, CCOP/SOPAC, Technical Bull. 3:37-45.
Craig, H. 1986. Papatua 86: A grand tour of the Havre Trough and the Lau, North Fiji, Woodlark and Manus
Basins, Eos 67:378.
Daniel, J. 1982. Morphologie et structures superficielles de la partie sud de la zone de subduction des Nouvelles-
H6brides, in Contribution a l'etude geodynamique du Sud-Ouest Pacifique, Equipe de G60logie- G60-
physique du Centre ORSTOM de Noum6a, Travaux et Documents de l'ORSTOM 147:39-60.
Dubois, J. 1971. Propagation of P waves and Rayleigh waves in Melanesia: structural implications, J. Geophys.
Res. 76:7217-7240.
Dubois, 1., Launay, J., R6cy, J., and Marshall, J. 1977. New Hebrides Trench: Subduction rate from associated
lithospheric bulge, Can. J. Earth Sci. 14:250-255.
Dubois, J., Pascal, G., Barazangi, M., Isacks, B. L., and Oliver, 1. 1973. Travel times of seismic waves between the
New Hebrides and Fiji Islands: A zone of low velocity beneath the Fiji Plateau, J. Geophys. Res. 78:3431-
3436.
Eguchi, T. 1984. Seismotectonics of the Fiji Plateau and Lau Basin, Tectonophysics 102:17-32.
Eissen,1. P., Lefevre, C., Maillet, P., Morvan, G., and Nohara, M. 1991. Petrology and geochemistry of the central
North Fiji Basin spreading center (SW Pacific) between 16°S and 22°S, Mar. Geol. 98:201-239.
Eissen, J. P., Morvan, G., Lefevre, C., Maillet, P., Urabe, T., Auzende, J. M., and Honza, E. 1990. Petrologie et
geochimie de la zone d'accretion du centre du bassin Nord Fidjien (SW Pacifique), C.R. Acad. Sci. Paris
310(11):771-778.
Eissen, J. P., Nohara, M., and Cotten, J. 1994. North Fiji Basin basalts and their magma sources. 1: Incompatible
element constraints, Mar. Geol. 116:163-178. '
Falvey, D. A. 1975. Arc reversals, and a tectonic model for the North Fiji Basin, Austr. Soc. Explor. Geophys. Bull.
6:47-49.
Falvey, D. A. 1978. Analysis of paleomagnetic data from the New Hebrides, Austr. Soc. Explor. Geophys. Bull.
9(3):117 -123.
Francheteau, J., Yelles-Chaouche, A., and Craig, H. 1987. The Juan Fernandez microplate north of the Nazca-
Pacific-Antarctic plate junction at 35°S, Earth Planet. Sci. Lett. 86:253-286.
Gente, P., Mevel, C., Auzende, J. M., Karson, J. A., and Fouquet, Y. 1991. An example of recent accretion on the
Mid-Atlantic Ridge: The Snake Pit neovolcanic ridge (MARK area, 23°22'N), Tectonophysics 190:1-29.
Gill, 1. B. 1981. Orogenic Andesites and Plate Tectonics, Springer-Verlag, Berlin.
Gill, 1. B. 1984. Sr-Pb-Nd isotopic evidence that both MORB and OIB sources contribute to oceanic island arc
magmas in Fiji, Earth Planet. Sci. Lett. 68:443-458.
Gill, J. B., and Gorton M. 1973. A proposed geological and geochemical history of eastern Melanesia, in The
Western Pacific: Island Arcs, Marginal Seas and Geochemistry (P. J. Coleman, ed.), pp. 543-566, University
of Western Australia Press.
Gracia-Mont, E. 1991. Etude morphostructurale du segment N160° la dorsale du Basin Nord-Fidjien, Analyse des
donnees de la campagne Yokosuka 90, Rapport de D.E.A., Universite de Bretagne Occidentale, Brest.
Gracia-Mont, E. 1992. EI segment N160 de la Conca Nord-Fijiana (Pacific Sud-Oest): Morfoestructura d'un eix
d'acrecio oceanica d'edat quaternaria dins una conca marginal, Tesi de Licenciatura, Universitat de Bar-
celona, Barcelona.
Gracia-Mont, E., Ondreas, H., Bendel, v., and STARMER Group. 1994. Multiscale morphologic variability of the
North Fiji Basin Ridge (Southwest Pacific), Mar. Geol. 116:\33-162.
Grimaud, D., Ishibashi, J. I., Lagabrielle, Y., Auzende, J. M., and Urabe, T. 1991. Chemistry of hydrothermal fluids
from the 17°S active site on the North Fiji Basin Ridge (SW Pacific), Chern. Geol. 93:209-218.
Gutenberg, B., and Richter C. F. 1954. Seismicity of the Earth, 2nd ed., Princeton University Press, Princeton, NJ.
Halunen, A. 1., Jr. 1979. Tectonic history of the Fiji Plateau, Ph.D. thesis, University of Hawaii, Honolulu.
Hey, R. N., Kleinrock, M. C., Miller, S. P., Atwater, T. M., and Searle, R. C. 1986. Sea BeamlDeep-Tow
investigation of an active oceanic propagating rift system, Galapagos 95.5°W, J. Geophys. Res. 91:3369-
3393.
Hamburger, M. W., and Isacks, B. L. 1988. Diffuse backarc deformation in the Southwestern Pacific, Nature
332:599-604.
Hamburger, M. w., and Isacks, B. L. 1994. Shallow seismicity in the North Fiji Basin, in Basin Formation, Ridge
Crest Processes, and Metallogenesis in the North Fiji Basin (L. W. Kroenke and J. V. Eade, eds.), pp. 21-32,
Circum-Pacific Council for Energy and Mineral Resources. Earth Science Series, Vol. 15.
Hawkins, 1. w., and Melchior J. T. 1985. Petrology of Mariana Trough and Lau Basin basalts, J. Geophys. Res.
90:11,431-11,468.
Huchon, P., Gracia, E., Ruellan, E., Joshima, M., and Auzende, J. M. 1994. Kinematics of active spreading in the
central North Fiji Basin (Southwest Pacific), Mar. Geol. 116:69-88.
Hughes Clarke, 1. E., Jarvis, P., Tiffin, D., Price, R., and Kroenke, L. 1993. Tectonic activity and plate boundaries
along the northern flank of the Fiji Platform, Geo-Mar. Lett. 13:98-106.
Ishibashi, J.-I., Grimaud, D., Nojiri, Y., Auzende, 1. M., and Urabe, T. 1994. Fluctuation of chemical compositions
of the phase-separated hydrothermal fluid from the North Fiji Basin Ridge, Mar. Geol. 116:215-226.
James, A., and Falvey, D. A. 1978. Analysis of paleomagnetic data from Viti Levu, Fiji, Austr. Soc. Explor.
Geophys. Bull. 9:15-123.
Jarvis, P., Hughes-Clarke, J., Tanahashi, M., Tiffin, D., and Kroenke, L. 1994. The western Fiji Transform Fault
and its role in the dismemberment of the Fiji Platform, Mar. Geol. 116:57-68.
Jarvis, P., Kroenke, L., Price, R, and Maillet, P. 1993. GLORIA imagery of sea floor structures in the northern
North Fiji Basin, Geo-Mar. Lett. 13:90-97.
Johnson, D. P., Maillet, P., and Price, P. 1993. Regional setting of a complex backarc: New Hebrides arc, northern
Vanuatu-eastern Solomon Islands, Geo-Mar. Lett. 13:82-89.
Johnson, K. T. M., and Sinton, J. M.1990. Petrology, tectonic setting and the formation of backarc basin basalts in
the North Fiji Basin, Geol. Jb. D 92:517-545.
Kappel, E. S., and Ryan, W B. E 1986. Volcanic episodicity and a non-steady state rift valley along Northeast
Pacific spreading centers: Evidence from SeaMarc I, J. Geophys. Res. 91(13):13,925-13,940.
Karson, J. A. 1990. Seafloor spreading of the Mid-Atlantic Ridge: Implications for the structure of ophiolites and
oceanic lithosphere produced in slow-spreading environments, in Proceedings of the Symposium "Troodos
1987" (J. Malpas, E. M. Moores, A. Panayiotou, and C. Xanophontos, eds.), Geological Survey Department,
Nicosia, Cyprus.
Karson, J. A., Thompson, G., Humphries, S. E., Edmond, J. M., Bryan, W B., Brown, J. R, Winters, AT.,
Pockalny, R A., Casey, 1. E, Campbell, A. C., Klinkhammer, G., Palmer, M. R., Kinzler, R J., and
Sulanovska, M. M. 1987. Along axis variations in seafloor spreading in the MARK area, Nature 328:
681-685.
Kleinrock, M. C. in press. The southern boundary of the Juan Fernandez microplate: braking microplate rotation
and deformation of the Antarctic plate, J. Geophys. Res.
Kroenke, L. W, Eade, 1. V., Yan, C. Y., and Smith, R. 1994. Sediment distribution in the north-central North Fiji
Basin, in Basin Formation, Ridge Crest Processes, and Metallogenesis in the North Fiji Basin (L. W.
Kroenke and J. V. Eade, eds.), pp. 63-73, Circum-Pacific Council Energy and Mineral Resources, Springer,
Heidelberg. Earth Science Series, Vol. 153.
Kroenke, L. W., Jarvis, P., and Price, R C. 1987, Morphology of the Fiji Fracture Zone: Recent reorientation of
plate boundaries in the vicinity of the North Fiji Basin, Eos 68(44):1445.
Kroenke, L. W, Price R. c., and Jarvis P. A. 1991. North Fiji Basin, 1:250,()()(), Pacific Seafloor Atlas, Hawaii
Institute of Geophysics, Honolulu, HI, sheets 10 to 17.
Lafoy, Y., Auzende, 1. M., Gente, P., and Eissen, J. P. 1987. L'extremite occidentale de la zone de fracture
Fidjienne et Ie point triple de 16°40S, Resultats du Leg III de la campagne SEAPSO du N.O. Jean Charcot
(Decembre 1985) dans Ie bassin Nord Fidjien, SW Pacifique, C.R. Acad. Sci. Paris 304:147-152.
Lafoy, Y., Auzende, J. M., Ruellan, E., Huchon, P., and Honza, E.1990. The 16°40S triple junction in the North Fiji
Basin (SW Pacific), Mar. Geophys. Res. 12:285-296.
Lagabrielle, Y., Auzende, J. M., Eissen, J. P., Janin, M. c., and Cotten, J. 1994. Geology and geochemistry ofa 800
m section through young upper oceanic crust in the North Fiji Basin (Southwest Pacific), Mar. Geol. 116:
113-132.
Larson, R L., Searle, R c., Kleinrock, M. c., Shouten, H., Bird, R T., Naar, D. E, Rusby, R I., Hooft, E. E., and
Lasthiotakis, H. 1992. Roller-bearing tectonic evolution of the Juan Fernandez microplate, Nature 356:
571-576.
Larue, B. M., Pontoise, B., Malahoff, A, Lapouille, A., and Latham, G. V. 1982. Bassins marginaux actifs du Sud-
Ouest Pacifique: Plateau Nord-Fidjien, bassin de Lau, in Contribution ii I' etude geodynamique du Sud-Ouest
Pacifique, Equipe de Geologie-Geophysique du Centre ORSTOM de Noumea. Travaux et Documents de
I'ORSTOM 147:363-406.
Louat, R 1982. Seismicite et subduction de la terminaison sud de l'arc insulaire des Nouvelles Hebrides, in
Contribution ii ['etude geodynamique du Sud-Ouest Pacifique, Equipe de Geologie-Geophysique du Centre
ORSTOM de Noumea. Travaux et Documents de l'ORSTOM 147:179-185.
Louat, R, and Pelletier, B. 1989. Seismotectonic and present-day relative plate motions in the New Hebrides-
North Fiji Basin region, Tectonophysics 167:41-55.
Luyendyck, B. P., Bryan, W B., and Jezek, P. A. 1974. Shallow structure of the New Hebrides island arc, Geol.
Soc. Am. Bull. 85:1287-1300.
Macdonald, K. C. 1983. Crustal processes at spreading centers, Rev. Geophys. 21:1441-1453.
Macdonald, K. c., Luyendyck, B. P., and Von Herzen, R P. 1973. Heat flow and plate boundaries in Melanesia, J.
Geophys. Res. 78:2537-2546.
Macdonald, K. c., Sempere, J. c., and Fox, P. J. 1984. The East Pacific Rise from the Siqueiros of the Orozco
fracture zone: Along strike continuity of the neovolcanic zone and the structure and evolution of overlapping
spreading centers, J. Geophys. Res. 89:6049-6069.
Maillet, P., Eissen, J. P., Lapouille, A., Monzier, M., Baleivanualala, v., Butscher, J., Gallois, E, and Lardy, M.
1986. La dorsale active du bassin Nord-Fidjien entre 200S et 20053'S: Signature magnetique et morpho-
logique, C.R. Acad. Sci. Paris 302(11):135-140.
Maillet, P., Monzier M., Eissen, 1. P., and Louat R 1989. Geodynamics of an arc ridge junction: The case of the
New Hebrides arc-North Fiji Basin, Tectonophysics 165:251-268.
Malahoff, A., Feden, R. H., and Fleming, H. F. 1982a. Magnetic anomalies and tectonic fabric of marginal basins
north of New Zealand, J. Geophys. Res. 87:4109-4125.
Malahoff, A., Hammond, S. R, Naughton, J. 1., Keeling, D. L., and Richmond, R N. 1982b. Geophysical evidence
for post-Miocene rotation of the island of Viti Levu, Fiji, and its relationship to the tectonic development of
the North Fiji Basin, Earth Planet. Sci. Lett. 57:398-414.
Malahoff, A., Kroenke, L. w., Cherkis, N., and Brozena, J. 1994. Magnetic and tectonic fabric of the North Fiji
Basin and Lau Basin, in Basin Formation, Ridge Crest Processes, and Metallogenesis in the North Fiji Basin
(L. W. Kroenke and 1. V. Eade, eds.), Earth Science Ser., Vol. 15, pp. 49-61, Circum-Pacific Council for
Energy and Mineral Resources.
Mammerickx, J., Chase, T. E., Smith, S. M., and Taylor, I. L. 1971. Bathymetry of the South Pacific, chart n012.
Scripps Institution of Oceanography.
Massoth, G. J., Butterfield, D. A., Lupton, J. E., McDuff, R E., Lilley, M. D., and Jonasson, I. R 1989. Submarine
venting of phase-separated hydrothermal fluids at axial volcano, Juan de Fuca Ridge, Nature 340:702-705.
Monjaret, M. C., Bellon, H., and Maillet, P. 1991. Magmatism of the troughs behind the New Hebrides island arc
(RV. Jean Charcot SEAPSO 2 cruise): K-Ar geochronology and petrology, J. Volcanol. Geotherm. Res.
46:265-280.
Monzier, M., Collot, J. Y., and Daniel, J. 1984. Carte bathymetrique des parties centrale et meridionale de I'arc
insulaire des Nouvelles-Hebrides. Office de la Recherche Scientifique et Technique Outre-Mer, Editions de
l'ORSTOM, Paris.
Monzier, M., Daniel, 1., and Maillet, P. 1990. La collision ride des LoyauteJarc des Nouvelles-Hebrides (Pacifique
Sud-Ouest), Oceanol. Acta 10:43-56.
Monzier, M., Maillet, P., and Dupont, J. 1991. Carte bathymetrique des parties meridionales de l'arc insulaire des
Nouvelles-Hebrides et du bassin Nord-Fidjien, Institut Fran~ais de Recherche Scientifique pour Ie Deve-
loppement en Cooperation, Editions de l'ORSTOM, Paris.
Nohara, M., Hirose, K., Eissen, J. P., Urabe, T., and Joshima, M. 1994. North Fiji Basin basalts and their magma
sources. II: Sr-Nd isotopic and trace element constraints, Mar. Geol. 116:179-196.
Nojiri, Y., Ishibashi, J. I., Kawai, T., Otsuki, A., and Sakai, H. 1989. Hydrothermal plumes along the North Fiji
Basin spreading axis, Nature 342:667-670.
Ondreas, H., Ruellan, E., Auzende, J. M., Bendel, v., de Alteriis, G., Gracia-Mont, E., Lagabrielle, Y., and
Tanahashi, M. 1993. Variabilite morpho-structurale 11 l'echelle kilometrique de la dorsale du Bassin Nord
Fidjien: Exploration in situ du segment compris enter 18°50'S et 19°5, C.R. Acad. Sci. Paris 316(11):115-122.
Pelletier, B., Charvis, P., Daniel, J., Hello, Y., Jamet, F., Louat, R., Nanau, P., and Rigolot, P. 1988. Structure et
lineations magnetiques dans Ie coin nord-ouest du bassin Nord-Fidjien: Resultats preliminaires de la
campagne EVA 14 (aofit 1987), C.R. Acad. Sci. Paris, 306(11):1247-1254.
Pelletier, B., Lafoy, Y., and Missegue, F. 1993a. Morphostructure and magnetic fabric of the northwestern North
Fiji Basin, Geophys. Res. Lett. 20(12):1151-1154.
Pelletier, B., and Louat, R 1989. Mouvement relatif des plaques dans Ie Sud-Ouest Pacifique, C.R. Acad. Sci.
Paris 308:123-130.
Pelletier, B., Missegue, F., Lafoy, Y., Mollard, L., Decourt, R, Dupont, J., Join, Y., Perrier, J., and Recy, J. 1993b.
Extremites nord du bassin Nord-Fidjien et des fosses arriere-arc des Nouvelles-Hebrides: Morphostructure et
signature magnetique, C.R. Acad. Sci. Paris 316(11):637-644.
Price, R. C., Johnson, L. E., and Crawford, A. J. 1990. Basalts of the North Fiji basin: The generation of back arc
basin magmas by mixing of depleted and enriched mantle sources, Contrib. Mineral. Petrol. 105:106-121.
Price, R e., and Kroenke, L. W. 1991. Tectonics and magma genesis in the northern North Fiji Basin, Mar. Geol.
98:241-258.
Price, R e., Maillet, P., and Johnson, D. A. 1993. Interpretation of GLORIA side-scan sonar imagery for the
Coriolis Troughs of the New Hebrides backarc, Geo-Mar. Lett. 13:71-81.
Price, R. C., Maillet, P., McDougall, I., and Dupont, J. 1991. The geochemistry of basalts from the Wallis Islands,
Northern Melanesian Borderland: Evidence for a lithospheric origin for Samoan-type basaltic magmas, J.
Recy, J., Pelletier, B., Charvis, P., Gerard, M., Monjaret, Me., and Maillet, P. 1990. Structure, age et origine des
fosses arriere-arc des Nouvelles-Hebrides (Sud-Ouest Pacifique) Oceanol. Acta 10:165-182.
Ruellan, E., Auzende, 1. M., Honza, E., and the STARMER Group. 1989. L'accretion dans Ie bassin Nord Fidjien
meridional: Premiers resultats de la campagne franco-japonaise STARMERlKAIYO 88, C.R. Acad. Sci.
Paris 309(11):1247-1254.
Ruellan, E., Huchon P., Auzende, 1. M., and Gracia, E. 1994. Propagating rift and overlapping spreading center in
the North Fiji Basin, Mar. Geol. 116:37-56.
Saunders, A. D., and Tarney, J. 1984. Geochemical characteristics of basaltic volcanism within backarc basins, in
Marginal Basins: Geology, Volcanism and Associated Sedimentary and Tectonic Processes in Modern and
Ancient Marginal Basins (B. P. Kokelaar and M. F. Holwells, eds.), pp. 59-76, Blackwell, Oxford.
Sc1ater, J. G., and Francheteau, 1. 1970. The implications of terrestrial heat flow observations on current tectonic
and geochemical models of the crust and upper mantle of the earth, Geophys. l. R. Astron. Soc. 20:509-542.
Sc1ater, J. G., and Menard, H. W. 1967. Topography and heat flow of the Fiji Plateau, Nature 216:991-993.
Sinton, J. M., and Fryer, P. 1987. Mariana lavas from 18°N: Implications for the origin of back arc basin basalts, l.
Geophys. Res. 92(BI2):782-802.
Sinton, J. M., Johnson K. T. M., and Price R. C. 1985. Petrology and geochemistry of volcanic rocks from the
Northern Melanesian Borderland, in Investigations of the Northern Melanesian Borderland (T. M. Brocher,
ed.), Earth Science Ser., Vol. 3, pp. 35-65, Circum-Pacific Council for Energy and Mineral Resources,
Houston, TX.
Sinton J. M., Price, R. c., Johnson, K. T. M., Staudigel, H., and Zindler, A. 1994. Petrology and geochemistry of
submarine lavas from the Lau and North Fiji Back-Arc Basins, in Basin Formation, Ridge Crest Processes,
and Metallogenesis in the North Fiji Basin (L. W. Kroenke and J. V. Eade, eds.), Earth Science Ser., Vol. 15,
pp. 119-135, Circum-Pacific Council for Energy and Mineral Resources.
Stackelberg, U. von, and Rad, U. von. 1990. Geological evolution and hydrothermal activity in the Lau and North
Fiji basins (Sonne Cruise SO-35)-a synthesis, Geol. lb. 92:629-660.
Stackelberg, U. von, and Shipboard Scientific Party. 1985. Hydrothermal sufide deposits in backarc spreading
centers in the Southwest Pacific, BGR Bundesandstalt flir Geowissenschaften und Riistoffe Circular 2:3-14.
Sun, S. S., and McDonough, W. F. 1989. Chemical and isotopic systematics of oceanic basalts: Implications for
mantle composition and processes, in Magmatism in the Ocean Basins (A. D. Saunders and M. 1. Norry,
eds.), Geol. Soc. Spec. Publ. 42:313-345.
Sykes, L. R. 1966. The seismicity and deep structure of islands arcs, l. Geophys. Res. 71:2981-3006.
Sykes, L. R., Isacks, B. L., and Oliver, 1. 1969. Spatial distribution of deep and shallow earthquakes of small
magnitudes in the Fiji-Tonga region, Bull. Seismol. Soc. Am. 59:1093-1113.
Urabe, T., Auzende, 1. M., et al. 1992. Bathymetric map of the central part of the North Fiji Basin, Southwest
Pacific, between 14°20S and 21°50S, 2 colored maps, scale: 11500 000, Published under the Special
Coordination for Promoting Science and Technology (Japan).
Von Damm, K. L., and Bischoff, J. L. 1987. Chemistry of hydrothermal solutions from the Juan de Fuca Ridge, l.
Geophys. Res. 92:11,334-11,346.
Wright, E., and White, W. M. 1987. The origin of Samoa: New evidence from Sr, Nd, and Pb isotopes, Earth
5
Tectonics, Magmatism, and Evo/ution

of the New Hebrides Backarc Troughs
(Southwest Pacific)
Patrick Maillet, Etienne Ruel/an, Martine Gerard, Alain Person,
Herve Bel/on, Joseph Cotten, Jean-Louis Joron, Setsuya Nakada,
and Richard C. Price
ABSTRACT
In the southwest Pacific, a discontinuous series of narrow and elongated troughs separates
the New Hebrides island arc from the adjacent active marginal basin, the North Fiji Basin.
This chapter reviews the structural, geophysical, geochronological, and petrological data
available for the New Hebrides backarc troughs (NHBAT) and discusses the significance of
these structures.
A diffuse horst-and-graben morphology, partly obscured in some places by recent
volcanic complexes, characterizes the northern Jean-Charcot troughs (JCT). By contrast,
the southern Coriolis troughs (CT) show well-developed flat-bottomed grabens. Moreover,
no backarc troughs are observed in the central backarc area, adjacent where the d'Entre-
casteaux zone collides with the arc.
Volcanic rocks dredged in the NHBAT show a wide range of Si02 contents, with high-
Al 20 3 and low-Ti0 2 contents, features typical of their arclbackarc environments. Trace
element analyses indicate a much stronger subduction component in the volcanics of the
southern CT than in those of the northern JCT. However, large-ionic-radius-lithophile-
element (LILE) (Ba, Rb, Sr) enrichments and high-field-strength-elements (HFSE) (Ta, Nb,
Zr, Ti, Y, Yb) depletions, relative to N-MORB (mid-ocean ridge basalts), are generally
observed in most NHBAT volcanics and are features characteristic of island-arc basic and
Patrick Maillet • ORSTOM Centre de Brest-GDR "GEDO" 910, 29280 Plouzane, France. Etienne
Ruellan • CNRS Sophia-Antipolis, 06560 Valbonne, France. Martine Gerard • ORSTOM, 93143
Bondy, France. Alain Person • Laboratoire de Geologie des Bassins Sedimentaires, Universite Pierre et
Marie Curie, 75252 Paris, France. Herve Bellon and Joseph Cotten • CNRS URA 1278-GDR
"GEDO" 910, Universite de Bretagne Occidentale, 29287 Brest, France. Jean-Louis Joron • Groupe
des Sciences de la Terre, Laboratoire Pierre-Sue, CEN Saclay, 91191 Gif sur Yvette, France. Setsuya
Nakada • Kyushu University, Fukuoka 812, Japan. Richard C. Price • La Trobe University, Bun-
doora, Victoria 3083, Australia.
177
178 PATRICK MAILLET et al.
intermediate volcanics. Backarc basin basalts (BABB) are scarce; the only occurrence was
found in the very northern JCT, indicative of an aborted tendency toward oceanic spread-
ing, between 3.9 and 1.1 Ma (KiAr dating).
Geological long-range inclined asdic (GLORIA) seafloor imagery, manned submers-
ible observations, water chemistry analyses, and sediment heat flow measurements do not
provide evidence for widespread hydrothermal activity, or oceanic spreading, in the
NHBAT. However, some ferromanganese crusts coating volcanic and volcano-sedimentary
formations have a hydrothermal origin (todorokite, birnessite, phillipsite). These hydro-
thermal crusts are mainly located on the eastern faulted border of the NHBAT.
The volcanic-tectonic evolution of the New Hebrides backarc troughs primarily
results from the concomitant effects of nearby subduction (along the New Hebrides
subduction zone) and spreading (in the central North Fiji basin) and secondarily from the
after-effects of the collision of the d'Entrecasteaux zone with the arc. The NHBAT
represent the very first stage of backarc crustal extension, characterized by volcanics with
predominantly island-arc tholeiite and some BABB compositional features.
1. INTRODUCTION
Located between an island arc and a marginal basin, backarc troughs may record the
concomitant or successive influences of crustal distension, rifting, oceanic spreading, with
or without any arc-induced magmatic contamination. A number of recent papers have
discussed the structural, petrological, and geochronological characteristics of the New
Hebrides backarc troughs (Recy et al., 1986, 1990; Monjaret et aI., 1987, 1991; Charvis and
Pelletier, 1989; Monjaret, 1989; Sage and Charvis, 1991; Matsumoto et al., 1992; Johnson
et al., 1993; Pelletier et al., 1993b; Price et al., 1993; Nakada et al., 1994). Yet the
significance of these troughs in the regional tectonic framework of the southwest Pacific
remains controversial. Do these troughs represent old and inactive structures or incipient
spreading features-that is, a nascent backarc basin with recent active submarine volca-
nism? Is their formation primarily linked to the evolution of the New Hebrides island arc or
to that of the North Fiji Basin?
To answer these questions, we review and discuss available data on the New Hebrides
backarc troughs and propose a new interpretation accounting for the occurrence of these
unusual submarine structures.
2. GEOLOGICAL AND TECTONIC FRAMEWORK OF THE NEW HEBRIDES

ISLAND ARC AND THE NORTH FIJI BASIN
In the area between 100 S and 25°S and 165°E and 1700 W, two subduction zones with
opposite directions of convergence dip toward each other (Fig. 5.1). Along the New
Hebrides subduction zone (i.e., beneath the Santa Cruz, or the eastern outer Solomon, and
Vanuatu islands), the consumption of the India-Australia plate is marked by a rapid
subduction in an E-NE direction (9-16 cm/yr; N76°E ±11; dip 70°). Along the Tonga
subduction zone, the geodynamic scheme is almost symmetrical for the subducting Pacific
plate (16.5-18 cm/yr along a W-NW direction, with a dip of 80°) (Isacks et aI., 1981; Louat
et al., 1988; Louat and Pelletier, 1989; Pelletier and Louat, 1989; Pelletier and Dupont,
1990a).
NEW HEBRIDES BACKARC TROUGHS 179
FIGURE 5.1. Tectonic setting of the New Hebrides (Vanuatu) arc in the southwest Pacific. Northern troughs =
Jean-Charcot troughs (JCT) in the text.
Plate consumption along these two subduction zones is accompanied by the opening
of two active marginal basins: the North Fiji Basin, between Vanuatu and Fiji (Auzende et
at., Chapter 4 this volume), and the Lau Basin, between Fiji and Tonga (Hawkins, Chapter 3
this volume).
Interestingly, both subduction zones are affected by the collision of major submarine
ridges. The d'Entrecasteaux zone is colliding with the central New Hebrides arc, and the
collision zone is moving northward (Maillet et at., 1983; Collot et at., 1985; 1992a,b),
whereas the Louisville Ridge is colliding with and sweeping southward along the Tonga arc
(Dupont and Herzer, 1985; Pelletier and Dupont, 1990b; Lallemand et at., 1992).
2.1. The New Hebrides Island Arc

The New Hebrides island arc comprises three volcanic chains (Carney et at., 1985): a
western belt (late Oligocene to middle Miocene) forming the islands of the Torres Group,
Espiritu Santo and Malakula; an eastern belt (late Miocene to early Pliocene), including the
islands of Maewo and Pentecost; and a central chain (late Miocene to Holocene), which
runs more than 1500 km from Tinakula-Nendo (Santa Cruz Islands) in the north to
Matthew and Hunter volcanoes in the south (Fig. 5.2).
The following summary is from Macfarlane et at. (1988) and Gorton (1974, 1977).
2.1.1. Western Belt

Volcanism in the islands of the Torres Group, Espiritu Santo and Malakula, com-
menced in latest Oligocene time but occurred largely during the early to middle Miocene
(25-14 Ma), along the former Vitiaz island arc, overlying a southwest dipping Benioff
zone. Petrographically, lavas of the western belt range from olivine basalt to hornblende-
165' 170'
PACIFIC PLATE
10 cm/yr
IS'
NORTH FIJI BASIN
Volcanic aS$oeial~
tale Miocene 10 Holocene

(Central Chain)
Late Miocene 10 urly Pliocene
(Eastern Se!l)
Lale OIiQocene 10 middle Miocene
(WeSlern Belt}
~~!!
.,"/,/' Uplifted ridge
FliUed basin
IA~1 Intra-a.rc sedimentary basin
Major fracture
-A- Trace of plale contact. dashed
where bIJried
==::> Rel~~~~:~~ ~o::~o\:~~ricafl

Geologists. 1981}
Central Sasin SMFZ Santa. Maria fraClure zone
Sanks Sasin ABFZ Aoba tractu:e zone
Vanmolo Basin AMFZ Ambrym fraclure ZOnG
EFZ Epi fracture zone
0 100 mkm VFZ Vulcan fracture zone

I I
FIGURE 5.2. The principal geological and tectonic features of the New Hebrides island arc (after Greene et aI.,
1988a). Vot Tande trough = Jean-Charcot troughs (JCT) in the text.
bearing dacite and rhyodacite pumice, but andesite (including both two-pyroxene and
hornblende-phyric varieties) is most abundant.
The western belt lavas show transitional calc-alkaline/tholeiitic characteristics and in
this respect resemble the Lau Ridge volcanics and lavas from the lower-~O suite presently
erupting in the New Hebrides arc central chain.
2.1.2. Eastern Belt

The island of Pentecost is the only island of the New Hebrides arc on which a
basement complex crops out. This is composed of slices of ultramafic to basaltic rocks
interpreted as a fragmented ophiolite, possibly representing sections of the oceanic crust
upon which the eastern belt arc volcanics accumulated. Most of these rocks are too altered
and metamorphosed for KlAr dating and have low ~O contents. Still, two mineral KlAr
ages (35 ± 2 Ma for a hornblende from an amphibolite, and 28 ± 6 Ma for a plagioclase
from a gabbro) place younger limits on the age of this basement (Gorton, 1974).
Most of the Maewo and Pentecost volcanics erupted between late Miocene and lower-
middle Pliocene time (7 -5 Ma for Maewo, 6-3 Ma for Pentecost) from submarine fissures.
Pillow lavas, associated intrusions, and pyroclastic rocks range in composition from basal
ankaramites, picrites, and mafic-enriched porphyritic basalts to an upper series of feldspar-
phyric basalts and basaltic andesites. They range across the medium-K20 to high-~O
fields and are differentiated to high-K20 calc-alkaline andesite.
2.1.3. Central Chain

Volcanism began in the central chain in latest Miocene time (Erromango island: 5.8-
5.3 Ma; Bellon et aI., 1984) and thus overlapped the eastern belt activity that ceased at about
3.5 Ma. Activity has been virtually continuous, at least from the earliest Pleistocene
(1.8 Ma). From Pleistocene to present times, volcanism, which was mainly subaerial,
developed extensively along the length of the active arc. Lavas range from picrite to dacite
or rhyodacite, with a peak frequency distribution in the basalt range, a trough between 52%
and 62% Si02, and a smaller peak at 63% Si02• This distribution was considered to be
representative of all the lavas in the exposed arc, but recent work seems to indicate that the
compositional range of the central chain volcanics may be much more continuous (Robin
et aI., 1991, 1994; Eissen et at., 1992; Picard et aI., 1995).
Two broad lava suites can be distinguished, although a compositional spectrum
undoubtedly exists between the two. A higher-~O suite is represented by picritic to
rhyodacitic lavas on the islands of Tongoa, Ambrym, Aoba, and Santa Maria. These
volcanics are strongly light-rare-earth-element (LREE) enriched and have 87Srj86Sr ratios
ranging from 0.7038 to 0.7043 (Briqueu and Lancelot, 1983). Lavas of the lower-K20
group dominate the islands of Matthew, Hunter (Maillet et aI., 1986), Anatom, Tanna,
Erromango, Efate, Lopevi, active submarine volcanoes around Epi, and all the Banks
islands except Santa Maria. This lower-K 20 suite characteristically exhibits LREE-
enriched REE patterns, but the degree of LREE enrichment is notably less than in the
higher-KzO suite, and 87Srj86Sr ratios are significantly lower (0.7030-0.7033; Briqueu and
Lancelot, 1983). The lower-~O suite might have been generated by higher degrees of
partial melting at shallower levels in the upper mantle than the higher-~O suite (Crawford
et at., 1988), although the respective mantle sources were clearly different isotopically.
The structural consequences of the collision of the seismically inactive d'Entrecas-
teaux zone with the central part of the New Hebrides arc (Fig. 5.2) are now well docu-
mented (Collot et aI., 1985, 1991, 1992a,b; Greene et at., 1988a,b, 1992). This collision,
which began 4-3 Ma (Macfarlane et aI., 1988), likely affects the central chain as well as the
backarc area, tectonically and petrologically. This essential point, though still controversial
(Roc a, 1978; Monjaret, 1989), will be discussed later.
2.2. The North Fiji Basin

Auzende et at. (Chapter 4 this volume) discuss the structure and evolution of the North
Fiji basin (NFB) in detail, and the reader is referred to that chapter for a complete
description. A summary of the geology and petrology of the NFB is presented here and is
taken from Auzende et at. (1990), Huchon et at. (1994), Nohara et al. (1994), and Eissen
et al. (1994).
The NFB is an active marginal basin bounded by the New Hebrides arc to the west,
by the Hunter ridge and fracture zone to the south, by the Fiji Platform to the east, and by
the Vitiaz paleosubduction zone to the north (Fig. 5.1). This basin was created by backarc
spreading, which commenced behind the New Hebrides arc 10 to 8 Ma ago in response to a
clockwise rotation of the arc from an initial position close to the present location of the
Vitiaz zone.
The formation of the NFB can be summarized as follows.
10 to 8 Ma (or even in early Miocene time, according to Kroenke, 1984): Collision
of the Ontong Java Plateau near the Solomon arc (volcanism ceased on New Hebrides
western belt ca. 15-14 Ma) and reversal of the polarity of the subduction from southwest-
ward along the Vitiaz zone to northeastward beneath the New Hebrides arc.
8 to 3 Ma: Beginning of NFB opening, along an axis roughly parallel to the Vitiaz
zone; clockwise rotation of the New Hebrides arc (the spreading axis trend rotates pro-
gressively from N120° to NI500), and counterclockwise rotation of the Fiji Platform.
Active volcanism occurs on the New Hebrides eastern belt until 3.5 Ma, and on the central
chain from 5.8 to 5.3 Ma onward.
3 to 0.7 Ma: The spreading axis jumps into the central NFB from a N150° to a N-S
direction; synchronously, spreading starts in the Lau Basin. Active volcanism occurs on the
New Hebrides central chain.
0.7 Ma to present: Modification of the spreading geometry in the central NFB between
15°S and 18°30'S; migration of a triple junction (ridge-ridge-transform) near 16°40'S,
with a southern branch oriented NI5°, a northern branch oriented NI60°, and the prolonga-
tion of the North Fiji fracture zone, oriented N60° in this area. Active volcanism continues
on the New Hebrides central chain.
As demonstrated by Eissen et at. (1991,1994), the dominant magma type produced in
the NFB is N-MORB, showing depleted large-ion lithophile (UL), high-field strength
(HFS), and light rare-earth (LRE) element patterns. This magma type is the only one
present along the most mature N-S spreading segment of the NFB (i.e., between 18°20'S
and 21°S).
Several ridges of various origin transect the NFB crust. In particular, the Hazel
Holme-South Pandora Ridge is significant, since it may influence the volcano-tectonic
evolution of the northern New Hebrides backarc area (Figs. 5.1 and 5.3). As indicated by
Pelletier et at. (1993a), this seismically active ridge is considered to be an extensional zone
in its western part (Hazel Holme extensional zone) and a slow spreading ridge in its
easternmost part (South Pandora Ridge; Kroenke et al., 1994; Price and Kroenke, 1991).
The western part ofthe Hazel Holme Ridge trends N85°-N90oE over 100-120 krn in width.
It is composed of a series of volcanic ridges, scarps, and parallel narrow troughs deeper than
3000 m (Fig. 5.4). To the west of 168°30'E, the width of the whole structure decreases;
lateral troughs disappear and a single 2500-3500 m deep E-W trough remains, bounded by
symmetrical ridges culminating at a depth of 1700 m (Figs. 5.5,5.6). This trough abruptly
1\'
S
p
13'
IA
IS' p
IA
.SHHA~'/.:
,
'
WNFB
•
20' ~ .\ IA
I~
•
•
I
I
165' E 170' 110'
FIGURE 5.3. Present-day relative motions in the New Hebrides-North Fiji Basin region, as proposed by Louat
and Pelletier (1989). P: Pacific plate; IA: Indo-Australian plate; DR: d'Entrecasteaux ridge; M and H: Matthew
and Hunter volcanoes. WNFB, ENFB, and SNFB: western, eastern and southern North Fiji Basin microplates,
respectively; NNHA, CNHA, and SNHA: northern, central and southern segments of the New Hebrides arc
microplate, respectively. Numbers and arrows beside plate boundaries indicate the rates (in cmJyr) and trends of
the relative motions. Very thick divergent arrows show the motion along the main spreading center of the North
Fiji basin. Line with filled barbs corresponds to the New Hebrides Trench. Line with open barbs marks the New
Hebrides backarc compressive belt.
terminates at 168°1O'E, at the very southernmost tip of the N-S-trending northern New
Hebrides backarc area (Pelletier et aI., 1993a).
3. THE NEW HEBRIDES BACKARC TROUGHS
3.1. Previous Work and Recent Investigations

A series of narrow and elongated troughs separates the New Hebrides island arc from
the adjacent, active, marginal North Fiji Basin (Fig. 5.2). These troughs, however, do not
extend along the entire length of the arc, since they are absent in the central backarc area
(and, in particular, to the east of Maewo and Pentecost islands). Interestingly, this central
area also lies opposite the d'Entrecasteaux zone.
16S'E 170' 172'
0 0
'~~I d
o e
FIGURE 5.4. Structural map of the northern New Hebrides arc/northwestern North Fiji Basin (adapted from
Pelletier et ai., 1993a). (a) troughs and depressions; (b) structural trends and fractures; (c) structural highs;
(d) structural lows; (e) volcanic highs. NHA: New Hebrides arc; VT: Vitiaz Trench; BAT: backarc troughs domain
(Jean-Charcot troughs); 9°30'SR: 9°30'S ridge; DR: Duff Ridge; WTR: West Tikopia ridge; TR: Tikopia Ridge;
HH-SPR: Hazel Holme-south Pandora ridge; CNFB 160o ER: central North Fiji Basin 1600 E ridge; TT: Tikopia
Trough; SCT: Santa Cruz Trough; WTP: west Torres plateau; ER: d'Entrecasteaux Ridge. Islands and reefs are in
black; D: Duff Islands; R: Reef Islands; Tn: Tinakula Island; N: Nendo Island; V: Vanikoro Island; Tk: Tikopia
Island; A: Anuta Island; F: Fatutaka island; C: Charlotte Bank; P: Pandora Bank. Areas detailed in Figs. 5.5-5.7
are outlined.
Thus, between lOoS and 21°S, it is possible to distinguish three main areas in the New
Hebrides backarc: (1) the Jean-Charcot troughs (JCT, named after the RN Jean Charcot),
or northern troughs, extend between lOoS (Pelletier et al., 1993a,b) and 13 °30' S, where they
abut the western termination of the Hazel Holme Ridge (Fig. 5.4); a central area, extending
from 13°30'S to 17°30'S-that is, roughly from the east ofVanua Lava Island to the east of
Tongoa Island, is totally devoid of troughs (Fig. 5.2); (3) the Coriolis troughs (CT; named
after the RN CorioUs), or southern troughs, extend between 17°30' Sand 21°S, where their
southern en echelon termination merges into the structural grain of the island arc (Monzier
et al., 1991; Figs. 5.2, 5.9).
Since the first mention of the Coriolis troughs in the literature (Puech and Reichenfeld,
line 18
2s
5~~_ _~_ _~~_ _~~~~~~~~~

14hr12hr 8hr 10hr 12hr
FIGURE 5.5. Bathymetric map of the Jean-Charcot troughs (after Charvis and Pelletier, 1989), showing
positions of volcanic cones (stars) and sample localities (dots). Data from SEAPSO 2, KAIYO 89, and SAVANES
cruises. Depths in km, with 250 m isobaths. Bottom diagrams: seismic cross sections along survey lines shown
above (adapted from Nakada et al., 1994). See Fig. 5.4 for location.
1969), several oceanographic cruises have been partly or fully devoted to the study of the
New Hebrides backarc area (Dubois et al., 1978). The most recent, which provide most of
the data used in this chapter, include SEAPSO 2 (Nov. 1985, RN Jean Charcot; Recy et al.,
1986); EVA 13 (1986, RN CorioUs; Sage and Charvis, 1991); MULTIPSO (May 1987, RN
Jean Charcot; Daniel et aI., 1989); EVA 14 (Aug. 1987, RN Le Noroft; Pelletier et al.,
1988); GLORIA survey (Aug. 1989, HMAS Cook; Johnson et al., 1993; Price et al., 1993;
Tiffin, 1993); KAIYO 89 (Dec. 1989-Jan. 1990, RN Kaiyo; KAIYO 89 Cruise Report,
1990; Nakada et aI., 1994; this chapter); YOKOSUKA 90 (Jan.-Feb. 1991; RN Yokosuka;
FIGURE 5.6. Structural map of the Jean-Charcot troughs (from Charvis and Pelletier, 1989). 1: normal fault; 2:
small and/or inferred normal fault; 3: transverse tectonic lineation (arrows indicate strike-slip fault); 4: inclined
bedding; 5: horizontal bedding; 6: structural high; 7: volcano; 8: lava flow; 9: major graben. TSCB: Torres-Santa
Cruz sedimentary basin; BAT: backarc troughs (Jean-Charcot troughs); V1T: Vot Tande Trough; ER: Eastern Ridge
(Duff Ridge of Fig. 5.4); HHFZ: Hazel Holme fracture zone; NFB: North Fiji Basin. See Fig. 5.4 for location.
Eissen et ai., unpublished information, 1994); SANTA CRUZ (Nov.-Dec. 1991; RN Le

Noroft; Pelletier et ai., 1993b); SAVANES (Dec. 1991-Jan. 1992; RN Le Noroft and Cyana
submersible; Savanes 91-92 Cruise Report, 1992; Nakada et ai., 1994; this chapter).
The KAIYO 89, YOKOSUKA 90, and SAVANES cruises were parts of the STAR-
MER French-Japanese Joint Project (Joint Research Program on Rift System in the Pacific
Ocean; Auzende and Urabe, 1994), endorsed by SOPAC (the South Pacific Applied
Geoscience Commission, Suva, Fiji).
3.2. Structure and Tectonics

3.2.1. The Jean-Charcot Troughs
The Jean-Charcot troughs (JCT) are a succession of discontinuous and, in some areas,
anastomosing, horsts, grabens, and half-grabens, trending slightly oblique to the submeri-
dional northern New Hebrides central chain (Fig. 5.4). Between 12°20'S and 13°20'S, in an
area 50-55 km wide (E-W) and 100-120 km long (N-S), at least six individual troughs have
been recognized, each 3-9 km wide and 20-35 km long (Figs. 5.5, 5.6; Recy et ai., 1986,
1990; Matsumoto et ai., 1992; Johnson et ai., 1993). The JCT are limited to the west by the
submarine extensions of the New Hebrides central chain; to the south by the Hazel Holme
ridge; and to the east by the N-S Duff ridge, culminating around 1250 m depth, which
separates the JCT from the North Fiji Basin. The northern extension of the JCT, however, is
still conjectural. North of a shallow (up to 460 m deep) volcanic complex centered near
12°lO'S, the JCT domain widens up to 100 km and abuts the intra-arc E-W-trending Santa
Cruz Trough (Fig. 5.4; Pelletier et ai., 1993a).
The average depth of the JCT is 3000-3500 m (Fig. 5.5), similar to the mean depth
of the North Fiji Basin (Pelletier et ai., 1993a; the maximum depth in the whole New
Hebrides backarc, 3658 m, was recorded in the JCT, during the KAIYO 89 cruise, at
12°46.9'S-167°41.9'E).
Seismic reflection profiles across the JCT show thick sedimentary layers (more than
I-s twt), which are usually tilted eastward and dip more steeply downsection. Unconfor-
mities or tectonic discordances are frequent. The most recent deposits also dip more steeply
downsection, indicating coeval tectonics and sedimentation during the entire structural
history of the troughs (Charvis and Pelletier, 1989; Recy et ai., 1990).
During the period 1977-1987 no shallow earthquake (0-70 km) with magnitude large
enough to have a focal mechanism determination has been recorded in the entire JCT
area or in the western end of the Hazel Holme Ridge (Charvis and Pelletier, 1989). Yet
Louat and Pelletier (1989) suggested a surprisingly high rate of extension in the JCT
(5.5 crn/yr along a N45°E direction, at 13°S; 7 crn/yr along a N37°E direction at nos; Fig.
5.3), using the RM-2 plate model of Minster and Jordan (1978) combined with a large set of
regional data (i.e., shallow seismicity, focal mechanism solutions, and magnetism). Louat
and Pelletier (1989) argue that horst-and-graben structures like the JCT can be formed in an
aseismic but tectonically active environment. However, GLORIA imagery data show no
evidence of any neovolcanism in these troughs (Figs. 5.7, 5.8). The eroded nature of at least
some of the seamounts behind the arc, as well as the thick sediment fills in the grabens
of the Hazel Holme Ridge, led Johnson et ai. (1993) to suggest that these structures are not
young. However, this interpretation is at odds with the evidence cited for recent faulting
producing rotated sedimentary sequences.
188 PATRICK MAILLET at al.
FIGURE 5. 7. GLORIA imagery of northern New Hebrides-southeastern Solomon Islands area (from Johnson
et aI., 1993). Lineations indicated by strong white reflections are associated with the Hazel Holme Ridge in lower
right and the Jean-Charcot troughs in upper diagram. See Fig. 5.4 for location.
3.2.2. The Coriolis Troughs

The morphology of the CT looks comparatively simpler (Fig. 5.9). They are made
up of well-delineated grabens (Daniel, 1982) discontinuously paralleling the eastern flank
of the southern New Hebrides central chain between Efate and Anatom Islands. SeaBeam
bathymetry (Recy et al., 1986, 1990) and GLORIA seafloor imagery (Price et aI., 1993)
allow three main grabens to be distinguished in the CT, namely the Efate, Erromango, and
Futuna troughs.
The Efate Trough is a complex double graben, trending NNW-SSE, 10-25 km wide
and about 100 km long. A horst at a depth of 1200 m separates a western, small, asymmetri-
cal graben (mean depth: 2150 m) from a larger, deeper, eastern graben (mean depth: 2500-
2600 m). The latter is bounded eastward by a very steep normal fault and terminates
southward on a structural threshold about 1500 m deep, occupied to the west at 18°32'S-
169°35'E by a 1000 m high, less than 500 m deep volcanic seamount (Monzier et aI., 1984a;
, - - - - - - - - - - - - - - - - - - - , , - - - - - - - - - -.... 11·'0·'
I'·S
I'·:SO',
I-"r-.•-----"~r.-.o-.• ----:!:-=:--"'----l'--'-'-==------=".:::.•:-----:,::
..::.•~O.E 14·'0',
FIGURE 5.8. Geological interpretation of the GLORIA imagery of Fig. 5.7 (from Johnson et al., 1993).
Recy et aI., 1990). Downfaulted sediment-draped blocks can be identified in seismic

profiles across the Efate Trough, but substantial areas appear on GLORIA imagery to be
unsedimented basement outcrop (Fig. 5.10).
Southeastward, the Efate Trough is succeeded by another depression, the Erromango
Trough, which is about 75 km long, 30 km wide, and 2500-3100 m deep. Its northeastern
and southwestern margins, interpreted as fault scarps, are roughly linear and parallel and
rise to about 1500 m depth. An unsampled bathymetric high (2100 m bsl) in the center of the
trough (l9°05'S-169°52'E) is highly reflective on GLORIA imagery and may correspond
to a young, unsedimented volcanic feature (Price et al., 1993).
There is no major structural discontinuity between the Erromango Trough and the
Futuna Trough, and the latter can be considered as the southeastward extension of the
former (Daniel 1982; Monzier et al. 1984a). The Futuna Trough is approximately 75 km
long, 20-30 km wide, and 3200 m deep. It is therefore the deepest graben of the CT (the
maximum depth recorded during SEAPSO 2 cruise is about 3400 m, at 19°42.5'S-
1700 09'E, about 20 km south-southwest of Futuna Island). Its asymmetrical morphology is
•
FUTUNA ISLA 0
NORTH FW BASI
o 10 20
FIGURE 5.9. The Coriolis troughs (from Recy et ai., 1990). (Top) Bathymetric map. Depths in lan, with 500-m
isobaths (from Monzier et al., 1984a). SeaBeam surveyed outlined areas. See Fig. 5.2 for location (Note: Vate =
Efate). (Bottom) The Futuna Trough. Block diagram based on SeaBeam bathymetry. Vertical exaggeration: 5/1;
mesh size: 300 m; each nuance represents a 150 m depth unit.
obvious, with the eastern wall steeper and shallower (above sea level at Futuna Island) than
the western wall (Fig. 5.9). Rough topography, steep fault scarps trending NW-SE, and a
thin (less than 80 m) sediment cover characterize most of its floor.
The shallower depth of the Efate Trough may indicate thermal uplift, consistent with
extensive recent volcanism apparent on GLORIA imagery (Fig. 5.10). The deeper Erro-
mango and Futuna troughs seem to be relatively immature backarc structures, formed as a
consequence of lithospheric extension accompanying uplift or doming (Price et aI., 1993).
Numerous shallow earthquakes border the western limit of the CT between 18°50'S
FIGURE 5.10. GLORIA imagery of the Coriolis troughs (from Price eta!., 1993). See Fig. 5.2 for location. South
of 19°15'S, the line R to W marks the northern wall of the northern Futuna trough. P is an isolated, sediment-filled
basin, perched above the Futuna Trough. South of 19°45'S, dark area labeled S is an acoustic shadow on the
eastern wall of the southern Futuna Trough. R indicates a series of scarps marking the southwestern wall of the
trough. Note: Vate = Efate; Anetium = Anatom.
and 21°S, and normal faulting mechanisms indicate a NE-SW extension. Louat and
Pelletier (1989) calculated a rate of 2 cmlyr along a N37°E direction, at 20 0 S (Fig. 5.3).
Gravity data show a clear positive anomaly paralleling both western and eastern
borders of the CT, without any deep-rooted compensation (Collot and Malahoff, 1982),
indicative of rift-flank flexural uplift (Weissel and Karner, 1989).
The formation of intra-arc sedimentary basins may have partly influenced the evolu-
tion of the New Hebrides backarc area. Among them, the Aoba and Vanikoro-Torres basins
deserve special mention.
3.2.3. The Aoba Basins

The central part of the New Hebrides island arc, between Banks Islands and Epi
Island, is structurally complex and has recorded the effects of the ongoing collision of the
arc with the d'Entrecasteaux zone (Fig. 5.11; Collot et al., 1985, 1992a; Collot and Fisher,
1988; Greene and Johnson, 1988). Deep and thick intra-arc sedimentary basins (North and
South Aoba basins) separate the old western belt islands (Espiritu Santo, Malakula) from
the younger eastern belt islands (Maewo, Pentecost). Preliminary results of Ocean Drilling
Program (ODP) Leg 134 in the North Aoba Basin (Fig. 5.11) indicate that this basin is the
product of island-arc volcanism and tectonic deformation-that is, a piggyback subsiding
basin pinched and overthrust between two ancient volcanic arcs (Gerard, 1993). At Site
832, an unconformity marks the uplift time of the central part of the western belt in response
to the collision of the d'Entrecasteaux zone. The age of this sedimentary hiatus is near the
Upper Pliocene-Lower Pleistocene boundary. Recent (Pleistocene) basin-filling recorded
at Sites 832 and 833 comes from effusive products of the central chain volcanoes (Aoba,
Santa Maria-Gaua, and Mere Lava islands). At Site 833, Pleistocene basaltic sills intruded
the Lower Pliocene sedimentary sequence. This recent sill complex indicates that volcan-
ism was active along the eastern belt during the early Pliocene as well as during the
Pleistocene (Collot et al., 1992a).
Compression (North Aoba Basin; Greene and Johnson, 1988; Daniel et al., 1989),
distension (South Aoba Basin; Greene and Johnson, 1988), and folds and faults with
different orientations characterize the Aoba basins area. Moreover, compressive stresses
are relayed to the very eastern margins of Maewo and Pentecost, since westward-dipping
thrust faults appear on seismic reflection records (Louat and Pelletier, 1989; Recy et al.,
1990). Shallow earthquake focal mechanisms confirm this observation (Louat and Pelletier,
1989).
On GLORIA seafloor imagery (Fig. 5.12), the structural limit between the central New
Hebrides arc platform and the North Fiji Basin clearly appears as anastomosing fault
scarps, between the midpoint of Pentecost Island and the northern end of Maewo Island
(Price et al., 1993; Tiffin, 1993).
3.2.4. The Vanikoro-Torres Basin

The poorly known Vanikoro-Torres basin, which extends between Vanikoro and
Vanua Lava islands, east of Torres Islands (Fig. 5.13), is a very thick sedimentary basin
(up to 6 km thick in its center; Holmes, 1988) lying on the summit platform of the northern
New Hebrides arc (Falvey and Greene, 1988). On seismic reflection profiles, three uncon-
formities are distinguished (early-middle Miocene; late Miocene; late Pliocene-early
Pleistocene), as well as faulting and folding extending up to the surface (Falvey and
~
ffi
~
m
I ~
i ~
~
:xl
§
~
r.:::3 n ~HI
~ . ... m ..
[§, _13 ell -21 -~-- 2l.!
A......-...a3
.'
~~; ~I/j 8"
FIGURE 5.11. (Left) Location of sites drilled during OOP Leg 134. Bold lines indicate location of cross sections shown at right. NOR: North d'Entrecasteaux Ridge; SOC: south
d'Entrecasteaux chain; NAB: North Aoba Basin; SAB: South Aoba Basin; VB: Vanikoro Basin. Bold line with teeth indicates approximate position of subduction zone; teeth are
on upper plate. Arrows indicate direction of plate convergence. Bathymetry in km. (Right) Geologic columns and cross sections, OOP Leg 134. I: oceanic crust; 2: western belt
volcanic rocks; 3: eastern belt volcanic rocks; 4: central chain volcanic rocks; 5: basin fill; 6: guyot volcanic rocks; 7: volcanic sand/sandstone; 8: volcanic silt/siltstone; 9: volcanic
sandstone/siltstone/claystone; 10: sed-lithic breccia; II: volcanic breccia; 12: basalt chalk; 13: multiple slivers of siltstone and chalk; 14: foraminiferal ooze; 15: nanofossil ooze; 16:
foraminiferal chalk; 17: nannofossil chalk; 18: calcareous chalk; 19: pelagic limestone; 20: lagoonal limestone; 21: unconformity; 22: ash; 23: thrust fault. NOR: north ~
d'Entrecasteaux Ridge; BG: Bougainville guyot; NAB: North Aoba Basin; NFB: North Fiji Basin (From eollot et aI., 1992a).
FIGURE 5.12. GLORIA imagery along the New Hebrides backarc between 14°30'S and 17°30'S (from Price
et al., 1993). See Fig. 5.2 for location.
Greene, 1988). The easternmost part of the Vanikoro-Torres Basin (i.e., the area closest
to the northern NHBAT) is intruded by young (Pleistocene-Holocene) volcanic intrusions
from the New Hebrides central chain.
3.3. Volcanic Petrology, Geochronology, and Geochemistry

3.3.1. Volcanic Petrology
Before the SEAPSO 2 cruise (1985), the petrological knowledge of the NHBAT was
limited to two studies. Dugas et al. (1977) described and analyzed a few volcanic samples
from the Futuna Trough (GEORSTOM III CENTRE cruise, 1975), while Vallot (1984)
presented the first compilation of dredged samples from along the whole southern New
Hebrides arc (i.e., inner trench slope, arc substratum, and Coriolis backarc troughs) during
GEORSTOM III CENTRE and EVA V (1977) cruises.
FIGURE 5.13. The central and northern New Hebrides are,

with its intra-arc sedimentary basins (North Aoba, South
Aoba and Vanikoro-Torres basins) and fault systems. The
Vanikoro-Torres basin appears as Torres-Santa Cruz sedi-
mentary basin in Fig. 5.6, and as Vanikoro Basin in Fig. 5.11.
The Jean-Charcot troughs area is outlined (see Fig. 5.5).
Vanikolo =Vanikoro. Bathymetry in km. Islands toponymy:
VA: Vanikoro; VL: Vanua Lava; VT: Vot Tande; MT: Mota
Lava; TK: Tikopia; MO: Maewo; PT: Pentecost; AO: Aoba;
ES: Espiritu Santo; MA: Malakula; AM: Ambrym (From
Falvey and Greene, 1988).
In order to get a closer view of the structure and volcanology of these troughs, a
comprehensive geological, geophysical, and geochronological study was initiated during
the 1985 SEAPSO 2 cruise (Recy et al., 1986; Monjaret et aI., 1991). Most of the petrologi-
cal data discussed here come from this study. A few cruises with more specific targets
followed more recently (see above), and some data from these are also included in this
chapter (Table I).
The dredged samples and samples picked up in situ (via submersible) from the
NHBAT have been classified geographically from north to south according to neighboring
islands or structures as follows: Vanikoro (VAN), Tikopia (TIK), and Vot Tande (VOT), for
the Jean-Charcot troughs (JCT); Hazel Holme (HAZ); Efate (EFA), Erromango (ERR),
Futuna (FUT), and Anatom (ANA) for the Coriolis troughs (CT). The coordinates of these
samples and their structural settings are shown in Table Ia-Ib and Figs. 5.5,5.14, and 5.15.
As summarized by Monjaret et al. (1991), basalts (45-53% Si02) clearly predominate
(60% oflavas dredged during the SEAPSO 2 cruise), whereas basaltic andesites (53-60%)
and dacites (Si0 2 > 60%) represent 13% and 27% of dredged samples, respectively. These
relative abundances did not drastically change after sampling carried out during KAIYO
89, YOKOSUKA 90 and SAVANES cruises (Table II).
Lavas from the JCT are usually aphyric or weakly porphyritic and show high ves-
icularity (generally 20% of the rock volume, sometimes up to 40-50%). By contrast,
volcanics from the CT are generally porphyritic (15% of phenocrysts: plagioclase,

clinopyroxene ± olivine, orthopyroxene, Fe-Ti oxides) or, in some cases, highly porphyritic
(up to 50% phenocrysts).
Nine petrological types (Table II: types a to i) can be distinguished in the NHBAT
volcanics (Monjaret et aI., 1991). Mafic lavas are represented by six types:
• Type a: MORB (Mid-ocean ridge basalts)
• Types b, c, d: IAT (island-arc tholeiites), which may be enriched in Ti02 (type c), in MgO (type
d), or in Ti0 2 and MgO (type c+d)
• Type e: CAB (calc-alkaline basalts and andesites), which are also in some cases enriched in
Ti0 2 (type e+c) or in MgO (type e+d)
• Type f: BABB (backarc basin basalts)-basalts with a composition intermediate between
MORB and IAT
Felsic lavas are represented by three types:
• Type g: high-Na/low-K dacites (Nakada et al., 1994)
• Type h: high-K dacites
• Type i: hyper-K dacites
MORB (a), BABB (f), and high-Na/low-K dacites (g) have only been found in the
northern troughs (JCT) and in the Hazel Holme area: MORB (a) are restricted to VAN and
HAZ areas, BABB (f) to the VAN area, high-Na/low-K dacites (g) to the VAN and TIK
areas (Table II); all other petrological types recovered in the NHBAT resemble the ones
found on the central chain islands.
3.3.2. Geochronology
Most K-Ar ages presented in Table III originated from SEAPSO 2 cruise samples and
were obtained in France (UBO, Universite de Bretagne Occidentale, Brest). They have been
discussed elsewhere (Monjaret et aI., 1987, 1991; Monjaret, 1989; Recy et al., 1990). Com-
plementary data from samples 3152 (Vanikoro area), 3981X and 415X (Tikopia area), recov-
ered during the KAIYO 89 cruise, were analyzed in Japan and have been added to this list.
Sample 7M2 is a MORB dredged on the Duff Ridge (Table I; Fig. 5.14) (i.e., on the
eastern border of the Vanikoro area). It gave a KJAr age of 12.4 ± 0.9 Ma (Table III). This
MORB sample represents the oldest known remnant of the North Fiji Basin oceanic crust
(Monjaret et aI., 1991) and is not considered to be related to oceanic spreading in the
NHBAT. K-Ar ages of BABB dredged in the northern JCT (type f in Table III) range
between 3.9 and 1.1 Ma. Other geochemical types encountered in the NHBAT are charac-
terized by a rather large spectrum of isotopic ages, most of them, however, being younger
than 4 Ma. To Monjaret et al. (1991) this succession reveals a polyphased and diachronous
trough formation.
3.3.3. Geochemistry
3.3.3.1. Major Elements. As noted by Fryer et al. (1990) for the volcanics from the
Izu-Bonin backarc rifts, the wide range of SiOz contents together with the high-Al z0 3
contents of the NHBAT volcanics (Table II; Fig. 5.16) clearly emphasize their arclbackarc
environment, and this is confirmed by their TiO z contents, which are usually lower than
1.2%. However, the SiOz contents of the NHBAT volcanics cover a compositional contin-
uum, which is not common in such a backarc environment.
TABLE I
Location (a) and Structural Setting (b) of Volcanic Samplesa
Location Start End Depth (m)
(a)
Jean Charcot Troughs
Vanikoro area (VAN)
1 12° 12.9'S-167°34.6'E 12°11.9'S-167°34.9'E -1470 to -940
2 12°15.3'S-167°38.5'E 12°14.9'S-167°38.8'E -1400 ± 100
3 12° 13.5 'S-167°40.6'E 12°1O.9'S-167°42.1 'E -1350 to -900
5 12°09.1 'S-167°48.5'E 12°08.8'S-167°48.8'E -1850 to -1650
6 12°14.7'S-167°50.3'E 12°15.8'S-167°51.0'E -2450 to -2120
7 12°16.3'S-167°51.8'E 12°16.7'S-167°52.7'E -2165 to -1930
3321 12°1O.3'S-167°47.0'E 12°1O.3'S-167°46.3'E -2038 to -1658
3151 12° 11.6'S-167°34.9'E 12°11.6'S-167°35.4'E -915 to -459
3154 12°11.6'S-167°34.9'E 12°11.6'S-167°35.4'E -915 to -459
3152 12°11.6'S-167°34.9'E 12°11.6'S-167°35.4'E -915 to -459
3155 12°11.6'S-16r34.9'E 12°11.6'S-167°35.4'E -915 to -459
Tikopia area (TIK)
3981 12°51.5'S-167°50.5'E 12°52.2'S-167°51.0'E -2984 to -2130
3981X 12°51.5'S-167°50.5'E 12°52.2'S-167°51.0'E -2984 to -2130
3982 12°51.5' S-167°50.5 'E 12°52.2' S-167°51.0'E -2984 to -2130
3983 12°51.5' S-167°50.5 'E 12°52.2'S-167°51.0'E -2984 to -2130
4150 12°33.3'S-167°40.6'E 12°33.2'S-167°40.8'E -1425 to -1284
4151 12°33.3'S-167°40.6'E 12°33.2'S-167°40.8'E -1425 to -1284
4153 12°33.3 'S-167°40.6'E 12°33.2'S-167°40.8'E -1425 to -1284
4155 12°33.3 'S-167°40.6'E 12°33.2'S-167°40.8 'E -1425 to -1284
4152 12°33.3'S-167°40.6'E 12°33.2'S-167°40.8'E -1425 to -1284
415X 12°33.3'S-167°40.6'E 12°33.2'S-167°40.8'E -1425 to -1284
CY11 12°32.0'S-167°40.4'E -1905
CY31 12°31.7'S-167°45.1 'E -2412
CY34 12°31.5'S-167°45.2'E -2203
CY36 12°32.4'S-167°47.2'E -1550
Vot Tande Area (VOT)
10 13°23.9'S-167°59.7'E 13°24.8'S-167°59.6'E -2130 to -1880
11 13°20.9'S-167°57.1 'E 13°21.7'S-167°55.9'E -2000 to -1550
12 13°20.9'S-167°49.2'E 13°21.4'S-167°48.1 'E -2200 to -1600
4294 13°19.9'S-167°55.8'E 13°19.9'S-167°56.5'E -2140 to -1650
4295 13°19.9'S-167°55.8'E 13°19.9'S-167°56.5'E -2140 to -1650
Hazel Holme Area (HAZ)
14 13°40.0'S-168°30.0'E 13°41.6'S-168°28.8'E -3800 to -2870
15 13°41.0'S-168°29.7'E 13°41.9'S-168°29.6'E -2900 to -2500
Coriolis Troughs
Efate area (EFA)
26 17°38.6'S-169°24.7'E 17°39.4'S-169°25.6'E -2080 to -1850
27 17°39.8'S-169°25.5'E 17°39.5'S-169°26.3'E -1960 to -1200
28 17°38.4'S-169°26.4'E 17°38.2'S-169°25.8'E -1270 to -700
29 17°38.4'S-169°25.6'E 17°38.2'S-169°26.1 'E -980 to -600
30 17°23.3'S-169°02.5'E 17°23.2'S-169°02.1 'E -1270 to -1200
31 17°23.5'S-169°08.0'E 17°23.6'S-169°09.0'E -1570 to -1250
Erromango area (ERR)
22 18°49.6'S-169°39.8'E 18°48.8'S-169°37.9'E - 2800 to - 2400
24 18°47.8'S-169°35.1 'E 18°47.9'S-169°34.9'E -1420 to -900
(continued)
TABLE I
(Continued)
Location Start End Depth (m)
(a)
Coriolis Troughs (cont.)
Erromango area (ERR) (cont.)
25 18°32.4'S-169°34.3 'E 18°31.9'S-169°34.8'E -910 to -750
Futuna area (FUT)
16 19°47.8'S-170016.5'E -3320 to -2500
17 19°47.6'S-170017.2'E 19°46.8'S-170018.3'E -2900 to -2150
19 19°46.3'S-170019.1 'E 19°46.2' S-170020.6'E -1750 to -1550
20 19°25.0'S-169°54.6'E 19°25.6'S-169°55.5'E -1400 to -1000
21 19°54.1 'S-170017.0'E 19°55.5'S-170018.6'E -3280 to -3150
Anatom area (ANA)
4311 20031.0'S-170000.5'E 20030.9'S-170000.7'E - 2288 to - 2066
4312 20°31.0' S-170000.5 'E 20030.9'S-170000.7'E - 2288 to - 2066
4313 20°31.0' S-170000.5 'E 20030.9'S-170000.7'E -2288 to -2066
4314 20°31.0' S-170000.5 'E 20030.9'S-170000.7'E -2288 to -2066
4315 20°31.0' S-170000.5 'E 20030.9'S-170000.7'E -2288 to -2066
4316 20031.0'S-170000.5'E 20 030.9'S-170000.7'E -2288 to -2066
4318 20°31.0' S-170000.5 'E 20030.9'S-170000.7'E -2288 to -2066
43110 20°3 1.0'S-1 70000.5'E 20030.9'S-170000.7'E -2288 to -2066
43112 20031.0'S-170000.5'E 20030.9'S-170000.7'E -2288 to -2066
4421 20052.6'S-170000.5'E 20°52.3' S-170000. 9'E -979 to -771
4422 20052.6'S-170000.5'E 20052.3'S-170000.9'E -979 to -771
4423 20052.6'S-170000.5'E 20°52.3 'S-170000.9'E -979 to -771
4424 20 052.6'S-170000.5'E 20052.3'S-170000.9'E -979 to -771
4531 20 034.0'S-170002.6'E 20033.9'S-170002.8'E -2054 to -1923
4532 20 034.0'S-170002.6'E 20033.9'S-170002.8'E -2054 to -1923
4533 20 034.0'S-170002.6'E 20033.9'S-170002.8'E -2054 to -1923
4534 20034.0'S-170002.6'E 20033.9'S-170002.8'E -2054 to -1923
4641 20002.6'S-170039.8'E 20 002.8'S-170040.3'E -1144 to -1055
(b)
Location Cruiseb
Jean Charcot Troughs

Vanikoro area (VAN)
I central volcanic complex
2 central volcanic complex
6 Duff Ridge (eastern limit of JCT)
7 Duff Ridge (eastern limit of JCT) I
3321 central volcanic complex (station 33) 2
Tikopia area (TIK)
3981 central graben (eastern wall) (station 39) 2
3981X central graben (eastern wall) (station 39) 2
4150 western seamount (W slope) (station 41) 2
(continued)
TABLE I
(Continued)
Location Cruise
(b)
Jean Charcot Troughs (cant.)
Tikopia area (TIK) (cant.)
415X western seamount (W slope) (station 41) 2
CYlI western seamount (NW slope) (station 41) 3
CY31 eastern seamount (NW slope) (station 41) 3
Vot Tande area (VOT)
10 small isolated seamount (east of central horst)
11 central horst (E slope)
12 Vot Tande island basement (SE flank)
4294 central horst (W slope) (station 42) 2
4295 central horst (W slope) (station 42) 2
Hazel Holme Area (HAZ)
14 southern scarp of western termination (basis)
15 southern scarp of western termination (top)
Coriolis Troughs
Efate area (EFA)
26 eastern flank (basis) of eastern graben
27 eastern flank (basis) of eastern graben
28 eastern flank (middle) of eastern graben
29 eastern flank (top) of eastern graben
30 western volcanic cone to the north of the trough
31 eastern volcanic cone to the north of the trough
22 western scarp of northwestern termination (basis)
24 western scarp of northwestern termination (top)
25 isolated volcanic cone to the north of the trough
Futuna area (FUT)
16 southeastern scarp of the trough (basis)
17 southeastern scarp of the trough (middle)
19 southeastern scarp of the trough (top)
20 volcanic cone on the NW flank of the trough
21 small southern relief on the trough bottom
Anatom area (ANA)
4311 to 43112 small ridge, within the en echelon termination of the trough (station 43) 4
4421 to 4424 small seamount outside the trough, on the central chain axis (station 44) 4
4531 to 4534 eastern flank of the en-echelon termination of the trough (station 45) 4
4641 volcanic cone (Mt Yokosuka) to the east of the trough (station 46) 4
"The CY samples were picked up in situ by Cyana submersible.
hi: SEAPSO 2 (1985); 2: KAIYO 89 (1989), with STARMER station number; 3: SAVANES (1991), with STARMER
station number; 4: YOKOSUKA 90 (1991), with STARMER station number.
FIGURE 5.14. Location of volcanic samples dredged

during the SEAPSO 2 cruise in the Jean-Charcot troughs
(top) and Coriolis troughs (bottom). Coordinates and
structural setting of these samples are listed in Tables la-
Ib (sample Dl in Fig. 5.14 = sample I in Tables I-III;
sample D25 in Fig. 5.14 = sample 25 in Tables I-III,
etc.). Bathymetry in km. Vate = Efate, Gaua = Santa
Maria in Fig. 5.2 (From Monjaret et ai., 1991).
The ~o versus Si02 diagram (Fig. 5.16) emphasizes two trends, similar to those for
the New Hebrides central chain volcanics. A lower-~O suite encompasses most samples
from the northern Jean-Charcot troughs: all the VAN volcanics (except the 7M4 andesite),
including the high-Na/low-K dacites studied by Nakada et al. (1994); most of the TIK
volcanics, except for four samples with higher-Kp contents (3981, 3981X, 3982, 3983;
Table II), which are calc-alkaline andesites; and the island-arc tholeiites of the VOT area. A
higher-K20 suite is represented in the southern Coriolis troughs by the volcanics from the
,..----,..----.----,;-,.--"'T""'"---r-r--------, 19'
FIGURE 5.15. Location of samples dredged during the YOKOSUKA 90 cruise in the southern Coriolis troughs.
Coordinates and structural setting of these samples are listed in Tables Ia-Ib (ST 43 =samples 4311 to 43112, ST
44 =samples 4421 to 4424, ST 45 =samples 4531 to 4534, ST 46 =sample 4641 in Tables I and II). Bathymetry in
kIn (From Eissen et aI., unpublished information, 1994).
Efate (EFA) and Anatom (ANA) areas. The ANA backarc volcanics, however, differ from
the Anatom Island volcanics, the latter belonging mainly to a lower-K20 suite; in contrast,
hyper-K dacites from Efate area (EFA; i group on Table II) show a close chemical similarity
with acidic pumices of Efate island (Coulon et al., 1979).
All other groups of volcanics from the backarc areas display an intermediate K 20
trend with respect to Si02 contents.
3.3.3.2. Trace Elements. Various trace elements are plotted against Si02 in Fig. 5.17.
LILE (e.g., Rb, Ba, and Th) and HFSE (e.g., Nb, Ta, Hf, Sm, Zr and Y) classically show a
steady increase with increasing Si02• In contrast, Ni behaves compatibly. Rb vs. Si02, Ba
vs. Si02, and Th vs. Si02 diagrams (Fig. 5.17) look quite similar, and all emphasize the
distinction between the lower-Kz0 suites (most of JCT samples) and the higher-K20 suites
(e.g., EFA samples). Contrasting with other LILE, Sr contents decrease with increasing
Si02 contents. This trend, likely due to plagioclase fractionation, is more pronounced in the
felsic lavas (Si0 2 >60%).
The Ba/Zr ratio, often considered as an indicator of the extent of "arc signature"
(Fryer et al., 1990), varies greatly in the NHBAT volcanics. VAN and HAZ samples (Table
II) have the lowest Ba/Zr ratio, usually between 0.4 and 1. These values are still well above
the ratio for N-MORB (0.08; Sun and McDonough, 1989). All other samples have a Ba/Zr
ratio ranging between 1 and 5 (or more), typical of volcanics from island-arc environments.
On the Th/Yb vs. Ta/Yb diagram (Pearce, 1983) the distinction between northern and
southern troughs is neat (Fig. 5.18). The geochemistry of all Coriolis samples (EFA, ERR,
FUT; southern troughs) is influenced by a subduction component, these samples having
relatively low Ta/Yb ratios and high Th/Yb ratios. In contrast, this subduction component is
less pronounced, or even absent, in the Jean-Charcot (VAN, VOT) and Hazel Holme (HAZ)
samples (northern troughs).
The chondrite-nonnalized REE patterns and N-MORB nonnalized "spider diagrams"
of the different petrological types encountered in the NHBAT are shown in Fig. 5.19.
MORB and BABB dredged in the VAN area present similar flat, or slightly LREE-depleted,
patterns, as typically observed in N-MORB or BABB (Saunders and Tarney, 1984; Eissen
et ai., 1991; 1994; Sinton et ai., 1994). MORB from the Hazel Holme area are significantly
LILE- and LREE-enriched, compared to the VAN N-MORB, and are therefore classified
as E-MORB.
The LILE (Ba, Rb, Sr) enrichment and HFSE (Ta, Nb, Zr, Ti, Y, Yb) depletion, relative
to N-MORB, characteristic of island-arc basic and intennediate volcanics (Gill, 1981;
Thorpe, 1982; Wilson, 1989), is particularly clear in Fig. 5.19 (see, for example, IAT types
b, c, d of Table II from the Coriolis troughs). As noted, K-rich felsic lavas-that is, high-
and hyper-K dacites (types hand i of Table II)-usually show a strongly-LREE-enriched
pattern, often accompanied by a Sr depletion due to plagioclase fractionation.
In summary, even though no oceanic spreading actually occurs in the New Hebrides
backarc troughs, some BABB dredged in their very northern part (central volcanic complex
ofthe Vanikoro area, Jean-Charcot troughs) still show evidence of a limited trend toward an
aborted marginal basin, with an age between 3.9 and 1.1 Ma (Tables I-III). A neat
petrological dichotomy marks these troughs. Volcanics from the southern troughs (CT)
present all the classical characteristics of arc-related, orogenic magmas. In contrast, the
northern troughs (JCT) are floored with volcanics reflecting more complex influences, with
compositions trending between BABB and more typical arc lavas.
The significant extension and thick sedimentary input that characterize the central and
northern New Hebrides arc platfonn have been noted, the Aoba and Vanikoro-Torres
basins (Fig. 5.13) having no equivalent in the southern part of the arc. The geographical
proximity of the Vanikoro-Torres Basin and JCT, both affected by extensional stresses,
raises therefore the question of their possible genetical relationship. No petrological or
geophysical argument, however, relates BABB fonnation in the JCT between 3.9 and 1.1
Ma to extension in the Vanikoro-Torres Basin.
3.4. Backarc Hydrothermal Activity and Ferromanganese Crusts

3.4.1. Backarc Hydrothermal Activity
During the KAIYO 89 cruise, five hydrocasts were perfonned in the JCT for
conductivity-temperature-depth (CTD) profiles and methane (CH4) analysis. Following
onboard analyses, one hydrocast gave a distinct CH4 anomaly, but this was not accom-
panied by a thennal anomaly. This water sampling was located in the Vanikoro area, at
station 31 (12°12.50'S-167°38.77'E), between 1617 m depth and sea level (Fig. 5.5). The
CH4 anomaly, which is possibly indicative of an hydrothennal plume, was found at 1200 m
depth-that is, 400 m above the sea bottom (KAIYO 89 Cruise Report, 1990). It was
characterized by an extremely high methane/manganese ratio of 8, similar to that observed
in the Okinawa Trough (Nojiri and Ishibashi, 1991).
Five heat flow stations were also run during this cruise, using a gravity corer, in small
sedimentary basins within the JCT (three measurements in the Vanikoro area, one in the
Tikopia area, one in the Vot Tande area). Heat flow values vary between 23 and 161 mW/m 2 ,
TABLE II
Chemical Whole-Rock Analyses of Volcanic Samplesa
Location VAN VAN VAN VAN VAN

Sample 7M2 7M3 2MI 2M5 3M4
Type a a f f f
Si02 (%) 47.00 48.20 49.00 49.00 49.40

Ti0 2 1.36 1.27 1.25 1.15 1.01
AIP3 17.60 17.39 16.43 16.50 15.78
Fe20 3 1.56 1.43 1.57 1.52 1.46
FeO 7.96 7.30 7.99 7.76 7.44
MoO 0.14 0.16 0.15 0.14 0.20
MgO 8.30 8.87 6.96 8.55 7.15
CaO 12.22 12.11 12.51 12.40 12.15
Nap 3.20 3.04 3.00 2.30 2.80
Kz°
P 20 5
0.11
0.05
0.09
0.02
0.31
0.05
0.28
0.05
0.15
0.10
H2O+ -0.08 -0.14 0.11 0.12 0.25
HP- 0.24 0.11 0.10 0.14 0.23
Total 99.66 99.85 99.43 99.91 98.12
Source INA ICP AAJICP INA ICP INA ICP INA ICP
Rb (ppm) 1.1 2 4.0 2.0 1.26
Sr 297 180 288 314 251
Ba 7.6 20 54.2 40.6 35.3
Hf 1.9 1.5 1.2 1.59
Zr 86 76 75 83 51 60 45 81 51
Ta 0.053 0.12 0.094 0.130
Tb 0.35 0.34 0.32 0.348
U 0.20 0.17 0.12 0.25
Cs 0.05 0.06 0.05 0.049
Sb 0.02 0.Q2 0.01 0.087
As 2.84
La 2.3 2.7 2.5 3.7 4.0 3.4 3.8 3.66 3.8
Ce 6.6 9 9.5 7.6 II 7.0 8 7.6 10
Nd 8 8.0 8 7 8
Sm 2.6 2.3 2.3 2.39
Eu 1.1 1.10 1.05 0.97 1.00 0.93 0.85 1.13 0.95
Gd
Tb 0.58 0.53 0.46 0.514
Dy 4.0 4.2 3.4 3.0 3.8
Er 2.8 2.6 2.3 2.1 2.8
Yb 2.4 2.50 2.55 2.4 2.10 1.95 1.85 1.95 2.25
Lu
Sc 34.6 39.9 37.3 38.6
V 167 218 265 216 234
Cr 255 238 148 274 150
Co 49.3 50 36.8 41.7 38.6
Ni 159 159 51 115 53
Y 26.5 27 22 20 24.5
Nb 1.5 1.5 2.2 1.8 2.3
(continued)
TABLE /I
( Continued)
Location VAN VAN VAN VAN VAN VAN

Sample 5M5 5MI 6MI 5M4 3321 3M3
Type f f c f b f
Si02 49.50 50.15 50.80 51.10 49.28 52.20

Ti02 0.85 0.79 1.03 0.76 0.55 1.22
A1 20 3 16.65 16.53 16.90 15.90 18.01 15.83
Fe20 3 1.52 1.37 1.67 1.39 9.02 1.59
FeO 7.76 6.99 8.53 7.08 8.10
MnO 0.16 0.16 0.16 0.16 0.17 0.18
MgO 7.35 7.51 5.18 7.15 7.11 5.51
CaO 13.05 12.83 10.80 12.11 13.70 9.86
Nap 2.44 2.33 3.02 2.30 1.74 3.51
~O 0.31 0.32 0.61 0.28 0.28 0.36
P 20 5 0.05 0.05 0.10 0.05 0.10 0.15
H2O+ 0,03 0.42 0.49 0.13 0.16
H 2O- 0.11 0.19 0.27 0.18 0.21
Total 99.78 99.64 99.56 98.59 99.96 98.88
Source INA ICP AAlICP INA ICP INA ICP XRFIICP INA ICP
Rb 3.4 4 7.6 3 3 4.3
Sr 295 227 291 288 298 258
Ba 42 60 78 45.8 167 72.7
Hf 1.2 1.8 1.15 2.19
Zr 40 42 41 66 63 49 43 27 84 74
Ta 0.079 0.14 0.073 0.195
Th 0.23 0.52 0.237 0.470
U 0.06 0.26 0.103 0.24
Cs 0.11 0.17 0.110 0.083
Sb 0.05 0.07 0.037 0.070
As 1.1 0.44 1.66
La 2.5 3.1 3.0 4.9 5.8 2.67 2.7 3.0 4.80 4.9
Ce 7.0 9 10 10.9 13 5.4 10 6.8 9.6 14
Nd 6 6 10 6.5 5.2 10
Sm 1.8 2.8 1.77 2.0 2.95
Eu 0.72 0.80 0.75 0.95 1.05 0.73 0.75 0.71 1.17 1.25
Gd 2.0
Tb 0.39 0.51 0.366 0.628
Dy 2.8 2.8 3.4 2.8 2.2 4.4
Er 1.6 2.0 2.2 1.8 1.4 2.9
Yb 1.9 1.65 1.70 2.3 2.10 1.42 1.75 1.3 2.67 2.70
Lu 0.19
Sc 38.4 34.0 37.8 32.0
V 196 243 258 240 240 278
Cr 141 144 26 153 61 13
Co 38.1 34 32.2 37.3 32.8
Ni 52 48 26 48 52 23
Y 18 18 23 18 11 29
Nb 1.8 2 2.7 1.2 1 3
(continued)
TABLE /I
( Continued)
Location VAN VAN VAN VAN VAN

Sample 7M4 3Ml 3M2 IMI IM9
Type e+c g g g g
Si0 2 55.10 57.90 60.50 62.60 64.00

Ti02 1.50 1.82 1.63 1.25 0.77
AI 20 3 14.52 15.75 14.78 15.93 15.50
Fe20 3 1.74 1.39 1.24 1.02 0.78
FeO 8.90 7.11 6.33 5.21 4.00
MnO 0.20 0.19 0.21 0.16 0.18
MgO 3.24 2.41 2.48 1.73 2.19
CaO 6.77 5.58 4.71 4.42 4.53
Nap 4.09 5.34 5.07 6.00 5.80
Kp 1.25 0.71 0.70 0.75 0.71
P 20 S 0.35 0.18 0.40 0.20 0.15
H 2O+ 0.58 0.79 0.97 0.70 0.78
H 2O- 0.27 0.11 0.25 0.07 0.09
Total 98.51 99.28 99.27 100.04 99.48
Source INA ICP INA ICP AAlICP INA ICP INA ICP
Rb 21.0 8.7 7 10.6 11.1
Sr 268 204 215 207 529
Ba 202 125 104 158 158
Hf 3.9 4.1 4.7 4.8
Zr 157 139 140 149 142 190 168 204 181
Ta 0.54 0.41 0.50 0.51
Tb 1.43 0.83 1.22 1.28
U 0.65 0.37 0.42 0.50
Cs 0.45 0.21 0.25 0.25
Sb 0.12 0.07 0.08 0.07
As 0.74
La 12.2 13.0 9.2 9.6 9.2 11.9 12.7 11.7 12.6
Ce 29.7 32 25.9 26 24 31.0 30 30.1 31
Nd 22 21 20 22 22
Sm 5.3 5.6 5.8 5.7
Eu 1.8 1.70 2.1 1.90 1.90 2.1 1.80 2.0 1.85
Od
Tb 0.95 1.16 1.20 1.16
Oy 6.3 7.9 7.7 8.3 7.9
Er 4.1 5.3 5.0 5.4 5.3
Yb 3.7 3.80 4.7 5.00 4.85 5.3 5.40 5.2 5.30
Lu
Sc 30.3 21.4 16.1 14.7
V 337 73 67 70
Cr 3 5 3 8 41
Co 26.5 8.9 7.0 5.9 9.6
Ni 7 2 0 2 16
Y 40 53 51 54 53
Nb 8.0 5.4 5.0 7.2 7.5
(continued)
TABLE /I
(Continued)
Location VAN VAN VAN VAN VAN VAN VAN

Sample IM8 IM5 1M3 3151 3154 3152 3155
Type g g g g g g g
Si02 65.20 67.00 67.50 66.43 66.96 66.77 66.09

Ti02 0.90 0.86 0.78 0.85 0.86 0.95 0.96
AI 20 3 16.20 14.99 14.60 15.33 15.52 15.38 15.28
Fe20 3 0.75 0.68 0.68 1.40 1.20 5.23 5.26
FeO 3.84 3.46 3.48 3.00 3.22
MnO 0.15 0.15 0.16 0.16 0.16 0.20 0.20
MgO 1.21 1.17 1.04 1.19 1.17 1.16 1.17
CaO 3.56 3.27 2.79 3.18 3.23 3.26 3.29
Nap 6.55 6.12 6.12 5.56 6.21 6.76 6.71
Kp 0.88 0.83 0.93 0.85 0.86 0.92 0.91
P 20 5 0.15 0.25 0.15 0.24 0.25 0.20 0.22
H 2O+ 0.72 0.73 0.90 0.73 0.60
HP- 0.05 0.11 0.15 1.28 0.08
Total 100.16 99.62 99.28 100.20 100.32 100.83 100.09
Source INA ICP AAJICP AAJICP XRFIICP XRFIICP XRFIICP XRFIICP
Rb 12.4 11 13 13.5 14.2 13 13
Sr 148 178 149 172 174 164 164
Ba 186 165 163 193 193 211
Hf 5.4
Zr 213 205 204 214 220 223 227 229
Ta 0.60
Til 1.53 1.4
U 0.50
Cs 0.30
Sb 0.08
As 1.3
La 13.7 14.5 14.7 15 14.8 15.0 15.4 14.8
Ce 28.0 34 37 37 32 33 36.3 33.9
Nd 24 25 25 24 25 22.8 21.7
Sm 6.1 7.6 7.2
Eu 1.94 2.05 1.95 2.00 2.00 1.95 2.2 2.0
Od 7.9 7.2
Til 1.23
Dy 8.6 8.8 10.0 8.6 8.8 8.7 7.9
Er 5.6 5.9 6.1 5.5 5.8 5.9 5.2
Yb 6.2 5.85 5.95 6.10 5.70 5.75 6.1 5.3
Lu 0.89 0.80
Sc 11.1
V 38 15 22 20 29 24
Cr 3 0 2 2
Co 4.9 4 2
Ni 2.3 0 0 5 5
Y 59 60 61 52.2 52.6 52 46
Nb 8.7 8.8 9.1 8.5 8.5 10 9
(continued)
TABLE /I
(Continued)
Location TIK TIK TIK TIK TIK TIK TIK

Sample 3981 3981X 3982 3983 4150 4151 4153
Type e e e e g g g
Si02 51.92 52.09 52.62 56.97 59.07 59.15 59.12

Ti0 2 0.70 0.73 0.76 0.77 0.95 0.92 0.93
AIP3 19.53 20.03 16.01 17.92 17.57 17.59 17.68
Fe20 3 3.68 7.76 11.67 7.87 2.55 2.50 2.55
FeO 3.78 3.95 3.93 3.89
MnO 0.14 0.16 0.21 0.18 0.17 0.17 0.17
MgO 2.86 2.65 3.99 2.34 3.07 3.06 3.03
CaO 8.80 8.85 9.19 7.30 6.68 6.70 6.68
Nap 2.81 3.59 2.71 3.66 3.60 3.62 3.70
K 20 2.32 2.43 1.30 1.81 0.94 0.92 0.93
P 20 S 0.51 0.49 0.22 0.24 0.26 0.26 0.26
H 2O+ 1.29 0.90 0.77 0.81
H 2O- 1.70 0.21 0.18 0.19
Total 100.04 98.78 98.68 99.06 99.92 99.77 99.94
Source XRFIICP XRFIICP XRFIICP XRFIICP XRFIICP XRFIICP XRFIICP
Rb 33.5 38 20 31 12.7 12.1 12.4
Sr 1020 999 357 376 286 288 288
Ba 311 313 515 270 180 180 183
Hf
Zr 92 109 58 88 154 154 155
Ta
Th 1.5
U
Cs
Sb
As
La 13.8 14.3 6.3 9.0 12.9 12.5 12.5
Ce 29 31.7 12.3 19.4 25 25 25
Nd 19.5 17.8 8.5 10.9 17 17 17
Sm 4.8 2.8 3.4
Eu 1.35 1.4 0.9 1.1 1.40 1.55 1.45
Gd 3.6 2.8 3.5
Tb
Dy 3.2 3.2 3.2 3.6 5.2 4.8 5.0
Er 2.0 1.9 2.1 2.3 3.2 3.2 3.2
Yb 1.75 1.8 2.0 2.3 3.40 3.40 3.50
Lu 0.22 0.34 0.32
Sc
V 269 140 133 135
Cr 9 3 2 2
Co
Ni 12 6 6 6
Y 19.7 31.7 31.5 31.7
Nb 2.5 9.5 9.5 9.0
(continued)
TABLE /I
(Continued)
Location TIK TIK TIK TIK TIK TIK TIK

Sample 4155 4152 415X CY11 CY31 CY34 CY36
Type g g g g c c C
Si0 2 59.10 58.48 58.48 61.55 53.66 53.71 54.79

Ti02 0.93 1.01 0.99 0.62 1.34 1.37 0.92
AIP3 17.54 17.54 17.60 17.60 16.18 17.76 17.79
Fe20 3 2.57 7.29 7.22 5.26 8.91 9.23 9.00
FeO 3.88
MnO 0.17 0.20 0.20 0.16 0.17 0.17 0.16
MgO 3.07 2.90 2.92 2.19 5.66 5.22 4.28
CaO 6.70 6.77 6.74 5.12 9.07 9.60 9.28
Nap 4.41 4.64 4.73 5.27 3.90 3.29 3.40
Kp 0.94 0.97 0.98 1.06 0.72 0.70 0.65
P 20 S 0.26 0.23 0.24 0.21 0.22 0.15 0.14
H2O+ 0.73
H2O- 0.22
Total 100.52 100.03 100.10 99.04 99.83 101.20 100.41
Source XRFIICP XRFIICP XRFIICP XRFIICP XRFIICP XRFIICP XRFIICP
Rb 12.5 13 11 17 11 11 9
Sr 287 273 265 238 178 238 286
Ba 188 217 196
Hf
Zr 155 167 163
Ta
Th 1.30 3 2
U 2
Cs
Sb
As
La 12.7 12.6 12.2 13.3 8.4 4.9 5.3
Ce 27 28.7 27.2 27.1 22.4 11.9 11.9
Nd 17 15.8 15.5 14.6 16.5 9.4 8.6
Sm 4.8 4.9 4.0 6.2 3.5 3.0
Eu 1.50 1.5 1.5 1.2 1.8 1.2 1.0
Gd 4.6 4.6 3.6 6.2 4.0 3.0
Tb
Dy 5.1 5.1 5.1 3.9 6.9 4.3 3.3
Er 3.4 3.4 3.2 2.5 4.2 2.7 2.1
Yb 3.50 3.5 3.3 2.7 4.1 2.6 1.9
Lu 0.51 0.51 0.42 0.60 0.38 0.30
Sc
V 135 176 189
Cr 3 5 8
Co
Ni 7 9 7
Y 31.5 29 30 22 38 24 17
Nb 9.5 10 10
(continued)
TABLE /I
(Continued)
Location VOT VOT VOT VOT VOT

Sample llMl 10M 1 10M2 12Ml 11M2
Type e+c c c d e+c
Si02 47.50 49.40 50.70 50.70 51.00

Ti0 2 1.40 1.03 0.94 0.77 1.36
AIP3 15.79 18.25 17.90 16.25 16.51
Fe20 3 1.94 1.61 1.51 1.39 1.96
FeO 9.90 8.22 7.73 7.11 9.99
MnO 0.17 0.16 0.17 0.15 0.19
MgO 5.57 5.85 5.79 9.60 4.18
CaO 10.62 10.82 10.53 10.62 8.80
Nap 2.62 2.97 2.89 2.32 2.60
K2 0 0.96 0.61 0.64 0.66 1.06
PP5 0.15 0.10 0.15 0.10 0.20
H 2O+ 1.32 0.37 0.28 0.33 0.51
HP- 1.01 0.28 0.26 0.13 0.46
Total 98.95 99.67 99.49 100.13 98.82
Source INA ICP INA ICP INA ICP INA ICP INA ICP
Rb 19.6 7.7 7.5 9.4 20.6
Sr 344 296 415 457 281
Ba 127 84 89 115 144
Hf 1.85 1.5 1.6 1.4 1.8
Zr 87 69 47 56 67 55 50 50 66 65
Ta 0.31 0.17 0.17 0.26 0.23
Th 0.43 0.30 0.26 0.56 0.30
U 0.17 0.11 0.12 0.20 0.18
Cs 2.0 0.11 0.16 0.22 1.30
Sb 0.20 0.11 0.11 0.09 0.05
As 1.3 1.2 0.16
La 5.7 6.0 4.2 4.9 4.0 4.6 5.0 6.0 4.1 5.4
Ce 15.0 15 8.1 12 8.3 12 12.9 13 9.0 14
Nd 11 9.5 9 8 11
Sm 3.1 2.4 2.3 2.1 3.1
Eu 1.1 1.10 0.90 0.85 0.89 0.95 0.76 0.80 1.05 1.10
Od
Tb 0.64 0.47 0.48 0.38 0.65
Dy 4.2 3.5 3.3 2.8 4.6
Er 2.7 2.4 2.3 1.8 3.2
Yb 2.5 2.55 2.2 2.2 2.2 2.05 1.5 1.65 2.8 2.75
Lu
Sc 38.2 31.0 31.1 35.1 35.3
V 294 238 260 208 370
Cr 54 31 13 410 6
Co 41.8 34.3 34.5 38.3 37.4
Ni 35 39 39 179 17
Y 28 25 22 17 29
Nb 4.9 3.1 2.6 4 3.3
(continued)
TABLE /I
( Continued)
Location VOT VOT HAZ HAZ HAZ HAZ

Sample 4294 4295 14M2 14MI 14M7 14MI4
Type e e a a a a
Si0 2 54.57 56.87 45.20 46.10 48.20

Ti02 0.95 0.80 1.52 1049 2.06
AI 20 3 15.15 15.71 16.05 15.09 14.54
Pe20 3 9.85 9.77 1.63 1.55 1.88
PeO 8.31 7.93 9.60
MnO 0.18 0.16 0.14 0.16 0.19
MgO 4.39 3.85 7.26 7.92 7.25
CaO 8.54 7.60 10.85 10.59 9.05
N~O 2.86 2.57 2.84 2.88 3.82
K20 1.75 2.31 0.21 0.26 0040
P 20 S 0.87 0.33 0.15 0.10 0.15
H2O+ 2.06 2.02 1.94
H2O- 3.00 3.16 0.28
Total 99.11 99.97 99.22 99.25 0.00 99.36
Source XRFIICP XRPIICP INA ICP INA ICP INA INA ICP
Rb 37 47 204 1.7 2.0 404
Sr 395 395 318 304 190 466
Ba 270 326 36.7 5204 17.5 118
Hf 2.4 2.1 2.0 3.0
Zr 80 76 106 93 76 89 77 139 116
Ta 0.60 0.55 0.29 1.54
Th 0.68 0.59 0.44 1.74
U 0.22 0.30 0.20 0.45
Cs 0.29 0.11 0.02 0.07
Sb 0.01 0.10 0.07 0.09
As 1.1
La 8.5 7.6 6.7 6.9 6.3 7.3 4.1 14.2 15
Ce 18.7 17.1 16.5 17 16.6 17 11.7 24.7 32
Nd 11.1 9.6 12 II 19
Sm 304 3.0 2.9 3.0 2.3 4.1
Eu 1.0 0.84 1.12 1.10 1.23 1.25 0.78 1.45 1.55
Od 304 2.6
Tb 0.63 0.59 0048 0.73
Dy 3.2 2.8 4.1 4.0 5.0
Er 2.1 1.8 3.0 2.5 3.2
Yb 2.1 1.7 2.9 2.55 2.2 2040 1.8 2.8 2.55
Lu 0.32 0.29
Sc 37.6 34.2 23.0 40.6
V 213 195 297
Cr 356 390 39 115
Co 48.7 44.8 24.4 41.7
Ni 199 204 23 48
Y 27 27 30
Nb 7.2 7.5 20
(continued)
TABLE /I
(Continued)
Location HAZ HAZ HAZ EFA EFA

Sample l5M6 14M5 15M12 27M12 29M6
Type b a d e+d
Si02 51.20 52.50 62.80 46.50 47.50

Ti0 2 0.72 0.75 0.62 0.85 0.69
A1 20 3 17.05 16.30 16.07 15.10 13.65
Fe 20 3 1.50 1.44 0.70 1.36 1.21
FeO 7.65 7.32 3.60 6.95 6.19
MnO 0.17 0.16 0.11 0.15 0.18
MgO 5.45 6.08 1.64 9.45 10.30
CaO 10.74 4.11 0.73 9.55 7.30
NazO 2.42 6.08 5.99 2.47 2.32
KzO 0.68 0.10 4.66 0.70 1.87
P 20 5 0.12 0.15 0.18 0.10 0.20
H2O+ 0.78 3.81 1.59 3.40 2.89
HzO- 0.63 0.69 0.39 2.71 4.65
Total 99.11 99.49 99.08 99.29 98.95
Rb 8.8 1.4 66.8 7.7 19.1
Sr 356 158 80 354 757
Ba 88 6.0 600 54.0 677
Hf 1.1 1.24 4.6 1.25 0.95
Zr 48 30 46 47 188 158 46 34 30 33
Ta 0.24 0.15 0.35 0.068 0.052
Th 0.40 0.27 3.4 0.314 0.368
U 0.14 0.11 1.48 0.15 0.39
Cs 0.16 0.01 0.06 0.068 0.46
Sb 0.13 0.10 0.20 0.061 0.Q18
As 2.5 0.68
La 4.6 5.2 3.9 4.3 17.6 16.0 2.34 2.7 2.92 3.3
Ce 10.2 10.5 10.4 12.5 34.3 33 5.2 9 6.3 9
Nd 7.5 8 20 6 6
Sm 2.0 2.0 4.5 2.02 1.79
Eu 0.86 0.80 0.75 0.75 0.98 1.05 0.90 0.90 0.73 0.75
Gd
Tb 0.36 0.39 0.65 0.422 0.328
Dy 2.5 2.9 4.0 3.0 2.5
Er 1.6 1.8 2.8 2.0 1.8
Yb 1.3 1.35 1.51 1.65 I 3.2 3.00 1.51 1.75 1.12 1.35
Lu
Sc 36.1 29.7 7.3 38.6 32.6
V 249 163 237 240
Cr 53 51 255 317
Co 35.2 33.6 6.1 34.6 44.9
Ni 35 32 5 65 160
Y 16 18 28 20 14
Nb 3.7 3.0 5.6 1.0 1.6
(continued)
TABLE /I
( Continued)
Location EFA EFA EFA EFA EFA

Sample 31M2 31MI 30M2 30MI 27M 17
Type d d d d
Si02 51.70 52.00 53.00 54.40 61.80

Ti02 0.74 0.72 0.53 0.60 0.80
AIP3 15.50 15.35 13.22 15.30 15.89
Fe 20 3 1.33 1.33 1.42 1.35 0.85
FeO 6.76 6.76 7.25 6.87 4.34
MnO 0.15 0.15 0.16 0.15 0.Q7
MgO 7.34 7.37 10.28 7.72 2.01
CaO 1l.50 1l.55 9.18 8.97 2.82
N~O 2.35 2.27 1.92 2.16 4.50
~O 0.38 0.41 0.62 0.75 5.60
P20 S 0.15 0.15 0.05 0.10 0.30
H2O+ 0.30 0.09 0.32 0.27 0.41
H2O- 0.12 0.10 0.12 0.10 0.18
Total 98.32 98.25 98.07 98.74 99.57
Rb 4.8 5.6 10.2 11.5 96
Sr 395 402 352 351 257
Ba 82.4 81.3 178 200 818
Hf 1.03 1.12 0.98 1.16 4.2
Zr 38 34 37 35 43 34 47 39 171 90
Ta 0.068 0.Q75 0.040 0.052 0.21
Th 0.418 0.438 0.680 0.777 4.1
U 0.14 0.22 0.23 0.13 2.2
Cs 0.22 0.22 0.30 0.34 0.68
Sb 0.037 0.058 0.044 0.050 0.10
As 1.04 1.41 1.00 1.31 0.8
La 3.64 3.7 3.43 3.9 3.57 4.0 3.95 4.7 12.6 14.3
Ce 10.3 9 10.7 10 5.8 9.5 7.7 9.5 27.3 31
Nd 7 6.5 5.5 6.5 20
Sm 1.84 1.88 1.47 1.62 3.9
Eu 0.75 0.70 0.78 0.80 0.52 0.55 0.59 0.60 0.98 1.05
Gd
Tb 0.339 0.355 0.266 0.298 0.53
Dy 2.5 2.5 2.2 2.3 3.3
Er 1.6 1.7 1.5 1.7 2.1
Yb 1.36 1.55 1.46 1.50 1.15 1.40 1.24 1.45 2.5 1.90
Lu
Sc 33.8 35.7 32.3 27.7 8.8
V 264 252 214 218 140
Cr 209 236 388 178 42
Co 33.7 34.9 44.7 34.8 6.7
Ni 72 74 188 121 17
Y 16 16 14 15 21
Nb 1.6 1.7 1.0 1.2 3.7
(continued)
TABLE /I
( Continued)
Location EFA EFA EFA EFA EFA EFA ERR

Sample 27MI 27M4 29M3 28Ml 26M6 26M7 22Ml
Type h h d
Si02 63.00 63.40 64.00 64.80 65.70 66.30 49.00

Ti02 0.43 0.50 0.50 0.65 0.55 0.61 0.79
AIP3 14.93 15.09 15.08 14.95 14.36 14.16 14.73
Fe20 3 0.65 0.68 0.66 0.60 0.71 0.69 1.54
FeO 3.30 3.46 3.38 3.08 3.63 3.53 7.88
MnO 0.12 0.12 0.12 0.09 0.12 0.12 0.17
MgO 1.00 0.85 0.81 2.47 1.05 1.01 6.51
CaO 2.33 2.32 2.21 0.94 2.58 2.37 12.94
N~O 4.64 4.55 4.51 3.47 4.83 4.76 2.08
K20 4.96 4.92 4.79 6.25 3.70 3.62 0.81
PP5 0.15 0.15 0.15 0.40 0.15 0.20 0.20
H2O+ 2.73 2.65 2.26 1.54 2.16 1.91 1.95
H2O- 1.59 0.44 0.99 0.61 0.33 0.07 0.68
Total 99.83 99.13 99.46 99.85 99.87 99.35 99.28
Source INA ICP AAlICP AAlICP AAlICP INA ICP AAlICP AAlICP
Rb 57.6 97 65 114 65.4 70 13
Sr 297 272 256 153 228 272 405
Ba 912 825 819 1045 1017 950 108
Hf 6.0 6.1
Zr 235 205 214 218 150 204 210 210 54
Ta 0.34 0.30
Th 7.4 5.1
U 2.7 1.85
Cs 1.78 2.2
Sb 0.23 0.23
As 3.7 6.1
La 23.8 24.5 25 25 15.2 23.6 25 25 4.9
Ce 48.1 50 50 51 34 44.2 51 50 12
Nd 25 24 25 19 28 28 8
Sm 5.0 6.1
Eu 1.00 1.10 1.10 1.10 0.95 1.3 1.30 1.45 0.90
Gd
Tb 0.62 0.84
Dy 4.1 4.1 4.5 3.4 5.3 5.6 2.5
Er 2.8 3.0 3.0 2.3 3.7 4.0 1.7
Yb 3.3 2.85 3.00 3.05 2.20 4.2 3.90 4.05 1.60
Lu
Sc 8.2 10.3
V 40 49 50 83 45 52 306
Cr 7 2 3 10 2 2 298
Co 7.7 7 7 7 6.3 5 38
Ni 1.9 2 10 2 I 101
Y 27 28 29 22.5 37 38 17
Nb 5.1 5.3 6.2 3.7 4.5 4.8 1.3
(continued)
TABLE 1/
(Continued)
Location ERR ERR ERR ERR ERR PUT

Sample 25M2 25M4 24M6 24M4 24M3 19M1
Type c c b d h c+d
Si02 50.30 51.30 52.50 65.00 50.00

Ti0 2 0.94 0.79 0.60 0.62 1.00
AI 20 3 16.84 18.13 14.80 14.60 14.90
Fe 20 3 1.85 1.50 1.55 0.98 1.21
FeO 9.45 7.66 7.90 5.00 6.16
MnO 0.16 0.16 0.14 0.14 0.24
MgO 5.91 4.31 7.21 1.11 9.62
CaO 10.55 10.58 10.33 3.60 10.46
Nap 2.47 2.85 2.42 4.19 2.29
K20 0.62 0.85 0.83 2.96 0.83
P20 S 0.15 0.10 0.10 0.25 0.25
H 2O+ -0.14 0.72 0.12 0.92 1.12
H 2O- 0.13 0.76 0.23 0.27 0.61
Total 99.23 0.00 99.71 98.73 99.64 98.69
Source INA ICP INA ICP AAlICP INA INA ICP INA ICP
Rb 10.8 7.7 10 12.8 49.4 12.2
Sr 606 620 410 424 244 789
Ba 168 152 283 218 626 106
Hf 1.4 1.3 1.4 4.7 1.8
Zr 52 42 49 42 43 54 176 153 65 68
Ta 0.13 0.094 0.065 0.25 0.46
Tb 0.71 0.58 0.55 2.0 0.85
U 0.25 0.26 0.27 1.18 0.32
Cs 0.35 0.19 0.55 2.2 0.36
Sb 0.08 0.08 0.12 0.47 0.10
As 1.5 1.9 6.9 0.9
La 5.7 5.8 5.4 3.8 6.4 3.7 11.8 12.5 8.6 9.2
Ce 12.8 15 12.9 10 14 6.9 26.1 28 16.2 19.5
Nd 10 7 12 19 12
Sm 2.8 2.8 2.0 4.6 2.8
Eu 0.95 0.95 0.93 0.70 1.30 0.70 1.14 1.25 0.97 1.00
Gd
Tb 0.51 0.47 0.40 0.84 0.45
Dy 3.2 2.5 4.4 5.8 3.1
Er 2.1 1.8 2.7 3.8 2.0
Yb 2.2 1.90 1.9 1.60 2.45 2.1 4.3 3.90 1.7 1.60
Lu
Sc 37.4 35.4 37.7 14.1 32.9
V 368 368 255 35 231
Cr 40 42 58 256 943
Co 40.2 39.1 24 37.4 9.3 39.2
Ni 33 33 24 75 0.9 324
Y 20 17 33 38 18.5
Nb 1.2 1.3 1.3 3.4 6.3
(continued)
TABLE /I
( Continued)
Location FUT FUT FUT FUT FUT

Sample 21M7 20M4 21MI 16MI 17M3
Type d b e+d e e
Si0 2 50.20 50.50 50.75 53.45 54.10

Ti0 2 0.50 0.86 0.64 0.82 0.68
AIP3 13.16 18.30 14.17 14.72 13.97
Fe 20 3 1.44 1.40 1.45 1.44 1.34
FeO 7.37 7.14 7.39 7.37 6.85
MnO 0.16 0.16 0.16 0.18 0.17
MgO 10.66 4.52 8.73 4.57 5.18
CaO 11.58 11.48 11.70 6.71 7.67
Nap 1.73 2.57 1.92 3.41 3.19
Kp 0.32 0.46 0.96 1.84 1.64
PP5 0.08 0.10 0.30 0.30 0.30
H 2O+ 1.13 1.59 0.50 1.99 1.86
HP- 0.43 0.22 0.53 2.23 1.34
Total 98.76 99.30 99.20 99.03 98.29
Rb 4.8 5.8 13.7 19.4 28.8
Sr 321 336 618 547 437
Ba 52.7 110 157 339 277
Hf 1.0 1.2 2.1 3.1 2.5
Zr 28 30 52 37 87 84 135 109 86 87
Ta 0.043 0.021 0.091 0.28 0.22
Th 0.12 0.29 1.09 2.2 1.76
U 0.09 0.13 0.80 0.74 0.79
Cs 0.25 0.17 0.24 0.48 0.84
Sb 0.06 0.04 0.04 0.14 0.11
As 0.6 0.6 1.1 1.9
La 1.4 1.7 2.5 3.3 16.7 17.5 16.5 16 12.7 13.4
Ce 3.9 4 6.1 9 44.3 45 28.0 34 22.8 30
Nd 3 8 32 20 17
Sm 1.1 2.3 6.7 4.4 4.1
Eu 0.48 0.50 0.92 0.90 1.89 2.00 1.3 1.40 1.12 1.20
Gd
Tb 0.29 0.44 0.65 0.69 0.61
Dy 1.9 2.9 3.6 4.2 3.8
Er 1.4 2.0 2.2 2.8 2.6
Yb 1.6 1.35 1.8 1.75 1.58 1.80 3.1 2.7 2.6 2.25
Lu
Sc 39.7 32.9 41.3 25.5 28.4
V 216 279 361 225 195
Cr 574 22 400 59 139
Co 47.0 30 39.9 28.1 29.3
Ni 146 32 100 30 46
Y 14 19 22 28 25
Nb 1.5 1.0 2.3 5.2 3.4
(continued)
TABLE /I
( Continued)
Location FUT FUT FUT ANA ANA ANA ANA ANA

Sample 20M6 20M 1 20M3 4311 4312 4313 4314 4315
Type c c c h h h h h
Si0 2 54.70 55.10 55.70 60.10 59.40 60.35 59.45 58.70

Ti0 2 1.09 1.23 0.95 0.85 0.89 0.87 0.83 0.83
Al 20 3 19.46 17.60 17.30 14.35 15.16 14.70 13.95 14.09
Fez03 1.09 1.17 1.13 9.70 9.41 9.17 9.85 9.78
FeO 5.58 5.95 5.77
MnO 0.13 0.14 0.14 0.20 0.19 0.17 0.21 0.21
MgO 2.98 4.79 4.56 2.04 1.92 1.93 1.99 1.99
CaO 8.68 8.32 8.30 5.11 5.10 5.14 5.10 5.04
NazO 3.79 3.87 3.38 4.23 4.73 4.47 4.17 4.34
KzO 0.80 0.56 0.79 2.28 2.44 2.35 2.37 2.04
PzO s 0.15 0.30 0.15 0.42 0.45 0.42 0.41 0.43
HzO+ 0.10 0.09 0.50 0.87 0.15 0.07 1.47 1.62
H20- 0.08 0.06 0.14
Total 98.63 99.18 98.81 100.15 99.84 99.64 99.80 99.07
Source INA ICP AAlICP INA ICP ICP ICP ICP ICP ICP
Rb 10.1 7 11.6 25 30 42 30 24
Sr 329 249 363 408 420 413 398 400
Ba 147 68 145 270 284 270 273 257
Hf 24 22
Zr 92 90 142 86 86 107 112 108 102 105
Ta 0.13 0.13
Th 0.71 0.55
U 0.29 0.30
Cs 0.32 0.38
Sb 0.06 0.10
As 1.3 1.7
La 5.9 6.7 7.5 5.5 6.20 10.85 12.45 11.35 12.10 11.35
Ce 15.0 17- 21 16.7 17
Nd 13 15 12 20 21 20 20 20
Sm 3.6 3.5
Eu 1.21 1.30 1.30 1.2 1.15 1.55 1.65 1.50 1.70 1.50
Od
Th 0.64 0.62
Dy 4.2 4.8 3.9 5.5 6.1 5.4 5.7 5.5
Er 2.7 3.4 2.7 3.8 4.0 3.6 3.7 3.7
Yb 3.1 2.60 3.10 2.7 2.55 3.55 3.80 3.30 3.50 3.50
Lu
Sc 22.0 29.9 19.0 19.0 19.5 19.0 18.5
V 206 240 218 150 150 160 157 150
Cr 15 60 89 2 3 5 3 3
Co 18 23 22.8 20 19 19 20 19
Ni 16 26 50 4 5 5 5 5
Y 28 32 27 36 39 36 36 36
Nb 2.1 2.6 2.3 1.75 1.95 1.95 1.45 0.85
(continued)
TABLE /I
(Continued)
Location ANA ANA ANA ANA ANA ANA ANA ANA ANA
Sample 4316 4318 43110 43112 4421 4422 4423 4424 4531
Type h h h h b b b c h
Si02 58.65 58.60 58.60 58.85 48.60 48.80 48.80 51.00 60.15
Ti02 0.83 0.82 0.81 0.82 0.70 0.70 0.71 1.12 0.76
AI 20 3 13.95 13.86 13.90 14.03 17.10 17.05 17.20 14.50 14.20
Fe 20 3 9.67 9.60 9.65 9.70 11.70 11.71 11.90 14.55 8.79
FeO
MnO 0.21 0.21 0.21 0.21 0.19 0.19 0.20 0.24 0.20
MgO 1.94 1.95 1.93 1.95 7.13 6.90 6.98 5.22 1.90
CaO 5.03 4.96 5.00 4.98 11.40 11.35 11.52 9.85 4.73
Nap 4.30 4.40 4.33 4.27 2.45 2.39 2.43 2.80 4.36
Kp 2.15 2.10 2.05 2.17 0.23 0.21 0.21 0.32 2.34
PPs 0.40 0.41 0.40 0.40 0.12 0.12 0.12 0.17 0.41
H2O+ 1.95 3.40 2.73 2.30 -0.27 -0.35 -0.31 -0.20 1.91
H2O-
Total 99.08 100.31 99.61 99.68 99.35 99.07 99.76 99.57 99.75
Source ICP ICP ICP ICP ICP ICP ICP ICP ICP
Rb 30 32 31 45 4 4 4 4 31
Sr 397 390 396 398 390 410 398 402 382
Ba 260 260 260 264 41 42 41 59 273
Hf
Zr 104 102 102 102 22 23 23 34 116
Ta
Th
U
Cs
Sb
As
La 10.85 10.70 10.70 11.25 2.45 2.40 2.45 3.70 11.95
Ce
Nd 19 20 19 20 6 6 6 9 20
Sm
Eu 1.45 1.55 1.45 1.35 0.75 0.75 0.65 1.10 1.45
Gd
Tb
Dy 5.5 5.4 5.5 5.5 2.4 2.4 3.0 3.4 5.3
Er 3.7 3.5 3.6 3.5 1.4 1.5 1.5 2.3 3.7
Yb 3.45 3.45 3.38 3.43 1.43 1.45 1.45 2.08 3.45
Lu
Sc 18.5 18.5 18.0 18.5 38.0 38.0 37.0 45.0 17.0
V 153 150 150 153 371 378 375 520 137
Cr 3 4 3 5 73 63 63 30 9
Co 20 19 18 19 41 40 40 38 18
Ni 5 5 5 6 54 49 50 20 8
Y 36 35 35 37 15 15 15 22 36
Nb 1.45 1.5 1.3 1.7 0.6 0.5 0.5 0.65 1.5
(continued)
TABLE 1/
(Continued)
Location ANA ANA ANA ANA

Sample 4532 4533 4534 4641
Type h h h e
Si02 60.20 60.40 60.10 48.65
Ti02 0.76 0.76 0.76 0.77
AIP3 14.15 14.15 14.05 14.15
Fe 20 3 8.65 8.64 8.75 9.62
FeO
MnO 0.19 0.19 0.21 0.13
MgO 1.87 1.88 1.86 5.72
CaO 4.73 4.64 4.67 12.50
Nap 4.54 4.44 4.42 2.11
K20 2.32 2.31 2.32 2.23
pps 0.41 0.40 0.40 0.45
H2O+ 2.00 1.75 1.89 3.33
H2O-
Total 99.82 99.56 99.43 99.66
Source ICP ICP ICP ICP
Rb 32 31 32 51
Sr 379 373 375 665
Ba 269 274 272 165
Hf
Zr 115 117 115 73
Ta
Th
U
Cs
Sb
As
La 12.30 12.45 12.80 10.75
Ce
Nd 21 21 22 17
Sm
Eu 1.45 1.50 1.55 1.25
Gd
Tb
Dy 5.6 5.5 5.6 3.2
Er 3.8 3.9 3.5 1.8
Yb 3.50 3.50 3.48 1.70
Lu
Sc 17.0 17.0 16.5 35.0
V 136 136 138 310
Cr 9 II 15 264
Co 18 17 17 36
Ni 8 9 13 76
Y 37 37 37 18
Nb 1.6 1.9 1.25 1.05
TABLE 11/
KI Ar Ages of Volcanic Samplesa
Sample Age (Ma) Type Sample Age (Ma) Type
lean-Charcot Troughs Coriolis Troughs

Vanikoro area (VAN) Efate area (EFA)
7M2 12.4 ± 0.9 a 27M12 3.5 ± 0.3 d
3M4 1.1 ± 0.2 f 29M6 3.2 ± 0.2 e+d
5Ml 2.9 ± 0.4 f 31M2 1.5 ± 0.2 d
6Ml 2.6 ± 0.5 c 31Ml 1.4 ± 0.2 d
5M4 3.9 ± 0.6 f 30M2 1.4 ± 0.2 d
3M3 1.8 ± 0.3 f 30M 1 1.1 ± 0.2 d
7M4 2.3 ± 0.2 e+c 27M 17 3.4 ± 0.2
3Ml 1.8 ± 0.1 g 27Ml 2.2 ± 0.1
3M2 1.5 ± 0.4 g 27M4 2.4 ± 0.1
IM9 1.5 ± 0.1 g 29M3 3.0 ± 0.2
IM8 1.1 ± 0.2 g 28Ml 2.2 ± 0.2
IM5 <0.3 g 26M6 0.4 ± 0.05 h
1M3 1.1 ± 0.2 g 26M7 0.5 ± 0.1 h
3152 <0.3 g
Tikopia area (TIK) 25M2 4.1 ± 0.3 c
3981X 3.7 ± 0.2 e 25M4 4.0 ± 0.6 c
415X <0.3 g 24M6 3.6 ± 0.2 b
24M4 4.1 ± 0.2 d
Vot Tande area (VOT)
24M3 2.7 ± 0.1 h
IIMl 4.9 ± 0.2 e+c
10M 1 2.8 ± 0.1 c Futuna area (FUT)
10M2 2.7 ± 0.1 c 19M1 2.6 ± 0.2 c+d
12Ml 2.8 ± 0.1 d 21M7 6.5 ± 0.5 d
11M2 4.8 ± 0.2 e+c 21Ml 6.1 ± 0.3 e+d
16Ml 5.2 ± 0.3 e
Hazel Holme area (HAZ)
17M3 6.1 ± 0.3 e
14M2 5.2 ± 0.8 a
20M6 0.7 ± 0.2 c
14Ml 5.5 ± 0.4 a
20M 1 0.7 ± 0.3 c
15M6 4.1 ± 0.2 b
15M12 3.5 ± 0.3
"See Monjaret (1989) and Monjaret et al. (1991) for analytical procedure.
Footnote to Table II.

"Petrological types a to i correspond to mid-ocean ridge basalts (MORB) (a). IAT (island-arc tholeiites) (b, c, d), calc-alkaline
basalts and andesites (CAB) (e), backarc basin basalts (BABB) (f), high-Nallow-K dacites (g), high-K dacites (h), and hyper-K
dacites (i), respectively (see text for discussion).
= =
Analytical methods: AA atomic absorption (J. Cotten, UBO). XRF X-ray fluorescence (R. C. Price, La Trobe University;
S. Nakada, Kyushu University). INA = instrumental neutron activation O. L. Joron, Pierre-Siie; A. Fujinawa, Ibaraki University,
for 3321, 3152, 4152, 415X, 3981X, 3982, 3983, 4294, 4295, CYII, CY31, CY34, CY36; see Nakada etal., 1994). ICP = inductively
coupled plasma emission spectrometry (ICP-ES, J. Cotten, UBO). Note: For 3151,3154,3981,4150,4151,4153, and 4155, Nh and
REE by ICP O. Cotten, UBO); for 3MI, IOMI, 10M2, 27MI, and 24M3, Cr by AA O. Cotten, UBO); for all analyses, V by AA (J.
Cotten, UBO) or XRF (S. Nakada, Kyushu University, for the 13 analyses listed in INA).
Analytical procedure for inductively coupled plasma emission spectrometry (/CP-ES) and atomic absorption spectrometry
(AA) analyses (J. Cotten, URO): Rock powders were digested in closed vessels with 4 ml of a concentrated hydrofluoric/nitric acid
mixture. H3B03 was then ad~ed to dissolve the precipitated fluorides and to neutralize the excess HE International standards OBI,
JB2, BEN, Mica-Fe, GSN, ACE) were used for calibration.
Major elements except P20 5: AA analyses with a relative standard deviation (RSO) close to 2%. PP5: colorimetry with a standard
deviation of 0.05%.
Trace elements: Rb, Sr (AA), limit of detection (LOO): I ppm, RSO: 5%; Ba, V (AA), LOO: 25 ppm; RSO: 10%; Cr, Co, Ni (AA):
LOO: 2 ppm; RSO: 5%; Ce, Nd, Zr (ICP-ES), LOO: 2 ppm; RSO: 5%; Nb, La, Er (ICP-ES), LOO: I ppm; RSO: 5%; Sc, Y, Oy
(ICP-ES), LOO: 0.5 ppm, RSO: 5%; Eu, Yb (ICP-ES), LOO: 0.2 ppm; RSO: 5%.
2.0~
a -20
Ti0 2 a 1\12 0 3
1.8
-
o 8~. a 0•• a 0&
o• •~
0°0 0
__ a 0 °0 o'b 0
a a • d' a
00 ,
oOa~ ., .,~ • .,,5'0-
.. .
• • •• t
,." '.
• 0 0
• a
.. . -
•• a
o !II '. • • •
3 • f-' .. • -1
FeO· '0 MgO
1
a
!II.. a a
-
.0 0 0 •
,
b- a
lIIO
0.. .._
8~o.<1t
. '"
a- 0 o •••
• o(!lJo. ~ a '. -
·
.0 00
oO • • • q.<iJ 0 •
o a
~••• ·~8a
000
...
..,.-
.,~o.
a
.00
o o-clo 0 CaO P2 0 ,
10~·. 0 0-": o· 0 -
III o~a.o~••
0._. . ...
00
0 0 0 - a a
·
.-. .
• t1.Q:j1D.~o •• 0 0 0 o ~ • ~ 0
J ...
- ·0.JD'. •
0111 o~. • • oe 0
.
0 0 0000 •
.
00
a
·::-.
Na20 80 K2 0 6
a
a
· ... .'.
0 a ~ - 5
a
-
".p
8 •
. .
4
"
a 8
. a 0
a .'.
3
00. 0
0 a a a a «JO' 1
45 50 55 60 65 50 55 60 65
SiO 2 SiO 2
FIGURE 5.16. Major elements (%) versus Si0 2 (%) diagrams for volcanic samples from the NHBAT. Open
circles: northern Jean-Charcot troughs (JCT); filled circles: southern Coriolis troughs (CT); asterisks: western
termination of the Hazel Holme Ridge. Data from Table II.
with an average of 74 mW/m2, indicating a rather low heat flow. The highest value (161
mW/m2) was recorded at station 40 in the Tikopia area (12°46.033'S-167°49.504'E; water
depth: 3027 m; Fig. 5.5). This is comparable to the mean of heat flow values on the North
Fiji Basin seafloor (150-250 mW/m2) (KAIYO 89 Cruise Report, 1990).
One of the aims of the SAVANES cruise (1991) was to conduct dives with the Cyana
submersible on possible active hydrothermal sites in the JCT, previously surveyed during
the KAIYO 89 cruise. Diplomatic restrictions imposed by Solomon Islands authorities,
relating to their 200 nautical mile boundary, prevented dives at station 31, and consequently
the KAIYO 89 observation could not be confirmed. However, a living biological commu-
nity was discovered southward, at station 41 in the Tikopia area, near sampling site
~
III
~
::x:
110 Rb Ba Hl
- 1000 II
.. .. 8m • Hf
\5
800 o 0 q.. 0 •
> •
. .. m
.. if--
500 ~
0
.. 400
• 0
08· 0
. • • 0
- •
0 I
0 e •• • 00 0 .0 °0 00 cB-l
~
0 0 • .0
o.
.. ... .-. 80 oood8Q 200 ~
eo ~ ••• ~ ~<f_1I •• oo •• ~'• •
80 0 o 0 0 <:08° _ o··o~
0 o
. . -..... :xl0
900
" '0
8r Th 300
y ~ ~ ~
700 o 0 as> Vi
II-
200
500 ..
300
I:. -.-
• 0 # ,_ c., . . . e
~··o -'e
>0 • 00 ~ • • • 8° ". ~ • 100
~ 0 CO 0 Ii 0° .
o • 0 0° a oft ~ ....
100
.-.... 00 <$>90
8 o 0
> , . '~o<>~. 0 o~~~ ••
.00 cq,
. .... ••• •• 'rod'· .: _........ C!,~ .. ~
OR> 26 1.5
•• 0 ~ 0
Nb Zr o • 0 La •• Ta
200 .
14
..
8 If-- -
12 0 150 1.0
10 0 0 • crlQO
8 B 0
0°08
100 ) ·0 8,~l\o
o 0
00 0 .-. 0
0 ".
• eeC()
~ etal
.°
~
0 0
...-- f-- 00
•
00'
0
- • 00 0.5
50
.,..
o°d'c._ • 0 •• ~ 0° 0 ~ • q,: • • •
, f 0 .-. ..0.%:,..... eo.oeilt~o"" • £fir. tI?> 0. ••. .
.000 ,'\r- •• , f'~ I __ ~_ ~ eO. ~'b_• • • I
45 50 55 60 65 50 55 60 65 45 50 55 60 65 50 55 60 65
SiO 2 SiO 2 SiO 2 SiO 2
FIGURE 5.17. Trace elements (ppm) versus Si0 2 (%) diagrams for volcanic samples from the NHBAT. Same symbols as in Fig. 5.16. Data from Table II.
~
-
0.8
0.6
)(
Th/Yb
0.4
0.2
o 0 0.1 0.2 0.3

Ta/Yb
FIGURE 5.18. ThlYb versus TalYb diagram for volcanic samples from the NHBAT. Data from Table II. JCT
samples: filled triangles (Vanikoro area, VAN) and filled squares (Vot Tande area, VOT). Hazel Holme samples:
asterisks. CT samples: open crosses (Efate area, EFA), oblique open crosses (Erromango area, ERR) and oblique
filled crosses (Futuna area, FUT). An arbitrary correlation line for each area is indicated. The extended diagram is
shown in inset. JCT samples: filled circles; Hazel Holme samples: asterisks; CT samples: open circles. ARC and
MORB fields are indicated as arbitrary lines.
CY36 (Table I; Fig. 5.5). On the summit of the eastern seamount of station 41
(12°32.6'S-167°47.3'E), at a depth of 1425 m, the Cyana submersible team discovered and
videotaped a flourishing colony of algae, sponges, corals, galatheas, and shrimp resting on a
50 x 50 m peak of massive to blocky lava. Shimmering water was not observed on the spot
during the dive nor on the video records, but a local increase in seawater temperature
(3.27°C instead of 2.52°C in the surroundings) might suggest a hydrothermal environment
(SAVANES 91-92 Cruise Report, 1992). However, chemical analyses of seawater sampled
on this site did not confirm this hypothesis (D. Grimaud, personal communication).
3.4.2. Ferromanganese Crusts

The study of ferromanganese crusts in island arcs gives good environmental criteria to
distinguish hydrothermal impact from sedimentary processes.
During the RIV Jean Charcot SEAPSO 2 cruise, 18 sites bearing ferromanganese
crusts were dredged in the NHBAT, in water depths ranging from 500 to 3000 m, along fault
scarps or volcanic cones (Fig. 5.20; Gerard et ai., 1987; Gerard, 1993). The crusts were
dredged in five Sea Beam-surveyed areas, two in the northern JCT, Vanikoro (VAN) and Vot
Tande (VOT), and three in the southern CT, Efate (EFA), Erromango (ERR), and Futuna
(FUT). The more complex ferromanganese crusts are essentially located on volcanic cones
in the Vanikoro area and along the eastern faulted border of the Futuna Trough. Most of the
crusts show coatings 0.1 to 1 cm thick on volcaniclastic rocks. However, the complex crusts
from dredges SPS2D4, SPS2D5 (Vanikoro area) and SPS2D19 (Futuna area) cover and
impregnate the volcano-sedimentary deposits up to 10 cm thick or show massive figures.
The main oxyhydroxide manganiferous phases present in these crusts are vernadite (a
low-crystalline manganese oxide mineral), todorokite, and buserite. Birnessite is less
common. The mineralogical structure of todorokite, buserite, and birnessite refers to the
evolution of lO-A and 7-A manganates (Person, 1980; Usui et al., 1989). Buserite is an
instable lO-A manganate which transforms to a 7-A manganate upon dehydration; todoro-
kite is a stable lO-A manganate, and birnessite is a stable 7-A manganate. These minerals
are good criteria to distinguish hydrothermal and hydrogenous processes. Based on micro-
structural and mineralogical data, two major genetical types of ferromanganese encrusta-
tions are found in the NHBAT, a thalassic type and a hydrothermal type (Gerard, 1993).
In the thalassic type, the common crusts consist of dMn0 2 (vernadite) and amorphous
FeOOH,xHp. Samples from SPS2D4 and SPS2D5 dredges (Vanikoro area; Fig. 5.20)
show such coatings on volcaniclastic rocks which are interpenetrated by buserite with
dendritic structures (Fig. 5.21). These crusts of hydrogenous precipitations display botry-
oidal microstructures associated with bacteriomorph occurrences (Fig 5.21), and are char-
acterized by simultaneous precipitation of Fe and Mn (Table IV).
The hydrothermal type is characterized by stable todorokite (which resists transforma-
tion upon dehydration), celadonite (Fe-phyllosilicate), amorphous Fe hydroxide (goethite),
or birnessite. Paleohydrothermalism is characterized by a Fe-Mn segregation on a macro-
scale, with patches of todorokite and celadonite, and amorphous Fe-phase. This todorokite
displays a well-crystallized microstructure, namely spheres of todorokite with lamellar
shape (Fig. 5.22). Celadonite may be the result of the evolution of hydrothermal, "meta-
stable" nontronite (Alt, 1988; Weaver, 1989); it also may result from the slow evolution of
biogenic silica and hydrothermal iron or from direct precipitation associated with low
hydrothermal activity (Odin and Desprairies, 1988). Birnessite crusts similar to those of the
vents of Teahitia submarine volcano in the central Pacific (Hoffert et al., 1987) are rarer.
They are attributed to a more recent hydrothermal phase. A latter stage of hydrothermal
activity is also noticeable. Thin cracks of micritic calcite and phillipsite cut the Mn deposits
in some todorokite crusts. They do not contain any biogenic fraction, in contrast to the
major and older cracks. These crystallizations in the thin cracks are attributed to a last phase
of hydrothermal activity.
Chemically, these two types of ferromanganese encrustations can also be distin-
guished (Table IV): high MnlFe ratios and low Co+ Ni +Cu contents characterize the
hydrothermal encrustations, whereas low MnlFe ratios and "higher" Co+ Ni +Cu contents
characterize the thalassic (hydrogenous) coatings and impregnations. Based on correlation
values between major, trace, and rare-earth elements (REE) within the two types of crusts
(Gerard, 1993), the principal differentiations are as follows:
1. Si or Fe and REE are positively correlated in the hydrothermal (todorokite) crusts,
which is attributed to the adsorption of REE on clays (celadonite), and negatively correlated
in the thalassic crusts.
2. P and REE are positively correlated in the thalassic (hydrogenous) group; no
significative correlation appears for P in the hydrothermal (todorokite) group. The behavior
of REE is mostly attributed to oxidation processes, the todorokite phase being a relatively
oxidant phase. The two types of crusts (thalassic and hydrothermal) are therefore essen-
tially controlled by redox conditions. Other determining factors are the low mobility of Al
and Ti during the thalassic alteration phase. The high Siffi ratio characterizes the hydro-
thermal influence.
In summary, the ferromanganese crusts dredged along the NHBAT can be classified
into two types: thalassic and hydrothermal. The common Fe-Mn crust is a vernadite
coating. The more complex crusts are made up of todorokite, buserite, and birnessite,
associated with a goethite-celadonite facies.
HAZEL HOLME
VANIKORO
MORB(a) ~
MORB(a)
,,~
VANIKORO
BABB(f)
FIGURE 5.19. (a,b). N-MORB normalized "spiderdiagrams" and chondrite-normalized REE patterns (insets)
for volcanic samples from the NHBAT. Data from Table II. Normalizing values: Sun and McDonough (1989) for
"spiderdiagrams," Nakamura (1974) for REE. Letters in parentheses (a to i) refer to the different petrological
types of Table II. Note on Fig. 5.J9b the change of vertical scale for the two bottom "spiderdiagrams" (high-K and
hyper-K dacites). See text for discussion.
The most evolved thalassic crusts (vernadite and buserite) were dredged in the
Vanikoro area (northern Jean-Charcot troughs) on a lateral cone of the 12°lO'S central
volcanic complex (05; Fig. 5.20; Table IV).
The hydrothermal crusts were dredged in the Jean-Charcot troughs on a volcanic cone
located at the eastern border of the same area (04; Fig. 5.20; Table IV) and in the southern
Coriolis troughs on the eastern faulted border of the Futuna Trough (019; Fig. 5.20; Table
IV). Three successive stages of hydrothermal activity have been identified, each with
VOTTANDE
IAT(c,d)
FUTUNA ANATOM
IAT (b,c,d) IAT (b,c)
--- h!:_:'",:S I..-Ce Pl'Nd Pm8mEuGdlb Ho ErTftlVbu.
b
.1 Ba Rb K Ta Nb La Ce Sr Nd P Zr Eu Ti Y Oy Yb Ba Rb K Ta Nb La Ce Sr Nd P Zr Eu Ti Y Oy Yb .1
FIGURE 5.19. (Continued)
specific mineralizations. An ancient stage is characterized by todorokite and celadonite; a

more recent stage is characterized by birnessite; a late hydrothermal stage, marked by
micritic calcite and phillipsite in todorokite crusts, is thought to be relatively recent. This
low hydrothermal activity seems essentially limited to the eastern border of the New
Hebrides backarc troughs.
4. DISCUSSION AND CONCLUSIONS
This review of the NHBAT emphasizes a few regional and thematic points, which are
summarized as follows.
FIGURE 5.20. Location of ferromanganese crusts dredged in the Jean-Charcot and Coriolis troughs (Gerard.
1993). Acronym SPS2 stands for SEAPSO 2 cruise. See also Fig. 5.14 for comparison.
FIGURE 5.21. (Left): Macroscopic view of vemadite crusts with diagenetic buserite dendrites on volcanic
sandstone. (Right): SEM view of bacteriomorphs in the vemadite crust.
1. The NHBAT correspond to a discontinuous alignment of tensional structures,

bordering the eastern flank of the New Hebrides island arc. A diffuse horst-and-graben
morphology, partly obscured in some places by recent volcanic complexes, characterizes
the northern JCT. The general orientation of these northern troughs is slightly oblique to the
arc central chain. In contrast, the southern CT show a more mature morphology (i.e., well-
developed and clearly delineated flat-bottomed grabens), which strictly parallel the arc.
2. Several characteristics of these troughs are paradoxical.
(a) Seismic refraction and gravimetry studies (Collot and Malahoff, 1982; Pontoise
et aI. , 1982; Sage and Charvis, 1991) indicate that the NHBAT partly developed on an
ancient oceanic crust similar to that of the NFB . N1300E magnetic lineations recognized
in the NHBAT (JCT and CT) are typical of the magnetic pattern of the oldest part of the
NFB and corroborate this point (Charvis and Pelletier, 1989; Sage and Charvis, 1991).
(b) The respective positions of the northern and southern troughs in the regional
tectonic environment are different (Charvis and Pelletier, 1989). The northern Jean-Charcot
troughs abut the western termination of the extensional Hazel Holme Ridge, from where
they diffusely extend and widen northward. No subaerial volcanic edifice exists on the arc
central chain between Vot Tande and Vanikoro Islands (i.e., opposite the troughs area). On
the other hand, the southern Coriolis troughs strictly lie opposite Efate, Erromango, Tanna,
and Anatom Islands. They terminate south of the latter island, where they merge into the
island arc substratum. Furthermore, they are located well away from the N-S axis of the
North Fiji Basin active spreading ridge (Monzier et aI., 1984b; Maillet et aI., 1989), and,
consequently, any direct influence of this expanding ridge on the petrological evolution of
these troughs can be excluded.
(c) As noted, two ridges of regional scale frame the backarc area; the d'Entrecasteaux
zone on the subducting plate, and the Hazel Holme Ridge on the North Fiji Basin. Their
tectonomagmatic influences on the formation of the NHBAT still remain to be deciphered.
Two remarks can be made. First, the collision of the d'Entrecasteaux zone with the New
Hebrides arc started around 4-3 Ma (Macfarlane et aI., 1988); second, the Hazel Holme
Ridge acts as an extensional structure in its westernmost termination (Pelletier et aI.,
1993a), namely, at the southern tip of the Jean-Charcot troughs. It is then possible that the
structural and petrological differences observed between the northern and southern troughs
may be due, at least in part, to the influences of these two ridges.
TABLE IV
Major Element Contents (wt%), Minor and Rare Earth
Element (REE) Contents (ppm) of Some 1Ypical
Hydrothermal and Thalassic Ferromanganese Crusts
Hydrothermal crusts Thalassic crusts
Samples D409 D1903 D502X D502Y 011206
Fe (wt%) 0.27 2.36 19.17 9.89 18.93

Mn 38.03 38.57 11.33 14.71 10.16
Si 1.31 4.69 6.8 7.8 4.16
AI 0.36 1.7 2.58 3.45 2.75
Mg 3.02 2.27 1.36 1.73 1.08
Ca 5.49 1.89 2.41 4.8 2.87
Na 1.37 2.16 1.73 1.7 1.79
K 0.32 0.68 0.56 0.51 0.5
Ti tr 0.1 0.89 0.58 1.03
p 0.03 0.04 0.15 0.08 0.16
Co (ppm) 29 51 1238 499 1433

Cr 1294 1227 231 428 160
Cu 22 45 462 199 563
Nb 24 14 44 31 46
Ni 227 157 1829 731 1352
Zn 44 40 695 563 555
MnlFe 140.85 16.34 0.59 1.49 0.54
MnfTi 3803 385.7 12.73 25.36 9.86
Fe/fi 27 23.6 21.54 17.05 18.38
CofZn 0.66 1.28 1.78 0.89 2.58
Co+Ni+Cu 278 253 3529 1429 3348
Y (ppm) 16.5 7.7 162.1 82.4 185.7

La 2.5 4.1 140.5 64 171.2
Ce 7.1 11.8 214.1 107.3 286
Nd 2 4.5 106.8 51 132.5
Sm 0.8 1.7 24.6 11.9 31.3
Eu 0.2 0.3 6.4 3.1 7.8
Od 0.9 1 27.9 13.5 33.9
Dy 5.8 9.3 28.5 15.7 35.1
Er 1.1 0.9 16 8.1 19.1
Yb 0.8 0.7 15.7 7.9 18.9
Lu 0.4 0.4 2.2 1.2 2.8
SUMREE 21.48 34.76 582.76 283.7 738.48

La/Yb 3.2 5.9 9 8.1 9.1
CelLa 2.9 2.9 1.5 1.7 1.7
LalSm 3.15 2.46 5.72 5.37 5.47
(d) Using seismological arguments, Charvis and Pelletier (1989) proposed to link the
formation of the NHBAT to a general NE-SW extensional stress regime affecting the whole
western part of the North Fiji Basin as well as its borders (i.e., the Hazel Holme Ridge and
the northern and southern NHBAT). This interpretation is contrary to the view of Collot
et al. (1985), who argued that the aftereffects of the collision-subduction of the d'Entre-
casteaux zone is the main factor responsible for the formation of the NHBAT. Thus, there is
511----
FIGURE 5.22. Todorokite facies. SEM views.
still debate concerning the respective influence of compressional and tensional stresses on
the formation of the NHBAT.
3. Petrologically, no pronounced difference exists between the recent volcanic prod-
ucts of the New Hebrides central chain and the volcanics which crop out on the floor or on
the faulted edges of the NHBAT. A ubiquitous orogenic, arc-related geochemistry prevails
within the whole backarc area. Yet in the very northern part of the northern Jean-Charcot
troughs, the presence of some BABB indicates an aborted tendency toward oceanic
spreading, effective between 3.9 and 1.1 Ma. These unusual volcanics seem to be spatially
restricted to a central volcanic complex in the northern Vanikoro area, where they neighbor
coevallow-K, high-Na dacites (Nakada et al., 1994). Manned-submersible surveys, water
chemistry analyses, and heat flow measurements do not indicate any widespread modem
hydrothermal activity. Yet some ferromanganese crusts coating volcanic and volcano-
sedimentary formations result from a recent low hydrothermal activity, which is mainly
limited to the eastern faulted border of the NHBAT.
4. The date of initiation of the formation of the NHBAT still remains imprecise, in
spite of reliable KlAr ages (Table III; Monjaret et al., 1991) and micropaleontological
determinations (Gerard, 1993). There is a consensus that the age and nature of the volcanic
and volcano-sedimentary formations sampled in the troughs or on their flanks are linked
to the historical evolution of the New Hebrides central chain. However, Monjaret et al.
(1991) argued that the formation of the NHBAT progressed from south to north, through
successive volcano-tectonic phases. The Coriolis troughs developed first (the Futuna
Trough around 6.5-6.1 Ma; the Erromango Trough around 4.1 Ma; the Efate Trough around
3.5 Ma) with the formation of the Jean-Charcot troughs (Vot Tande area, 2.7 Ma; Vanikoro
area, 2.3 Ma) being more recent. Collision of the d'Entrecasteaux zone with the arc
commenced around 4-3 Ma. This last tectonic phase affected the entire backarc region. In
contrast, Recy et al. (1990) observed that arc tholeiitic eruptions on the eastern scarps of the
NHBAT mostly ceased between 2.8 and 2.3 Ma. Accordingly, they argued that only one
major tectonic phase was responsible for the formation of the NHBAT, between those two
dates.
5. The NHBAT can be compared with the series of backarc structures and marginal
basins from the western Pacific. If these are collectively considered to represent various
stages of an evolving process leading to oceanic spreading through crustal extension and
rifting, the NHBAT stand at the very least evolved point; that is, they are essentially
characterized by crustal extension. Rifting without spreading occurs in the Okinawa

Trough (Sibuet et aI., 1987; Chapter 9 this volume) and in the Bonin Trough (Sumisu and
Torishima rifts) (Leg 126 Scientific Drilling Party, 1989; Leg 126 Shipboard Scientific
Party, 1989; Taylor et al., 1990, 1991, 1992; Taylor, 1992). Oceanic spreading occurs in the
Lau Basin (Hawkins and Melchior, 1985; Hawkins et aI., 1990; Chapter 3 this volume; Leg
135 Scientific Party, 1992) and the North Fiji Basin (Auzende et aI, 1990; Chapter 4 this
volume).
In the NHBAT, arc lavas overwhelmingly predominate on BABB, as they do in the
Okinawa Trough (Ishizuka et al., 1990). In contrast, rifting in the Bonin Trough is accom-
panied by a more pronounced BABB signature (Ikeda and Yuasa, 1989; Fryer et al., 1990;
Hochstaedter et aI., 1990a,b; Tatsumi et al., 1992; Taylor, 1992). Predominating BABB and
MORB petrology is found in the Lau and North Fiji basins, where arc influences tend to
diminish (Frenzel et al., 1990; Price et al., 1990; von Stackelberg and von Rad, 1990;
Johnson and Sinton, 1990; Eissen et aI., 1991; 1994; Sigurdsson et al. 1993).
In conclusion, the volcanic-tectonic evolution of the New Hebrides backarc troughs
results primarily from the concomitant effects of nearby subduction (along the New
Hebrides subduction zone) and spreading (in the central North Fiji Basin), and secondarily
from the aftereffects of collision (between the d'Entrecasteaux zone and the New Hebrides
island arc). The long and complex history of the New Hebrides island arc, the maturity of
the North Fiji Basin and its nonrigid plate behavior (Jarvis et aI., 1993) are thought to
account for the progressive formation of such extensional structures.
Acknowledgments
Thanks are due to captains and crews of research ships and to chief scientists and
scientific parties of numerous cruises, who helped, all together, collect data used in this
chapter. Thanks also to Jean-Philippe Eissen (ORSTOM Brest), whose competence in
computer drafting was appreciated, and to Kevin Speer (CNRS, Brest), who read an early
draft of this paper. A. J. Crawford (University of Tasmania, Hobart) and an anonymous
reviewer helped to improve this chapter.
REFERENCES
Alt, J. C. 1988. Hydrothennal oxide and nontronite deposits on seamounts in the eastern Pacific, Mar. Geol.
81:227-239.
Auzende, J. M., Honza, H., and Scientific Party. 1990. Active spreading and hydrothennalism in North Fiji Basin
(SW Pacific): Results of Japanese French cruise Kaiyo 87, Mar. Geophys. Res. 12:269-283.
Auzende, J. M., and Urabe, T. 1994. The STARMER French-Japanese Joint Project, 1987-1992, Mar. Geol.
116:1-3.
Bellon, H., Marcelot, G., Lefevre, C., and Maillet, P. 1984. Le volcanisme de l'ile d'Erromango (Republique de
Vanuatu): Calendrier de l'activite (donnees 40K-40Ar), C. R. Acad. Sci. Paris 299(11):257-262.
Briqueu, L., and Lancelot, J. R. 1983. Sr isotopes and K, Rb, Sr balance in sediments and igneous rocks from the
subducted plate of the Vanuatu (New Hebrides) active margin, Geochim. Cosmochim. Acta 47:191-200.
Carney, J. N., Macfarlane, A., and Mallick, D. I. J. 1985. The Vanuatu island arc: an outline of the stratigraphy,
structure and petrology, in The Ocean Basins and Margins, Vol. 7A: The Pacific Ocean (A. E. M. Nairn, F. G.
Stehli, and S. Uyeda, eds.), pp. 683-718, Plenum Press, New York.
Charvis, P., and Pelletier, B. 1989. The northern New Hebrides backarc troughs: History and relation with the
North Fiji basin, Tectonophysics 170:259-277.
Collot, J. Y., Daniel J., and Burne, R. V. 1985. Recent tectonics associated with the subduction/collision of the
d'Entrecasteaux zone in the central New Hebrides, Tectonophysics 112:325-356.
Collot, 1. y, and Fisher, M. A 1988. Crustal structure, from gravity data, of a collision zone in the central New
Hebrides island arc, in Geology and Offshore Resources of Pacific Island Arcs-Vanuatu Region (H. G.
Greene and E L. Wong, eds.), Earth Science Ser., Vol. 8, pp.125-139, Circum-Pacific Council for Energy and
Mineral Resources, Houston, TX.
Collot, J. Y., Greene, H. G., Stokking, L., et I'equipe du Leg 134. 1991. Resultats preliminaires du Leg 134 de
l'Ocean Drilling Program dans la zone de collision entre I'arc insulaire des Nouvelles-Hebrides et la Zone
d'Entrecasteaux, C. R. Acad. Sci. Paris 313(11):539-546.
Collot, J. Y., Greene, H. G., Stokking, L. B., et al. 1992a. Proc. ODP, Init. Repts., 134, Ocean Drilling Program,
Collot,1. Y., Lallemand, S., Pelletier, B., Eissen, J. P., Gla~on, G., Fisher, M. A., Greene, H. G., Boulin, J., Daniel,
J., and Monzier, M. 1992b. Geology of the d'Entrecasteaux-New Hebrides Arc collision zone: Results from
a deep submersible survey, Tectonophysics 212:213-241.
Collot, 1. y, and Malahoff, A 1982. Anomalies gravimetriques et structure de la zone de subduction des
Nouvelles-Hebrides, in Contribution Ii I 'etude geodynamique du Sud-Ouest Pacifique, pp. 91-109, Travaux
et Documents de I'ORSTOM n0147.
Coulon, c., Maillet, P., and Maury, R. C. 1979. Contribution a I'etude du volcanisme de I'arc des Nouvelles-
Hebrides: Donnees petrologiques sur les laves de l'i1e d'Efate, Bull. Soc. Geol. France 7, 21(5):619-629.
Crawford, A 1., Greene, H. G., and Exon, N. E 1988. Geology, petrology and geochemistry of submarine
volcanoes around Epi Island, New Hebrides island arc, in Geology and Offshore Resources of Pacific Island
Arcs-Vanuatu Region (H. G. Greene andE L. Wong, eds.), Earth Science Ser., Vol. 8, pp. 301-327, Circum-
Daniel, J. 1982. Morphologie et structures superficielles de la partie sud de la zone de subduction des Nouvelles-
Hebrides, in Contribution Ii l'etude geodynamique du Sud-Ouest Pacifique, pp. 39-60, Travaux et Docu-
ments de I'ORSTOM n0147.
Daniel, J., Gerard, M., Mauffret, A, Boulanger, D., Cantin, B., Collot, J. Y., Durand, J., Fisher, M. A, Greene, H.
G., Michaux, P., Pelletier, B., Pezzimenti, A, Renard, Y, Schaming, M., and Tissot, J. D. 1989. Deformation
compressive d'un bassin intra-arc dans un contexte de collision ride-arc: Le bassin d' Aoba, arc des
Nouvelles-Hebrides, C. R. Acad. Sci. Paris 308(11):239-245.
Dubois, 1., Dugas, E, Lapouille, A., and Louat, R. 1978. The troughs at the rear of the New Hebrides island arc:
Possible mechanisms of formation, Can. J. Earth Sci. 15:351-360.
Dugas, E, Carney, J. N., Cassignol, C., Jezek, P. A., and Monzier, M. 1977. Dredged rocks along a cross-section in
the southern New Hebrides island arc and their bearing on the age of the arc, in International Symposium on
Geodynamics in South-West Pacific, Noumea (New Caledonia), 27 August-2 September 1976, pp. 105-116,
Editions Technip, Paris.
Dupont, J., and Herzer, R. H. 1985. Effect of subduction of the Louisville Ridge on the structure and morphology
of the Tonga arc, in Geology and Offshore Resources of Pacific Island Arcs-Tonga Region (D. W. Scholl and
T. L. Vallier, eds.), Earth Science Ser., Vol. 2, pp. 323-332, Circum-Pacific Council for Energy and Mineral
Eissen, J. P., Lefevre, C., Maillet, P., Morvan, G., and Nohara, M.1991. Petrology and geochemistry of the central
North Fiji Basin spreading centre (Southwest Pacific) between 16°S and 22°S, Mar. Geol. 98:201-239.
Eissen,1. P., Nohara, M., Cotten, J., and Hirose, K. 1994. North Fiji Basin basalts and their magma sources. I:
Incompatible element constraints, Mar. Geol. 116:153-178.
Eissen, J. P., Robin, c., and Monzier, M. 1992. Decouverte et interpretation d'ignimbrites basiques a Tanna
(Vanuatu, SO Pacifique), C. R. Acad. Sci. Paris 315(11):1253-1260.
Falvey, D. A., and Greene, H. G. 1988. Origin and evolution of the sedimentary basins of the New Hebrides arc, in
Geology and Offshore Resources of Pacific Island Arcs-Vanuatu Region (H. G. Greene and E L. Wong,
eds.), Earth Science Ser., Vol. 8, pp. 413-442, Circum-Pacific Council for Energy and Mineral Resources,
Houston, TX.
Frenzel, G., Miihe, R., and Stoffers, P. 1990. Petrology of the volcanic rocks from the Lau Basin, Southwest
Pacific, Geol. Jb D 92:395-479.
Fryer, P., Taylor, B., Langmuir, C. H., and Hochstaedter, A 1990. Petrology and geochemistry of lavas from the
Gerard, M. 1993. Bassins d'arc et fosses arriere-arc dans un contexte de collision-subduction: I'arc des Nouvelles-
Hebrides (Vanuatu). Hydrothermalisme, neogeneses, diagenese d'une serie volcanosedimentaire, These de
Doctorat, Universite de Paris-Sud, Orsay, France.
Gerard, M., Person, A., Recy, J., and Dubois, J.1987. Preliminary results of petrological and mineralogical studies
of manganesiferous encrustations dredged over the New Hebrides back arc (Vanuatu), E.U.G.-E.G.S. VII,
Strasbourg, 13-16 April 1987, Terra Cognita 7:2-3.
Gill, J. B. 1981. Orogenic Andesites and Plate Tectonics, Springer-Verlag, Berlin, Heidelberg, New York.
Gorton, M. P. 1974. The geochemistry and geochronology of the New Hebrides, Ph.D. thesis, Australian National
University.
Gorton, M. P. 1977. The geochemistry and origin of Quaternary volcanism in the New Hebrides, Geochim.
Cosmochim. Acta 41:1257-1270.
Greene, H. G., and Johnson, D. P. 1988. Geology of the central basin region of the New Hebrides arc inferred from
single-channel seismic-reflection data, in Geology and Offshore Resources of Pacific Island Arcs-Vanuatu
Region (H. G. Greene and F. L. Wong, eds.), Earth Science Ser., Vol. 8, pp. 177-199, Circum-Pacific Council
for Energy and Mineral Resources, Houston, TX.
Greene, H. G., Macfarlane, A., and Wong, F. L. 1988a. Geology and offshore resources of Vanuatu-Introduction
and summary, in Geology and Offshore Resources of Pacific Island Arcs-Vanuatu Region (H. G. Greene,
and F. L. Wong, eds.), Earth Science Ser., Vol. 8, pp. 1-25, Circum-Pacific Council for Energy and Mineral
Greene, H. G., Macfarlane, A., Johnson, D. P., and Crawford, A. 1. 1988b. Structure and tectonics of the central
New Hebrides arc, in Geology and Offshore Resources of Pacific Island Arcs-Vanuatu Region (H. G.
Greene, and F. L. Wong, eds.), Earth Science Ser., Vol. 8, pp. 377-412, Circum-Pacific Council for Energy
and Mineral Resources, Houston, TX.
Hawkins, J. w., Lonsdale, P. F., Macdougall, J. D., and Volpe, A. M. 1990. Petrology of the axial ridge of the
Hawkins, J. W., and Melchior, J. T. 1985. Petrology of Mariana Trough and Lau Basin basalts, J. Geophys. Res.
90:11,431-11,468.
Hochstaedter, A. G., Gill, 1. B., Kusakabe, M., Newman, S., Pringle, M., Taylor, B., and Fryer, P. 1990a. Volcanism
in the Sumisu Rift. I: Major element, volatile, and stable isotope geochemistry, Earth Planet. Sci. Lett.
100:179-194.
Hochstaedter, A. G., Gill, J. B., and Morris, J. D. 1990b. Volcanism in the Sumisu Rift. II: Subduction and non-
Hoffert, M., Cheminee, J. L., Larque, P., and Person, A. 1987. Depot hydrothermal associe au volcanisme sous-
marin "intraplaque" oceanique. Prelevements effectues avec Cyana sur Ie volcan actif de Teahitia, c. R.
Acad. Sci. Paris 304(11):829-833.
Holmes, M. L., 1988, Seismic refraction measurements in the summit basins of the New Hebrides are, in Geology
and Offshore Resources of Pacific Island Arcs-Vanuatu Region (H. G. Greene, and F. L. Wong, eds.), Earth
Science Ser., Vol. 8, pp. 163-176, Circum-Pacific Council for Energy and Mineral Resources, Houston, TX.
Huchon, Ph., Gracia, E., Ruellan, E., Joshima, M., and Auzende, J. M. 1994. Kinematics of active spreading in the
central North Fiji Basin (Southwest Pacific), Mar. Geol. 116:69-87.
Ikeda, y', and Yuasa, M. 1989. Volcanism in nascent back-arc basins behind the Shichito Ridge and adjacent areas
in the Izu-Ogasawara are, northwest Pacific: Evidence for mixing between E-type MORB and island arc
magmas at the initiation of backarc rifting, Contrib. Mineral. Petrol. 101:377-393.
Isacks, B. L., Cardwell, R. K., Chatelain, J. L., Barazangi, M., Marthelot, J. M., Chinn, D., and Louat, R. 1981.
Seismicity and tectonics of the central New Hebrides island arc, in Earthquake Prediction: An International
Review (D. W. Simpson and P. G. Richards, eds.), Vol. 4, pp. 93-116, American Geophysical Union Maurice
Ewing Series.
Ishizuka, H., Kawanobe, Y., and Sakai, H. 1990. Petrology and geochemistry of volcanic rocks dredged from the
Okinawa Trough, an active back arc basin, Geochem. J. 24:75-92.
Jarvis, P. A., Kroenke, L. W., Price, R. C., and Maillet, P. 1993. GLORIA imagery of sea floor structures in the
northern North Fiji Basin, Geo-Mar. Lett. 13:90-97.
Johnson, D. P., Maillet, P., and Price, R. C. 1993. Regional setting of a complex backarc: New Hebrides Are,
northern Vanuatu-eastern Solomon islands, Geo-Mar. Lett. 13:82-89.
Johnson, K. T. M., and Sinton, J. M. 1990. Petrology, tectonic setting, and the formation of backarc basin basalts in
the North Fiji Basin, Geol. Jb 0 92:517-545.
KAIYO 89 Cruise Report. 1990. (KAIYO 89 cruise in the North Fiji Basin and Vanuatu backarc troughs, 14
December 1989-13 January 1990), STARMER Cruise Report Vol. V, unpublished.
Kroenke, L. W. 1984. Cenozoic tectonic development of the Southwest Pacific, U. N. ESCAP, CCOP/SOPAC
Tech. Bull. 6.
Kroenke, L. W., Smith, R., and Nemoto, K. 1994. Morphology and structure of the seafloor in the northern part of
the North Fiji Basin, in Basin Formation, Ridge Crest Processes, and Metal/ogenesis in the North Fiji Basin
(L. W. Kroenke and 1. V. Eade, eds.), Earth Science Ser., Vol. 12, pp. 15-25, Circum-Pacific Council for
Energy and Mineral Resources, Springer, Heidelberg.
Lallemand, S. E., Malavieille, J., and Calassou, S. 1992. Effects of oceanic ridge subduction on accretionary
wedges: experimental modeling and marine observations, Tectonics 11(6):1301-1313.
Leg 126 Scientific Drilling Party. 1989. ODP Leg 126 drills the Izu-Bonin arc, Geotimes 34(10):36-38.
Leg 126 Shipboard Scientific Party. 1989. Arc volcanism and rifting, Nature 342:18-20.
Leg 135 Scientific Party. 1992. Evolution of backarc basins: ODP Leg 135, Lau Basin, EOS, Trans. AGU 73:22.
Louat, R., Hamburger, M., and Monzier, M. 1988. Shallow and intermediate-depth seismicity in the New Hebrides
arc: constraints on the subduction process, in Geology and Offshore Resources of Pacific Island Arcs-
Vanuatu Region (H. G. Greene and E L. Wong, eds.), Earth Science Ser., Vol. 8, pp. 329-356, Circum-Pacific
Council for Energy and Mineral Resources, Houston, TX.
Louat, R., and Pelletier, B. 1989. Seismotectonics and present-day relative plate motions in the New Hebrides-
North Fiji Basin region, Tectonophysics 167:41-55.
Macfarlane, A., Carney, J. N., Crawford, A. J., and Greene, H. G. 1988. Vanuatu-A review of the onshore
geology, in Geology and Offshore Resources of Pacific IslandArcs- Vanuatu Region (H. G. Greene and E L.
Wong, eds.), Earth Science Ser., Vol. 8, pp. 45-91, Circum-Pacific Council for Energy and Mineral
Maillet, P., Monzier, M., Eissen, J. P., and Louat, R. 1989. Geodynamics of an arc-ridge junction: the case of the
New Hebrides ArclNorth Fiji Basin, Tectonophysics 165:251-268.
Maillet, P., Monzier, M., and Lefevre, C. 1986. Petrology of Matthew and Hunter volcanoes, south New Hebrides
island arc (southwest Pacific), J. Volcanol. Geotherm. Res. 30:1-27.
Maillet, P., Monzier, M., Selo, M., and Storzer, D. 1983. The d'Entrecasteaux zone (southwest Pacific), a
petrological and geochronological reappraisal, Mar. Geol. 53:179-197.
Matsumoto, T., Iwabuchi, Y., and Maillet, P. 1992. Tectonics in the Vanuatu backarc basin as derived from precise
bottom topography, International Geological Congress, Kyoto, Japan, August 1992, Abstracts Volume, p. 36.
Minster,1. B., and Jordan, T. H. 1978. Present day plate motions, J. Geophys. Res. 83:5,331-5,354.
Monjaret, M. C. 1989. Le magmatisme des fosses 11 l'arriere de l'arc des Nouvelles-Hebrides (Vanuatu) (Cam-
pagne SEAPSO 2 du N. O. Jean Charcot). Implications geodynamiques. Chronologie, petrologie, geochimie,
These de Doctorat, Universite de Bretagne Occidentale (UBO), Brest, France.
Monjaret, M. C., Bellon, H., and Maillet, P. 1991. Magmatism of the troughs behind the New Hebrides island arc
(RV Jean-Charcot SEAPSO 2 cruise): K-Ar geochronology and petrology, J. Volcanol. Geoth. Res. 46:
265-280.
Monjaret, M. c., Bellon, H., Maillet, P., and Recy, 1. 1987. Le volcanisme des fosses arriere-arc des Nouvelles-
Hebrides (campagne SEAPSO Leg 2 du N/O Jean-Charcot dans Ie Pacifique Sud-Ouest): Datations K-Ar et
donnees petrologiques preliminaires, C. R. Acad. Sci. Paris 305(11):605-609.
Monzier, M., Collot, J. Y., and Daniel, J. 1984a. Carte bathymetrique des parties centrale et meridionale de l'arc
insulaire des Nouvelles-Hebrides, ORSTOM, Paris.
Monzier, M., Maillet, P., and Dupont, J. 1991. Carte bathymetrique des parties meridionales de l'arc insulaire des
Nouvelles-Hebrides et du bassin Nord-Fidjien, Institut Fran~ais de Recherche Scientifique pour Ie Deve-
loppement en Cooperation (ORSTOM), Paris.
Monzier, M., Maillet, P., Foyo Herrera, J., Louat, R., Missegue, E, and Pontoise, B. 1984b. The termination of the
southern New Hebrides subduction zone (southwestern Pacific), Tectonophysics 101:177-184.
Nakada, S., Maillet, P., Monjaret, M. c., Fujinawa, A., and Urabe, T. 1994. High-Na dacite from the Jean-Charcot
Trough (Vanuatu), Southwest Pacific, Mar. Geol. 116:197-213.
Nakamura, N., 1974, Determination of REE, Ba, Fe, Mg, Na and K in carbonaceous and ordinary chondrites,
Nohara, M, Hirose, K., Eissen, 1. P., Urabe, T., and Joshima, M. 1994. The North Fiji Basin basalts and their
magma sources. II: Sr-Nd isotopic and trace element constraints, Mar. Geol. 116:179-195.
Nojiri, Y., and Ishibashi, J. 1991. Hydrothermal plumes observed in the North Fiji Basin, in STARMER Symposium,
Geology and Biology of the Rift System in the North Fiji and Lau Basins, 7-11 Feb. 1991, Noumea, New
Caledonia, Abstracts Volume, p. 41.
Odin, G. S., and Desprairies, A. 1988. Nature and geological significance of celadonite, in Green Marine Clays (G.
S. Odin, ed.), Developments in Sedimentology 45, Elsevier, Amsterdam.
Pearce, J. A. 1983. The role of sub-continental lithosphere in magma genesis at destructive plate margins, in
Continental Basalts and Mantle Xenoliths (C. 1. Hawkesworth and M. J. Norry, eds.), pp. 230-249,
Nantwich, Shiva.
Pelletier, B., Charvis, P., Daniel, J., Hello, Y., Jamet, E, Louat, R., Nanau, P., and Rigolot, P. 1988. Structure et
lineations magnetiques dans Ie coin Nord-Ouest du bassin Nord-Fidjien: resultats preliminaires de la
campagne Eva 14 (aoftt 1987), C. R. Acad. Sci. Paris 306(11):1247-1254.
Pelletier, B., and Dupont, J. 1990a. Erosion, accretion, extension arriere-arc et longueur du plan de subduction Ie
long de la marge active des Kermadec, Pacifique Sud-Ouest, C. R. Acad. Sci. Paris 310(11):1657-1664.
Pelletier, B., and Dupont, J. 1990b. Effets de la subduction de la ride de Louisville sur I'arc des Tonga-Kermadec,
Oceanol. Acta 10:57-76.
Pelletier, B., Lafoy, Y., and Missegue, E 1993a. Morphostructure and magnetic fabric of the northwestern North
Fiji Basin, Geophys. Res. Lett. 20(12):1151-1154.
Pelletier, B., and Louat, R. 1989. Mouvements relatifs des plaques dans Ie Sud-Ouest Pacifique, C. R. Acad. Sci.
Paris 308(11): 123-130.
Pelletier, B., Missegue, E, Lafoy, Y., Mollard, L., Decourt, R., Dupont, J., Join, Y., Perrier, J., and Recy, J. 1993b.
Extremites nord du bassin Nord-Fidjien et des fosses arriere-arc des Nouvelles-Hebrides: morphostructure et
signature magnetique, C. R. Acad. Sci. Paris 316(11):637-644.
Person, A. 1980. Concretions polymetalliques des sediments de l'ocean Pacifique equatorial-zone nord est: etude
diffractometrique du comportement aux contraintes thermiques de la todorokite, Bull. Soc. Fr. Mineral.
Cristallogr. 103(2).
Picard, C., Monzier, M., Eissen, J. P., and Robin, C. 1995. Concomitant evolution of tectonic environment and
magma geochemistry, Ambrym volcano (Vanuatu-New Hebrides arc), in Volcanism Associated with Exten-
sion at Consuming Plate Margins (J. L. Smellie, ed.), Geological Society Special Publication No. 81, pp.
135-154.
Pontoise, B., Latham, G. V., and Ibrahim, A. B. K. 1982. Sismique refraction: structure de la croftte aux Nouvelles-
Hebrides, in Contribution a l'etude geodynamique du Sud-Ouest Pacifique, pp. 79-90, Travaux et Docu-
ments de l'ORSTOM, n0147.
Price, R. c., and Kroenke, L. W. 1991. Tectonics and magma genesis in the northern North Fiji Basin, Mar. Geol.
98:241-258.
Price, R. C., Johnson, L. E. , and Crawford, A. 1. 1990. Basalts of the North Fiji Basin: the generation of back arc
basin magmas by mixing of depleted and enriched mantle sources, Contrib. Mineral. Petrol. 105:106-121.
Price, R. C., Maillet, P., and Johnson, D. P. 1993. Interpretation of GLORIA side-scan sonar imagery for the
Coriolis troughs of the New Hebrides backarc, Geo-Mar. Lett. 13:71-81.
Puech, J. L., and Reichenfeld, C. 1969. Etudes bathymetriques dans la region des lJes Erromango, Tanna et
Anatom (Nouvelles-Hebrides), C. R. Acad. Sci. Paris 208:1259-1261.
Recy, J., Charvis, P., Ruellan, E., Monjaret, M. c., Gerard, M., Auclair, G., Baldassari, c., Boirat, J. M., Brown,
G. R., Butscher, J., Collot, J. Y., Daniel, J., Louat, R., Monzier, M., and Pontoise, B. 1986. Tectonique et
volcanisme sous-marin It I'arriere de l'arc des Nouvelles-Hebrides (Vanuatu): resultats preliminaires de la
campagne SEAPSO (leg 2) du N/O Jean Charcot, C. R. Acad. Sci. Paris 303(11):685-690.
Recy, 1., Pelletier, B., Charvis, P., Gerard, M., Monjaret, M. C., and Maillet, P. 1990. Structure, age et origine des
fosses arriere-arc des Nouvelles-Hebrides (Sud-Ouest Pacifique), Oceanol. Acta 10:165-182.
Robin, C., Eissen, J. P., and Monzier, M. 1994. Ignimbrites of basaltic andesite and andesite compositions from
Tanna (New Hebrides Arc), Bull. Volcanol. 56:10-22.
Robin, C., Monzier, M., Eissen, J. P., Picard, c., and Camus, G. 1991. Coexistence de lignees HK et MK dans les
pyroclastites associees It la caldera d' Ambrym (Vanuatu-Arc des Nouvelles-Hebrides), C. R. Acad. Sci. Paris
313(11):1425-1432.
Roca, J. L.1978. Contribution It l'etude petrologique et structurale des Nouvelles-Hebrides, These de 3eme cycle,
Universite des Sciences et Techniques du Languedoc, Montpellier, France.
Sage, E, and Charvis, P. 1991. Structure profonde de la transition arc insulaire-bassin marginal dans Ie nord des
Nouvelles-Hebrides (Vanuatu, Pacifique sud-ouest), C. R. Acad. Sci. Paris 313(11):41-48.
Saunders, A. D., and Tarney, J. 1984. Geochemical characteristics of basaltic volcanism within backarc basins, in
Marginal Basins Geology: Volcanism and Associated Sedimentary and Tectonic Processes in Modern and
Ancient Marginal Basins (B. P. Kokelaar and M. E Howells, eds.), pp. 59-76, Blackwell Scientific,
Cambridge, MA.
SAVANES 91-92 Cruise Report. 1992. (SAVANES 91-92 cruise with Cyana submersible in the northern Vanuatu
backarc troughs), 19 December 1991-12 January 1992, STARMER Cruise Report Vol. VIII, unpublished.
Sibuet, J. C., Letouzey, J., Barbier, E, Charvet, J., Foucher, J. P., Hilde, T. W. C., Kimura, M., Ling-Yun, c.,
Marsset, B., Muller, C., and Stephan, J. F. 1987. Backarc extension in the Okinawa trough, J. Geophys. Res.
92:14,041-14,063.
Sigurdsson,I. A., Kamenetsky, V. S., Crawford, A. J., Eggins, S. M., and Zlobin, S. K. 1993. Primitive island are
and oceanic lavas from the Hunter Ridge-Hunter fracture zone. Evidence from glass, olivine and spinel
compositions, Mineral. Petrol. 47:149-169.
Sinton, J. M., Price, R. C., Johnson, K. T. M., Staudigel, H., and Zindler, A. 1994. Petrology and geochemistry of
submarine lavas from the Lau and North Fiji back-are basins, in Basin Formation, Ridge Crest Processes,
and Metallogenesis in the North Fiji Basin (L. W. Kroenke and J. V. Eade, eds.), Earth Science Ser., Vol. 12,
pp. 155-177, Circum-Pacific Council for Energy and Mineral Resources, Springer, Heidelberg.
Sun, S.-s., and McDonough, W. F. 1989. Chemical and isotopic systematics of oceanic basalts: implications for
mantle composition and processes, in Magmatism in the Ocean Basins (A. D. Saunders and M. J. Norry,
eds.), pp. 313-345, Geological Society Special Publication No. 42.
Tatsumi, Y., Murasaki, M., and Nohda, S. 1992. Across-are variation of lava chemistry in the Izu-Bonin are:
identification of subduction components, J. Volcanol. Geotherm. Res. 49:179-190.
Taylor, B. 1992. Rifting and the volcanic-tectonic evolution of the Izu-Bonin-Mariana are, in Proc. ODp, Sci.
Results Vol. 126 (B. Taylor, K. Fujioka et aI., eds.), pp. 627-651, Ocean Drilling Program, College Station,
TX.
Taylor, B., Brown, G., Fryer, P., Gill, J. B., Hochstaedter, A. G., Hotta, H., Langmuir, C. H., Leinen, M.,
Nishimura, A., and Urabe, T. 1990. ALVIN-Sea Beam studies of the Sumisu Rift, Izu-Bonin are, Earth
Taylor, B., K. Fujioka, K., et al. 1992. Proc. ODp, Sci. Results, Vol. 126: College Station, TX (Ocean Drilling
Program).
Taylor, B., Klaus, A., Brown, G. R., and Moore, G. F. 1991. Structural development of Sumisu rift, Izu-Bonin are,
J. Geophys. Res. 96:16,113-16,129.
Thorpe, R. S (Editor). 1982. Andesites, Orogenic Andesites and Related Rocks, Wiley, New York.
Tiffin, D. L. 1993. Tectonic and structural features of the PacificlIndo-Australian plate boundary in the North Fiji-
Lau Basin regions, southwest Pacific, Geo-Mar. Lett. 13:126-131.
Usui, A., Mellin, T. A., Nohara, M., and Yuasa, M. 1989. Structural stability of marine 10 A manganates from the
Ogasawara (Bonin) are: implication for low-temperature hydrothermal activity, Mar. Geol. 86:41-56.
Vallot, 1. 1984. Volcanites dragu~s au large de I'are insulaire des Nouvelles-Hebrides. Implications petrologiques,
These de 3eme cycle, Universite de Paris-Sud, Orsay, France.
Von Stackelberg, U., and von Rad, U. 1990. Geological evolution and hydrothermal activity in the Lau and North
Fiji basins (SONNE cruise SO-35)-a synthesis, Geol. Jb D 92:629-660.
Weaver, C. E. 1989. Clays, Muds and Shales, Developments in Sedimentology, Vol. 44, Elsevier, Amsterdam.
Weissel, J. K., and Karner, G. D. 1989. Flexural uplift of rift flanks due to mechanical unloading of the lithosphere
during extension, J. Geophys. Res. 94:13,919-13,950.
Wilson, M. 1989. Igneous Petrogenesis, A Global Tectonic Approach, Unwin Hyman, London.
6
Geology of the Mariana Trough

Patricia Fryer
ABSTRACT
The Mariana Trough is an actively spreading, crescent-shaped, backarc basin located in the
western Pacific between the active Mariana volcanic island arc and the West Mariana
Ridge, a remnant arc. The geologic evolution of the Mariana Trough varies along strike of
the basin from the initial opening phase in the north to the mature seafloor spreading phase
in the central latitudes. The opening of the basin began with an initial period of stretching
and collapse of the preexisting arc followed by development of ridge! transform structures
along an active volcanic and tectonic zone on the eastern side of the basin. Eventually a true
spreading center developed within the basin as the principal volcanic and tectonic zone
diverged from the active volcanic front. The current along-strike variations in rifting!
spreading history define distinct geographic regions: the northern rifting apex, the central
spreading basin, and the southern platform.
The northern apex region of the Mariana Trough typifies the early rifting stage. The
zone of principal volcanism and tectonic deformation is close to the active volcanic front,
and the western margin is characterized by numerous normal fault blocks downstepping to
the east. Magmatic activity has remained primarily on the eastern side of the basin close to
its present position, although occasional outbreaks of volcanism occur in the western
portion of the basin along boundaries offault blocks. A block-faulted region approximately
60 km wide exists along the western margin of the Mariana Trough reflecting the early stage
of rifting of the basin.
The central spreading region of the Mariana Trough represents the mature spreading
section of the basin. The seafloor morphology is typical of slow spreading, with a series of
north-south trending valleys and ridges similar to abyssal hills offset by transform fault
Valleys. Volcanism is predominantly confined to the spreading center. Graben and horst
structures at the margin of the abyssal hill fabric in the western part of the basin represent
the early phase of opening of the basin. Preexisting structural elements that cut across the
arc system persist throughout both the early rifting and the later spreading stages of the
basin. Chains of submarine volcanoes (volcanic cross-chains) extend into the backarc basin
Patricia Fryer • Hawaii Institute of Geophysics and Planetology. School of Ocean and Earth Sciences and
Technology, University of Hawaii at Manoa. Honolulu. Hawaii 96822.
Backarc Basins: Tectonics and Magmatism, edited by Brian Taylor. Plenum Press, New York, 1995.
237
238 PATRICIA FRYER
along some of these lineaments, pennitting the leakage of arc magmas into the backarc
basin.
The southern basin platfonn is a shallow (less than 3 km), relatively low relief region,
except where it is separated from the active volcanic front and from the remnant arc by
deep, narrow troughs. Perturbations of the typical spreading center morphology occur
where the active spreading zone of the platfonn intersects the volcanic front at about 13°N,
144°E in the southernmost portion of the backarc basin. This position coincides with a
major north-south trending fault that crosscuts the forearc, arc, and backarc basin. It has
influenced the spreading axis and the eastern margin of the trough.
The distribution of volcanism within the basin varies according to the developmental
stages of the spreading center. Volcanism and hydrothennal activity are also influenced by
cross-arc and along-arc defonnation of the Mariana platelet. Most lavas from the backarc
basin spreading ridges show systematic differences from mid-ocean ridge basalt as a
consequence of infusion of the source v>,ith subduction-related constituents. The composi-
tional variations of lavas within the basin indicate complex local tectonic control over
magma genesis and an intricate interplay of mixing from several sources with crystalliza-
tion at various pressures.
1. INTRODUCTION
The Mariana Trough is an actively opening, crescent-shaped, backarc basin that lies at
the eastern edge of the Philippine Sea plate (Fig. 6.1). It is bounded by the West Mariana
Ridge (a remnant arc) and the Mariana island arc (an active volcanic arc) (Fig. 6.2). The
Mariana Trough is one of several backarc basins that have fonned in this position since the
Eocene (Karig, 1971a,b, 1974, 1975). Klein and Kobayashi (1980) and Lewis et al. (1982)
suggested that the Philippine Sea plate, from the Oki-Daito Ridge to the Palau-Kyushu
Ridge and the intervening basins, fonned as a consequence of southward convergence, arc
fonnation, and episodes of backarc rifting. They suggested that this region moved north-
ward during a translation/rotation episode prior to about 21 Ma. From about 30 to 15 Ma the
Parece Vela (Mrozowski and Hayes, 1979) and Shikoku basins (Kobayashi et aI., Chapter
10 this volume) were opening and separating the MarianalIzu-Bonin arcs from the Palau-
Kyushu remnant arc.
The Mariana Trough is widest at about 18°N and narrows northward to 24°N where the
remnant arc and active arc join at the Volcano Islands (Smoot, 1990). Thick volcaniclastic
sediments derived from the active volcanic arc fonn a wedge-shaped blanket over the
eastern side of the trough, and the western side has complex fault-controlled topography
with variable sediment cover (Karig et aI., 1978; Bibee et aI., 1980; Fryer and Hussong,
1981; Baker, 1992). North of 18°N the azimuth of the basin fault scarps lies between 330°
and 345° (21°-23° = 330°; 19°-20° = 345°; 19°-18° = 340°) (Baker, 1992; Martinez et al.,
in press; Smoot, 1990). The zone of rifting intersects the volcanic arc at about 23°N (Fryer,
1986; 1989; Baker et aI., 1989; Baker, 1992), marking the northern tenninus of the basin.
The azimuth of basin fault scarps indicates that the basin is opening in an east-west
direction from 18°N to 12°N (Smoot, 1990; Lange, 1992; Fryer, 1993). The axis of spreading
intersects the volcanic arc at the southern tenninus of the backarc basin at about BON and
may extend into the forearc (Fryer, 1993).
The first attempts to detennine an age for the opening of the Mariana Trough met with
GEOLOGY OF THE MARIANA TROUGH 239
CHINA
PACIFIC
35'
PLATE
30'
FIGURE 6.1. Bathymetry and geologic

features in the Philippines Sea region. Ba-
sins and ridges are outlined by the 4-km
bathymetric contour, except for the Izu-
Bonin arc, West Mariana Ridge and Mariana
are, which are outlined by the 3-km contour.
Barbed lines locate axes of trenches; me-
dium double lines locate active spreading
centers; dashed lines locate relict spreading
centers, single dashed line is a proposed new
plate boundary. Arrow shows convergence
direction between the Pacific plate and the
Philippine Sea plate. /50'
limited success. A search for symmetric spreading anomaly patterns (Karig, 1971a,b;
Anderson, 1975; Karig et al., 1978) was unsuccessful because of the proximity of the region
to the magnetic equator, the nearly north-south strike of the arc, and the high relief of the
seafloor. Later, Bibee et al. (1980) correlated observed and synthetic magnetic anomalies
near 18°N with a spreading half-rate of 1.5 cm/yr for the last 3 m.y., although the fit was not
universally accepted (Eguchi, 1984). Results from the Deep Sea Drilling Project (DSDP)
Leg 60 suggest a symmetric spreading rate of 2.15 cm/yr implying initiation of rifting in the
late Miocene, about 6.5 Ma (Hussong and Uyeda, 1981). Other workers have suggested
initiation of rifting of the Mariana Trough at 3 Ma (Karig, 1975), 8 Ma (Bibee et aI., 1980),
and less than 10 Ma (Eguchi, 1984). Clearly, the age of initiation of rifting in the basin was
not well constrained in these early attempts.
Several models proposed as mechanisms for rifting of the Mariana Trough accounted
for the asymmetry of the basin, its crescent shape, and peculiarities of its morphology,
compared with typical ocean basins. Karig (1971a,b) invoked nonrigid behavior of the
volcanic arc and forearc region (between the trench axis and the volcanic arc) to account for
the shape of the basin and apparent variations in spreading rate along strike. Bracey and
Ogden (1972) suggested that the basin formed when arc segments rifted apart, beginning
in the center of the crescent and migrating northward and southward simultaneously. A
240 PATRICIA FRYER
138"E 140'E 142'E 144'E 148'E 148'E
24'N 24"N
22"N 22'N
2O'N 20'N
18'N 18'N
16'N 16'N
14'N 14'N
12"N 12'N
10'N 10'N
138'E 140'E 142'E 144'E 148'E 148'E
FIGURE 6,2, Bathymetric map of the Mariana system based on the DBDB5 data set of NODC, Contour interval
500 m, Shading darkens with depth,
rigid-plate model with a pole of rotation about a point at 21°Q4'N, 138°0l'E was suggested
by Le Pichon et ai, (1985), Detailed studies of the match of the east and west bounding
margins of the basin (Karig et ai" 1978, Fryer and Hussong, 1981; Stern et aI., 1984) show
that a single pole of rotation cannot exist, Stern et al. (1984) suggested that the basin is
unzipping northward with the apex of the zipper located at the Volcano Islands, the site of
eruption of highly alkalic lavas, These alkalic lavas are suggested to represent volcanism
precursory to backarc basin rifting (Stern et ai" 1984; Bloomer et aI., 1989a), Similar
magma types are commonly observed in continental regions undergoing incipient or early
stages of rifting, Hsui and Youngquist (1985) suggested that the curvature of the Mariana
arc is a consequence of Pacific plate rollback and pinning of the northern and southern
apices of the Mariana arc system by the Ogasawara Plateau and the Caroline Ridge
respectively. None of these models, however, took into account the details of the basin
structures or systematic changes in basin morphology.
A two-stage evolution of the Mariana Trough was first suggested using morphological
data from the central part of the Mariana Trough (from about 17°30' to 18°30'N; Fryer,
1981). Fryer (1981) observed a zone of block-faulted terrain on the western side of the
backarc basin, 60 km wide at 18°N. East of the block-faulted zone there is a pronounced
change in morphology. The region closest to the spreading axis is characterized by closely
spaced ridges and troughs, typical of abyssal hill fabric, and aligned parallel to the strike
of the spreading axis (Fryer, 1981). An initial stage of stretching and collapse of the arc in
the first few million years of basin extension probably produced the block-faulted terrain.
The collapse resulted in normal faulting and down-dropping of fault-controlled terraces
within a region at least 60 km wide and, assuming symmetrical rifting, produced a
maximum total width of rifted backarc basin of 120 km. A transition to true seafloor
spreading resulted in the formation of the ridge segment and transform fault morphology
observed in the central Mariana Trough today (Fryer, 1981; Fryer and Hussong, 1981). Early
block-faulted terrain of the Mariana Trough is similar in scale and structural style to
continental rifts (Cochran and Martinez, 1988; Martinez and Cochran, 1988; Baker, 1992)
and to that of incipient rift grabens in the Izu-Bonin arc, located immediately to the north of
the Mariana arc (Brown and Taylor, 1988; Brown, 1991; Klaus et aI., 1991; Taylor et aI.,
1991). A two-stage model for opening of the Mariana Trough is reflected in the changes in
morphology of the backarc basin along strike. This paper presents the details of the along-
strike and across-strike changes in the basin in terms of stages of opening of the Mariana
Trough. It also presents a summary of the volcanism and hydrothermal activity in the basin
and of the petrology and petrogenesis of basin lavas.
2. STRUCTURE OF THE MARIANA TROUGH
2.1. Initiation of Rifting

Several geophysical studies, side-scan sonar surveys, dredging cruises, and submers-
ible dives indicate that the northern third of the basin is rifting, not spreading. Between
21°N and 23°N the basin is made up of three morphologically distinct regions from north to
south (Figs. 6.3 and 6.4), referred to in this section as the northern basin, the central
volcanic platform, and the southern basin (southwest rifted basin (SWB) and fault block
terrain (FBT)). Each of these three regions is bounded on the east by a zone of active
volcanism and tectonism.
SeaMARC II side-scan sonar and bathymetry data suggest that the intersection of the
zone of active backarc rifting and the active volcanic front is southeast of the Volcano
Islands (Fryer, 1986; Beal, 1987; Baker, 1992). The zone of volcanism and tectonic
deformation that represents the center of the rifting region passes through the body of
Nikko seamount, located at about 23°N (Fryer, 1986; Fryer et aI., 1986; Baker, 1992).
Normal faults on the forearc north of Nikko seamount are oriented approximately
northwest-southeast, trending along the strike of the eastern boundary fault of the backarc
basin. Nikko seamount is a composite of numerous side vents, flow fields, and satellite
cones that reflect the interaction between the volcanic front and the backarc rift on which
it is built (Fryer, 1986; Baker, 1992).
The northern basin, extending from 23°N to about 22°25'N, is in the early part of the
242 PATRICIA FRYER
FIGURE 6.3. Bathymetry map of the northern half of the Mariana Trough. The volcanic-tectonic zone, an area of
concentration of recent faulting and volcanism within the trough, follows roughly the dark gray region of deeper
valleys that lies in the center of the basin at 17°N and is progressively closer to the volcanic front with increasing
latitude. A color version of this figure appears opposite page 246.
first stage of development proposed for the Mariana Trough (Fryer, 1981, 1986; Baker,
1992). Side-scan sonar survey of the region shows the diversity of the seafloor terrain in this
region (Fig. 6.5). The western margin of the basin is a block-faulted terrain, principally a
series of normal faults downdropped to the east (Fig. 6.6), with a variable sediment cover.
Similar terrain extends along the entire length of the western margin of the basin. The
northern basin lacks a well-developed spreading ridge/transform fault system. However,
the eastern side of the basin is characterized by a zone of recent tectonic disturbance and
volcanic activity. This zone of recent activity is referred to as the volcanic and tectonic zone
(VTZ) and differs from a spreading center in that it does not represent a plate boundary
FIGURE 6.4. Location map of structural provinces and features of the northern Mariana Trough, superimposed
on bathymetry. Seamounts of the volcanic arc are N-Nikko, I-Ichiyo, S-Syoyo, F-Fukujin, and K-Kasuga cross-
chain seamounts. The long narrow region between the two arrows is the volcanic-tectonic zone, an area of
concentration of recent faulting and volcanism within the trough. Major backarc volcanic seamounts along the
volcanic-tectonic zone are designated A, B, and C. The major structural divisions of the survey area are the NPR
(northern province), CPR (central volcanic platform), SWB (southern basin), and FBT (fault block terrain). Lines
lettered A-D show the location ofGSJ 3.5-kHz profiles in Fig. 6.7. Black symbols illustrate rock dredge locations.
(dredge data sources: circles, Jackson, 1989; squares, Stern and Bloomer, 1989; triangles, P. Fryer, unpublished
data).
(Baker, 1992). A deeply sedimented graben lies close to the volcanic front ofthe arc. Recent
volcanics have intruded into the sediments or emanate from the walls of faults that bound
the graben. A particularly large lava flow has emanated from a fault bounding the western
edge of this graben.
The central platform of the northern Mariana Trough (Figs. 6.4 and 6.5) was built up
by a locally persistent volcanic center that extends diagonally across the basin between
22°45'N and 22°N (Beal, 1987; Baker, 1992). Volcanism is currently active only on the
eastern edge of the platform, and a narrow (5-20 kIn wide) graben separates the platform
from the line of the active volcanic frontal arc. The morphology of the platform indicates an
increase in volcanism to the east (Baker, 1992). There are no discrete volcanic edifices
along the volcanic front line between Syoyo and Nikko seamounts (Fig. 6.4). The locus of
arc volcanism was probably shifted westward, and the volcanoes that lie on the eastern edge
of the central volcanic platform probably represent volcanism diverted from the volcanic
front. The composition of lavas recovered in dredges from seamount A (22°39'N) and from
basin lava flows supports this suggestion (both backarc basin basalt and island-arc basalt
sources (Jackson, 1989)). The large lava flow (see Fig. 6.5) from the volcanic ridge in the
basin north of this seamount is of island-arc tholeiite composition (Jackson, 1989).
The southern basin from 22°lO'N to 21 °40'N contains both the early stair-stepped fault
blocks on the west (SWB) and the later fault blocks that comprise a basin-and-range-type
244 PATRICIA FRYER
FIGURE 6.5. "Ramped" SeaMARC II sidescan sonar acoustic image of the northern Mariana backarc basin
(Sender et ai., 1989). Data are plotted so that acoustic signals with strong amplitudes are dark. The ship track is
represented by the blank stripe down the center of each IO-km-wide swath. The ramped processing scheme
maximizes gray-level contrast to enhance local structural details. The major tectonic and structural elements of the
survey region are superimposed on the side-scan imagery.
terrain (FBT) in the center of the backarc basin. The eastern margin of the southern basin
contains a series of volcanic ridges on the east (Figs. 6.4 to 6.6). The faulting in the deep
western and central parts of the basin was apparently accompanied by volcanism increasing
in volume toward the east (Baker, 1992). The stair-step faults of the westernmost part of the
southern basin were probably the sites of scattered intrusions or flows in the earliest phases
of backarc basin formation (Baker, 1992; Martinez et aI., in press). The ridges and troughs
of the fault-block terrain located between the western stair-stepped fault blocks and the
no no
~\
i
r :..... ,-
, , ,
,-~ ft\)j YEA ,.,.
. .. "'"
v........................ .
- '!'V -
- r-\ 1_
-
~dY\~, '~~,.fl 7 ~ (:
e
'-n ,... _ /
;-dm, 4~"
- ~
-
~I:~ W~~
• • ~"7'Q'~;'
~':' ~ ';',It.'\
..
. '1,'
1-
....... ,; " '"
. .' . tT
-1~ ~~
~~ '~. . ~~...;.r~~~,_~<P\~' ~1 -
.111
- I. . . .
R;,. JIfI..
- ,-,~ -
r.t; " __
. Jl.l1.-9~' ~r ~., :~ l:l-!,. • /"'., .... .... _.
~
--
''\.\
--.~,l<l-...~ ~" ~.
-... ~ -"'V,,;\..~/ -
FIGURE 6.6. Profiles A to D across the northern Mariana Trough traced from GSJ 3.5-kHz data. Profile locations are shown in Fig. 6.4. These profiles illustrate the change in width
and morphology of the basin from north (profile D) to south (profile A). Large tilted fault blocks characterize the north part of the basin, volcanic seamounts and fault blocks in the
central region, and intensely faulted terrain and volcanism in the south, shown by an increasing intensity of hyperbolic echoes. The east end of profile A crosses the south flank of the
-
arc volcano, Fukujin seamount.
246 PATRICIA FRYER
VTZ are only partially constructional volcanic ridges (Baker, 1992). Volcanism is super-
imposed on fault blocks, even in the eastern part of the basin (Baker, 1992,) although recent
Alvin observations of fissures, numerous vents, and fresh pillow ridges suggest that
volcanism dominates the central part of the VTZ (Fryer, 1990; Stem et ai., 1990). Based on
SeaMARC II interpretations, the VTZ lies in a 6- to ll-km-wide graben from 143°E to
143°lO'E between 22°lO'N and 21°40'N and is characterized by fault-bounded Valleys and
ridges, lava flows that show up as high-backscatter regions on sonar imagery, and small
volcanic cones (Fig. 6.7). From southeast to northwest the zone widens and deepens to
>4500 m at the base of a fault scarp that defines the boundary with the volcanic platform
to the north at 22°20'N (see Fig. 6.4). At 22°20'N several small seamounts have developed
on the edge of a fault block or accommodation zone (Beal, 1987; Baker, 1992) against
which this segment of the VTZ abuts. Several lava flow fields can be seen on the uplifted,
south-facing wall of this accommodation zone in the side-scan images (Figs. 6.5 and 6.7).
The southward increase in width of the basin and the general relationships of the structures
can be seen in seismic reflection profiles across the basin (Fig. 6.6).
2.2. From Rifting to Spreading

The VTZ remains on the eastern (arc) side of the basin during the early rifting stage of
the Mariana Trough (Baker, 1992; Martinez et al., in press). It begins to diverge from the
volcanic arc once the basin extends to a width of about 120 km (at about 22°N; Fig. 6.3).
The VTZ eventually becomes dominated by development of new backarc basin lithosphere
at a seafloor spreading center. The transition from rifting to spreading would therefore be
marked by a shift from volcanism confined to fault boundaries to development of volcanic
ridges and abyssal hill fabric. We can investigate this model for transition either by locating
the current point of transition along strike of the basin or by examining the change from
rifted, graben-and-horst terrain to abyssal-hill fabric on transects across the basin.
Along strike the basin must undergo a change from rifting to spreading somewhere
between 21°40'N, where side-scan data indicate the basin is still rifting, and 19°45'N,
where side-scan data show true seafloor spreading. Only bathymetry, magnetics, and
gravity data (Martinez et ai., in press) are available to trace the continuation of the VTZ
between these latitudes. The VTZ progressively diverges from the active volcanic arc and
terminates at about 21°20'N, approximately 75 km west of the eastern boundary fault ofthe
basin. Yamazaki et al. (1991, 1993) interpreted lineated magnetic anomalies in the basin
between 21°N and 22°N as seafloor spreading anomalies. More recently, Martinez et ai. (in
press) argued that the magnetic anomalies in the whole of the region from 21°N to 23°N are
the result of structurally controlled intrusions, implying rifting, not seafloor spreading.
Bouguer gravity data in the northern Mariana Trough show positive anomalies, implying
thinned crust (Martinez et ai., in press) and discontinuous lows that suggest low-density
material (possibly partial melt zones) underlies the currently active volcanic and tectonic
zone. The discontinuous nature of the lows implies that the volcanic segments are not
organized in the form of a spreading axis anywhere between 21°N and 23°N (Martinez
et ai., in press).
Rates of extension for the northern part of the Mariana Trough have generally been
estimated to be slow, consistent with the morphology. Slow spreading ridges may be
structurally complex; cycles of alternating tectonic and volcanic activity can occur within a
given location. The structures produced during tectonic cycles of a slow spreading ridge
may not differ significantly from those produced by rifting of preexisting crust. A qualita-
NDR1l4EAN IIIARIAHA ARC
BAnMIETRY
CIOWI.£O SY
,.
F'C1INN.rJO ItINITrEZ
"""
IMNM NSTTTVTE OF~')IIH\"SjI:S
FIGURE 6.3. Color contoured bathymetry map of the northern half of the Mariana Trough. The volcanic-tectonic
zone, an area of concentration of recent faulting and volcanism within the trough, follows roughly the dark blue
region of deeper valleys that lies in the center of the basin at l7°N and is progressively closer to the volcanic front
with increasing latitude.
FIGURE 6.7. Geologic interpretation map of the southernmost part of the volcanic and tectonic zone surveyed
by SeaMARC II. The interpretation is based on the combination of side-scan sonar, bathymetric, single-channel
seismic, and 3.5-kHz data (Baker, 1992). The major tectonic and structural elements of the survey region are
superimposed on the side-scan imagery. Faults are shown as solid lines with hachures indicating the downthrown
side. Solid lines are for faults with throws of over 100 m. Fine lines are faults with throws less than 100 m.
Lineaments visible in side-scan data, but not apparent on seismic profiles as having detectable offsets, are mapped
using a solid line lacking hachures. Volcanic cones are designated as solid black regions. Volcanism is concen-
trated in a fault-bound valley trending 3300 which widens into a 4S00-m graben at the top of this figure.
248 PATRICIA FRYER
tive structural description of the Mariana Trough between 21 040'N and 22°1O'N does not
adequately distinguish rifting and spreading, and the distinction is important because it
bears on the question of general evolution of backarc basins. If the northern portion of the
Mariana Trough (north of 210N) is not spreading, and by spreading it is meant that new
lithosphere of backarc basin origin is forming, then there is no new plate boundary in the
northern Mariana Trough.
The geology changes markedly between 210N and 200N. A series of -5-km-deep half-
grabens oriented roughly N-S occupies the center of the basin in these latitudes (see Fig.
6.3; Smoot, 1990). There is no side-scan imagery of these grabens, and thus the exact
distribution and type of volcanic activity associated with them is unknown at present;
however, dredging in these basins in 1984 (Shibata and Segawa, 1984; Ishii et aI., 1985)
recovered fresh basalts compositionally similar to the backarc basin basalts of the ridge
segments farther south. Gabbros recently dredged from the grabens are also of backarc
basin basalt composition (Stem, pers. comm., 1992). The major bounding faults of the half-
grabens are alternately on the east and west sides of the half-grabens, and thus the sense of
offset between the half-grabens alternates between right and left lateral. The length of the
half-grabens is approximately 20 to 30 km between offsets. This type of half-graben is
similar to that observed in the rift grabens of the Izu-Bonin arc to the north (Klaus et aI.,
1991), and their distribution is similar to the distribution of half-grabens in continental rift
areas. The magnetics and gravity data in this part of the basin differ significantly from that
of the basin to the north (Martinez et aI., in press). There is a break in the continuous
magnetic lineaments at about 21°N, which implies that this part of the basin is rifting in a
less well-defined, regular manner (Martinez et aI., in press). The half-grabens thus repre-
sent a different tectonic regime superimposed on the rifting backarc basin. There is
insufficient data at present to define the origin of the half-grabens; however, it is possible
that cross-arc structural trends and forearc tectonics influence their formation. Hussong and
Fryer (1983) suggested that Fukujin and Kasuga seamounts formed within a N-S trending
cross-arc rift (represented by a regional bathymetric low and a pronounced reentrant in the
inner trench slope). They suggested that this rift cuts across the northern Mariana forearc
from the trench axis to the volcanic front and, together with the eastern bounding fault of
the basin, controls distribution of the arc volcanoes along its trend. The extension of this rift
into the backarc basin would intersect the backarc basin and encompass the half-grabens.
Such a cross-arc rift could influence the structural development of the backarc basin to the
extent that it would perturb a gradual change from the rift morphology north of 21°30'N
to the spreading morphology that characterizes the basin to the south.
A SeaMARC II survey of the central part of the Mariana Trough between 19° and
19°45'N shows a well-developed spreading center with symmetric abyssal hill fabric
developed at the margins of a spreading valley that lies in a zone of high backscatter (Figs.
6.8a,b). The high-backscatter region indicates recent volcanism. The fault-bounded spread-
ing valley is less than 10 km wide and is flanked by abyssal hill fabric extending to a total
width of about 40 km (widest in the south). Two first-order discontinuities are evident in
this image at latitudes of 19°22'N and 19°40'N separating three ridge segments. Bathyme-
try coverage of the region is not complete; therefore, bathymetric details of secondary
discontinuities along the ridge segments are not obvious. However, it is apparent from
inspection of the SeaMARC II imagery (Fig. 6.8b) that the volcanic ridges within the three
segments surveyed contain several discontinuities along strike (solid strips in Fig. 6.8a).
The mean azimuthal trend of the faults bounding the axial valley is about 345°; however,
the fault scarps converge on about 19°45'N and are terminated in a V-shaped valley
(Fig. 6.8a), suggesting a propagating rift. The bathymetry of the region suggests that the rift
tip may extend to 20 01O'N (Fig. 6.3), but it will require side-scan data to verify this
suggestion. A change from rifting to true mid-ocean spreading in the Mariana Trough takes
place near this point. South of 200N new backarc basin lithosphere forms along mid-ocean-
style volcanic ridges.
If gradual transition from rifting to spreading is the correct model for evolution of the
Mariana Trough, this transition in the northern part of the basin is obscured by two abrupt
structural changes in the style of opening of the basin; from rifted basin to half-grabens and
from half-grabens to propagating ridge tip. It may be possible to identify the transition from
rifting to spreading by examining the distinct change in morphology across the basin from
rifted block-faulted crust to abyssal hill fabric. The best-studied part of the basin where
seafloor spreading is taking place is at about 18°N.
2.3. The Central Mariana Trough

The central latitudes ofthe Mariana Trough (from 14° to 200N; see Figs. 6.3, 6.8-6.10)
show a series of spreading valley segments offset by a complex series of faults and fault
systems comprising transform fault valleys. In general, the center of spreading is about 40-
50 km closer to the active Mariana Islands chain than to the West Mariana Ridge at these
latitudes (Fig. 6.9a). Fryer and Hussong (1981) estimated that the central portion of the
Mariana Trough at 18°N contains the widest section of newly generated backarc basin
lithosphere formed by seafloor spreading (approximately 110 km). They also assume that
the seafloor spreading is symmetric. Based on a spreading half-rate of 1.5 cm/yr (Seama and
Fujiwara, 1993), the newly generated backarc basin lithosphere would have begun forming
about 3.67 Ma. The change from block-faulted terrain to abyssal-hill-type topography
occurs at a maximum of 80 km east of the West Mariana Ridge. If the total width of the
basin at about 18°N is 240 km, assuming that Pagan Island marks the base of the eastern
bounding fault of the basin, there must be approximately 50 km of rifted arc lithosphere
underlying the volcaniclastic wedge of the arc on the eastern side of the basin (some of
which is visible in Fig. 6.9a). The total rifted portion of the basin is about 130 km wide.
Estimates of a maximum possible width of 120 km for extension by rifting prior to develop-
ment of true seafloor spreading in the basin (Fryer, 1981; Fryer and Hussong, 1981) are
consistent with this calculation. The only reliable age date on the rifted portion of the basin,
5 Ma, is from DSDP Site 453 (Hussong and Uyeda, 1981), located about 25 km east of the
West Mariana Ridge. Thus, it is not possible to date accurately the inception of rifting in the
Mariana Trough at this latitude or to determine the rate of extension in the rifting stage.
The locus of spreading between about 17°30'N to 19°45'E is a well-defined valley
with symetrically disposed fault blocks (Fig. 6.8b) and containing an elongate volcanic
ridge or ridges (Figs. 6.9b,c). The maximum width of the spreading center valley is about
20 km, and it ranges in depth from about 4.0 to 5.3 km (Fig. 6.9c). Broad ridges and troughs
also strike ENE across the width of the Mariana Trough (Figs. 6.3 and 6.9b). The ridges are
from 20 to 30 km wide (north to south), with intervening troughs about 20 km wide. The
troughs are collinear with offsets in spreading rift segments and are interpreted as fracture
zone valleys (Fryer and Hussong 1981; Hussong and Sinton, 1983; Kong, 1993). Hussong
and Sinton (1983) suggested that there are small offsets in ridge segments within the Pagan
fracture zone at 17°30'N. The transform fault valleys are generally 4.2 to 4.4 km deep. The
deepest portions of the valley segments (maximum depths are slightly greater than 5 km)
occur where they intersect ridge offsets. Two of these intersections contain in excess of
250 PATRICIA FRYER
2O' N
19' N
a 144' E 145'E
FIGURE 6,8. a. Bathymetry of the region 18°30' to 200N. This is not SeaMARC II bathymetry but a compilation
of data from several sources (DBDBS, Sea Beam and 3.5-kHz data). (b) SeaMARC II side-scan sonar imagery of a
survey of the Mariana backarc basin spreading center from l8°4S'N to about 19°4S'N. The high-backscatter
regions (dark) in the center of the survey denote the active volcanic ridges of the spreading center valley and walls.
The small arrows point to several lava flows that have broken out on the flanks of the central valley and to a small
cone (second from the top) that has developed on the eastern flank of the valley. Note the symmetric distribution of
abyssal hill topography and the sinuosity of the abyssal hill terrain. This is typical of mid-ocean ridge topography.
Bathymetry data were not collected on this early survey.
300 m of sediment in the active transfonn fault zones between major rift segments
(Hussong and Sinton, 1983).
The 20- to 30-lan-wide ridges that strike east-west across the Mariana Trough contain
a series of fault blocks that parallel the strike of the ridge segment (nearly north-south).
These fault blocks have symmetrically (toward the rift valley) dipping fault scarps similar
to abyssal hill fabric (Figs. 6.9a,b) typical of mid-ocean ridge systems (e.g., Perram and
Macdonald, 1990). Although the eastern portion of the basin is covered with a wedge of
b
FIGURE 6.8. (COrll;ml~d)
volcaniclastics from the active arc, basement protrudes through the sediment wedge close
to the rift valley (Fig. 6.9a). The distribution of these faults is shown on the SeaMARC II
image of the rift segments between 19°N and 19°45'N (Fig. 6.8b). The sinuousity of the
faults throughout this survey area is an artifact.
The flanks of the ENE ridges have frequent reentrants (Fig. 6.9b), which give them a
jagged, sawtooth appearance. The ridges vary in depth along strike, rising to minima
adjacent to intersection highs. Intersection highs have been noted on a number of slow
spreading mid-ocean ridges where major fracture zones intersect rift segments (e.g.,
Atlantis, Vema, and Oceanographer). A study of seismicity at the intersection of the Pagan
fracture zone with the spreading valley at 18°N in the Mariana Trough (Hussong and
Sinton, 1983) indicates that the formation of the intersection high, and the associated
relative uplift of the adjacent ridge edge, may result from high-angle dip-slip faulting which
predominates along the scarps bordering the intersection high. The occurrence of these
intersection highs and associated steep ridge flanks produces a complex bathymetric grain
in the central part of the basin. Between 17° and 19°30'N the ridges have steep, sediment-
free flanks, with frequent reentrants. This "abyssal hill" fabric is absent about 80 km from
the remnant arc, where the change from the initial rifting terrain to the subsequent
252 PATRICIA FRYER
a -""---,, I , I
10002 *OZ I I I I
_~ 01002 _Z OlOOZ 02002 0,002 00002 2:1002 2200Z
, , - " '... '711
FIGURE 6.9. a. Seismic reflection profile at 17°50'N across the Mariana Trough from the West Mariana Ridge
(left of upper profile) to the volcaniclastic apron south of Pagan Island (right of lower profile). The center of the
spreading axis of the basin is located at 0700Z. Note angular unconfonnities in the sediments of the western
grabens and nonnal faulting at 1415Z. (b) Bathymetry of the central Mariana Trough and reflection seismic profiles
showing major structural features (Fryer and Hussong, 1981). Profile A (B-B') shows the transition from the
abyssal hills topography to the graben and horst topography beginning with the large block at 1530Z. Profile B
(C-C') shows the morphology along strike of the abyssal hill fabric. (c) Bathymetry of the central Mariana Trough
showing the morphology of the central spreading ridge and the locations of dredge samples and dives (Hawkins
et al., 1990).
spreading morphology takes place (Fryer and Hussong, 1981). Also, at this distance from
the West Mariana Ridge, several of the transform valleys change strike from ENE to NE-
SW, suggesting a change in the direction of extension (Fryer and Hussong, 1982).
The locus of spreading in the central Mariana Trough is defined on the basis of
SeaMARC II side-scan imagery, bottom photography, detailed bathymetric data, the
recovery of fresh, glassy basalt fragments in dredge hauls, seismicity studies, and direct
observation using submersibles between about 17°30' to 19°45'N. A detailed description of
the rift valley between 17°30' and 18°30'N is given by Hawkins et al. (1990) and Kong
(1993). At these latitudes the complexity of the volcanic axial ridge is apparent (see Fig.
6.9c). There are three major offsets greater than 5 km, which comprise the first-order
segmentation of the axial ridge (Kong, 1993), and more than a dozen minor offsets, which
FIGURE 6.9. (Conllfllll'd)
comprise the second-order segmentation (Hawkins et aI., 1990). These second-order dis-
continuities are expressed by a series of both right- and left-lateral en echelon offsets ofless
than 1 kIn (Hawkins et aI., 1990). The ridge segments are broadly shallowest at the centers
of the segments and slope steeply toward the deep offset valleys. The shoal regions have
maximum relief on the order of 300 m with local highs between the en echelon offsets
marking the second-order discontinuities in the ridges.
Interpretation of the location of the spreading center valley from 14° to I7°N is based
solely on bathymetry data (Barone et aI., 1990; Smoot, 1990; Lange, 1992). The bathymetry
coverage of the Mariana Trough south of 15°40'N is the most complete (Fig. 6.10). A series
of short spreading valleys is offset westward by a series of right-lateral transform fault
valleys (Smoot, 1990; Lange, 1992). The spreading valleys maintain a distance of about 100
kIn from the active frontal arc between about 14°35' and 16°N by stepping westward as the
curve of the frontal arc swings southwest. The spreading valley segments are about 15 kIn
wide. Only a few identifiable central ridges or volcanic cones are in these valleys. The most
prominent of these is from 15°45'N to 16°N. A region of small cones and ridges lies along
144°30'E between about 15°25' and 15°30'N (Fig. 6.10). From 15°30' to 15°45'N and from
14°45' to 15°N there are two deep (>4 km) valleys in which the transform motion between
254 PATRICIA FRYER
•• ~ ALVIN DIVE TRACK AND DIVE NUMBER (18 :..: I

\...
@ DREDGE STATION
SirES OF EARLIER SAMPLES DESCRIBED BY

HAWI<INS AND MELCHIOR (1985)
,,
\
AXIAL VOLCANIC \ SPREADING
RIDGE AXIS
TURBIDITE APRON MAJOR RIFT· VALLEY

\
AND BASINS FAULT SCARP
SHALLOWER CONTOUR INTER! At

:5U",
THAN 3500 m nf'.PTH,~ IN l",
c
the spreading valley segments probably takes place. South of 14°40'N the spreading valley
is located at about 144°08'E. There is a gap in published data between 14°49' and 14°. At
14° the spreading valley is located between -144°40'E and 144°E, and therefore there must
be a transform valley in the data gap. The azimuth of the faults bounding the valleys from
16° to -14°30'N is about 0°; thus spreading is taking place in almost an east-west direction
in this part of the basin.
FIGURE 6.10. Bathymetry data from the Mariana Trough contoured every 250 m (data from Lange, 1992 and
Smoot, 1990).
The zone of active faulting along a mid-ocean ridge is restricted to a narrow region
adjacent to the spreading axis, regardless of spreading rate (e.g., Searle and Laughton, 1977;
Macdonald and Atwater, 1978; Kong et ai., 1988/89; Edwards et ai., 1991). Progressive
focusing of extensional deformation to a narrow region during the transition to seafloor
spreading (Chorowicz et ai., 1987; Cochran and Martinez, 1988; Martinez and Cochran,
1988) is characteristic of continental rifting models. In contrast, the faulting in the Mariana
Trough is not restricted to the region immediately adjacent to the volcanic ridges of the
volcanic and tectonic zone in the eastern part of the basin. Sediments in the western
portions of the valleys show numerous angular unconformities (Fig. 6.9a), implying a
history of tectonic disturbance (Fryer and Hussong, 1981). The angle of the unconformities
generally indicates westward rotation of fault blocks, and the gradual increase in angle of
the sediment layers argues for continuous extension, at least at the margins of the Mariana
Trough, throughout its rifting history. Faulting along the base of the West Mariana Ridge
has created a deep trough at the western edge of the basin (Fryer and Hussong, 1981; Baker,
1992; Wessel et ai., in press). These faults at the basin margins disrupt recent sediments
(Fig. 6.9a), indicating that deformation is still active at the outer margins of the basin. The
most marked of these faulted basins lies at the eastern base of the West Mariana Ridge.
There is a deep trough (average depth 4.5 km) at the western margin of the Mariana Trough
along most of its length. This trough lies between the western block-faulted terrain and the
West Mariana Ridge (see Fig. 6.2) at the base of the western bounding fault of the backarc
basin. Data recently collected in the southern part of the Mariana Trough show structures
that are either unsedimented fault blocks or igneous piercement structures in the deep
trough at the western margin of the backarc basin (Hagen et ai., 1992). The presence of
256 PATRICIA FRYER
unsedimented, faulted ridges implies recent tectonism in the westernmost Mariana Trough.
If, however, the structures are igneous bodies, then either volcanism was rejuvenated near
the remnant arc or extension may be taking place in the far western margin of the southern
Mariana Trough. As we shall see in the next section, tectonism and volcanism are taking
place in a spreading valley in the eastern half of the southern Mariana Trough (Fryer, 1993).
If volcanism is also active on the western margin of the southern part of the basin it
represents a unique phenomenon for the Mariana Trough.
2.4. Extension in the Southern Mariana Trough

The Mariana Trough undergoes a regional change in morphology at about 14 ON (Figs.
6.2 and 11). South of 14°N the central part of the basin is uniformly shallower by about 1 kIn
than the central basin north of that latitude, and the relief of this part of the basin is much
less. Bathymetry data from prior to 1991 show a complex distribution of numerous small
ridges (a few kilometers wide and tens of kilometers long), many small volcanoes, and
14°,------,--------,---------,--------,-----,-------:--,------,
142°E
FIGURE 6.11. Bathymetry of the southern Mariana Trough. Contour interval 250 m [data from Bloomer and
Hawkins, 1983; Smoot, 1990; Hagen et at., 1992; Fryer, 1993, and DBDB5 data set of National Geophysical Data
Center (NGDC)].
several prominent bounding troughs in the southern basin (Smoot, 1988, 1990). The ridges
parallel the strike of the arc, for the most part, changing strike southward as the arc curves
around to the west (Smoot, 1990). Karig and Ranken (1983) suggested that the active rift
zone from BON to 14.5°N was located in the approximate center of the basin. Smoot (1990)
suggested, as had Hawkins (1977), that the southern part of the Mariana Trough has formed
by diffuse spreading and that there is no well-defined spreading axis.
Recent SeaMARC II and Sea Beam data from about 13°45'N indicate a narrow, low-
relief spreading axis between 143°50' and 144°E (Hagen et aI., 1992; Fig. 6.12). The
transect skirts Tracey seamount, part of the active Mariana arc (Dixon and Stem, 1983). The
location of the volcanic rift, suggested by Hagen et al. (1992), is a zone of high backscatter
with numerous small volcanic cones between about 143°50'E and 144°E at 13°45'N (Fig.
6.12). This rift valley is oriented approximately N-S and is about 10 km wide. SeaMARC II
and Sea Beam data at this latitude (Hagen et aI., 1992) show this spreading valley is about
400 m deep and flanked by fault-block ridges. The relief (Fig. 6.13) of the faulted boundary
of the spreading valley is less than the relief of the faults bounding the spreading valley over
most of the rest of the basin (generally over 1000 m and up to -3000 m; see Figs. 6.3, 6.8a,
and 6.9). The high-backscatter area ends abruptly to the east where volcaniclastic sediments
from the active arc are ponded against a volcanic ridge. The high backscatter decreases
gradually to the west as increasing sediment cover blankets the topography. The side-scan
images of the region (Hagen et aI., 1992) show sediments particularly in the western half of
the basin (Fig. 6.12). Sediments also partially fill the two deep troughs that lie at the margins
of the platform, but are lacking from some of the ridges within the western trough. The
14T~~~_______'4~~rS~_______
~______~~~____~'~4~rS~_______ '~~______--.'.,~
FIGURE 6.12. SeaMARC II side-scan data from the southern Mariana Trough and a sketch of tile structures
of the swath interpreted by Hagen et at. (1992).
258 PATRICIA FRYER
deepest basins lie in the eastern bounding trough south of Tracey seamount and in the
broader parts of the western bounding trough (Hagen et al., 1992). The trough at the base of
the West Mariana Ridge is a continuation of the trough that exists over much of the length
of the basin. The trough on the east, however, is restricted to the southern part of the basin.
It is related to a major north-south trending fault that cuts across the entire forearc and arc
and well into the backarc basin (Fryer, 1993).
Hagen et al. (1992) estimated that spreading has been going on for about 3 m.y. in this
part of the basin, assuming an extension rate of 2.15 cm/m.y. NNE trending abyssal hill
fabric extends from the western boundary of the spreading valley westward to -143°20'E.
The ridges and troughs are spaced 8 to 10 km apart and have relief of 200 to 400 m. A
similar set of ridges, but trending NNW, lies at about 143°15'N to 143°20'E. Hagen et al.
(1992) suggested the change in orientation of fabric may relate to a complex multiple
spreading history involving propagation of rifting toward the south in the earliest stage of
basin formation. If we accept the model of an initial stage of rifting, not spreading, as the
norm for the Mariana Trough, the Hagen et al. (1992) suggestion of an early southward
propagating ridge does not fit. Rather, the change in orientation of the basin fabric may
represent a boundary between the early rifted arc lithosphere and seafloor created over the
last 3 m.y. by seafloor spreading.
A 10-km-wide trough separates the eastern slope of the West Mariana Ridge from the
southern Mariana Trough platform (Hagen et aI., 1992). This trough varies in depth from
4600 to more than 5100 m within the survey area, and the floor of the trough exhibits low
backscatter, indicating a sedimented surface. A second, smaller and shallower graben,
located just to the west, is separated from the first by a shallow ridge that rises above
3000 m (Hagen et aI., 1992). The relief of the fault blocks is greater than that observed in
existing data from any other part of the western bounding fault graben along the length of
the Mariana Trough. The presence of high-relief, narrow ridges and troughs in the western
boundary graben is in sharp contrast to the stair-stepped, fault blocks draped with sediment
that are characteristic of the western boundary of the central and northernmost parts of the
backarc basin. This difference in morphology possibly reflects a difference in the tectonic
style of the basin. It is possible that extension in the Mariana Trough at this latitude is
expressed both in the formation of a subdued spreading rift and in the formation of the large
grabens that bound the central platform. The abrupt change in depth of the Mariana Trough
at about 14°N (shallower to the south) may mark the location of a basinwide structural
change, as yet undefined, that facilitates a change in the nature of rifting and volcanism
south of that latitude. It is not possible with currently available data to distinguish whether
activity in the bounding rift grabens is accompanied by volcanic processes, like that
occurring at the periphery of the rift grabens in the northernmost portion of the Mariana
Trough, or whether the ridges are being formed solely by continued faulting of the margin
of the basin. Some combination of the two is possible.
The pronounced graben at the base of the eastern margin of the southern platform (i.e.,
south of 13°N) is absent in the rest of the basin. This eastern graben is, on average, about 15
km wide, has a flat-sediment-covered floor, and is generally 3.8-4.0 km deep, with local
depths to 4.4 km (Hagen et al., 1992). Piercement structures are apparent within the eastern
boundary graben (Hagen et aI., 1992). If this graben does project further north, it is
apparently filled with volcaniclastics from the active-arc volcanoes. Recent bathymetry
data collected during cruise Y9204 of the RN Yokosuka in the southern platform of the
Mariana Trough indicate that the eastern graben extends southward to the trench axis and
that recent defonnation has occurred along its length (Fryer, 1993). Bathymetry data
collected from near 144°lO'E between 12°lO'N and 12°37'N by the RN Yokosuka define the
extent of this fault (Fig. 6.13). The position of the fault is consistent with earlier regional
structural interpretations (Karig and Ranken, 1983). The bathymetry data show an
eastward-facing fault with a throw of nearly 2500 m that strikes approximately north-south
at about 144°lO'E. Submersible observations of nearly unsedimented talus slopes on the
scarp, relatively fresh lava outbreaks on the fault scarp, and unusual fluids in sediments at
the base of the scarp suggest recent movement on the fault (Fryer, 1993). The lavas
recovered during dives are of island-arc basalt composition; thus it is possible that the fault
is an avenue for migration of arc magmas into the forearc region (Fryer, 1993).
Although the data from Hagen et at. (1992) suggested that the spreading center in the
backarc basin is narrow and of low relief, bathymetry data (Fig. 6.11) collected during the
recent RN Yokosuka surveys show that the southern extension of the spreading center
broadens southward from 15 km at 13°40'N to over 40 km at 13°lO'N (Fryer, 1993). The
volcanic edifices located within the broad spreading valley are large and active both
tectonically and volcanically (Fryer, 1993). Volcanoes are located along several north-south
trending ridges on the eastern margin of the spreading region. A dive on a small donnant
volcano on the extreme western edge of the broad rift valley demonstrated that faulting is
active throughout the spreading valley (Fryer, 1993). A dive on one of the larger active
volcanoes of the eastern side of the rift showed recent volcanism and active hydrothennal
activity (Johnson et at., 1993).
If spreading extends farther to the south, the exact location of the spreading center is
poorly constrained, particularly in the southwestern part of the basin. Spreading may
simply be failing as the spreading axis intersects the arc line. The elevated southern
platfonn may be an expression of diffuse volcanism, as suggested by Hawkins (1977) and
Smoot (1990), but this volcanism is not necessarily related to diffuse "spreading," as
suggested by these authors. It may relate to intrusion and volcanism of arc magmas within a
rifting backarc basin, with true spreading taking place only in the eastennost part of the
basin. If spreading does persist to the south, there are at least two possible locations for
spreading centers in the southernmost Mariana Trough, and spreading may be taking place
in more than one of them.
(1) The spreading center may curve southwestward and lie close to the southwestern
extension of the volcanic arc. The general morphology of the southernmost part of the
Mariana Trough shows ridges and several localized highs (arc volcanoes) curving to the
southwest parallel to the Mariana Trench (Bloomer and Hawkins, 1983; Smoot, 1990). The
ridge that bounds the western margin of the spreading valley follows this trend (see Fig.
6.13). Such a phenomenon is consistent with the observations we have made of the
tendency of the VTZ to remain close to the arc in the northern part of the Mariana Trough. If
the spreading center does follow this trend, the expression of the spreading valley becomes
extremely subdued toward the southwest until at least 143°E, where a pair of >4-km-deep
troughs lies immediately north of the inner trench slope (at about 12°N, see Figs. 6.2 and
6.11). We have very few data in this part of the basin to help constrain this suggestion.
(2) A far-better-substantiated interpretation is that the spreading valley is offset to the
east at 13°lO'N and becomes collinear with the major north-south trending fault in the
Mariana forearc at 144°lO'E. East-west faulting has influenced the development of the
volcano at 13°lO'N, 143°50'E in the spreading valley. The morphology and marked
curvature of the deep graben at 144°E, between 13°N and 13°lO'N, toward the southeast are
260 PATRICIA FRYER
similar to the morphology of transform fault valleys farther north in the backarc basin. An
eastward offset of the spreading center along this transform would result in its lying nearly
collinear with the north-south forearc fault. Arc lavas were recovered along the fault well
into the forearc (Fryer, 1993). Such a scenario is similar to the extension of backarc
spreading from the Three Kings triple junction into the forearc in the northeastern comer of
the Lau Basin (Parson and Tiffin, 1993). With currently available data the interpretation that
spreading in the southern Mariana Trough extends through the arc line and into the forearc
is the best constrained of the two and best fits the regional morphology and the distribution
of volcanism in the southernmost part of the Mariana Trough.
3. VOLCANISM OF THE MARIANA TROUGH
3.1. The Volcanic Front and Cross-Chains

Most of the subaerial and submarine volcanoes of the arc have formed immediately
west of the eastern boundary fault zone of the Mariana Trough (Fryer and Hussong, 1981,
1982; Hussong and Fryer, 1983; Smoot, 1988). It has been shown recently, however, that arc
magmatism has occurred as recently as late Pliocene (1.7 Ma) in the forearc region of the
Mariana system at less than 100 km from the Mariana Trench (Marlow et aI., 1992; Fryer,
1993). In addition, one of the striking features of the Mariana Trough is a series of chains of
volcanoes ("cross-chains"; Fryer and Hussong, 1982) that lie within the backarc basin.
These cross-chains of volcanoes are almost all oriented nearly orthogonal to the strike of
the volcanic front along structural lineaments that cut across the volcanic front. Cross-
chains of volcanoes in the Mariana Trough probably reflect interaction between preexisting
cross-arc structures, forearc tectonic activity, magmatic activity at the volcanic front, and
backarc basin rifting (Fryer and Hussong, 1982; Fryer, 1993). Fryer and Hussong (1982)
suggested that cross-arc structures related to strain between the trench and the volcanic
front may be inherited by each successive volcanic arc and backarc basin as an intraplate
convergent margin evolves. The locations of volcanoes along island arcs may be influenced
by preexisting cross-arc structures. The definition of "frontal volcanoes" in the Mariana
arc system must, therefore, be flexible enough to accommodate both forearc and backarc
volcanic features.
The subaerial volcanoes of the Mariana arc from 14° to 20 0 N, called the central island
province (CIP) by Bloomer et al. (1989b), are active volcanic islands of composite
volcanoes (Dixon and Batiza, 1979; Woodhead, 1987, 1989; Stem et aI., 1988; Bloomer
et aI., 1989b). These volcanoes are very similar morphologically. Several of the volcanoes
are made up of multiple vents lying along structural lineaments that are oriented approx-
imately orthogonal to the strike of the arc (e.g., Guguan, Pagan, Anatahan). Several of the
volcanoes also have structures, grabens or faults, that are oriented parallel to the strike of
the arc. For example, Agrigan (Stem, 1979) and Maug (Smoot, 1988) both have north-south
striking grabens in their flanks. These arc parallel normal fault grabens indicate that
deformation along the eastern boundary of the Mariana Trough continues today.
The largest of the CIP volcanoes are Pagan and Agrigan at about 18°N. The size of
these edifices is probably related to regional deformation of the arc at this latitude. These
volcanoes lie adjacent to a forearc basin. Studies of the forearc at this latitude indicate
forearc-wide tensional deformation at all scales (Mrosowski et aI., 1981). There is a distinct
break in the inner forearc between 17°30' and 18°30'N (Fig. 6.3). If this region represents a
zone of faulting that cuts across the entire forearc and backarc, the rate of eruption of
volcanoes in the gap would be expected to be greater, thus the greater volume of Pagan and
Agrigan. Such a cross-arc structure might also explain the more pronounced development
of the east-west trending transform fault valleys of the Mariana Trough at this latitude.
Cross-arc structures of such magnitude would also explain the presence of arc magmas in
the backarc basin even on the active spreading ridge segments (Fryer et aI., 1981; Hawkins
et aI., 1990) by providing avenues for migration of arc magmas into the backarc basin.
The volcanoes north of Uracas, in a region termed the northern seamount province
(NSP) by Bloomer et al. (1989b), are generally smaller than those of the CIP. The spacing
between volcanoes is less regular in the NSP than in the CIP, possibly reflecting the
distribution of cross-arc structures. The development of cross-chains of volcanoes extend-
ing from the arc line into the backarc basin is more marked in the NSP than elsewhere in the
Mariana Trough because the individual cross-chain volcanoes are much larger (Figs.
6.2 and 6.3).
The Kasuga volcanic cross-chain, at about 22°N, has been studied in detail (Figs. 6.4
and 6.13a,b) by side-scan sonar and bathymetry surveys (Hussong and Fryer, 1983) and has
been extensively sampled by dredging and submersible (Fryer et aI., 1987; Jackson, 1989;
McMurtry et aI., 1993) It trends nearly north-south and intersects the arc at an angle of
about 60° to the strike of the arc (Figs. 6.3, 6.4, and 6.13a,b). A recent analysis of numerous
faults and lineaments in the inner forearc at about 21°30'N, adjacent to the Kasuga
seamount chain (Mahoney and Fryer, 1988; Wessel, 1991; Wessel et ai., in press), indicates
that arc-parallel tensional stress is causing rifting in the forearc. This rifting is orthogonal to
and tectonically decoupled from the extensional forces involved in backarc basin forma-
tion. The forearc rifting may be related to a local tectonic event, the gradual increase in
curvature of the forearc with time, or to both (Fryer 1986; Wessel et aI., in press). The
forearc rifting has apparently crosscut the entire forearc region at 143°30'E and is respons-
ible for development of a bathymetric trough oriented roughly north-south extending from
the trench axis at least to the volcanic front. The rifting causes a significant offset in the
eastern bounding fault of the backarc basin, and thus does affect the arc front. Activity on
this rift may also be responsible for the development of the graben immediately east of the
Kasuga volcanoes, for the numerous side vents and lava flow fields on the east flanks of
these volcanoes, and ultimately for the growth of the major edifices of the Kasuga cross-
chain.
The Kasuga volcanoes are complex edifices reflecting the interplay between cross-arc
and arc-parallel structures. There are numerous lava flow fields and regions mantled with
volcaniclastic debris on the two southernmost of the three volcanoes imaged (Hussong and
Fryer, 1983). The volcanoes were probably built on sundered arc lithosphere, during cross-
arc rifting. The magmas from which they were constructed were derived from at least two
separate sources, a backarc basin basalt source and an arc source. These volcanoes
demonstrate the interrelationships between magma sources and magma mixing processes
(Jackson, 1989). Magma mixing processes are evident in the composition of the summit
lavas from both volcanoes. Hydrothermal systems were discovered on the summits of the
two southernmost volcanoes (McMurtry et aI., 1993). A large volcaniclastic debris deposit
mantles the southern flank of the second volcano in the chain, and observations of the upper
half of the deposit indicate that it may have been emplaced hot in the submarine environ-
ment. A comparison of the small grain-size fraction of this deposit with that of Kick-em-
Jenny volcano of the Lesser Antilles arc shows remarkable similarity (Ballance et aI., 1988).
The presence of flank ridges on the two southern volcanoes of the Kasuga chain
262 PATRICIA FRYER
1474O'E
\
21°
,-~=--...
b ~--------------~15'
a ....
N
FIGURE 6. 13. (a). SeaMARC II side-scan image of the three Kasuga volcanoes of the Mariana backarc basin at
22°N (Hussong and Fryer, 1983). (b) Bathymetry of the three Kasuga volcanoes of the Mariana backarc basin at
22°N (Hussong and Fryer, 1983).
suggests volcanic rift development. Examination of the regional structures, however,

indicates instead that these flank ridges represent structural elements of the backarc basin
on which the later lavas of the Kasuga volcanoes have been superimposed. It is entirely
possible that these structures subsequently also furnished avenues for egress of Kasuga
magmas.
The Kasuga volcanoes lie at the top of a fault scarp bounding a graben to the east of the
chain (Hussong and Fryer, 1983). Lobate regions of high-backscatter on side-scan images
of the eastern flanks of the volcanoes indicate recent volcanism along the face of the fault
scarp (Hussong and Fryer, 1983). Satellite cones on the Kasuga volcanoes are most
prevalent on the eastern side of the chain, also suggesting a control over distribution of
volcanism by the fault underlying the eastern flanks of the volcanoes. The fact that the
eastern flanks of the volcanoes are the regions of highest relief on the edifices suggests
that the bounding fault of the graben is still active and that deformation of the flanks is
ongoing.
The volcanoes south of 16°N comprise the southern seamount province (SSP) of
Bloomer et al. (1989b). As in the north, these volcanoes are smaller than those of the CIP.
Cross-chain development is present, but the individual volcanic edifices are smaller (e.g.
see Fig. 6.10). Several of the volcanoes of the SSP have a north-south elongation of vents
and satellite cones (Bloomer et al., 1989b). This suggests that the fault system bounding the
eastern portion of the basin also plays a significant role in the development of these
volcanoes. Until recently, Tracey seamount was considered to be the southernmost active
volcano in the frontal arc. The results of recent bathymetric mapping and submersible work
in the Mariana backarc basin (Fryer, 1993) show that the arc volcanoes extend farther to the
southwest and intersect the spreading ridge of the southernmost backarc basin (Fig. 6.11).
Although detailed bathymetry data are not available for the southwest comer of the
Mariana Trough, several shoal areas extend to the southwest of the rift intersection and
probably represent additional arc volcanoes that lie along a line roughly parallel to the
curvature of the Mariana Trench (see Figs. 6.2 and 6.11). The composition of the lavas
collected on Shinkai 6500 dives on two of the volcanoes of this southwestern group at about
144°E have backarc basin basalt signatures (Johnson et al., 1993). Several of the volcanoes
have north-south alignment of rift zones. These rift zones are probably related to the
intersection of the backarc rifting system with the line of the volcanic arc. To date no
samples have been collected from the southwestern volcanoes.
3.2. Intrabasin Faults and the Spreading Ridges

Throughout the basin side-scan sonar images indicate that fault boundaries of major
structures are the most likely sites of volcanism. Many of the fault scarps, like that inferred
to the east of the Kasuga cross-chain, are draped with lobate high-backscatter regions
indicative of recent volcanism. The distribution of volcanism on faults oriented orthogonal
to the strike of the arc or backarc basin trends is in discrete edifices, creating cross-chains.
Volcanic ridges oriented at high angles to the arc line are nonexistent. Along arc-parallel
lineaments within the backarc basin, however, the lava outbreaks tend to occur in fissures.
Small volcanic cones or ridges lie parallel to the faults or coalesce to form larger volcanic
ridges parallel to local structures. This tendency contributes to the growth of arc-parallel
ridges that eventually form significant structural elements within the volcanic regions of the
backarc basin. Magnetization of the Mariana Trough from the northern apex to about 21°N
shows a series of well-defined strips indicating periods of intrusion and volcanism as the
basin developed (Martinez et al., in press). The process of concentration of volcanism along
arc-parallel fissures eventually results in the development of true mid-ocean ridge spread-
ing segments and the evolution of new backarc basin lithosphere. On the regional scale
most outbreaks oflavas occur on the eastern side of the basin. This preference for alignment
on the frontal arc side of the basin persists over the length of the Mariana Trough and is
probably related to rheological characteristics of the arc lithosphere. It is likely that the
combination of high heat flow and more silicic compositions at the frontal arc tend to
facilitate rifting on the eastern side of the Mariana Trough (Martinez et al., in press).
Although most volcanism occurs in the east, individual volcanic edifices persist in the off-
axis regions of the western part of the basin as edifices and volcanic ridges aligned parallel
to basin bounding fault activity, as hydrothermal fields, or as scattered edifices (Leinen and
264 PATRICIA FRYER
Anderson, 1981; Hobart et al., 1983; Lonsdale and Hawkins, 1985; Martinez et al., in press).
This broad distribution implies that volcanic activity can playa significant role along the
boundaries between axis-parallel fault blocks at some distance from the currently active
ridge segments. The broad distribution of volcanism further supports the suggestion that
deformation can continue basinwide throughout the evolution of the Mariana Trough.
The better-formed ridge segments south of 19°45'N show concentration of volcanism
along ridges in the center of roughly symmetric valleys. At 19°45'N there appears to be a
propagating ridge tip (Figs. 6.8a,b) with a northward- (and possibly a southward-) facing
propagator. Farther south along the spreading axis the volcanic ridges are disposed in a
series of en echelon segments offset by small displacements (Hawkins et aI., 1990) or in
segments offset by large transform valleys (Hussong and Sinton, 1983). Volcanism is
distributed along these ridges much as it is on normal mid-ocean spreading centers. The
ridges are shallowest in the center between offsets (Hawkins et al., 1990). Between 17° and
14°N the spreading center is essentially a series of volcanic fissure ridge segments offset by
lateral shifts; however, no side-scan images are available for this region, and thus the details
of distribution of volcanism are unknown.
The southern backarc basin has been surveyed recently with SeaMARC II and Sea
Beam (Hagen et al., 1992), and the eastern portion has been swath-mapped producing
detailed bathymetry (Fryer, 1993). Volcanism there is a combination of spreading (or
possibly rifting) fissure activity and eruptions at discrete volcanic edifices. Details of the
volcanism of the southwestern extension of the Mariana arc are poorly known at present.
Certainly volcanism has occurred in the forearc on an extension southward of the volcanic
and tectonic zone in the southernmost part of the backarc basin. Whether this volcanism
represents an extension of the spreading axis of the backarc basin, as does the volcanism
along the continuation of the zone of volcanism and tectonic deformation north of Nikko
seamount, cannot be determined without side-scan sonar mapping of the seafloor.
4.1. Cross-Chains and Volcanic Front

Active hydrothermal systems have been discovered on several of the seamounts of the
volcanic front and within the backarc basin. Fukujin seamount is located on the volcanic
front, immediately north of the Kasuga volcanic cross-chain. A small satellite cone on the
northwest flank at a depth of about 1100 m yielded dredge samples of a hydrothermally
altered basalt and one of a vuggy silica deposit containing disseminated pyrite (McMurtry
et aI., 1993).
The summits of the two southern Kasuga volcanoes (Kasuga 2 and Kasuga 3 sea-
mounts) show hydrothermal activity and low-temperature solfatara-type deposits as ob-
served from Alvin submersible surveys. The deposits on Kasuga 2 consist of crusty patches
of nontronite and sulfur compounds. Very small (less than 10 cm) chimney structures of
nontronite are scattered in patches about the fields. The only biological associations with
the hydrothermal fields that have been identified are small (10-15 cm long) fish similar to
plecostamus. These are abundant in the hydrothermal areas and are probably feeding on
microbiological communities associated with the hydrothermal fields. The maximum
temperature measured in the hydrothermal solfatara field was 38.8°C. The details of the
fluid composition and of the hydrothermal deposits are discussed elsewhere (McMurtry
et aI., 1993). They indicate a hydrothermal system unique among those sampled in the
submarine environment to date. Apparently, addition of significant magmatic CO2 and S02
has contributed to the unusual composition.
Hydrothermal deposits at the summit region of Kasuga 3 consist of a crusty pavement
and small (up to 20 cm high) chimneys of amorphous iron oxides and nontronite and linear
ridges of birnessite and vernadite about 3 to 5 cm high. Some of these small ramparts are
nearly 20 cm high. A few yellowish green chimney structures of nontronite, up to about
30 cm high, were observed along some of the small manganese ramparts. A small vent of
shimmering water was discovered at 1140 m on a dacite dome at the summit of the volcano.
Two temperature measurements recorded a dTof 4.67°C and 5.53°C. The mineralogy of the
hydrothermal deposits is consistent with the low temperatures measured (McMurtry, et al.,
1993). The seafloor around the vent is encrusted with patchy reddish deposits of amorphous
iron oxide that are very vuggy. White flocculent material, probably bacterial mats, was
observed on much of the exposed rock and hydrothermal crusts, but no other biological
communities.
Sulfide deposits were recovered recently from dredging the summit of Esmeralda
Bank in the southern Mariana seamount province (Stuben et aI., 1992).
4.2. Central Spreading Basin

A summary of preliminary results of Alvin submersible studies of hydrothermal
systems along the Mariana Trough spreading ridge segments of 18°N was presented by
Fryer (1990). Recent publications have refined some of the earlier observations. Hydrother-
mal vent activity was first detected as plumes of methane and high 3He/4 He ratios (Poreda,
1985; Horibe et aI., 1986; Craig et al., 1987) 800 m above the seafloor at 18°13'N.
Submersible observations show that the axial volcanoes of the spreading axis support two
large hydrothermal fields at depths of 3600-3700 m (Craig et al., 1987; Kastner et al.,
1990). The vent waters are similar to Loihi vents in 3He/4He ratio (8.6 x atmospheric (RiR A
= 8.6» with CH/3He = (0.5 ± 0.2) x 106 . These values are very different from the plumes
(RiR A = 3.2 and CH/3He > 100 x 106 ) which had led to their investigation (Craig et aI.,
1987). Several smaller fields were discovered on a subsequent dive series (Hessler et al.,
1987). The measured temperatures of the hydrothermal vents vary from -6°C to a maxi-
mum of 287°C. The high-temperature chimney structures produce "clear smokers" and
structures dominantly composed of barite and silica, with lesser amounts of sphalerite,
galena, pyrite, and chalcopyrite (Craig et aI., 1987; Kastner et aI., 1987). The composition
of the Mariana vent fluids and chimney structures differs from those of hydrothermal
regions of similar temperatures on mid-ocean ridge spreading centers in several respects
(Campbell et aI., 1987; Kastner et al., 1987; Kusakabe et al., 1990). The pH of the Mariana
vent fluids is higher (4.39) than that for mid-ocean ridge vent fluids (3.8), and the alkalinity
is higher (0.43 meqlL) compared with mid-ocean ridge vent waters (0.19 meqlL; Campbell
et al., 1987). Boron in the vent fluids of the Mariana chimneys is 200% that of ambient
seawater and has a l)lIB of + 20/mil, whereas mid-ocean ridge vents are usually only 10% to
20% that of seawater and have a l)lIB of +26 to + 32/mil (Campbell et al., 1987). It appears
from the compositions of the chimney structures and the Mariana vent fluids that precipita-
266 PATRICIA FRYER
tion of sulfides must be occurring at depth within the ridge crest hydrothermal system,
although no samples of such precipitates were recovered (Campbell et aI., 1987; Kastner
et al., 1987).
Biological studies of the vents indicated "rich and unusual" communities (Hessler
et aI., 1987). Abundant widespread anemones and clusters of barnacles around discrete vent
openings are characteristic of the low-temperature (-6°C) regions. Galatheid crabs and
polynoid polychaetes are associated with the barnacles and anemones in these low-
temperature regions. "Hairy" gastropods cluster around the slightly higher temperature
fields (from -15°C to 250°C). Grazing bresilid shrimp and brachyuran crabs, encrustations
of barnacles, paralvinellid polychaetes and associated harpactacoid copepods, many
limpets, occasional mussels, and dispersed anemones are also found in these locations
(Hessler et al., 1987).
A region of off-axis hydrothermal mounds was discovered approximately 50 km west
of the Mariana backarc spreading center by heat flow and deep-tow studies (Leinen and
Anderson, 1981; Hobart et al., 1983; Lonsdale and Hawkins, 1985). The mounds were
examined by Alvin submersible dives in 1987 (Leinen et aI., 1987; Wheat and McDuff,
1987). They show heat flow values exceeding 10 W/m. Nonlinear, concave thermal gradi-
ents indicate an active hydrothermal system, recharge of the system, and potential escape of
fluids along fault scarps within the region (Leinen and Anderson, 1981; Hobart et aI., 1983;
Leinen et aI., 1987).
The distribution of the high-heat-flow areas and the associated mounds appears to be
controlled by faulting, not by evenly disseminated hydrothermal activity. The mounds are
structurally very complex and associated heat flow values indicate variability even within a
given mound. The distribution of the mounds and the local topography confirm the
structural control over the formation of the mounds (Leinen et aI., 1987).
4.3. Southern Platform

Some of the largest volcanoes of the southern arc lie in the Mariana Trough at about
BON (Fig. 6.11). These volcanoes lie at the intersection of the island arc and the backarc
basin spreading center (Fryer, 1993). It may be that within this region both backarc basin
lavas and arc lavas are generated and erupt from an interrelated plumbing system, as has
been suggested for the northernmost Mariana backarc basin active rifting region (Baker,
1992). The north-south trending volcanic ridges that form at the eastern boundary of the
spreading center give clear evidence that magma from the major volcanic centers is leaking
along the fault zones, at least at the edge of the spreading center. An intersection of a north-
south trending fault mapped in the forearc (Fig. 6.11) with the rift axis would facilitate both
magma egress and permit arc or backarc basin magmas to follow the north-south trending
fault southward into the forearc region (Fryer, 1993). The volcano that lies at the southern
margin of the spreading valley is offset by an east-west striking fault with right lateral
displacement (Fryer, 1993). Fryer (1993) has noted that the latitude of the offset of this
volcano is nearly identical to that of a zone of east-west striking faults identified by Karig
and Ranken (1983) in the forearc east of Guam at BON. East-west trending faults extending
across the arc into the backarc basin may also facilitate magma egress at the spreading
center (Fryer, 1993). Most likely all three phenomena playa role in the development of the
volcanoes of the backarc spreading center at BON (Fryer, 1993).
Recent submersible dives with the Shinkai 6500 on volcanoes of the southern platform
revealed a region of hydrothermal chimneys at the summit region of the eastern volcano
observed using the Shinkai 6500 submersible (Fryer, 1993). The chimney samples are
composed principally of marcasite (Fryer, 1993). The temperature of the vent from which
this chimney sample was obtained was 204°C (Johnson et aI., 1993). Biological commu-
nities observed in association with this vent and others in the vicinity are dominated by the
same type of hairy gastropod (Alvinensis hesslerii) observed at the 18°N vent fields. Small
numbers of shrimps and of galatheid crabs were also observed at the vent sites. Additional
work on the composition of the hydrothermal deposits is underway.
5. PETROLOGY AND PETROGENESIS
5.1. Introduction
In the Mariana convergent margin system, arc magmas are generated as a consequence
of interaction of the mantle overlying the subducted slab with constituents driven off the
descending slab as subduction proceeds (e.g., Fryer, 1981; Fryer et aI., 1981; Sinton and
Fryer, 1987; Hawkins et aI., 1990; Stem and Bloomer, 1992). These arc magmas have
erupted within the forearc over the entire history of its formation, although predominantly
in the early stages (Marlow et al., 1992; Stem and Bloomer, 1992). Today the arc magmas
are predominant in the volcanic front of the arc and produce volcanic seamounts along
lineaments that crosscut the arc and extend into the backarc basin (Hussong and Fryer,
1983). Although arc magmas do occur along the rifting and spreading regions of the
Mariana Trough (Fryer, 1981; Macdougall et aI., 1987; Volpe et aI., 1990), magmas similar
to mid-ocean ridge basalt (MORB) predominate and constitute the newly forming
lithosphere of the spreading portions of the backarc basin.
The similarity between Mariana Trough basalts and MORB is widely recognized (e.g.,
Hart et aI., 1972; Hawkins, 1977; Natland and Tamey, 1981; Hawkins and Melchior, 1985;
Stem et al., 1990). Mariana Trough backarc basin basalt (BABB), as defined by Fryer
(1981), is indistinguishable from normal MORB on trace element provenance diagrams
(Fig. 6.14) used to discriminate between magmas generated in a variety oftectonic settings
(Johnson and Fryer, 1990). Glass selvages from basalts of the Mariana Trough at 18N,
however, do show consistent differences from MORB (Fig. 6.15). The definition of the term
&OO~-------------------------,
400
E
Co
FIGURE 6.14. TiN plot of lavas from the Mariana .S:

Trough, mid-ocean ridges and Izu-Bonin backarc > 200
regions (Fryer et al., 1990). Squares are Mariana
Trough glasses from Sinton and Fryer (1987), circles
are Izu-Bonin backarc basin samples, and triangles
are Izu-Bonin island-arc samples from Fryer et al.
(1990). The outlined fields are the fields of MORB 10000 20000 30000
and IAT from Shervais (1982). TI (ppm)
268 PATRICIA FRYER
3.8 3.0
3.4 2.5
3.0 2.0
Na 20 TIO 2
2.6 1.5
2.2 1.0
0.5 ~ _ _ _ _ _ _+-_ __
1.8
4 5 6 7 8 9 10 4 5 6 7 8 9 10
MgO (wt. ~)
18
16
14
Fe 0
2 3 12
10
8
4 5 6 7 8 9 10
FIGURE 6.15. MgO variation diagrams for Mariana Trough basalts (squares) compared with basalts from the
East Pacific Rise (stippled field) (Fryer et al., 1990).
BABB (Fryer, 1981) is based on the fact that, at equivalent MgO contents, glasses from
Mariana Trough BABB have higher AIP3' lower FeO, slightly lower Ti0 2, and slightly
higher Na20 than MORB. This subtle difference in composition is caused by a slab-related
component involved in formation of the Mariana BABB. The nature of this component is
the subject of some debate, as we shall see. The problem is in determining whether this
component contains sediments subducted with the slab, represents small amounts of
melting of the slab itself, or represents purely a volatile component (Meijer, 1976, 1980;
Hawkins and Melchior, 1985; Sinton and Fryer, 1987).
The relationship between BABB and island-arc magmas is also controversial. It is a
particularly challenging problem to determine the nature of this relationship in the initial
stages of formation of the backarc basin, because both types of magmas can be generated in
a backarc basin even in the earliest stages of arc rifting (Fryer et ai., 1990; Stem et ai.,
1990). The compositional variability in each can be substantial. The northern Mariana
Trough region is an excellent place to examine this problem because it is in the initial stages
of rifting, contains a well-developed VTZ on the eastern side, and has a large, active
volcanic cross-chain. In addition, it is well sampled and there exist detail,ed bathymetry and
side-scan data with which to place the results of the sample analysis in context.
5.2. Northern Rifting Basin and the Kasuga Cross-Chain

The characteristics of the lavas of the northern Mariana Trough are that they are
enriched in large-ion lithophile elements (LILE), are depleted in high-field-strength cat-
ions, and have more radiogenic Sr and less radiogenic Nd than MORB (Stem et ai., 1990).
The general similarity of the northern Mariana Trough lavas to arc lavas suggested to Stem
et ai. (1990) that either an arc component becomes incorporated in a MORB-like source by
mixing or that variable metasomatism of that source by subduction-related fluids produces
the northern Mariana Trough BABB. Although the lavas recovered in dredging and
submersible sampling from the northern Mariana Trough are similar to arc magmas for the
most part, the presence of lavas with BABB compositions and lavas with unusual enrich-
ments in subduction-related components (Jackson, 1989; Fryer et ai., 1990) argues for
caution in developing models for petrogenesis in this region. The necessity for multiple-
source compositions for the northern Mariana Trough lavas is clearly indicated by the
regional diversity of the magmas. This may explain local variations in the general composi-
tional trends that have been noted for some of the major volcanic edifices of the northern
Mariana Trough.
Arc magmas have erupted from fissures on the western boundary of the VTZ in the
eastern part of the northern Mariana Trough (Jackson, 1989; Stem et aI., 1990). Magmas of
both arc and backarc basin composition have erupted from the large volcano (seamount A)
at 22°39'N and from a volcanic ridge immediately north of this volcano at the northwestern
edge of the VTZ in the northern Mariana Trough (Jackson, 1989; Baker, 1992). Lavas from
the summit of seamount A are of two compositions: a BABB and a primitive, low-Ti02
basalt that lacks plagioclase phenocrysts (Jackson, 1989). The source of this low-Ti0 2
basalt was suggested by Jackson (1989) to be similar to that involved in production of
forearc boninitic suites. Either a metasomatic fluid rich in subduction-related components
or a low-degree melt of some exotic mantle component, or both, would be required for the
formation of such lavas. This is the only site in the Mariana Trough where lavas of both
a BABB and a "boninitic" source have been erupted in the same volcano.
Both arc basalts and BABB have erupted from the currently active fissures of the
eastern portion of the Mariana Trough at about 22°N. The rocks dredged in the northern
Mariana Trough (locations indicated in Fig. 6.4), revealed BABB both north and south of
22°N (Jackson, 1989; Stem, 1990). Jackson (1989) suggested that these lavas were formed
by a similar degree of melting, but at higher temperatures and pressures than MORB and
rocks collected at 18°N.
The proximity of the VTZ to the active volcanism of the frontal arc in the northern
Mariana Trough presents a challenge in unraveling the petrogenetic interrelationships of
the lavas. This problem is well demonstrated by the lavas of the Kasuga cross-chain. Most
lavas from these volcanoes are basalts, although andesites and dacites are present at the
summits. The most striking geochemical characteristic of the Kasuga volcanoes is that the
along-chain variability of the Kasuga lavas (Jackson, 1989) is as great or greater than the
regional variability of lavas from the NSP (Bloomer et aI., 1989b; Lin et ai., 1989, 1990;
Ikeda and Yuasa, 1989) from Uracas to Iwo Jima. For example, the Kasuga cross-chain
contains both lavas with the most alkalic composition of any lavas from the NSP and lavas
similar to the least alkalic lavas of the elP (Jackson, 1989; Stem et aI., 1990; Gill and
Williams, 1990). Lavas from Kasuga 1 at the volcanic front are most similar to the least
alkalic of the lavas from the NSP volcanoes and the high-K lavas of Kasuga 2 are most
similar to the most alkalic lavas of Iwo Jima.
If Kasuga lavas were to have been generated by the same processes and from the same
sources as the lavas erupting along the VTZ immediately north of the chain, they should
have similar compositions. Yet all Kasuga rock types differ strikingly from those dredged
at the pillow ridges along the volcanic and tectonic zone (Stem et aI., 1990). The wide
270 PATRICIA FRYER
range in compositions of lavas from the Kasuga seamounts of the northern Mariana Trough
probably reflects a combination of heterogeneous processes of melt production or source
composition, or both (Jackson, 1989). Several source components have been invoked to
explain the isotopic diversity of the northern part of the Mariana arc and trough (Bloomer
et aI., 1989a; Lin et al., 1989, 1990; Hickey-Vargas, 1991; Hickey-Vargas and MandaI,
1992).
The variation in K20 content of the Kasuga lavas west of the volcanic front partly
results from variations in the degree of partial melting and partially from a variety of
sources in the northern Mariana Trough region. Higher-K basalts of Kasuga 2 and Kasuga 3
probably preferentially sampled a local source component that shares distinctive geochemi-
cal features with some boninites: low ENd accompanied by high K, Rb, Zr, P, and LREE
relative to Sr, Ba, Pb, HREE, Y, Ti, and Nb (Bloomer et al., 1989a,b; Jackson 1989; Lin
et aI., 1989, 1990; Gill and Williams, 1990; Stem et aI., 1990). A component of this type
apparently contributes to the diversity oflavas in the NSP as a whole. The same component
appears to be further diluted in lavas within comparably near-front, active volcanic ridges
in the Mariana Trough at 22°N (Stem et aI., 1990). Seamount A on the northeast tip of the
volcanic platform of the Mariana Trough at 22°39'N yielded a lava similar in composition
to the boninitic series lavas of the forearc (Jackson, 1989). Thus, this component may be
better represented in the northern backarc basin than recognized by Stem et al. (1990). The
association of K, Zr and LREE enrichment with low BalLa and eNd ratios is common in
subduction zones, especially behind volcanic fronts (e.g., Carr et aI., 1990; Tatsumi et aI.,
1991). Atypical tectonics allows atypically pure melts of this exotic component to erupt.
The cross-arc structures such as the faults upon which volcanic cross-chains such as the
Kasuga volcanoes develop are likely to provide a better opportunity to explore the hetero-
geneity of the mantle wedge in the backarc basin settings. The diversity of rock types in
close proximity to each other argues for caution in generalizations correlating composi-
tion of the rocks with tectonic setting. The distributions of varying magma types in the
region probably reflect the interactions between the incipient spreading zones and the
rifting arc lithosphere both in the backarc basin and, in the case of the Kasuga seamounts,
across the forearc and arc.
5.3. Central Spreading Basin

Most of the detailed petrologic work in the central part of the Mariana Trough has been
done in the spreading valley at about 18°N. Dredges and DSDP samples defined the
principal characteristics of BABB and pointed out that arc magmas have been erupted in the
central spreading region ofthe Mariana Trough (Fryer et aI., 1981; Wood et aI., 1981, Sinton
and Fryer, 1987). An Alvin submersible dive series on the volcanic ridge that lies in the
central spreading valley provided structural and petrologic studies of the rift center volca-
noes (Hawkins et aI., 1990).
The controversy over the origin of BABB arises as a consequence of the fact that
compositional differences between BABB and MORB may have several causes. The major
element differences are easily explained. Higher H20 content in the source region for the
Mariana Trough backarc basin basalts displaces the olivine-plagioclase-liquid cotectic
toward plagioclase in the clinopyroxene-olivine-plagioclase-Si02 basalt system, resulting
in magmas with higher Al 20 3 content in olivine-saturated backarc basin liquids relative to
(olivine + plagioclase)-saturated basalt liquids from normal mid-ocean ridges (Fryer, 1981;
Fryer et al., 1981; Sinton and Fryer, 1987). The fact that the Mariana Trough basalts have
10w-FeO contents at equivalent MgO by comparison with normal MORB, probably reflects
lower temperature and pressures of melting (Sinton and Fryer, 1987). Although Stem et al.
(1990) have noted that the Mariana BABB lacks enrichment in LIL elements, Sinton and
Fryer (1987) observed enrichments in glasses from the central Mariana Trough (18°N) and
suggested that several possible phenomena may contribute to the source of the BABB
magmas, principal among these is a fluid component from the subducted slab that has been
active in metasomatism of an otherwise depleted mantle, as well as a small amount of slab
melting (Sinton and Fryer, 1987).
Fryer (1981) noted that some lavas from the spreading ridge segment to the north of the
Pagan transform (Fig. 6.9c) have compositions intermediate between the typical BABB
of the region and Mariana island-arc tholeiite; however, the spatial distribution of samples
from the dredge sites was too poorly constrained spatially to permit interpretations of this
observation. Alvin submersible investigations of the volcanic ridge in the Mariana Trough
spreading center valley at 18°N (Macdougall et aI., 1987; Volpe et al., 1990) recovered
basalts that span a range between BABB and island-arc tholeiite. Isotopic analyses and the
Rb and Sr concentrations of the basaltic glasses from some of the samples show an intimate
mixing of the MORB source with an arc1ike source component (Macdougal et al., 1987).
The lavas recovered during the dives confirm the general compositional trends described by
Fryer et al. (1981), but provide far greater detail in terms of distribution of various rock
types along axis. Lavas with greatest similarity to MORB occupy the ridge segment
immediately north of the Pagan transform, those of more typical BABB composition are
located on the ridge segment further north from the transform, and the arc1ike lavas occupy
one position on the northern segment of the spreading ridge (Macdougall et al., 1987; Volpe
et al., 1990). Volpe et al. (1990) interpret the diversity in the rock types on such small scale
as indicating that the source region for the lavas is equally diverse. They suggest that the
heterogeneity of the source may indicate the presence of fragments of previously subducted
lithosphere in the source region. This would imply variability on a very small scale for the
sources of backarc basin lavas, and, as we have shown, the studies of the northern portion of
the Mariana Trough would support this interpretation. Furthermore, the studies of the
Kasuga lavas support the contention that an unusual local source has contributed to the
magmas of the backarc basin. It may be that in the incipient phases of opening of the
basin and before new plate boundaries are established the heterogeneities of the backarc
basin sources are most prevalent.
5.4. Southern Platform

The seamounts of the frontal arc south of 16°30'N lie at the base of the fault scarp that
forms the eastern boundary of the Mariana Trough, but are better separated from that scarp
than are the islands of the CIP or the seamounts of the NSP. The lavas of the SSP edifices
show a compositional diversity similar to that in the NSP and greater than that of the CIP
(Dixon and Stem, 1983; Stem and Bibee, 1984; Stem et aI., 1989).
Bloomer and Hawkins (1983) show that arc-related rocks ranging from ultramafics to
dacites are exposed in the inner trench wall region south of 12°N. The petrologic and
geochemical results of their study indicate that no true mid-ocean ridge material is exposed
in the inner trench wall of the Mariana convergence zone within their sample area. This
prompted them to suggest that most of the crust of the southern backarc basin is composed
of pre-late Eocene arc complex. These findings are consistent with an evolutionary model
for the backarc platform in the southern part of the Mariana Trough that requires it to be
272 PATRICIA FRYER
composed of fractured, subsided, and possibly rotated blocks of the arc massif. Very little
information exists on the nature of the transition from the composition of rocks from this
deformed arc complex to the lavas of the active spreading zone to the north.
Additional rock samples were collected from the forearc fault scarp south of Guam
during recent Shinkai 6500 dives (Fryer, 1993). The lavas recovered are primarily massive
mafic rocks: basaltic and gabbroic. Details of the compositions are given by Fryer (1993),
but all are of island-arc composition.
6. SUMMARY AND CONCLUSIONS
The Mariana Trough is an active backarc basin that represents the latest in a series
of arc rifting and spreading events that have taken place at the eastern edge of the Philippine
Sea plate. During its opening, tectonic and, to some extent, volcanic activity sometimes
occurred basinwide, although the locus of greatest activity remained close to the volcanic
front.
An initial period of stretching and collapse of the preexisting arc is indicated by a
broad normal-faulted terrain, particularly at the western margin of the basin. During this
period of extension, deformation and volcanism were focused on a VTZ close to the
volcanic front. Subsequently, ridge transform structures developed as seafloor spreading
began. Arc construction continued on the eastern rifted crust. The basin thus has the
appearance of having spread asymmetrically, even in the more mature portion of the basin
(between 14° and 200N), but in reality the basin rifted asymmetrically. Compositional
variations of lavas within the basin suggest both local tectonic control over magma genesis
and an intricate interplay of mixing from several magmatic sources. The fundamental
aspects of basin development can be traced in each of three major geographic provinces of
the Mariana Trough, the northern rifting apex, the central spreading basin, and the southern
platform.
North of 200N, graben-and-horst structures predominate, and maximum depths are in
excess of 3500 m. Either the rifting of the arc in this basin is propagating northward and the
northern part of the Mariana Trough has only recently begun to separate, or rifting has been
going on much more slowly in this part of the basin, possibly pivoting about a point at the
northern apex of the basin. The volcanic-tectonic zone remains within 15 km of the frontal
arc between 234° and 22°N; southward of that latitude the zone diverges from the arc, and
by 21°30'N it is 70 km west of the arc line. At this point the volcanic-tectonic zone
undergoes a transition to a series of half-graben structures and at about 200N there is a
change to true seafloor spreading farther south. The composition of lavas within the
northern region is more variable than in any other portion of the basin studied to date. Both
island-arc tholeiite basalts and backarc basin basalts erupt from fissures within the currently
active VTZ, and one dredge recovered lavas that are similar to those from boninitic source
magmas from forearc regions. Some fault scarps within the basin are mantled with recent
lava flows. A few are also the sites of construction of major volcanic edifices in the backarc
basin and thus give rise to volcanic cross-chains. Some of these major edifices appear to
have supplanted the adjacent volcanic front as the primary locus of along-strike eruption of
arc magmas.
The spreading center is best developed between about 14° and 200N, in the widest part
of the basin. The bathymetric grain reflects the structures of the spreading axis and
transform fault valleys. Abyssal hill fabric, parallel to the ridge segments, is offset by broad
valleys that cut across the basin and are continuous with lateral offsets between the
spreading ridge segments. Structural lineaments exist even in the early block-faulted terrain
at the western margin of the basin. The tectonic control effected by these preexisting
structural elements apparently persisted throughout both the early rifting and the later mid-
ocean ridge/transform spreading stages of opening in the basin. Cross-chains of submarine
volcanoes extend westward from volcanic centers at the volcanic front into the backarc
basin along some of these lineaments, suggesting that they are avenues for migration of arc
magmas into the backarc basin. Magmas of the spreading ridge segments are composed of
basalts that are very similar to mid-ocean ridge basalts. They differ from normal mid-ocean
ridge basalts in having major element compositions affected by the higher volatile content
of the source and in having trace element signatures indicative of a component of slab-
related fluids. Arc lavas and basalts transitional to arc lavas have also been collected at a
few localities from the central spreading ridge segments. This diversity of lava types
suggests a complex interplay of sources even in the more mature central portion of the
basin.
Hydrothermal systems of the Mariana Trough are well developed along the volcanic
fissures that mark the spreading ridge segments. Well-developed hydrothermal systems are
also evident on the large volcanic centers of the southernmost intersection of the spreading
axis with the volcanic front. These systems produce hydrothermal deposits which reflect
the cooler temperatures (by comparison with the mid-ocean ridge systems) observed in the
basin hydrothermal fields. The presence of off-axis hydrothermal systems suggests that
magmatic activity is not confined strictly to the ridge crests even in the most mature portion
of the central spreading basin. Hydrothermal systems of the volcanoes along the frontal arc
and the cross-chains are similar to one another and differ from those of the backarc basin
spreading center only in subtle geochemical characteristics related to the individual edi-
fices.
South of 14°N the basin is marked by a shallow (less than 3 km) platform. The
southern platform is separated from the active arc and the remnant arc by deep fault-
controlled grabens. The southern platform is furrowed by a series of narrow ridges and
troughs. The strike of these features changes from roughly north-south to roughly east-west
with decreasing latitude. In the western margin of the southern platform there is a series of
ridges and troughs with a trend that suggests a possible earlier strain regime that is different
from the rest of the southern platform. The position of the spreading center south of 14 ON
has recently been defined by side-scan surveys, swath mapping, and submersible observa-
tions at 13°40'N to be close to the volcanic front. At that latitude the rift valley is a narrow
feature only approximately 15 km wide. Spreading at nOIO'N, 144°IO'E in the south-
ernmost portion of the backarc basin is taking place in a broader zone and may represent the
intersection of the spreading axis with the frontal arc line. Large volcanic centers lie atop
volcanic ridges that are only 30 km or less from the eastern bounding fault of the backarc
basin. Preliminary studies of lavas collected recently from these features indicate BABB
from two sources are being erupted from one of the edifices on the eastern portion of the
spreading ridge (Johnson et al., 1993). There is evidence of tectonic disturbance in sedi-
ments near the western boundary of the spreading valley, as observed from submersibles,
and thus the entire valley may be tectonically active. Southward of this position the
spreading axis is not well defined but may be influenced by forearc faulting. A major N-S
striking fault in the Mariana forearc lies south of the spreading axis and is offset about 15
274 PATRICIA FRYER
km to the east of the easternmost volcanic ridge. Exposures of fresh basalts in the forearc
along this fault scarp suggest that the magmas generated in the southern portion of the arc
may be involved to some extent in forearc volcanism.
Understanding the history of suprasubduction zone tectonics and magma genesis in
the Mariana Trough is important for the ultimate understanding of geochemical mass
balance in convergent margin systems and of evolution of continents. Profound changes in
the models of arc evolution have resulted from recent studies of the Mariana interoceanic
arc system. Vast portions of the suprasubduction zone have been subjected to metamor-
phism driven primarily by fluid flux through that region in response to dehydration of the
subducted oceanic slab. Clearly, variations in source composition as a consequence of
previous subduction episodes, possible incorporation of fragments of older lithosphere,
migrations of mantle sources, and influx from variations in convection play a part in the
generation of magmas in this region. The wide variety of rock types recovered from the
Mariana Trough and the discovery of a new model for evolution of the backarc basin
requires us to reevaluate our concepts of the petrogenetic evolution of the backarc crust and
mantle. The study of the lavas of the northern Mariana arc and of the Mariana backarc basin
requires us to evaluate the interaction of magmas generated by shallow partial melting of an
exotic source with those of the deeper-seated arc magma sources and to try to trace the
influence on both of the subduction derived constituents. The apparently shallow generated
backarc basin basalts are emplaced in close proximity to true arc-generated lavas. Mixing
and magma differentiation processes on cross-chain volcanoes in particular tend to obscure
the end-member source components in the lavas. The importance of understanding these
regions is becoming increasingly evident as more and more ophiolite terrains are identified
as having formed in an arc or suprasubduction zone environment. Thus, the comparisons of
subaerial terrains to the recent studies of the interoceanic convergent margins will continue
to add immeasurably to our understanding of plate interactions and of the evolution of the
volcanic arcs and backarc basins generated within such environments.
Acknowledgments
The author is indebted to F. Martinez for his invaluable assistance in producing several
of the bathymetry maps. Thoughtful reviews of this manuscript by J. Natland and A. Klaus
and comments by B. Taylor and R. Stem have improved it significantly. This manuscript is
SOEST contribution no. 3581 and HIGP contribution no. 767.
REFERENCES
Anderson, R. N. 1975. Heat flow in the Mariana marginal basin, J. Geophys. Res., 80:4043-4048.
Baker, N. 1992. Rifting and volcanism in the northern Mariana Trough: a SeaMARC II and seismic reflection
study, Master's thesis, Univ. Hawaii.
Baker, N., Fryer, P., and Brown, G. 1989. Rifting in the northern Mariana backarc basin (abstract), EOS Trans.
AGU, 70:1313.
Ballance, P. F., Carey, S., Sigurdsson, H., and Fryer, P. 1988. Production of basaltic sediment on Kick'em Jenny
submarine volcano, Lesser Antilles arc, and central seamount, Mariana arc (abstract), Abstracts with
program, Geo!. Soc. N.Z., Ann. Mtg., Geo!. Soc. N.Z. Misc Pub!. 4Ia:30.
Barone, A. M., Lonsdale, P. F., and Puteanus, D. 1990. Formation of axial volcanic ridges of the Mariana slow
spreading center from spatial magma sources (abstract), EOS Trans. AGU 71:1637-1638.
Beal, K. L. 1987. The opening of a backarc basin: northern Mariana Trough, Master's thesis, Univ. Hawaii.
Bloomer, S. H., and Hawkins, J. W. 1983a. Gabbroic and ultramafic rocks from the Mariana trench: an island arc
ophiolite, in The Tectonic and Geologic Evolution of Southeast Asian Seas and Islands: Part II. (D. E. Hayes,
ed.) Geophys. Monogr. Ser., Vol. 27, pp. 294-317, American Geophysical Union, Washington, DC.
Bloomer, S. H., and Hawkins, J. W. 1983b. Petrology and geochemistry ofboninite series volcanic rocks from the
Mariana trench, Contrib. Mineral. Petrol. 97:361-377.
Bloomer, S. H., Stern, R. J., Fisk, E., and Geshwind, C. H. 1989a. Shoshonitic volcanism in the northern Mariana
Arc. 1: mineralogic and major and trace element characteristics, l. Geophys. Res. 94 (B4):4469-4496.
Bloomer, S. H., Stern, J. R., and Smoot, N. C. 1989b. Physical volcanology of the submarine Mariana and volcano
arcs, Bull. Volcanol. 51:210-224.
Bracey, D. R., and Ogden, T. A. 1972. Southern Mariana arc geophysical observations and hypothesis of
evolution, Bull. Geol. Soc. Am., 83:1509-1522.
Brown, G. 1991. Rifting of the Bonin Island arc, Doctoral dissertation, Univ. Hawaii.
Brown, G., and Taylor, B. 1988. Seafloor mapping of the Sumisu Rift, Izu-Ogasawara (Bonin) island arc, Bull.
Geol. Surv. lpn. 39:23-3.
Bibee, L. D., Shor, G. G., Jr., and Lu, R.S. 1980. Inter-arc spreading in the Mariana Trough, Mar. Geol. 35:
183-197.
Campbell, A. C., Edmond, J. M., Colodner, D., Palmer, M. R, and Falkner, K. K.1987. Chemistry ofhydrothennal
fluids from the Mariana Trough backarc basin in comparison to mid-ocean ridge fluids (abstract), EOS Trans.
AGU 68(44):1531.
Carr, M. J., Feigenson, M. D., and Bennett, E. A. 1990. Incompatible element and isotopic evidence for tectonic
control of source mixing and melt exteraction along the Central American Arc, Contrib. Mineral. Petrol.
105:369-380.
Chorowicz, J., Le Fournier, 1., and Vidal, G. 1987. A model for rift development in eastern Africa, Geol. l.
22:495-5\3.
Cochran, 1. E, and Martinez, E 1988. Evidence from the northern Red Sea on the transition from continental to
oceanic rifting, Tectonophysics 153:25-53.
Craig, H., Horibe, Y., and Farley, K. A. 1987. Hydrothennal vents in the Mariana Trough: results of the first Alvin
dives (abstract), EOS Trans. AGU 68(44):1531.
Dixon, T. H., and Batiza, R. 1979. Petrology and chemistry of recent volcanics in the northern Marianas, Contrib.
Dixon, T. H., and Stern, R. 1. 1983. Petrology, chemistry and isotopic composition of submarine volcanoes in the
southern Mariana arc, Geol. Soc. Am. Bull. 94:1159-1172.
Edwards, M. H., Fornari, D. J., Malinverno, A., and Ryan, W. B. E 1991. The regional tectonic fabric of the East
Pacific Rise from 12°50'N to 15°IO'N, l. Geophys. Res. 96:7995-8017.
Eguchi, T. 1984. Seismotectonics around the Mariana Trough, Tectonophysics 102:33-5.
Fryer, P. 1981. Petrogenesis of basaltic rocks from the Mariana Trough, Doctoral dissertation, Univ. Hawaii.
Fryer, P. 1986. Tectonics of the BoninlMariana arc intersection and effects on alkaline magmatism of the volcano
islands (abstract), EOS, Trans. AGU 67(44):1277.
Fryer, P. 1989. Tectonic evolution of the Mariana backarc basin, Abstracts Vol. I, 28th Int. Geol. Cong.,
Washington, DC, p. 51.
Fryer, P. 1990. Recent marine geological research in the Mariana and Izu-Bonin island arcs, Pac. Sci. 44(2):
59-114.
Fryer, P. 1993. The relationship between tectonic defonnation, volcanism, and fluid venting in the southeastern
Mariana convergent margin, Proc. JAMSTEC Symp. Deep Sea Res., Vol. 9, pp. 161-179.
Fryer, P., Beal, K. L., and Jackson, M. C. 1986. Volcanism of the northern Mariana backarc rift (abstract), EOS
Trans. AGU 67(44):1239.
Fryer, P., Gill, J., Jackson, M., Fiske, R., Hochstaedter, A., McMurtry, G., Sedwick, P., Malahoff, A., Mouginis-
Mark, P., and Horikoshi, E. 1987. Results of recent Alvin studies of the Kasuga volcanoes, northern Mariana
arc (abstract), EOS Trans. AGU 68(44):153.
Fryer, P., Beal, K. L., and Jackson, M. C., 1986. Volcanism of the northern Mariana backarc rift (abstract), EOS
Trans AGU 67(44):1239.
Fryer, P., and Hussong, D. M. 1981. Seafloor spreading in the Mariana Trough: results of Leg 60 drill site selection
surveys, in In it. Repts. DSDP, 60 (D. M. Hussong and S. Uyeda et al., eds.), pp. 45-55, U.S. Govt. Printing
Office, Washington, DC.
Fryer, P., and Hussong, D. M. 1982. Arc volcanism in the Mariana Trough (abstract), EOS Trans. AGU 63:1135.
276 PATRICIA FRYER
Fryer, P., Saboda, K. L., Johnson, L. E., MacKay, M. E., Moore, G. F., and Stoffers, P. 1990. Conical seamount:
SeaMARC II, Alvin submersible, and seismic reflection studies, in Proc. OD?, Init. Repts. 125. (P. Fryer, J. A.
Pearce, L. B. Stokking, et al., eds.), pp. 69-80, Ocean Drilling Program, College Station, TX.
Fryer, P., Sinton, J. M., and Philpotts, J. A 1981. Basaltic glasses from the Mariana Trough, in Init. Repts. DSDP
(D. M. Hussong and S. Uyeda et aI., eds.), pp. 601-607, U.S. Govt. Printing Office, Washington DC.
Gill, J. B., and Williams, R 1990. Th isotopes and U-series disequilibrium in subduction-related volcanic rocks,
Geochim. Cosmochim. Acta 54(5):1427-1442.
Hagen, R, Shor, A, and Fryer, P. 1992. SeaMARC II evidence for the locus of seafloor spreading in the southern
Mariana Trough, Mar. Geol. 103:311-322.
Hart, S. R, Glassey, W. E., and Karig, D. E. 1972. Basalts and seafloor spreading behind the Mariana island arc,
Hawkins, J. W. 1977. Petrology and geochemical characteristics of marginal basin basalt, in Island Arcs, Deep Sea
Trenches, and Back-Arc Basins, (D. E. Hayes, ed.), Geophys. Monogr. Ser., Vol. 23, pp. 355-365, American
Geophysical Union, Washington, DC.
Hawkins, J. w., Lonsdale, P. F., Macdougall J. D., and A. M. 1990. Petrology of the axial ridge of the Mariana
Trough backarc spreading center, Earth Planet. Sci. Lett. 100(113):226-250.
Hawkins, 1. w., and Melchior, 1. T. 1985. Petrology of Mariana trough and Lau basin basalts, 1. Geophys. Res.
90:11,431-11,468.
Hessler, R R., France, S. C., and Boudrias, M. A. 1987. Hydrothermal vent communities of the Mariana backarc
basin (abstract), EOS Trans. AGU 68(44):1531.
Hickey-Vargas, R 1991. Isotope characteristics of submarine submarine lavas from the Philippine Sea: implica-
tions for the origin of arc and basin magmas of the Philippine plate, Earth Planet. Sci. Lett. 107:290-304.
Hickey-Vargas, R, and Mandai, G. K. 1992. The changing contribution of mantle and slab derived components in
arc and backarc basin magmas: a model based on the Eocene to Recent, Palau-Kyiushu-West Mariana-
Mariana arc system (abstract), EOS Trans. AGU 73:342.
Hobart, M. A, Anderson, R. N., Fujii, N., and Uyeda, S. 1983. Heat flow from hydrothermal mounds in two
million year old crust of the Mariana Trough which exceeds two watts per meter (abstract), EOS Trans. AGU
64:315.
Horibe, Y., Kim, K. Craig, H. 1986. Hydrothermal methane plumes in the Mariana backarc spreading center,
Nature 324:131-133.
Hsui, A. T., and Youngquist, S. 1985. A dynamic model of the curvature of the Mariana Trench, Nature
318(6045):455-457.
Hussong, D. M., and Fryer, P. 1983. Back-arc seamounts and the SeaMARC II seafloor mapping system (abstract),
EOS Trans. AGU 64(45):627-632.
Hussong, D. M., and Sinton, J. B. 1983. Seismicity associated with back-arc crustal spreading in the central
Mariana Trough, in The Tectonic and Geologic Evolution of Southeast Asian Seas and Islands: Part 2, (D. E.
Hayes, ed.), Geophys. Monog. Ser., Vol. 27, pp. 217-235, American Geophysical Union, Washington, DC.
Hussong, D. M., and Uyeda, S. 1981. Tectonic processes and the history of the Mariana arc, a synthesis of the
results of Deep Sea Drilling leg 60, in Init. Repts, DSDP, 60 (D. M. Hussong and S. Uyeda et aI., eds.) pp.
909-929, U.S. Gov!. Printing Office, Washington, DC.
Ikeda, Y., and Yuasa, M. 1989. Volcanism in nascent back-arc basins behind the Shichito Ridge and adjacent areas
in the Izu-Ogasawara arc, northwest Pacific: evidence for mixing between E-type MORB and island arc
magmas at the initiation of backarc rifting, Contrib. Mineral. Petrol. 101(4):377-393.
Ishii, T., Kobayashi, K., Shibata, T., Naka, 1., Johnson, K., Ikehara, K., Iguchi, M., Konishi, K.,Wakita, H., Zhang,
F., Nakamura, Y., and Kayane, H. 1985. Description of samples from Ogasawara forearc, Ogasawara Plateau
and Mariana Trough, during KH84-1 cruise, 1985, in Preliminary Report of the Hakuho Maru Cruise
KH84-1; Geological and Geophysical Investigation of Seafloor East of Ogasawara (Bonin) Islands, North-
ern Mariana Trough, Southern Mariana, Yap, and Palau Trench and Arcs (K. Kobayashi, ed.), pp. 89-168,
Ocean Research Insititute, University of Tokyo.
Jackson, M. C. 1989. Petrology and petrogenesis of recent submarine volcanoes from the northern Mariana arc and
back-arc basin, Doctoral dissertation, Univ. Hawaii.
Johnson, L. E., Fryer, P., Masuda, H., Ishii, T., and Gamo, T. 1993. Hydrothermal vent deposits and two magma
sources for volcanoes near 13°20'N in the Mariana backarc basin: a view from Shinkai 6500 (abstract), EOS
Trans. AGU Fall Meeting Suppl. 74(43):681.
Karig, D. E. 1971a. Structural history of the Mariana island arc system, Geol. Soc. Am. Bull. 83:323-344.
Karig, D. E. 1971b. Origin and development of marginal basins in the western Pacific, J. Geophys. Res. 76:2452-
2561.
Karig, D. E. 1974. Evolution of arc systems in the western Pacific, Ann. Rev. Earth Planet. Sci. 2:51-75.
Karig, D. E. 1975. Basin genesis in the Philippine Sea, in Init. Repts. DSDP, 31 (D. E. Karig et aI., ed.), pp. 857-
879, U.S. Govt. Printing Office, Washington DC.
Karig, D. E., Anderson, R. N., and Bibee, L. D., 1978, Characteristics of back arc spreading in the Mariana Trough,
J. Geophys. Res. 83:1213-1226.
Karig, D. E., and Ranken, B. 1983. Marine geology of the forearc region, southern Mariana island arc, in The
Tectonic and Geologic Evolution of Southeast Asian Seas and Islands, Part 2 (D. E. Hayes, ed.), Geophys.
Monogr. Ser., Vol. 27, pp. 266-280, American Geophysical Union, Washington, DC.
Kastner, M., Craig, H., and Sturz, A. in press. Hydrothermal deposits in the Mariana Trough: preliminary
mineralogical investigation (abstracted), EOS, Trans. Am. Geophys. Un. 68:1531.
Klein, G. deY., and Kobayashi, K. 1980. Geologic summary of the north Philippine Sea, based on Deep Sea
Drilling Project Leg 58 results, in Init. Repts. DSDP (G. deV. Klein and K. Kobayashi et aI., eds.) Vol. 58, pp.
951-962, U.S. Govt. Printing Office, Washington DC.
Klaus, A., Taylor, B., Moore, G. E, Murakami, E, and Okamura, Y. 1991. Structural evolution of Hachijo and Aoga
Shima rifts: backarc rifting in the Izu-Bbnin island arc, EOS Trans. AGU, Fall Meeting Suppl., 72(44):243-244.
Kong, L. S. 1993. Seafloor spreading in the Mariana Trough, in Preliminary Report of the Hakuho-Maru Cruise
KH92-1 (J. Segawa, ed.), pp. 5-16, Ocean Research Institute, Univ. Tokyo.
Kong, L. S., Detrick, R. S., Fox, P. 1., Mayer, L. A., and Ryan, W. B. E 1988/89. The morphology and tectonics of
the MARK area from Sea Beam and SeaMARC I: Observations (mid-Atlantic Ridge 23°N), Mar. Geophys.
Res. 10:59-90.
Kusakabe, M., Mayeda, S., and Nakamura, E. 1990. S, 0, and Sr isotope systematics of active vent materials from
the Mariana backarc basin spreading axis at 18°N, Earth Planet. Sci. Lett. 100(1/3):275-282.
Lange, K. 1992. Auswertung eines weitwinkelreflexion- und refraktionsseismischen Profils im sudlichen Mari-
anen Backarc Becken, Diplomarbeit, Institut fur Geophysik, Univ. Hamburg.
Le Pichon, X., Huchon, P., and Barrier, E. 1985. Geoid and the evolution of the western margin of the Pacific
Ocean, in Formation of Active Ocean Margins (N. Nasu, ed.), pp. 3-42, Terrapub, Tokyo.
Lewis, S. D., Hayes, D. E., and Mrozowski, C. L. 1982. The origin of the West Philippine Basin by interarc
spreading (thesis), in Geology and Tectonics of the Luzon-Marianas Region, in Philippine Seatar Committee
Special Publication 1 (G. R. Balce and A. S. Zanoria, eds.), pp. 31-51.
Lin, P.-N., Stem, R. 1., Bloomer, S. H. 1989. Shoshonitic volcanism in the northern Mariana arc. 2: Large-ion
lithophile and rare earth element abundances evidence for the source of incompatible element enrichments in
intraoceanic arcs, J. Geophys. Res. 94(B4):4497-4514.
Lin, P.-N., Stem, R. J., Morris, J., and Bloomer, S. H. 1990. Nd- and Sr-isotopic composition of lavas from the
northern Mariana and southern volcano arcs: implications for the origin of island arc melts, Contrib. Mineral.
Petrol. 105:381-392.
Leinen, M., and Anderson, R. N. 1981. Hydrothermal sediment from the Mariana Trough (abstract), EOS Trans.
AGU 62:914.
Leinen, M., MacDuff, R. E., Delaney, J., Becker, K., and Schultheiss, P. 1987. Off-axis hydrothermal activity in
the mariana mounds field (abstract), EOS Trans. AGU 68(44):1531.
Lonsdale, P., and Hawkins, J. W. 1985. Silicic volcanism at an off-axis geothermal field in the Mariana Trough
backarc basin, Geol. Soc. Am. Bull. 96:940-951.
Macdonald, K. c., and Atwater, T. M. 1978. Evolution of rifted ocean ridges, Earth Planet. Sci. Lett. 39:319-327.
Macdougall,1. D., Volpe, A., and Hawkins, J. W. 1987. An arc-like component in Mariana Trough lavas (abstract),
EOS Trans. AGU 68(44):1531.
Mahoney 1., and Fryer, P. 1988. Stress field re-orientation of the inner Mariana forearc at 22°N (abstract), EOS
Trans. AGU 69(44):1443.
Marlow, MS., Johnson, L. E., Pearce, J. A., Fryer, P., Pickthorn, L. G. and Murton, B. J. 1992. Upper Cenozoic
volcanic rocks in the Mariana forearc recovered from drilling at Ocean Drilling Program Site 781: implica-
tions for forearc magmatism, J. Geophys. Res. 97:15,085-15,098.
Martinez, E, and Cochran, J. R. 1988. Structure and tectonics of the northern Red Sea: catching a continental
margin between rifting and drifting, Tectonophysics 150:1-32.
Martinez, E, Fryer, P., Baker, N., and Yamazaki, T. in press. Evolution of backarc rifting Mariana Trough
20-24°N, J. Geophys. Res.
278 PATRICIA FRYER
McMurtry, G., Sedwick, P., Fryer, P., Vonderhaar, D. L., and Yeh, H.-W. 1993. Unusual geochemistry of
hydrothermal vents on submarine arc volcanoes: Kasuga seamounts, northern Mariana arc, Earth Planet. Sci.
Lett. 114:517-528.
Meijer, A. 1976. Pb and Sr isotope data bearing on the origin of volcanic rocks from the Mariana island arc system,
Geol. Soc. Am. Bull. 87(9):1359-1369.
Meijer, A. 1980. Primitive arc volcanism and a boninite series; examples from western Pacific island arcs, in The
Tectonic and Geologic Evolution of Southeast Asian Seas and Islands (D. E. Hayes, ed.) Geophys, Monogr.
Ser., Vol. 23, pp. 271-282, American Geophysical Union, Washington, DC.
Mrozowski, C. L., Hayes, D. E. 1979. The evolution of the Parece Vela Basin, eastern Philippine Sea, Earth
Mrozowski, C. L., Hayes, D. E., and Taylor, B. 1981. Multichannel seismic reflection surveys of Leg 60 Sites,
Deep Sea Drilling Project, in Init. Repts. DSDP, 60 (D. M. Hussong and S. Uyeda et aI., eds.), pp. 57-71, U.S.
Govt. Printing Office, Washington, DC.
Natland, J. H., and Tamey, J.1981. Petrologic evolution of the Mariana arc and backarc basin system-a synthesis
of drilling results in the south Philippine Sea, in Init. Repts. DSDP (D. M. Hussong and S. Uyeda, eds.), Vol.
60, pp. 877-907, U.S. Govt. Printing Office, Washington DC.
Parson, L. M., and Tiffin, D. L. 1993. Northern Lau Basin: Backarc extension at the leading edge of the Indo-
Australian plate, Geo-Mar. Lett. 13(2):107-115.
Perram, L. M., and Macdonald, K. C. 1990. A one-million-year history of the 11°45'N East Pacific Rise
discontinuity, J. Geophys. Res. 95:21,363-21,381.
Poreda, R. 1985. Helium-3 and deuterium in backarc basalts: Lau and Mariana troughs, Earth Planet. Sci. Lett.
73:244-254.
Seama, N., and Fujiwarra, T. 1993. Geomagnetic anomalies in the Mariana Trough 18°N, in Preliminary Report of
the Hakuho-Maru Cruise KH92-1 (1. Segawa, ed.), pp. 70-71, Ocean Research Institute, Univ. Tokyo.
Searle, R. C., and Laughton, A. S. 1977. Sonar studies of the Mid-Atlantic Ridge and Kurchatov fracture zone, J.
Geophys. Res. 82:5313-5328.
Sender, K. L., Shor, A. N., and Hagen, R. A. 1989. SeaMARC II sidescan processing techniques (abstract), EOS
Trans. AGU 70(43):1304.
Shervais, J. W. 1982. Ti-V plots and the petrogenesis of modem and ophiolitic lavas, Earth Planet. Sci. Lett.
59:101-118.
Shibata, T. and Segawa, K. 1984. Basaltic glasses from the northern Mariana Trough, in Preliminary Report of the
Hakuhu Maru cruise KH84-1 April, 16-May 30 1984 (K. Kobayashi, ed.), pp. 215-224, Ocean Inst., Univ.
Tokyo.
Sinton, J. M., and Fryer, P. 1987. Mariana Trough lavas from 18°N: Implications for the origin of backarc basin
basalts, J. Geophys. Res. 92(12):12,782-12,802.
Smoot, N. C. 1988. The growth rate of submarine volcanoes on the South Honshu and East Mariana ridges, J. Volc.
Smoot, N. C. 1990. Mariana Trough by multi-beam sonar, Geo-Mar. Lett. 10:137-144.
Stem, R. J. 1979. On the origin of andesite in the northern Mariana islands: implications from Agrigan, Contrib.
Stem, R. 1., and Bibee, L. D. 1984. Esmeralda Bank: geochemistry of an active submarine volcano in the Southern
Mariana island arc, Contrib. Mineral. Petrol. 86:159-169.
Stem, R. J., and Bloomer, S. H. 1992. Subduction zone infancy, Bull. Geol. Soc. Am. 104(12):1621.
Stem, R. 1., Bloomer, S. H., Lin, P.-N., Ito, E., and Morris, J. 1988. Shoshonitic magmas in nascent arcs: new
evidence from submarine volcanoes in the northern Marianas, Geology 16:426-430.
Stem, R. J., Bloomer, S. H., Lin, P.-N., Ito, E., and Smoot, N. C. 1989. Submarine arc volcanism in the southern
Mariana arc as an ophiolite analogue, Tectonophysics 168:151-170.
Stem, R. J., Morris, J. D., Jackson, M. C., Fryer, P., Bloomer, S. H., and Ito, E. 1990. Enriched backarc basin
basalts from the northern Mariana Trough: implications for evolution of backarc basins, Earth Planet. Sci.
Lett., 100(1/3):210-225.
Stem, R. 1., Smoot, N. C., and Rubin, M.1984. Unzipping of the volcano arc, Japan, Tectonophysics 102:153-174.
Stuben, D., Bloomer, S. H., Taibi, N. E., Neumann, T. Bendel, v., Puschel, U., Barone, A., Lange, A. Shiying, w.,
Cuizhong, L., and Deyu, Z. 1992. First results of study of sulphur-rich hydrothermal activity from an island-
arc environment: Esmeralda Bank in the Mariana arc, Mar. Geol. 103:521-528.
Tatsumi, Y., Murasaki, M., Arsade, E. M. and Nohda, S. 1991. Geochemistry of Quaternary lavas from NE
Sulawesi: transfer of subduction components into the mantle wedge, Contrib. Mineral. Petrol. 107:137-149.
Taylor, B., Klaus, A., Brown G. R., and Moore, G. F. 1991. Structural development of Sumisu rift, Izu-Bonin are,
l. Geophys. Res. 96(BIO):16,113-16,129.
Volpe, A. M., Macdougall, 1. D., Lugmair, G. w., Hawkins 1. W., and Lonsdale, P. 1990. Fine-scale isotopic
variation in Mariana Trough basalts: evidence for heterogeneity and a recycled component in backarc basin
mantle, Earth Planet. Sci. Lett. 100(1/3):251-264.
Wessel, 1. K. 1991. A structural analysis of the Mariana inner forearc at 22°N, Master's thesis, Univ. Hawaii.
Wessel, J., Fryer, P., Wessel, P., and Taylor, B. 1994. Extension in the northern Mariana inner forearc, l. Geophys.
Res., 99(B8):15,181-15,203.
Wheat, C. G., and McDuff, R. E. 1987. Advection of pore waters in the Marianas mounds hydrothermal region as
determined from nutrient profiles (abstract), EOS Trans. AGU 68(44):1531.
Wood D. A., Marsh, N. G., Tamey, I., Joron, 1.-L., Fryer, P., and Treuil, M. 1981. Geochemistry of igneous rocks
recovered from a transect across the Mariana Trough, are, forearc, and trench, Sites 453 through 461, Deep
Sea Drilling Project Leg 60, in [nit. Repts. DSDP (D. M. Hussong and S. Uyeda et al., eds.), Vol. 60, pp. 611-
645, U. S. Govt. Printing Office, Washington DC.
Woodhead, J. D. 1987. Geochemistry of volcanic rocks from the northern Mariana islands, west Pacific, Doctoral
dissertation, Oxford Univ., UK.
Woodhead, J. D. 1989. Geochemistry of the Mariana arc (western Pacific): source composition and process, Chern.
Geol. 76:1-24.
Yamazaki, T., Ishihara, T., Murakami, F. 1991. Magnetic anomalies over the Izu-Ogasawara (Bonin) are, Mariana
arc and Mariana Trough, Bull. Geol. Surv. lpn. 42(12):655-686.
Yamazaki, T., Murakami, F., and Saito, E. 1993. Mode of seafloor spreading in the northern Mariana Trough,
7
Tectonic Framework
of the East Scotia Sea
Peter F. Barker
ABSTRACT
This chapter reviews the tectonic evolution of the East Scotia Sea, testing and extending
previously published conclusions in light of the additional and expanded data sets now
available. The East Scotia Sea floor was generated behind the east-migrating South Sand-
wich Trench, at a spreading center now lying along 300 W. On its western flank, lineated
magnetic anomalies are identified out to at least anomaly 5 (10-11 Ma) and probably out to
anomaly 5B (ca. 15 Ma). Spreading was essentially symmetric at about 27 mm/year from
15 Ma to about 5-7 Ma, then slowly accelerated. From 4 Ma to 1.7 Ma, spreading was at
50 mmlyear and slightly asymmetric. Since 1.7 Ma, spreading has been up to 15% asym-
metric, favoring accretion to the arc flank, within an overall rate of 65 mm/year. Asym-
metry is confined within segments bounded by fracture zones that in some cases were
created only at 1.7 Ma. A relation between asymmetric spreading, segmentation, and ridge
migration seems likely. The median valley is between 6 and 20 km wide and exceptionally
is up to 1200 m deep, but usually is smaller and the ridge flanks smooth, as is typical of
faster spreading. The ridge crest depth is 500 m or more deeper than the global MOR
average. Before 3-4 Ma the ridge was rougher and probably the ridge crest shallower.
In the south, the extensional zone is narrower, and the present spreading probably
started only about 3 Ma, after an eastward ridge jump associated with ridge crest-trench
collision in the South Sandwich forearc. The ridge jump caused fragments of the previous
South Sandwich arc and forearc to be transferred to the Scotia and Antarctic plates, as
part of the inevitable adjustment of plate boundaries following ridge crest collision. The
detailed history of collision along the South Scotia Ridge is poorly known, but previous
collisions involved similar transfers of arc and forearc fragments and may have influenced
previous episodes of backarc extension.
On the eastern flank, volcanoes of the South Sandwich island arc lie on ocean floor
aged from about 10 Ma to as young as 3 Ma, formed during the present spreading episode.
Both island-arc and backarc extensional volcanic geochemistry seems to reflect varying
Peter F. Barker • British Antarctic Survey. Natural Environment Research Council. Cambridge CB3 OET.
England.
Backarc Basins: Tectonics and Magmatism. edited by Brian Taylor. Plenum Press. New York. 1995.
281
282 PETER F. BARKER
degrees of contamination of the mantle, by fluids from the subducting South American
lithosphere, and of partial melting. The rocks appear similar to but simpler than those from
other intraoceanic backarc environments, showing only minor prior source depletion.
However, the arc chemistry is geographically heterogeneous and does not reflect systematic
north-south variation in the age and sediment cover of the subducted slab, as might have
been expected.
The distribution of older magnetic anomalies on the western flank suggests that
congruent ocean floor of the eastern flank should occupy most of the present forearc. If so,
there has been significant tectonic erosion of the northeast comer of the forearc, where also
serpentinized ultrabasic rocks have been dredged. This comer lies directly above the locus
of tearing of the subducting slab at its northern end. In the southern forearc off Montagu
Island, dredge hauls have identified an abnormally elevated block as a fragment of a 31 Ma
calc-alkaline arc volcano, presumably associated with the early stages of Scotia Sea
evolution.
There is no correlation between the level of development of the accretionary prism in
the lower forearc and the sediment cover of the subducting slab. Variations along strike in
the elevation of the forearc mid-slope high have controlled the transport of arc volcaniclas-
tic sediment to the trench, and hence influenced accretionary prism development to some
extent, but the main control appears to have been tectonic.
1. INTRODUCTION
The East Scotia Sea backarc basin is in many respects an end member of the wide
range of backarc extensional environments, and its overall setting and behavior are suffi-
ciently unusual to merit close examination. However it is less well known than many
backarc basins because of its geographic isolation and the rumored unpleasantness of its
climate. A principal aim of this chapter is to expose its inherent interest to a wider audience
so as to encourage further work.
Knowledge of the East Scotia Sea and its surroundings has accumulated only slowly.
Historically, the two most significant data sets have been bathymetric and magnetic. These
were last reviewed and interpreted in detail by Barker and Hill (1981). Those findings were
incorporated with minor revision into a tectonic map of the entire Scotia Sea region and
review of tectonic evolution (Tectonic Map, 1985; Barker et ai., 1991). Meanwhile, these
data sets have grown by collection on passage and by specific survey of a young ridge crest
collision at the southern end of the backarc basin (Hamilton, 1989). Of other, lesser data
sets, earthquake distribution (Isacks and Molnar, 1971; Forsyth, 1975; Brett, 1977; Pelayo
and Wiens, 1989) and dredge hauls (Saunders and Tamey, 1979; Hamilton, 1989) have been
described and discussed at intervals. Seismic reflection profiles remain sparse. More
recently, a small geologic long-range inclined asdic (GLORIA) survey was accomplished
in 1989 in the area considered by Hamilton (1989), and release in 1992 of Geosat GM
altimetry of the region south of 300 S has provided a major new free-air gravity data set.
Since this region was last reviewed, significant additional data have accumulated, only
published in part, and it is not sufficient here merely to review published work. In Section 2
therefore I set in context the main conclusions drawn from published work. In Sections 3
and 4 I evaluate the new and enlarged data sets now available, then use them to reassess
and extend those conclusions.
TECTONIC FRAMEWORK OF EAST SCOTIA SEA 283
2. REGIONAL CONTEXT OF EAST SCOTIA SEA DEVELOPMENT
2.1. Early Exploration

The South Sandwich volcanic island arc and trench lie at the eastern end of a long
eastward loop of islands and submarine ridges, loosely termed the Scotia arc, that connects
southernmost South America and the Antarctic Peninsula via the North and South Scotia
ridges (Fig. 7.1). The southern islands were discovered by Cook (1777), the northern islands
by von Bellingshausen in 1819 (Debenham, 1945), and the trench by the Meteor expedition
of 1926 (Maurer and Stocks, 1933). The similar geology of the Antarctic Peninsula and
South America had long been recognized (e.g., Barrow, 1830; Suess, 1909) and had led
many to anticipate a continental connection between them, around the Scotia arc. Further
marine survey by RRS Discovery (Kemp and Nelson, 1931; Herdman, 1948) and USNS
Eltanin (Heezen and Johnson, 1965) improved the regional bathymetry and supported this
view. Rocks from the South Sandwich Islands, though showing none of the continental
affinities of subaerial components of the North and South Scotia Ridge such as South
Georgia and the South Orkney Islands (Matthews, 1959; Hawkes, 1962), were found to be
arc tholeiites akin to Lesser Antilles rocks (Tyrrell, 1945; Baker, 1968); the apparent
parallels between the Scotia and Caribbean arcs dominated regional tectonic models well
into the 1960s.
2.2. Marine Geophysics

Marine geophysical exploration of the Scotia Sea began in 1958 (Griffiths, 1963;
Griffiths et al., 1964; Heezen and Johnson, 1965; Allen, 1966; Ewing et al., 1971), also
FIGURE 7.1. Present plate boundaries in the Scotia Arc region. "Absolute" plate motion vectors shown by thick
arrows. Half-arrows show relative motion at plate boundaries. Major plates are South American (SAM), African
(AFR), and Antarctic (ANT). Smaller plates are Scotia (SeO), South Shetland (SHE), and Sandwich (SAN). East
Scotia Sea backarc region is shaded.
284 PETER F. BARKER
inspired partly by early successes in the Caribbean. The result, however, was to emphasize
the differences between those two regions and the similarities between the Scotia Sea and
western Pacific backarc regions. Active backarc extension in the East Scotia Sea was
proposed by Barker (1970) and confirmed by a more precisely positioned marine magnetic
survey shortly afterward (Barker, 1972). These were the first well-formed magnetic anoma-
lies to be identified in an active backarc, and they showed the essential identity of ocean-
floor and backarc spreading processes. Brett and Griffiths (1975) demonstrated attenuation
and low velocities of P-wave energy crossing the backarc basin, compatible with the
presence of high temperatures at shallow mantle depths and thus with active extension.
Work elsewhere in and around the Scotia Sea served to date the opening of Drake
Passage and the Central Scotia Sea (Barker and Burrell, 1977; Hill and Barker, 1980) and
map the structure of the North and South Scotia Ridge (e.g., Harrington et ai., 1972; Ludwig
and Rabinowitz, 1982; Simpson and Griffiths, 1982). It became possible to set the develop-
ment of the East Scotia Sea into context, as the latest in a series of overlapping extensional
episodes that had built the Scotia Sea and extended the North and South Scotia ridges
eastward overthe past 30-40 m.y. (Barker and Hill, 1981; Barker et aI., 1984; Fig. 7.2). That
this extension was essentially backarc became clear because it had created more ocean
floor, at faster spreading rates, than was required by major plate (South American-
Antarctic: SAM-ANT) motion alone. Collisions between the advancing trench and ridge
crest sections of the SAM-ANT plate boundary were likely triggers for rearrangement of
the mode of backarc extension, including the onset of East Scotia Sea opening (Barker
et aI., 1982, 1984).
2.3. Neotectonics: Regional Context

Present-day plate boundaries and motions in the southwest Atlantic, shown in Fig. 7.1,
demonstrate one of the main features of East Scotia Sea backarc extension: its virtual
independence of major plate motion. In general terms, the Scotia Sea region appears as a
complication of the SAM-ANT plate boundary. The eastern SAM-ANT boundary is a
simple, long-offset spreading center from the Bouvet triple junction to the southern end
of the South Sandwich Trench at 61°S, 26°W, and becomes simple again in the far west at
the slowly subducting Chile Trench between 52° and 46°S. In between lie the moderately
large Scotia (SCO) plate and the Sandwich (SAN) and Shetland (SHE) microplates. The
East Scotia Sea (shaded in Fig. 7.1) is developing along the divergent SCO-SAN boundary,
close to 300 W.
SAM-ANT relative plate motion is west-east and slow-about 18 mm/year, whether
estimated locally (Lawver and Dick, 1983; Barker and Lawver, 1988) or globally (NUVEL
1: De Mets et ai., 1990)-and has changed little over the past 20 m.y. (Barker and Lawver,
1988). SAM-ANT motion is shared between SAM-SCO and SCO-ANT boundaries, the
North and South Scotia ridges respectively. Earthquake focal mechanisms (Forsyth, 1975;
Pelayo and Wiens, 1989) indicate sinistral motion along each, so each is slow. Pelayo and
Wiens calculated motion along the South Scotia Ridge to be about twice as fast as along the
North Scotia Ridge.
"Absolute" motions of the South American and Antarctic plates (with respect to the
mean Pacific hot-spot reference frame HS-2: Minster and Jordan, 1978; Gripp and Gordon,
1990) are also slow. In the vicinity of the Sandwich plate, Antarctic plate motion in this
reference frame is about 12 mm/year toward 188° and South American plate motion 22 mm/
year toward 237° (Fig. 7.1). Because of the large uncertainties in hot-spot motion, the
o
FIGURE 7.2. Tectonic evolution of the Scotia Sea region over the past 30 m.y. (Barker et aI., 1984, 1991). Small
blocks are BB, Burdwood Bank; S, Shag Rocks; SG, South Georgia; H, Herdman Bank; D, Discovery Bank;
J, Jane Bank; SO, South Orkney; B, Bruce Bank; P, Pirie Bank. These are one or more of microcontinental
fragment, dissected intraoceanic island arc, accretionary prism. PB is Powell Basin. (A) Present: showing
identified magnetic anomalies in the Scotia Sea and northern Weddell Sea. South Sandwich Trench is shaded.
Reconstructions (B) To 10 Ma: showing trench shaded, and "ghost" of present S. Sandwich Trench for scale. East
Scotia Sea opening here assumed all younger than 10 Ma (but see Section 4 and Fig. 7.12). (C) To 20 Ma: showing
much smaller Scotia Sea, and Jane Bank ridge crest-trench collision current. (D) To 30 Ma: showing very much
smaller Scotia Sea and Powell Basin. Note persistence of South Georgia in the northern forearc.
286 PETER F. BARKER
Antarctic plate motion is barely significant. A small southwesterly vector of absolute

motion of the Scotia plate may also be deduced, but is equally uncertain.
Essentially, therefore, both relative and "absolute" motions of the major (SAM and
ANT) plates and of the Scotia plate are slow. Within this slow-motion environment, the
rapid eastward migration of the Sandwich plate stands out. Present East Scotia Sea
extension was estimated by Barker and Hill (1981) from magnetic anomalies at about 70
mmlyear, in a direction close to east-west. Assuming Pelayo and Wiens's (1989) estimate of
SCO motion is correct, then SAN-SAM convergence at the South Sandwich Trench may be
estimated at about 75 mm/year along 085°. "Absolute" motion of the Sandwich plate is
about 57 mm/year toward 095°. If the forearc is regarded as being in dynamic equilibrium,
then this last is the rate of eastward roll-back of the trench hinge. Revised estimates of these
motion vectors, using East Scotia Sea spreading rates determined in Section 4.3, are only
slightly different.
Estimates of plate motion are pertinent to a discussion of the reasons for backarc
extension in the East Scotia Sea. Chase (1978) first noted the correlation between backarc
extension and retreat (in the hot-spot reference frame) of the major plate behind the arc.
Following Elsasser (1971), he postulated that the subducting slab would not merely glide
along its own length, forming an anchor in the sublithospheric mantle, but would also tend
to sink vertically because of its weight, leading to rollback of the trench hinge. If the
overriding major plate was not advancing, then backarc extension might be induced to fill
the gap. In the case of the East Scotia Sea, rollback is remarkably fast (only the New
Hebrides is faster: Jarrard, 1986), and in the place of the major plate is the Scotia plate, not
particularly large but apparently retreating slowly southwestward. As noted elsewhere
(Section 2.5), it is considered that almost all of the Scotia plate has itself formed by backarc
extension during earlier episodes, and the present East Scotia Sea opening represents
merely the latest phase of a long-lived backarc extensional regime that presumably was also
reacting to rollback of the trench hinge. It has been suggested by Alvarez (1982) that
rollback here is facilitated by shallow eastward sublithospheric mantle return flow that is
balancing the net asthenospheric loss (to the lithosphere) in the growing Atlantic and Indian
Oceans and the net gain in the shrinking Pacific.
2.4. Neotectonics: The Subduction Zone

Oceanic lithosphere of the South American plate is being subducted at the South
Sandwich Trench, the eastern boundary of the Sandwich plate (Fig. 7.1). Ocean floor at the
trench ranges in age from 27 Ma (C8r) in the south to perhaps 80 Ma (C33r) in the north
(Barker and Lawver, 1988). It is likely that the oceanic lithosphere now at moderate depth
within the subduction zone (say less than 200 km) was all produced at slow spreading rates
at the South American-Antarctic Ridge and therefore has a similar age range (probably
younger in the far south) to that in the trench.
Isacks and Molnar (1971), Forsyth (1975), and Brett (1977) have analyzed the seis-
micity of the dipping slab. Along most of the trench, the slab dips at 45-55° down to
180 km (possibly to 254 km near 58°S; Brett, 1977). A zone of more intense activity with a
steeper apparent dip occurs at the northern end (between 55.8° and 56.3°S). Since the
period of activity examined in these studies, four deeper earthquakes have been reported in
NOAA PDE lists, at 196-, 217- and 288-km depth at the northern end, within a continuation
of the zone of steep apparent dip.
The steeply dipping northern zone was associated by Forsyth (1975) and Brett (1977)
with tearing of the downwarped South American plate, as the hinge of subduction has
advanced eastward (Fig. 7.3). At the southern end of the South Sandwich Trench, the long
eastward line of the South Sandwich fracture zone fonns the southern boundary of the
South American plate and facilitates future subduction and rollback, but at the northern end
the subducting slab must tear. Earthquake focal mechanisms within the steeply dipping
zone were interpreted by Forsyth as hinge-faulting or bending-stress release, compatible
with tearing of a slab downwarped to 30-120 km, along a west-east line close to 56°S.
Elsewhere within the subduction zone, earthquake focal mechanisms are distributed
similarly to those in other such regions in most respects, but there are notable differences.
While intennediate-depth earthquakes in the northern half are downdip extensional, those
in the southern half are downdip compressional. Forsyth (1975) and Brett (1977) attributed
this to the greater buoyancy of the younger, southern part of the sinking slab, which may be
being towed down by the older, denser northern part. Similarly, the greater age of the
northern part might be expected to result in deeper earthquake foci there, given that the
younger, thinner southern part should reheat sooner to temperatures too high to permit
failure by brittle fracture (e.g., Vlaar and Wortel, 1976; Jarrard, 1986). At the time, no such
effect was seen, but the distribution of the more recent, deeper events supports this notion.
Again, the age-related density difference could explain the greater slab dips seen by Brett
(1977) north of 58°S.
2.5. Scotia Sea Reconstructions

The continuity and vigor of backarc extension, driven by eastward rollback of the
trench hinge, are the key features of Scotia Sea evolution. Several spreading regimes have
been identified within the Scotia Sea, overlapping in time. For example, Drake Passage and
Central Scotia Sea opening (Barker and Burrell, 1977; Hill and Barker, 1980), though now
stopped, are thought to have continued for a while after East Scotia Sea opening started.
Other, short-lived spreading systems are inferred, in basins too small to produce unam-
biguously identifiable magnetic anomaly sequences. In total, far too much oceanic
lithosphere was being produced in the Scotia Sea, for at least the past 21 m.y. (Barker and
Hill, 1981) and most probably over its entire 30-40 Ma lifetime, to be explained without
recourse to the decoupling provided by subduction: even without a complete knowledge of
FIGURE 7.3. Sketch of tearing of 80-m.y.-old South American

oceanic lithosphere at the northern end of the South Sandwich
subduction zone, based on earthquake fault plane solutions (from
Forsyth, 1975). The young Sandwich plate (shown dashed) over-
lies the downwarped edge of the South American plate to the north
of the tear, preventing its complete recovery.
288 PETER F. BARKER
the spreading history, or of contemporaneous arc volcanism, the essential backarc setting of
Scotia Sea extension was apparent.
Figure 7.2 shows the regions of the Scotia Sea that are dated by marine magnetic
anomalies, and reconstructions of Scotia Sea evolution through 10 Ma, 20 Ma, and 30 Ma
(Barker et al., 1984, 1991). Several assumptions are necessary in order to make these
reconstructions, principally concerning the age of several undated small ocean basins and
the nature of coupling between contemporaneous spreading provinces. It is not appropriate
to discuss those further here (see Barker et al., 1991), but it is useful to consider Scotia Sea
evolution in order to understand the processes acting.
No single spreading regime persisted for more than a few million years, and it has been
suggested that collisions between ancestors of the present South Sandwich Trench and
more westerly spreading sections of the South American-Antarctic plate boundary trig-
gered changes of regime. The South Scotia Ridge is the locus of these past collisions,
younging eastward. To illustrate this process, it is clear from Fig. 7.1 that, if present-day
spreading rates and directions persist, the southern part of the South Sandwich Trench will
overtake the slower spreading ridge crest sections now at 19-20oW, in about 5 m.y. The
presumption is that spreading and subduction at the collision zone would then both stop,
but subduction would continue at the northern half of the trench, possibly with a different
rate, direction, and site of complementary backarc extension. Somehow, the Sandwich plate
would fracture, its southern part accreting to the Antarctic plate to produce a longer, more
complicated Scotia-Antarctic plate boundary. The key feature of past collisions seems to
have been that all of the southern oceanic part of the South American plate was subducted
as the arc and trench migrated eastward along the plate boundary, but none of the Antarctic
plate.
Attempts have been made to define and date previous ridge crest-trench collisions
along the South Scotia Ridge. Rocks dredged from Discovery Bank (Barker et al., 1982)
and Jane Bank (Barker et al., 1984)-0 and J in Fig. 7.2-show strong chemical sim-
ilarities to South Sandwich Island lavas: Discovery Bank rocks are of Miocene age (12-20
Ma). A magnetic anomaly sequence in the northern Weddell Sea that younged toward Jane
Bank was interpreted by Barker et al. (1984) to imply a collision there ca. 20 Ma (Fig.
7.2C). The magnetic anomaly identifications have been queried (Hamilton, 1989; Liver-
more and Woollett, 1993) in light of rate constraints placed on SAM-ANT motion by Barker
and Lawver (1988). However, it is possible that, for a while before ridge crest collision, the
subducting slab became decoupled from the South American plate (as happened repeatedly
off the Pacific margin of North America in similar circumstances; Menard, 1978; Atwater,
1989), thereby removing such constraints. Age uncertainties do not affect the reality of the
collisions or their backarc implications, merely their timing.
Further definition of the Jane Bank collision zone was hampered by sediment cover.
Hamilton (1989) surveyed and sampled the youngest collision zone, in the area directly
south of the present backarc spreading center, that promised to be more accessible. He
found the collision zone to be narrow, and probably oblique to the convergence direction,
because of complexities of the SAM-ANT boundary. Also, SAM-ANT spreading was slow
over that period, producing barely recognizable marine magnetic anomalies. The history of
collision was therefore difficult to extract. However, the additional data do allow an
improved description of the history of extension in the southernmost part of the backarc
(Section 4.6) and better definition of present plate boundaries. The GLORIA data from the
collision zone, which will not be considered in detail here, appear largely to confirm
Hamilton's (1989) conclusions.
2.6. East Scotia Sea Development and Structure

Several features of East Scotia Sea development and structure are significant in global
terms. Its location remote from other backarc basins, its well-formed magnetic anomalies
and simple chemistry, and the rapid rollback of the trench hinge make it a useful end
member among backarc environments. Other, lesser features gain significance from this
uniqueness. They include aspects of the spreading history; the bathymetric expression of
spreading, the forearc, arc, and remnant arc; and the anomalous southern margin. Because
of the primitive nature of knowledge of the East Scotia Sea at the time they were described
(Barker, 1972; Saunders and Tamey, 1979; Barker and Hill, 1981), compared with present
knowledge of this and other backarc regions, it is important to reexamine these features and
their significance in the light of the expanded data sets now available. In preparation for
this, I summarize them briefly below.
Most of East Scotia Sea spreading was asymmetric, with accretion favoring the
eastern, island-arc flank. Spreading accelerated from about 50 to 70 mmlyear, between 3
and 1.5 Ma. The oldest identified anomaly on the western flank was 4A. Little difference
was seen between the amount of ocean floor formed at 56°S and at 59°S, over the past 5-6
m.y., suggesting that spreading did not contribute to the curvature of the arc.
Ridge crest depth of the East Scotia Sea spreading center is about 3000 m, about 500 m
deeper than the global average for mid-ocean ridges (e.g., Parsons and Sclater, 1977). The
eastern arc flank is buried beneath a thick arc-derived volcaniclastic apron, but the western
flank is covered sparsely and unevenly by pelagic and hemipelagic sediment deposited
under the influence of the Antarctic Circumpolar Current (ACC). Although the western
flank shows rough, elevated topography, no simple grouping of elevations with east-facing
scarps could be identified with a remnant arc that might mark the beginning of East Scotia
Sea extension. East Scotia Sea spreading overlapped in time extent with the adjacent
Central Scotia Sea spreading, so a complicated early history of spreading might be
expected.
Many of the volcanic islands of the arc are built on ocean floor produced during the
present spreading episode, no more than a few million years old. The oldest radiometric
ages from the arc (3.1 Ma; Baker, 1990) are compatible with this. Although no marine
magnetic lineations were seen in the forearc (Barker and Hill, 1981), the upper forearc,
too, at least in the north, may comprise ocean floor formed during the present spreading
episode. The upper forearc in the south is significantly more elevated than in the north,
which some have linked to the greater elevation of the younger SAM ocean floor being sub-
ducted in the southern half of the trench. The lower forearc is narrow and poorly developed.
In the south, the distance between the arc and the elevated topography of the South
Scotia Ridge (specifically, Herdman Bank) is less than the east-west extent of magnetic
anomalies farther north. Either spreading started more recently in the south or (since the
orientation of the magnetic anomalies appeared to change) an additional Herdman Bank
microplate was involved, so spreading here has been slower.
All these published conclusions (Barker, 1972; Saunders and Tamey, 1979; Barker and
Hill, 1981) are reassessed in Section 4.
290 PETER F. BARKER
3. NEW OR ENLARGED DATA SETS
3.1. Bathymetry
The bathymetry of the Tectonic Map (1985) was based on all available soundings to
1980, including, with discretion, some early data that may not have been accurately located.
The amount of data now available for the backarc in general has increased by perhaps 10%,
which has not justified revision of the contour map to date. In the southern area around the
youngest ridge crest collision zone, the increase in soundings has been greater, and
Hamilton's (1989) recontoured chart of this area is incorporated in Fig. 7.4. Contours are at
500-m intervals. Ship tracks are not shown, to avoid confusion, but the magnetic tracks
(Section 3.3) show the locations of most of the post-1969 soundings.
3.2. Shipboard and Geosat GM Free-Air Gravity

The Geosat "Geodetic Mission" (GM) radar altimetric data for the ocean areas south
of 300 S were declassified by the U.S. Navy in late 1992. The long repeat period of the GM
orbit produced a ground track separation in these latitudes of 2-3 km, which is short
compared with the 2-9 km altimeter footprint (McAdoo and Marks, 1992) and implies
essentially complete coverage. Free-air gravity data at a grid spacing of 0.05 0 longitude by
FIGURE 7.4. Bathymetry of the East Scotia Sea region, replotted IImFtfti1ertetfflical 0HII6morphic projection,
from Tectonic Map (1985) with revisions by Hamilton (1989). Contours are at 500-m intervals, and the 2-, 3-,
5-, and 7-kIn contours are thickened to aid definition of tectonic provinces. Deepest contour is 8000 m in northeast
South Sandwich Trench. South Sandwich Islands are Z, Zavodovski; L, Leskov; V, Visokoi; C, Candlemas (with
Vindication); S, Saunders; M, Montagu; B, Bristol; T, Thule (with Cook and Bellingshausen). Other, submarine
arc volcanoes are P, Protector Shoal; K, Kemp; and N, Nelson seamounts.
FIGURE 7.5. Free-air gravity map of the East Scotia Sea region derived from Oeosat (OM and ERM) altimetric
data (NODC-NOAA, 1993), using Lambert Conical Orthomorphic projection. Contours at IO-mgal intervals. A
color version of this figure appears opposite page 310.
0.04 0 latitude are now available on CD-ROM (NGDC-NOAA, 1993). The limit of short-
wavelength response of the gridded data is generally considered to be about 12-20 kIn
(McAdoo and Marks, 1992), comparable with surface ship measurements in oceanic areas.
To produce Fig. 7.5, I projected the grid points onto a Lambert Conical Orthomorphic
projection and regridded at approximately the same spacing before contouring, using
proprietary minimum-tension software.
Gravity measurements were made along about one third of the ship tracks shown in
Fig. 7.6 (Section 3.3). Compared with the Geosat gravity data, therefore, the shipboard data
are few and their distribution uneven. Comparison of typical ship tracks with satellite data
shows that over almost all of the area of Fig. 7.5, except areas of shallow water, the
correspondence is very close. Discrepancies in shallow water stem from a combination of
the short-wavelength gravity anomalies from shallow sources, better reproduced by ship-
board measurements, and distortion introduced by gridding and contouring algorithms
close to land where there are no altimetry data.
The bathymetric chart and gravity contour map (Figs. 7.4 and 7.5) are best considered
together. In a purely oceanic region, the overwhelming contribution to the gravity anomaly
map at short to intermediate wavelengths comes from the topographic relief on oceanic
basement, either at the seafloor or, reduced in amplitude, buried beneath sediments
(McAdoo and Marks, 1992). Complications arise when the ocean floor is loaded locally (as
with volcanic islands) after gaining rigidity so that the load is regionally rather than locally
292 PETER F. BARKER
36'11" 35'11" 34'11" 33'11" 32'11" 31'11" 30'11" 29'11" 28'11" 27'11"
58'S 56'S
57'S 57'S
58'S 58'S
59'S
37'11" 36'11" 35'11" 34'11" 33'11" 32'11" 31 '11" 30'11" 29'11" 28'11" 27'11" 26'11" 25'11"
FIGURE 7.6. Magnetic anomalies in the East Scotia Sea region, projected along ship tracks (solid line ornament
for positive values). Identified anomalies are marked and major anomalies numbered. For ease of identification,
the ridge crest is dotted, where seen on bathymetric profiles or in Fig. 7.5, and anomalies 3A and 58 are dashed.
Comparisons with synthetic anomalies are shown in Figs. 7.10 and 7.11. Thicker, straight lines are transform faults
and fault traces. The lOOO-m and 6000-m contours from Fig. 7.4 are used to identify the South Sandwich Islands,
other submarine elevations, and the South Sandwich Trench. Lambert Conical Orthomorphic projection.
compensated and the crustal structure possibly modified. Other complications concern
continental fragments that may have an entirely different, and typically more hetero-
geneous, crustal structure. Nevertheless there are strong similarities between gravity and
bathymetric maps of largely oceanic areas, and the former can be used to infer bathymetric
features in areas where shipboard bathymetric data are sparse.
Almost all ofthe area shown in Figs. 7.4 and 7.5 is floored by oceanic lithosphere of
the South American, Antarctic, Scotia, or Sandwich plates. The South Georgia block is a
continental fragment, and the South Sandwich volcanic arc may lie on ocean floor modified
to a greater or lesser extent by previous subduction-related intrusion. Parts of the South
Scotia Ridge, such as Herdman Bank, may be continental fragments or parts of older
intraoceanic arcs, rendered inactive by ridge crest collision and left behind by subsequent
backarc extension. Both figures show the main tectonic features well. There are differences
(notably in the definition of ocean-floor fabric outside the Scotia Sea), but few are major, a
reflection perhaps of the high density of existing bathymetric data over much of the region.
Note, however, the gravity lows indenting the northern and southern margins of the South
Georgia block, interpreted in Section 4.8.
3.3. Magnetic Anomalies

Magnetic anomalies in the East Scotia Sea have been attributed to east-west backarc
extension over the past 8 m.y. (Barker, 1970, 1972; Barker and Hill, 1981). Since these
studies, the magnetic data set has increased about 40% by acquisition along passage tracks
(not always in a useful direction) and also by detailed study of ridge crest-trench collision
in the far south (Hamilton, 1989). Among the principal findings of previous work were the
fast rates of modem seafloor spreading in the backarc, the acceleration in spreading at 1.5
to 3 Ma, and a pronounced asymmetry of spreading. The spreading rates have been
combined with other similar data in computing unknown plate motions, both relative and
"absolute" (see, for example, Section 2.3), the acceleration bears on studies of driving
forces and causation, and the asymmetry may reflect conditions at the base of the litho-
sphere. Because of their significance, it is useful to reassess these original findings in light
of the enlarged data set.
Also, the spreading rates and particularly the acceleration depend upon the accuracy of
the magnetic reversal time scale (MRTS) used in their determination, which has been
revised meanwhile. Barker and Hill (1981) used the MRTS of LaBrecque et ai. (1977); De
Mets et ai. (1990), in determining NUVEL-l, used that of Harland et ai. (1982) to fix the
time of the center of anomaly 2A (C2An.2n). For this chapter I have used the MRTS of
Cande and Kent (1992), which is a significant improvement in accuracy over the previous
generation. However, further revision of the MRTS may follow detailed comparisons of
magnetic and orbital (Milankovich) time scales (e.g., Wilson, 1993).
Magnetic anomalies are displayed along track in Fig. 7.6 after removal of the Interna-
tional Geomagnetic Reference Field (IGRF) (Langel, 1992). No attempt has been made to
eliminate the diurnal variation, but data were checked for magnetic storm influence before
acceptance. Ship tracks for all profiles are satellite-positioned (transit or GPS), mostly to
better than 1.5 km.
3.4. Seismic Reflection Profiles

Few seismic reflection profiles have been acquired in the South Sandwich arc and
backarc region. Those known to me are the tracks shown in Fig. 7.7, which are almost all
single-channel profiles and largely unpublished, and the older, astronavigated RV Zapiola
Cruise II profile shown by Heezen and Johnson (1965). Data from two of the profiles
crossing the backarc are incorporated into Fig. 7.8 (from Barker and Hill, 1981), and others
have been described by Hamilton (1989). Most of those in the far south and the far north-
east were acquired in conjunction with the GLORIA survey during RRS Charles Darwin
Cruise 37 in 1989 and will not be considered specifically here.
3.5. Onshore Geology and Dredged Rocks

The South Sandwich Islands have been sampled several times during opportunistic
visits, but the most significant mapping and sampling exercise by far was that undertaken in
the period 1961-1964 from RRS Shackleton and HMS Protector and described by Baker
(1978), Holdgate and Baker (1979), and Tomblin (1979). Geochemical analyses of these
rocks by Luff (1982) have been supplemented or superseded by more extensive recent work
(Pearce et ai., in press).
294 PETER F. BARKER
FIGURE 7.7. Locations of seismic reflection profiles and dredge hauls in the East Scotia Sea region. Profiles
referred to in the text or shown in other figures are lettered, and dredge hauls discussed are numbered. No dredge
sites are shown where an in situ component could not be identified. PS is Protector Shoal. The East Scotia Sea
spreading center, where known from median valley or magnetic anomaly, is shown dotted. The South Sandwich
Trench (7-km contour) is shaded.
FIGURE 7.B. Bathymetric profiles crossing the East Scotia Sea (from Barker and Hill, 1981). Profiles have been
lettered where referred to in the text or shown in other figures. Vertical scale is km.
Dredging in the Scotia Sea region is made more difficult by the need to establish that a
dredged rock was in situ, against a background of abundant ice-rafted detritus. Fig. 7.7
shows the locations of successful dredges, in which an in situ component has been
established, that have been obtained during Birmingham University and British Antarctic
Survey (BAS) cruises over the past 20 years. These dredges were undertaken principally for
"mapping" purposes, to determine an environment from its chemistry rather than to
examine the chemistry of a known environment. For example, dredge hauls from Discovery
Bank showed it was an intraoceanic island arc ancestral to the South Sandwich arc (Barker
et ai., 1982). The main exception was a suite of four dredges from the backarc spreading
center, described by Saunders and Tamey (1979). Unpublished dredge hauls from the
northern, eastern, and southern margins of the Sandwich plate have been described, and
x-ray fluorescence (XRF) analyses were carried out at Birmingham University by P. L.
Barber (pers. comm.) and by Hamilton (1989). The unnumbered dredge hauls in the south
are from the youngest ridge crest collision zone (Hamilton, 1989) and will not be consid-
ered here because there is no independent evidence of the tectonic environment that they
sampled.
4. DATA INTERPRETATION
4.1. Context: The South Sandwich Island Arc and Trench

4.1.1. South Sandwich Island Arc
All of the volcanic islands, except Leskov Island, lie along a smooth arc (Baker, 1990;
Fig. 7.7). That arc continues in the north to include Protector Shoal, at 27 m the shoalest part
of a large bank that the gravity map suggests extends to Zavodovski Island, the north-
ernmost of the chain. Protector Shoal was considered by Gass et al. (1963) and Holdgate
and Baker (1979) the likely source of a 1962 pumice eruption. The arc extends southward
also to include Kemp Seamount, a submarine volcano near 59.5°S, 28°W with a least
measured depth of 89 m, dredge-sampled from RRS Discovery in 1985 (DR 112 in Fig. 7.7).
Hamilton (1989) considered this volcano to have been recently active because of its shal-
low depth: many small, isolated elevations in this area that are not actively being enlarged
are flat-topped at 500- to 8oo-m depths, the likely result of planation by icebergs from the
southern Weddell Sea.
Several submarine volcanoes lie, like Leskov Island, off the main line of the arc. The
gravity and bathymetric maps (Figs. 7.4 and 7.5) indicate subsidiary submarine elevations
west of Saunders, Montagu, and Bristol Islands that are almost certainly volcanic. East of
the main arc, in the forearc, there are also isolated elevations, but the two largest of these,
near 58.5°S, 25°W and 60.3°S, 27.5°W, are old (see Section 4.7). Others, notably that at
59.6°S, 26°W, have an uncertain origin. Only Nelson Seamount, the small peak at 60.3°S,
28.2°W (Fig. 7.4) with a least measured depth of 280 m, is known to be volcanic: Hamilton
(1989) notes similarities to South Sandwich Islands lava chemistry in rocks dredged from
this peak from RRS Discovery in 1985 (DR 111 in Fig. 7.7). Nelson Seamount was also
considered by Hamilton to be young (rather than an older volcano dissected by extension
following ridge crest collision), because it is shallow yet not flat-topped. This is borne out
by the GLORIA survey, which shows its apron overlapping young faults and ocean floor.
296 PETER F. BARKER
However, it is very close to the southern limit of the Sandwich plate and closer to the trench
than other arc volcanoes.
The dominant rock type of the islands is a low-K tholeiitic basalt, with successively
smaller amounts of more acidic equivalents. In these rocks the alkali metals are enriched
with respect to mid-ocean ridge basalts (MORB). The rare-earth elements (REE), partic-
ularly the light-rare-earth elements (LREE), and high-field-strength (HFS) elements are
depleted. They are thought to have been produced by high degrees of partial melting of
depleted mantle in the region above the subducting slab, fluxed by hydrous fluids or melts
from the slab (Luff, 1982; Pearce, 1982; Baker, 1990; Pearce et aI., in press). They are
similar to rocks from the Mariana, Izu, and Tonga arcs.
The systematic tectonic differences that one might expect to see reflected in arc
chemistry are the north-south age gradient in the subducted slab, that will have been
consistent for several million years, and a consequent thicker sediment cover in the north. A
secondary feature reflecting the difference in basement age and latitude of the subducting
slab will be an almost total dominance of biosiliceous sediments in the south, but a more
even calcareous/siliceous mix in the north. In addition, to the extent that the arc does not lie
on ocean floor of the present spreading episode, but on that produced during the earlier
Central Scotia Sea opening (Barker and Hill, 1981; Fig. 7.2B), the mantle beneath the
northern and southern South Sandwich arcs could be different. All of these factors might be
expected to produce systematic chemical differences between northern and southern parts
of the arc, but no such differences have been detected.
Inter-island chemical variation is significant but not obviously linked to regional
tectonic variation. About the clearest distinction is a low-K, LREE-depleted group of
Zavodovski, Candlemas, Vindication, Montagu, and Bristol Islands, most of which also
have low incompatible-element ratios and abundances. These islands do not form a
geographically distinct group. One possible exception identified by Pearce et al. (in press)
concerns the more calc-alkaline character of rocks from Leskov Island (which lies west
of the main arc) and Freezland Rock (5 km west of Bristol Island and possibly the oldest
exposed rocks, at 3.1 Ma; Baker, 1990).
Hamilton (1989) reported somewhat limited geochemical data from the Kemp and
Nelson seamount dredge hauls (DR 112 and 111 respectively; Fig. 7.7). Kemp Seamount,
which lies on a southern extension of the main arc, yielded a basalt most closely resembling
rocks from Zavodovski and Visokoi Islands. This result accords with the apparently
random geographic distribution of chemical diversity around the arc. A fresh dacite
dredged from Nelson Seamount most closely resembles dacites from Leskov Island and
Freezland Rock. This correspondence also reflects geographic heterogeneity since, unlike
these sites, Nelson Seamount is (a) young and (b) closer to the trench than the main line
of the arc.
4.1.2. South Sandwich Trench

The trench extends from 56°S, 33°W in the northwest to 60.7°S, 26°W in the southeast
(Figs. 7.4 and 7.5). The outer trench wall is uneven, the effect of a linear ocean-floor
topography oriented between east-west and southeast-northwest, created at the SAM-ANT
plate boundary between 25 and 80 Ma (Barker and Lawver, 1988) and clearly seen in Fig.
7.5. Oceanic basement of the outer wall is covered by up to 1.5 km of pelagic sediments,
distributed unevenly because of the rough basement topography (profile Q of Fig. 7.9) and
0.------------------------------------------------,0
4 P
M
6
0,------- -----,0
,0
FIGURE 7.9. Single-channel seismic reflection profiles P, Q, and R (located in Fig. 7.7) crossing the northern
and eastern (forearc) margins of the Sandwich plate. M is the mid-slope high in profiles Q and R, and the edge of
normal backarc in profile P. T is the trench axis. Vertical scale is seconds of two-way time. Horizontal bars are 20
km. Dredge hauls from the outer (trench) slope of M on each profile are reported in the text.
influence of the ACe. Sediments are thicker in the north because of the greater base-
ment age.
A mid-slope high (labeled M in Fig. 7.9, profiles Q and R) with a steep outer slope and
gravity gradient separates the upper and lower forearcs of the inner trench wall. The trench
is deepest, below 8000 m, in the northeast comer where also the trench is broadest, the
upper forearc narrowest, the slope to the arc steepest, and the lower forearc least well
developed (Figs. 7.4, 7.5, and 7.9). Southward, the trench shoals and narrows and both
lower and upper forearcs become broader, more continuous, and more elevated. Although
the outer slope of the mid-slope high forms a smooth curve along the forearc, its elevation
is highly variable along strike.
On reflection profiles (for example, profiles Q and R of Fig. 7.9) the lower forearc is
mainly acoustically opaque, as is common within accretionary prisms. Although the gravity
effect of the prism has not been modeled, it is clear that, paradoxically, the prism is least
well developed in the north, where the subducting ocean floor brings the most sediment
to the trench.
The other contributor to trench (thence prism) sediments is the arc. Some reflection
profiles (e.g., profile Q of Fig. 7.9) show arc-derived sediment ponded behind the mid-slope
basement high, but on others (e.g., profile D of Fig. 7.8) there appears to be no barrier to
direct downslope transport to the trench floor. The variable depth to the mid-slope basement
high therefore ensures that arc sediment is channeled to certain sections of the trench.
Seismic reflection profiles and the gravity map show a coincidence between the presence or
absence of a prominent, shallow mid-slope basement high and the local accumulation of
sediments on the trench floor. Fig. 7.5 shows that, where the mid-slope high is least well
developed, as between 56.8 and 57.8°S, there is also a gentler slope to the gravity profile
over the lower forearc and trench, indicating a better-developed accretionary prism. Sim-
298 PETER F. BARKER
ilarly, where the mid-slope high is most prominent, as between 55.3 and 56. ODS (Fig. 7.9,
profile Q), the accretionary prism is less well developed. These are second-order effects,
however, superimposed on the general north-south contrast in accretionary prism develop-
ment. The question of the origin of this contrast remains, and the possibility must be
considered of tectonic erosion, concentrated in the north because of the ready subsidence of
the older oceanic lithosphere (but also, possibly, in the far south associated with ridge crest
subduction).
Tectonic erosion of a forearc is generally inferred when transport of loose material
down the subduction zone regularly exceeds what is being supplied at the seabed in the
trench. It has been associated with ridge crest subduction off Chile (Cande et al., 1987) and
the Antarctic Peninsula (Barker, 1982), in the Woodlark Basin (Taylor, 1987), and along the
South Scotia Ridge (Barker et al., 1984). The term is vague, however, in the sense that the
processes involved may be active in all trenches, to some degree. Tectonic erosion can lead,
as off Peru and northern Chile (Von Huene et al., 1985), to a trench with a long subduction
history having only a poorly developed accretionary prism and exposing relatively old
rocks on the inner trench slope. The question arises, is erosion confined to the loose debris
of the accretionary prism, or is material from the edge of the rigid overriding plate also
being removed, and subducted? I return to this question in Section 4.7, after considering
also the backarc spreading history, which bears on the nature and extent of forearc
basement.
In the north along the east-west arm of the trench, the southern, inner trench slope is
the northern margin of the East Scotia Sea backarc basin. Its profile (for example, Fig. 7.9,
profile P) corresponds closely to a normal trench profile, despite its having undergone only
eastward transcurrent motion with respect to the South American plate. An elongated ridge
and trough occupy the lower part of the inner trench slope and resemble a poorly developed
lower forearc accretionary prism in their topography and gravity signature. The main upper
slope is very steep. Dredge hauls from this steep wall (described in Section 4.4) are mostly
fresh lavas, showing that it is original, the product of backarc extensional volcanism rather
than of any subsequent tectonic erosion.
4.2. Character of the Backarc

The East Scotia Sea backarc region is bounded in the north by the east-west prolonga-
tion of the South Sandwich Trench, and in the south by a complex boundary with the
Antarctic plate, lying approximately along 6O.5°S, and involving the elevated Herdman and
Discovery banks (both of which in places lie above 750 m). In the east, the formal boundary
is the island arc but, as has often been noted in this region and elsewhere, both arc and
backarc extension are secondary features of subduction, so the boundary of "backarc"
extension is not necessarily the arc. The nature of subarc and upper forearc basement is
therefore at issue.
In the west, the boundary between the East Scotia Sea and Central Scotia Sea, current
and previous backarc extensional provinces (e.g., Fig. 7.2A), is uncertain. There is a wide
gap in published magnetic lineations of each province. Topographically (seen also in the
gravity map, Fig. 7.5) there is a change at about 36°W from rough, elevated to smoother,
deeper seafloor, more pronounced in the south than in the north. As a boundary to the
modem backarc, however, this elevated area is more diffuse and irregular than the remnant
arc province of most backarc regions.
The dominant feature of the backarc is its spreading system, centered approximately
along 30o W, and here defined very precisely by its lineated oceanic magnetic anomalies.
Fig. 7.6 shows the available magnetic profiles and identified anomalies. Almost all of the
identifications previously made along older profiles by Barker and Hill (1981) over most of
the backarc region are confirmed. The newer data have filled gaps and in places have
allowed additional identifications to be made on existing profiles. In particular, some older
anomalies can be identified, and the history of the complicated southern end of the
spreading center is now perhaps clearer (Section 4.6). The very northern end, where the
spreading center meets the trench at a presumed RFF triple junction, remains poorly
known.
4.3. Magnetic Anomaly Identifications, Spreading Rates, Asymmetry

In the northern and central sections of the backarc (north of 59.5°S; Fig. 7.6), well-
formed magnetic anomalies are seen out to at least anomaly 5 (ca. 11 Ma) on the western
flank, but in the east the magnetic signature of the island arc overprints most profiles.
Farther south, anomalies can be identified only out to 2 or 2A, within a narrower band of
rough topography flanked in the west by the flat-topped Herdman Bank and in the east by
the volcanic arc. Spreading today is essentially east-west, in a simple two-plate system, but
the slight discordance between the older anomalies on each flank suggests the possibility of
significant earlier complexity. The positions of transform faults and fracture zones in Fig.
7.6 have been determined mainly from bathymetric data (Barker and Hill, 1981; Fig. 7.4)
and from the gravity map (Fig. 7.5, which shows the median valley well) and partly by the
need to accommodate offsets in the magnetic pattern. Farther from the ridge crest, partic-
ularly on the eastern flank, fracture zone locations and orientations are not well constrained.
Many transforms appear to have been short-lived.
Figure 7.10 shows a selection of long magnetic and bathymetric profiles across the
backarc, arc, and forearc together with distance plots reduced to 25 mm/year (Barker, 1979;
Barker and Hill, 1981) and a synthetic profile, generated using the MRTS of Cande and
Kent (1992). The reduced distance method is more sensitive to small variations in spread-
ing rate and was used by Barker and Hill (1981) (with the MRTS of LaBrecque et ai., 1977
and on profile A only) to identify anomalies out to 4A and to propose acceleration and a
persistent asymmetry. The better coverage of long profiles now available prompts a revised
and extended interpretation.
Spreading since 1.7 Ma or so has been fast, averaging about 65 mm/year over the full
length of the backarc. Between about 4 and 1.7 Ma it was slower, approximately 50 mm/
year, and before 5-7 Ma it was slower still, less than 30 mm1year. On most bathymetric
profiles crossing the spreading center, ocean floor produced during the past 3-4 m.y. is
smoother than that produced earlier at lower spreading rates. Figure 7.8 (from Barker and
Hill, 1981) shows several east-west bathymetric profiles across the backarc, arc, and forearc
that illustrate this feature. The profiles are aligned along western anomaly 2A, and the
positions of other anomalies and the median valley are marked.
There is a general asymmetry of spreading, favoring accretion to the arc flank, but not
as uniform or as pronounced as previously thought (Barker and Hill, 1980, 1981). It is not
seen in ocean floor produced before 4 Ma and is barely detectable until 1. 7 Ma. Thereafter,
at the fastest spreading rates, it is more obvious, but even so is not ubiquitous. It appears to
occur within distinct spreading segments bounded by transform faults (Fig. 7.6), but the
300 PETER F. BARKER
FIGURE 7.10. (A) Magnetic profiles A to F (with bathymetry for A and D) crossing the East Scotia Sea,
compared with a synthetic anomaly profile (SP) generated using the MRTS of Cande and Kent (1992). Located in
(B). Also shown (C) is a reduced distance plot (Barker, 1979; Barker and Hill, 1981) of identified anomalies from
the lettered profiles, and the set of spreading rates (SP) used to generate the synthetic anomaly profile, all reduced
to 25 mmlyear.
asymmetry in one segment is in the opposite sense. In the slower-spreading flank areas
in such cases, there is no initial anomaly offset; the transforms appear to have been created
at the time of the acceleration in order to accommodate asymmetric accretion.
Barker and Hill (1980) pointed to the correlation between sense of asymmetry and
"absolute" migration of the ridge crest in some areas of the main ocean basins and in
backarc regions as supporting a thermal origin for asymmetric spreading. In this explana-
tion, isotherms at the base of the lithosphere become asymmetrically dipping as a result of
the migration, and accretion favors the cooler, leading flank. This correlation is largely
supported by the new results, since the East Scotia Sea ridge crest will only migrate now
when spreading is faster than about 30 mm/year (from Section 2.3). However, it is clear that
ridge crest migration is not the whole story. An additional enabling mechanism for
asymmetric accretion might be involved. Although more detailed survey and direct sam-
pling would be required to examine this fully, there may be a causal relationship between
spreading rate, asymmetry, and ridge segmentation.
4.4. The Ridge Crest and Spreading Center Chemistry

The topographic expression of the ridge crest varies along strike. In Fig. 7.5, gravity
lows associated with a median valley can be recognized along most of the central 70% of
ridge crest length. Similarly, the median valley is seen on most bathymetric crossings (for
example, Fig. 7.8). The lengths of ridge crest where either bathymetric or gravity expres-
sion of the median valley can be detected are dotted in Fig. 7.6. Where seen, the median
valley reaches 1200 m in depth and 20 km in width but more typically is 500-800 m deep
and 6 to 10 km wide. Gravity anomalies associated with these typical dimensions are 20-30
mgal. In the northernmost ridge segment, there are no magnetic data to constrain the ridge
crest position, and more than one elongated gravity low. In the far south, crossings show
neither gravity nor topographic expression (except that a central ridge may occur at the
southern end; Section 4.6), and the average elevation at the ridge crest is shoaler than
farther north.
Four dredge hauls from close to the ridge crest (DR 20, 22, 23, 24 in Fig. 7.7, all
younger than 0.7 Ma in age from their positions) were reported by Saunders and Tamey
(1979). The lavas are all basalts and basaltic andesites. Major and trace element chemistry,
Nd and Sr, Pb and C isotopes (Hawkesworth et aI., 1977; Cohen and O'Nions, 1982; Mattey
et al., 1984 respectively) and volatiles (Muenow et aI., 1980) all show features intermediate
between MORB and arc tholeiites and point to an N-type MORB source variably contami-
nated by a slab-derived component and to varying degrees of partial melting (Saunders and
Tamey, 1991). The slab-derived component is similar to that responsible for island-arc
chemical compositions (generally considered to be an enriched aqueous fluid produced by
low-temperature dehydration of the subducting slab; Saunders and Tamey, 1991) but is less
abundant. These Scotia Sea rocks show no signs of additional complexity beyond a light
carbon-isotope ratio that might reflect a sedimentary component (Mattey et al., 1984).
The apparent simplicity of East Scotia Sea magma genesis is underlined by a compara-
tive study of high field strength and transition elements in arc-backarc systems by Wood-
head et al. (1993). The analyzed rocks of the East Scotia Sea show only minor signs of
depletion of the MORB source by melt extraction before the production of the backarc
melt: for the South Sandwich island arc, prior depletion is estimated at 2% (Pearce et aI., in
press).
There is a wide range of compositional variation between the four dredge sites, with
DR 24 being the most clearly subduction related, DR 22 rather less so, and the other two
only slightly. Sites of DR 24 and DR 22 are the central locations, and DR 23 lies within the
more recently rifted southern province (Section 4.6), but it would be unwise to read any
geometric significance into data from only four sites. The transect of sites dredged up the
northern wall of the backarc (DR 56, 57, 60, 61 in Figures 7.7 and 7.9) underlines this point.
The exact ages of these rocks are unknown, for lack of magnetic data from the most
northerly spreading segment, but they are probably less than 2 Ma. The rocks are fresh
basalts and basaltic andesites, and preliminary XRF major and trace element geochemical
data (P. L. Barber, pers. comm., 1992) point to a strong similarity to DR 24 with, in
particular, low Zr, Ti, P, and Nb and high (with respect to N-MORB) K, Rb, and Ba. The
scale of heterogeneity within the backarc cannot be established without a much more
detailed sampling grid. -
4.5. Western Flank: Older Anomalies

The age of the older part of the western flank of the East Scotia Sea spreading center
bears on several other matters-the nature of the forearc, possible tectonic erosion in the
forearc, interaction with Central Scotia Sea spreading, reconstructions-that are discussed
in subsequent sections. There are no published magnetic anomaly identifications in the
area between 34° and 38°W, and East Scotia Sea opening was only documented back
to 8 Ma.
Previous anomaly identifications out to 4A on the western flank (Barker and Hill,
1981) still hold, and it has been possible to identify western anomaly 5 on many profiles to
the north of about 59.5°S. On some of the profiles in Fig. 7.10 (profiles A and E in
302 PETER F. BARKER
particular), magnetic anomalies can be seen that are a reasonable fit to a synthetic magnetic
profile produced by extrapolating back in time using the same half spreading rate (14 mmI
year) out to anomaly 5B (about 15 Ma). One strong lineation is seen even farther west in
Fig. 7.6. As noted, the bathymetry in this area shows no clear sign of a remnant arc ridge
with a synchronous or consistent onset of spreading. There are topographic elevations
(Figs. 7.4 and 7.8) generally shoaler in the south, but they appear disordered, and all lie
eastward of the 36°W boundary already noted (Section 4.2). On the gravity map (Fig. 7.5),
this topographic roughness shows at first sight a strong linearity, trending about 300°. The
linearity could be associated with fracture zones, since magnetic anomaly sequences
separated by one or more lineations are offset, which would imply an early direction of
opening along 120°-300°. However, the magnetic anomaly orientations, though variable,
retain a mean north-south orientation and, on closer inspection, the gross lineation of the
gravity anomalies is not continuous. There may of course be additional complications that
could generate the kind of bathymetric complexity observed. According to Hill and Barker
(1980), the Central Scotia Sea did not stop opening until about 6 Ma (7 Ma by the MRTS of
Cande and Kent, 1992), after East Scotia Sea opening had started: The boundary between
these two synchronous, orthogonal spreading systems could itself have been chaotic. Also,
the synthetic magnetic profile shows that short sections of the time scale are not distinctive
at slow spreading rates. However, these older anomaly identifications out to C5b may well
be correct, and the rough, elevated topography (Fig. 7.8) may reflect the slower spreading
of that time coupled with a more "normal" ridge crest elevation than the present, anoma-
lously deep average value.
4.6. Southern Province: Ridge Crest Collision and Rifting

South of 59.5°S lies the narrower area of rough, ocean-depth topography between
Herdman Bank and the arc. South again lies the zone of the youngest collision between an
ancestral South Sandwich Trench and a segment of the South American-Antarctic Ridge.
In the area between Herdman Bank and the arc, magnetic anomalies are similar in
amplitude and wavelength to those to the north, and the evidence of dredge DR 23 (Saun-
ders and Tarney, 1979; Fig. 7.7) and earthquake epicenters (Forsyth, 1975; Pelayo and
Wiens, 1989) confirms it as an area of active backarc extension. However, Barker and Hill
(1981) failed to identify any magnetic lineations there, a measure of its greater complexity
and the sparse data base of the time. Hamilton (1989) has since investigated the youngest
collision zone in detail, using magnetic, bathymetric and gravity data, a short length of
seismic profile and several dredge hauls intended to identify the tectonic origin of the major
features from their chemical composition. Hamilton also tentatively identified magnetic
anomalies in the narrow, southern backarc area. Subsequently, a small GLORIA survey of
the collision zone was undertaken during Charles Darwin Cruise 37 in 1989 that also
included the southern margin of the backarc.
The collision zone is extremely complicated and not yet fully understood. It is well
beyond the scope of this chapter to present in full the evidence that supports our current,
imperfect understanding of its structure and evolution. However, it cannot be neglected:
ridge crest collision has interacted strongly with the developing backarc of the East Scotia
Sea and (as argued in Section 2.5) will do so again in the future. I therefore attempt in this
section to present the essentials: (a) a model of spreading history in the southernmost
backarc; (b) the relationship between backarc extension and ridge crest collision.
4.6.1. Backarc Spreading History

Figure 7.1 shows the likely disposition of active plate boundaries at the southern end of
the backarc basin. The backarc spreading center ends at an RFF triple junction: East from
there the dextral transcurrent SAN-ANT boundary is offset southward at a short ridge crest
section, on its way to another, TFF triple junction at the southern end of the South Sandwich
Trench. To the west, the sinistral transcurrent ANT-Sea boundary is also offset south at a
small, slowly spreading ridge section in the basin south of Discovery Bank. These features
appear also in the inset location diagram for profiles G, H, and J in Fig. 7.11. Also shown
in Fig. 7.11 (inset) are Herdman and Discovery banks, the deep trough (dashed line) striking
about 070° that divides them, and Hamilton's (1989) "Area B", an elevated (at 1000-2500
m), largely non-magnetic area with a likely collision suture along its southeast margin.
South of the suture lie long ridges striking slightly south of east that are old short-offset
fracture zones on the Antarctic plate, produced by SAM-ANT separation shortly before
collision (Tectonic Map, 1985; Barker and Lawver, 1988; Hamilton, 1989; Livermore and
Woollett, 1993; Figs. 7.4 and 7.5). These are the several elements essential to an under-
standing of tectonic evolution.
Profiles G and H in Fig. 7.11 show a central 160-2oo-km-wide zone of strongly
. magnetized rough topography at 2-3-km depth, typical of young ocean floor, flanked by
Herdman Bank in the west and part of the volcanic arc and forearc in the east; elevations at
l-km depth or less. These elevations are nonmagnetic, except for the young volcano of
G
4km
SP
r-----------,0
J~:km
H
~ h.rJlrh ~I\.AA SSP
'~V YJI lJ V~V~'
I 2
FIGURE 7.11. Magnetic, bathymetric, and shipboard gravity profiles along three tracks crossing the southern
part of the East Scotia Sea. The synthetic magnetic anomaly profile (SP) compared with profiles G and H is that
used in Fig. 7.10, but that (SSP) compared with profile J has a uniform half spreading rate of 16 mmlyear. Her is
Herdman Bank, Disc is Discovery Bank, N and K are Nelson and Kemp seamounts; B is Hamilton's (1989) "Area
B," an elevated forearc remnant. The gravity anomaly is the thin line in profiles G and H; zeroes of bathymetry and
gravity are common, and 2-km water depth is -100 mgal. The inset location map also shows the ridge crest and
active transforms. The dashed line off Discovery and Herdman banks is a sediment-free trough imaged in Fig.
7.5 and considered to have been the line of relative eastward motion of Herdman Bank before 3.2 Ma. Diamonds
and crosses south of Area B mark the sutures of two recent ridge crest collisions.
304 PETER F. BARKER
Nelson Seamount (N on profile H). At the edges of the oceanic province are deeps (buried
by sediment at the east end of profile G but revealed by the gravity low). The magnetic
anomalies on profile G match the central part of the synthetic magnetic profile (SP) well,
with spreading rates over the past 1-2 m.y. identical to those farther north. This shows that
Herdman Bank was part of the Scotia plate over that time. If that was the case over the full
period of opening of the basin, then opening would have started 3-3.5 m.y. ago. Kemp
Seamount appears to lie on ocean floor no more than 3 m.y. old. The outer parts of profile G
are not as good a match to the synthetic anomalies as the central part: Hamilton (1989)
explained the offset of the central, Brunhes-age anomaly from the basin center by a small
ridge jump before 2 Ma. Profile H is a much poorer match to SP and is included to show the
possibility of greater complexity. Hamilton's (1989) correlations are shown in Fig. 7.11 for
profile H: the central anomaly includes an unusually large axial magnetic low and a central
topographic high, which are not unknown at ridge crests but are unusual here. Alter-
natively, one or more ridge jumps may also have affected this section.
Profile J of Fig. 7.11 is from a different, more southerly spreading province, offset
eastward and nominally representing SAN-ANT motion. The rough topography is more
elevated, and there are no flanking topographic highs. To the east lies the South Sandwich
trench and to the west Hamilton's "Area B," which he concludes is a fragment of
precollision forearc. Both the magnetic province and the well-formed central anomaly are
narrower than on profile G. They match a synthetic anomaly (SSP) generated using a full
spreading rate of 32 mmlyear. At that rate, the magnetic region would also be generated in
3-3.5 m.y. However, the rate is only 60% of estimated SAN-ANT motion, about 20 mmJ
year too slow.
There are many possible explanations for this discrepancy. Some call into question the
correctness of the path of the present Scotia-ANT and Sandwich-ANT plate boundaries in
Figs. 7.1 and 7.11, and the present state of knowledge leaves open a range of alternatives.
However, the simplest and perhaps most plausible is active stretching of Area B and its
western flank. The six dredges from Area B considered by Hamilton (1989) yielded a wide
range of lithologies which, if dredged in situ, would signify a continental origin. Addi-
tionally, however, there were indications of young igneous activity, mostly acidic. More-
over, Area B is nonmagnetic except for a small area (marked "B" on profile J) directly
south of the present ridge crest (the central ridge on profile H). It is possible that Area Band
its western flank are being pervasively stretched, with thinning and remelting of continental
crust, and are thus accommodating part of the SAN-ANT motion.
4.6.2. Ridge Crest-Trench Collision

The relationship between rifting in the southern backarc and ridge crest-trench colli-
sion directly to the south is important. That rifting was caused by collision is a part of the
general hypothesis concerning the effects of successive collisions on backarc extension in
the Scotia Sea over the past 30 Ma or more. It would have been useful to examine the time
relationships of collision and rifting here, to test that hypothesis. Unfortunately the evi-
dence is incomplete, so I offer what is essentially a restatement of the hypothesis as it
applies in this particular case.
Hamilton (1989) identified two narrow ridge crest-trench collision zones in the south
of this area. Their ages could not be determined independently of data from the backarc, so
they were assumed to be virtually contemporaneous and to coincide with the time of rifting
of Herdman Bank and Area B from the forearc. The magnetic data discussed in Section 4.6
suggest this happened about 3.2 Ma. The extent ofthese two collision zones is marked (with
respect to Area B) by crosses and diamonds in Fig. 7.11. Strictly, Hamilton considers that
the ridge crest would have arrived at the trench earlier and that a short period of accelerated
tectonic erosion of the lower forearc accretionary prism would have preceded collision with
the rigid Sandwich plate: Area B and Herdman Bank are considered to have been parts of
the upper forearc.
It is assumed that prior to 3.2 Ma, as part of the Sandwich plate, Herdman Bank and
Area B were moving eastward away from Discovery Bank along the line of the deep trough
shown dashed in Fig. 7.11 (inset). A reconstruction to about 3.2 Ma (within the older part of
anomaly 2A) ofthe southern province ofthe backarc is shown in Fig. 7.12A. It incorporates
a restored Herdman Bank and Area B, west of the present forearc. The reconstruction is
referred to a presumed rigid Scotia plate.
The hypothesis that rifting followed ridge crest collision is attractive. Ridge crest
subduction could have provided the heat source needed for pervasive extension of Area B
and might have instigated the rifting to adjust the line of the Scotia-ANT and Sandwich-
ANT plate boundaries. Prior to collision, the southern backarc lay west of Herdman Bank
and south of Discovery Bank. That region is now being dissected further by Scotia-ANT
motion, as previously inferred (Tectonic Map, 1985; Pelayo and Wiens, 1989, Barker et ai.,
1991). Discovery Bank has yielded Miocene arc volcanic rocks (12-20 Ma; Barker et ai.,
1982). A forearc comprising the largely nonmagnetic and possibly continental Area Band
Herdman Bank to the southeast would have been matched by a backarc volcaniclastic apron
in the gently sloping sedimented area northwest of Discovery Bank (Fig. 7.4). Parts of both
Herdman Bank and the congruent eastern forearc flank ofthe backarc basin are flat-topped,
3-2Ma
mc~SAN
A ~:"-£-i.---- c
~
FIGURE 7.12. Revised reconstructions of East Scotia Sea evolution (A) to ca. 3.2 Ma (B) to 10.5 Ma. Anomalies
3A and 5B are dashed as in Fig. 7.6, and the ridge crest position is dotted where constrained by magnetic anomaly.
(C) Vectors of SAN motion with respect to Central Scotia Sea opening (SSCO and NSCO) before 7 Ma,
compatible with East Scotia Sea fabric and HerdmanJDiscovery Bank motion. In (A) the -60-mgal contour from
the forearc in Fig. 7.5 defines the eastern edge of the Sandwich plate, and the +80-mgal contour defines the
elevations within the arc and forearc. The 2-km contour from Fig. 7.4 defines the South Georgia block. Diamonds
and crosses mark the sutures of recent ridge crest-trench collisions south of Area B.
306 PETER F. BARKER
which Hamilton (1989) attributed to planation by Weddell Sea icebergs. It seems likely that
both bodies would have been elevated by the thermal effects of rifting.
Thus, in summary, the effects of the youngest collision(s) were to erode the lower
forearc, to cause subduction to cease at the collision zone, to transfer older forearc blocks
from the Sandwich to the Antarctic and Scotia plates, and to modify the locus of backarc
extension. This is compatible with the general hypothesis of the effects of collision.
For the next collision zone to the west, Hamilton (1989) correlated magnetic anoma-
lies with two alternative sections of the MRTS, implying collision ages of 3 Ma or 10 Ma,
both using Barker and Lawver's (1988) rotations to constrain spreading rates. If this
constraint is relaxed (by "pivoting subduction"; Section 2.5), other identifications are
possible. The older age makes more sense of the backarc basin history: a 10 Ma collision
opposite Discovery Bank could have triggered the dissection of its forearc, with eastward
migration of Herdman Bank as proposed above continuing until the 3.2 Ma rifting event.
This older, western collision zone lies west of the zone marked by diamonds in Fig. 7.12A.
Backarc extension has accelerated over the past few million years (Fig. 7.10). Is this
too related to the collisions? Certainly, to the extent that rollback drives backarc extension,
the cessation of extension in Drake Passage and the Central Scotia Sea (Barker and Burrell,
1977; Hill and Barker, 1980), about 6 m.y. ago (7 Ma using the Cande and Kent, 1992
MRTS), would have required faster extension within (most probably) the only backarc
spreading regime that remained. The acceleration apparent in Fig. 7.10 began at about
7 Ma. In addition, however, several factors associated with the collisions may have
contributed to the more recent (ca. 2 Ma) acceleration:
(a) The very young, buoyant ocean floor in the far south, being subducted with difficulty before
collision, would remelt and become detached from the sinking SAM plate, allowing the main
slab to sink more quickly.
(b) In the backarc, the long transform offset along the line S of Discovery Bank was eliminated,
reducing drag.
(c) The rifting process required by collision was completed, and new ridge segments, providing
ridge push, were put in place in the backarc.
4.7. Eastern Flank: Volcanic Arc, Forearc and Trench

On the eastern flank of the backarc basin, many members of the volcanic island arc lie
on ocean floor produced during the past 10 m.y. The main islands of the central section of
the arc may lie on ocean floor possibly as old as 11 Ma, but the northern and southern
extremities lie on younger ocean floor. Kemp and Nelson seamounts, already mentioned, lie
on or close to ocean floor no older than about 3 Ma, and the more westerly of the seamounts
inferred from the gravity map (Fig. 7.5) lie on ocean floor as young as 6 to 8 Ma. The
magnetic signature of the arc rocks overprints any seafloor spreading record, but the older
anomalies may perhaps be seen along some passage tracks through the arc. Profiles C and
D in Fig. 7.10 show a possible anomaly 5 on the eastern flank, and the sequence to anomaly
5B may even be preserved in the forearc (though this identification is highly speculative).
Previous studies (e.g., Barker and Hill, 1981) had suggested that before the start of
east-west extension in the East Scotia Sea, a north-south backarc extensional province in
the Central Scotia Sea, active between 21 and 7 Ma, may have reached as far east as the
trench (Fig. 7.2B,C).1t was thought that the northern half of the forearc province may have
been produced by this extensional regime, whereas the southern half was already in
existence and may even have been continental. This, rather than the lesser age of the
subducting South American ocean floor, was thought to account for the greater elevation of
the forearc in the south.
Although there are now more magnetic data from the forearc than were considered by
Barker and Hill (1981) it remains impossible to map magnetic lineations there with any
confidence. The identifications in Fig. 7.10 are tentative, although if the older anomalies on
the western flank are correctly identified then the presence of corresponding lineations in
the forearc is to be expected.
As important to an understanding of the forearc are the other data (single-channel
seismic profiles, dredge hauls, and Geosat gravity) that are now available. The gravity
signature of the forearc is dominated by the steep gradient, common in forearcs, that
separates the upper forearc, underlain by rigid lithosphere of the Sandwich plate, from the
deforming accretionary prism of the lower forearc resting directly on the subducting South
American lithosphere. In the upper forearc, gravity varies considerably along strike. The
simple concept of a north-south contrast needs revision: There is high gravity both in the far
northeast and in the center south, but there are also low areas in the south as well as the
center north. In discussing the trench (Section 4.1.2), I pointed to variations in the height of
the mid-slope basement high as the cause of variation in the amount of sediment that could
be retained on the upper forearc. These variations in basement elevation and sediment cover
probably account for most of the gravity variation.
The outer slopes of both sections of the mid-slope high that show the major gravity
highs have been dredged: the northeast corner east of Zavodovski Island and the large
gravity high east of Montagu Island (Fig. 7.7, and profiles Q and R of Fig. 7.9). Two
dredges from the northeast corner of the upper forearc (DR 52 and DR 53, 56.1 oS, 26.2°W:
4100-3100 m depth, and profile Q) recovered in situ serpentinized peridotite (P. L. Barber,
pers. comm., 1992). Two others from east of Montagu Island (DR 49 and DR 50, 58.6°S,
24.8°W: 2100-1100 m depth, and profile R) yielded calc-alkaline basalts, basaltic andesites
and andesites, both fresh and altered. Three samples from DR 49 and DR 50 yielded K-Ar
ages in the range 28.5 to 32.8 Ma (weighted average 31.0 Ma; Table I).
TABLE I
KIAr Ages of Rocks Dredged from the South Sandwich
Forearc (DR 49 and DR 50, Fig. 7.7),
Determined by R. J. Pankhurst and I. Millara
Sample K (%) 40Ar (nllg) Percent (atm) Age (Ma)
DR 49/8 0.950 1.0810 32.1 29.1 ± 0.8

1.0608 37.2 28.5 ± 0.9
DR 50/1 1.097 1.3604 55.8 31.7 ± 0.8
1.3530 45.4 31.5 ± 1.0
DR 50/3 0.668 0.8586 64.8 32.8 ± 3.2
0.8511 63.3 32.4 ± 1.5
aDecay constants as in Steiger and Jaeger (1977).
308 PETER F. BARKER
The origin of the peridotites from dredges DR 52 and 53 is uncertain. Among the
possibilities are
(a) a rafted fragment of a late Jurassic backarc basin at the Pacific margin (seen also on South
Georgia; Macdonald et al.. 1987); this is less likely than either
(b) serpentinite diapirs. seen also in the Marianas upper forearc and considered there to result
from diapiric uplift of buoyant serpentinized fault blocks. and serpentinite mud volcanoes
(e.g .• Fryer. 1992). within an extensional forearc environment; or
(c) part of the mantle of the Sandwich plate. tectonically exposed within the trench inner wall.
This would imply an additional tectonic unroofing mechanism. such as tectonic erosion
at the plate margin (Section 4.1) or uplift and subaerial erosion ofthe overlying oceanic crust
of the Sandwich plate at its northeast comer. It may be highly significant that the scarp from
which these rocks were dredged lies almost exactly on the line of the steeply dipping zone of
deep earthquake activity along 56°S. associated by Forsyth (1975) with tearing of the
downwarped South American plate at depth. at the northern end of the subducting slab
(Section 2.4 and Fig. 7.3).
The mid-Oligocene calc-alkaline basalts and andesites dredged from the elevation east
of Montagu Island may be explained in tenns of Scotia Sea evolution. The Scotia Sea has
grown by backarc extension over the past 30-40 Ma. and has carried eastward fragments of
an original continuous continental connection of South America and the Antarctic Penin-
sula (see. for example, Barker et at., 1991). The elevation sampled by these dredges may
have lain at that margin originally, having fonned as a result of subduction of Pacific ocean
floor. More probably, it was produced at an early stage of Scotia Arc development, before
the arc products of east-directed subduction had acquired their present intraoceanic low-K
tholeiitic character (Discovery Bank rocks, dated at 12-20 Ma, are low-K tholeiites: Barker
et al., 1982). It is plausible that successive episodes of backarc extension subsequently
could have allowed a fragment of that original arc to survive within the east-migrating
forearc.
Identification of the two main elevated areas within the upper forearc as (a) tec-
tonically exposed mantle and (b) a fragment of a much older arc is compatible with, but
does not prove, a simple ocean-floor origin for the remaining, less-elevated parts. Strong
support for this interpretation comes from the older magnetic anomalies identified on the
western flank of the backarc basin and from the coincidence within the forearc of sub-
sediment basement depths and free-air gravity anomalies similar to those of "nonnal"
ocean floor of the same likely age.
4.8. Reconstructions
Figure 7.12 shows reconstructions of East Scotia Sea evolution at (a) 3.2 Ma and
(b) 10.5 Ma. These times were chosen (a) to demonstrate the effects of the two most recent
ridge crest collisions and (b) to compare the results of this reassessment with the recon-
structions of Fig. 7.2B. The eastern boundary of the Sandwich plate in Fig. 7.12 is taken
from the free-air gravity map (Fig. 7.5) at the -60-mgal contour, within the steep gravity
gradient that separates the upper and lower forearc. Choice of a point on that gradient is
unimportant in most areas, provided it is unlikely to reflect "nonnal" oceanic crustal
structure. The island-arc and anomalous forearc elevations are defined by the +80-mgal
contour for the same reason. For other bodies, not within the Sandwich plate, the level
chosen varies to take into account their assumed structure and present surroundings. For
the boundary of the South Georgia microcontinent, the 2000-m bathymetric contour is
used.
The 3.2 Ma reconstruction (Fig. 7.l2A) closes the young southern basin of the
backarc, under assumptions discussed in Section 4.6. Anomalies 2A on either flank fit
precisely under a rigid two-plate assumption, and Hamilton's (1989) elevated Area Band
Herdman Bank become more southwesterly components of the forearc. Although this
reconstruction is not placed in major plate coordinates (the Scotia plate remains fixed), it is
assumed that two ridge crest collisions have taken place in the past 3.2 Ma, so that the
SAM-ANT boundary met the forearc then farther to the southwest than at present, and that
Scotia-ANT motion had been about 12 mmlyear (from Pelayo and Wiens, 1989) through-
out. The crosses and diamonds mark the extent of these young collisions.
Little is known of the northern limit of the backarc, except that dredges DR 56-61
suggest that most of the present northern margin is original. It seems most likely (by
comparison with the corresponding western ridge flank in Fig. 7.l2A) that the presently
elevated northeastern comer of the Sandwich plate was formed by simple backarc exten-
sion and has been subsequently elevated, eroded, and intruded by subduction-related
magma.
A reconstruction to 10.5 Ma is not so straightforward to achieve: anomalies older than
3A on either flank do not quite coincide under a similar simple two-plate assumption. The
reconstruction in Fig. 7.l2b is therefore speculative and involves additional complexity and
less constraint. Hill and Barker (1980) concluded that north-south spreading in the Central
Scotia Sea did not cease until about 6 Ma (7 Ma using the Cande and Kent, 1992 time scale).
Those Central Scotia Sea age determinations should be re-assessed in light of additional
data now available, but if the original ages are accepted then the western flank of the East
Scotia Sea backarc spreading center CQuld have been two separating plates until 7 Ma. As
mentioned, the ocean floor fabric on the western flank of the East Scotia Sea spreading
center is confused: The apparent 120°-300° gross orientation visible in the Geosat gravity
map (Fig. 7.5) is discontinuous in detail, and it is difficult to see where a triple junction,
implied by the coexistence of two orthogonal spreading centers before 7 Ma, could have
lain. The main fracture zones in the East Scotia Sea (Fig. 7.6) are oblique to the prominent
lineation south of Discovery Bank, which is presumed to show the direction of Sandwich
plate motion relative to a "southern Scotia plate" (SSCO) before the 3.2 Ma rifting. Fig.
7.l2C shows this obliquity to be roughly compatible with a three-plate system with the
observed spreading rates. One possible scheme for the period before 7 Ma would therefore
locate the triple junction in the south, somewhere close to Discovery Bank.
In the north, a 10.5 Ma reconstruction brings the forearc to the eastern end of South
Georgia. Although the South Georgia block seems likely to have occupied a similar forearc
or arc position through the previous 20-30 m.y. of Scotia Sea evolution (see Fig. 7.2), there
is no sign of Cenozoic arc magmatism on the island or continental shelf (Simpson and
Griffiths, 1982; Macdonald et al., 1987). A likely sinistral strike-slip fault with about 70 km
offset, crossing the continental shelf of the South Georgia block, is picked out by the gravity
anomaly (Fig. 7.5) and shown dashed in Fig. 7.l2B. It is not restored in the 10.5 Ma
reconstruction, since there are no constraints upon its age.
This reconstruction is essentially compatible with the 10 Ma reconstruction shown in
Fig. 7.2B. The backarc nature of the extensional environment is unchanged, and the history
of opening of Drake Passage farther west is common to both. The balance of ocean-floor
creation by the central and eastern Scotia Sea spreading regimes is different, because the
310 PETER F. BARKER
latter started earlier than previously realized. In combination, however, they had the same
effect. No attempt is made in Fig. 7.12B to place the reconstruction in a context of slow east-
west major plate motion, but it seems plausible that the north-south opening of the Central
Scotia Sea would have involved both a northward component of South Georgia motion with
respect to South America (until collision with the Northeast Georgia Rise, perhaps) and a
southward component of motion of Discovery Bank (leading to collision with sections of
the SAM-ANT plate boundary). These are also assumptions of earlier models but cannot be
tested until the detailed history of ridge crest collision along the South Scotia Ridge is
known.
In the 10.5 Ma reconstruction, islands of the present arc are not shown. None appears
likely to have been erupted on ocean floor older than 11 Ma. The Discovery arc to the south
(now exposed in Discovery Bank) would have been active before 10 Ma, and one un-
sampled elevation in the present southern forearc may also have been an older arc volcano.
Other possible components of a pre-IO Ma island arc are two seamounts lying to the south
of South Georgia (Figs. 7.4 and 7.5, 7.12). Less speculatively, if the identification of
anomaly 5B south of South Georgia is correct in the reconstruction in Fig. 7.12B, then some
subsequent tectonic erosion of the northeast comer of the Sandwich plate is implied, since
the congruent ocean floor of anomaly 5B age on .the Sandwich plate is missing.
The combination of implied tectonic erosion (above) with abnormal elevation and
exposure of serpentinized peridotite at shallow depth, at the present northeast comer of the
Sandwich plate above the northern end of the subducting slab, is quite compelling.
Moreover, it raises an additional question. The exact line of the tear of the subducting South
American oceanic plate (Fig. 7.3) lies beneath the newly formed Sandwich plate, so the
downwarped South American oceanic lithosphere to the north of the tear cannot return to its
original surface position; it is held down by the backarc lithosphere. The bathymetric
profile across the northern, east-west arm ofthe South Sandwich trench (profile P, Fig. 7.9)
closely resembles a subducting trench profile, although relative motion across it has been
sinistral strike slip. How close are these circumstances to those of the initiation of subduc-
tion? If ridge crest collisions have at several times in the past caused subduction at a
southern part of the subduction zone to stop, has a similar asymmetry at the northern end
of the trench ever developed into subduction so that subduction and complementary
backarc extension might continue?
5. CONCLUSIONS
This has been an opportunity to reexamine and extend a suite of conclusions concern-
ing the evolution of the East Scotia Sea, a site of active backarc extension, in the light of a
larger and more diverse data set than was previously available.
(a) The East Scotia Sea floor was generated at a spreading center now lying approx-
imately along 30 0 W longitude. Well-formed magnetic anomalies may be identified back to
anomaly 5, and probably to anomaly 5B (about 15 Ma) on the western flank. To the east, the
island arc lies on ocean floor aged from about 10 Ma to as young as 3 Ma, all produced
during the present spreading episode. The time extent of East Scotia Sea spreading and the
area of ocean floor produced are greater than previously thought.
(b) Spreading was largely symmetric at about 27 mm/year full rate, from 15 Ma to
about 5-7 Ma, when it started to accelerate. From about 4 Ma to 1.7 Ma spreading was at 50
Gravity Anoma/y (mgaI) by D. 5andwe11 . 5.1.0 .
• •
-- BAnmEnllCllAPOF BRANSFIELD STRAIT AND VICINITY

...... ~fW.A. .......... O'--C" .... Il...._r..~ ..... nc
a.-.~
~a.-.
~ U£A.
__ •
.... ...
FIGURE B.S. GeostatlERS-1 gravity data from Sandwell and Smith (1992) with digital bathymetric data of
Klepeis and Lawver (1994) superimposed. The gravity data color bar indicates gravity values from < -40 mgal to
> + 30 mgal. Bold black bathymetric contours are shown every 500 m, while thin black contours are observed
bathymetry contoured every 100 m. Thin blue contours are interpolated bathymetric data contoured every 100 m.
mm/year and slightly asymmetric, favoring the eastern, arc flank. At 1.7 Ma a further
acceleration to 65 mm/year occurred, with up to 15% asymmetry in many places, favoring
the arc flank in most cases. In places, transform faults appeared about 1.7 Ma, separating
segments having different degrees of asymmetric spreading. There may be a link between
spreading rate, ridge crest migration, asymmetry, and ridge segmentation.
(c) The ridge crest depth exceeds 3000 m, more than 500 m deeper than the global
average for mid-ocean ridges. A median valley is present along about 70% of the ridge crest
length; typical dimensions are 6-10 km wide and 500-800 m deep, with a free-air gravity
anomaly of 20-30 mgal. The ridge crest depth may have been closer to the global average
before the accelerations in spreading.
(d) Sediment cover on the east flank is a thick arc-derived volcaniclastic apron. On the
west flank, sediment is pelagic and hemipelagic, deposited under the influence of the
Antarctic Circumpolar Current. The west flank becomes rougher westward, reflecting a
slower spreading rate and perhaps interaction with contemporaneous north-south extension
in the Central Scotia Sea. This rough zone ends at about 36°W.
(e) In the south, the backarc basin is much narrower. This zone was formed after
Herdman Bank and another forearc fragment rifted from the Sandwich plate ca. 3.2 Ma,
probably in response to collision of the trench with two sections of the South American-
Antarctic Ridge to the west of any surviving part of the SAM-ANT plate boundary.
Herdman Bank became a part of the Scotia plate following rifting, and the other fragment a
part of the Antarctic plate. Before then, and following an earlier collision farther west
(possibly 10 Ma), Herdman Bank was moving eastward from its original position as forearc
to the Discovery arc, as part of the Sandwich plate. The precise timing of these collisions
remains uncertain.
(f) The present SCO-ANT boundary runs south of Herdman Bank. To the east lies the
RFF SAN-SCO-ANT triple junction, then a slow spreading (32 mm/year) SAN-ANT ridge
segment, and a TFF SAN-SAM-ANT boundary. Some crustal thinning and intrusion of
"Area B" may be required, to complete SAN-ANT motion.
(g) Ten previously unpublished dredge hauls are reported briefly: four from the
forearc, four from the backarc, and two from previously unknown seamounts related to the
island arc. The chemistry of all arc and backarc rocks appears to reflect production from
N-MORB, variably contaminated by fluids enriched in alkali metals from the subducted
slab. They are similar to rocks from other intraoceanic arc-backarc systems, but possibly
simpler, having suffered only minor prior source depletion. However, chemical variation in
the arc rocks is geographically heterogeneous and bears no visible relationship to known
systematic north-south variation in age and sediment cover of the subducting slab. The
sparse distribution of backarc samples permits no conclusion on the scale of spatial or
temporal heterogeneity.
(h) The upper forearc is probably ocean floor formed on the east flank during the early
stages of the present spreading episode, for the most part. An exception is a more elevated
area east of Montagu Island, shown from dredge hauls to be a fragment of a ca. 31 Ma calc-
alkaline volcanic arc, active during the early stages of Scotia Sea evolution.
(i) The other major forearc elevation, at the northern end, has yielded serpentinized
ultrabasic rocks from its steep northeastern scarp. The magnetic anomaly pattern and the
ultrabasic outcrop suggest an extensional forearc tectonic regime and likely tectonic
erosion. This scarp lies directly beneath the east-west line of tearing of the subducted slab,
defined by earthquake fault-plane solutions. Since the northern edge of the torn South
312 PETER F. BARKER
American lithosphere underlies newly created backarc, it cannot return to its original
position: the asymmetry may create the conditions necessary for incipient subduction.
U) The lower forearc accretionary prism is poorly developed. Variations in the
elevation of the lip to the upper forearc seem to control the transport of arc volcaniclastics
to the trench; basement elevation and sediment thickness dictate free-air gravity levels.
Acknowledgments
I am grateful to Paul Barber and Ian Hamilton for permission to summarize unpub-
lished petrographic and geochemical data from 10 dredge hauls from the East Scotia Sea
and South Sandwich forearc. Bob Pankhurst and Ian Millar produced K-Ar ages for rocks
from two of those dredge hauls. Julian Pearce and Peter Baker kindly provided an early
draft of their paper describing South Sandwich arc volcanic rocks.
REFERENCES
Allen, A. 1966. Seismic refraction investigations in the Scotia Sea, Br. Antarct. Surv. Sci. Rep. 55: 44 pp.
Alvarez, W. 1982. Geological evidence for the geographical pattern of mantle return flow and the driving
mechanism of plate tectonics, 1. Geophys. Res. 87:6697-6710.
Atwater, T.1989. Plate tectonic history of the northeast Pacific and western North America, in The Eastern Pacific
Ocean and Hawaii (E. L. Winterer, D. M. Hussong, and R. W. Decker, eds.), pp. 21-72, Geol. Soc. America,
Boulder.
Baker, P. E.1968. Comparative volcanology and petrology of the Atlantic island arcs, Bull. Volcanol. 32:189-206.
Baker, P. E. 1978. The South Sandwich Islands: iii. petrology of the volcanic rocks, Br. Antarct. Surv. Sci. Rept. 93:
34 pp.
Baker, P. E. 1990. E. South Sandwich Islands, in Volcanoes of the Antarctic Plate and Southern Oceans (W. E.
LeMasurier and J. W. Thomson, eds.), Vol. 48, pp. 361-395, Antarctic Res. Ser., American Geophysical
Union, Washington, DC.
Barker, P. E 1970. Plate tectonics of the Scotia Sea region, Nature 228:1293-1296.
Barker, P. E 1972. A spreading centre in the East Scotia Sea, Earth Planet. Sci. Lett. 15:123-132.
Barker, P. E 1979. The history of ridge-crest offset at the Falkland-Agulhas fracture zone from a small-circle
geophysical profile, Geophys. 1. R. Astron. Soc. 59:131-145.
Barker, P. E 1982. The Cenozoic subduction history of the Pacific margin of the Antarctic Peninsula: ridge crest-
trench interactions, 1. Geol. Soc. London 139:787-801.
Barker, P. E, Barber, P. L., and King, E. C. 1984. An early Miocene ridge crest-trench collision on the South Scotia
Ridge near 36W, Tectonophysics 102:315-332.
Barker, P. E, and Burrell, J. 1977. The opening of Drake Passage, Mar. Geol. 25:15-34.
Barker, P. E, Dalziel, I. W. D., and Storey, B. C. 1991. Tectonic development of the Scotia Arc region, in Geology
of Antarctica (R. J. Tingey, ed.), pp. 215-248, Oxford University Press, Oxford.
Barker, P. E, and Hill, I. A. 1980. Asymmetric spreading in back-arc basins, Nature 285:562-564.
Barker, P. E, and Hill, I. A. 1981. Back-arc extension in the Scotia Sea, Phil. Trans. R. Soc. Lond. A 300:249-262.
Barker, P. E, Hill, I. A., Weaver, S. D., and Pankhurst, R. J.1982. The origin of the eastern South Scotia Ridge as an
intra-oceanic island arc, in Antarctic Geoscience (C. Craddock, ed.), pp. 203-211, Univ. Wisconsin Press,
Madison.
Barker, P. E, and Lawver, L. A. 1988. South American-Antarctic plate motion over the past 50 Myr, and the
evolution of the South American-Antarctic Ridge, Geophys. 1. R. Astron. Soc. 94:377-386.
Barrow, J. Sir. 1830. Account of the island of Deception, one of the New Shetland Isles, 1. R. Geogr. Soc. 1:62-66.
Brett, C. P. 1977. Seismicity of the South Sandwich Islands region, Geophys. 1. R. Astron. Soc. 51:453-464.
Brett, C. P., and Griffiths, D. H. 1975. Seismic wave attenuation and velocity anomalies in the eastern Scotia Sea,
Nature 253:613-614.
Cande, S. C., and Kent, D. V. 1992. A new geomagnetic polarity time scale for the Late Cretaceous and Cenozoic,
1. Geophys. Res. 97:13,917-13,951.
Cande, S. c., Leslie, R. B., Parra, J. C., and Hobart, M. 1987. Interaction between the Chile Ridge and Chile
Trench: geophysical and geothermal evidence, J. Geophys. Res. 92:495-520.
Chase, C. G. 1978. Extension behind island arcs and motions relative to hotspots, J. Geophys. Res. 83:5385-5388.
Cohen, R. S., and O'Nions, R. K. 1982. Identification of recycled continental material in the mantle from Sr,
Nd and Pb isotope investigation, Earth Planet. Sci. Lett. 61:73-84.
Cook, J. 1777. A voyage towards the South Pole and around the World Performed in His Majesty's Ships the
"Resolution" and "Adventure" in the Years 1772-75, Shanan and Cadell, London.
Debenham, F. 1945. The Voyage o/Captain Bellingshausen to the Antarctic Seas, 1819-1821 (translated from the
Russian), Vols. 1,2, Hakluyt Society, London.
DeMets, c., Gordon, R. G., Argus, D. F., and Stein, S. 1990. Current plate motions, Geophys. J. Int. 101:425-478.
Elsasser, W. M. 1971. Sea floor spreading as thermal convection, J. Geophys. Res. 76:1I01-1112.
Ewing, 1.1., Ludwig, W. J., Ewing, M., and Eittreim, S. L.1971. Structure ofthe Scotia Sea and Falkland Plateau, J.
Geophys. Res. 76:7118-7137.
Forsyth, D. W. 1975. Fault plane solutions and tectonics of the South Atlantic and Scotia Sea, J. Geophys. Res.
80:1429-1443.
Fryer, P. 1992. A synthesis of Leg 125 drilling of serpentine seamounts on the Mariana and Izu-Bonin fore-arcs, in
Proc. ODP. Sci Results, 125 (P. Fryer, J. A. Pearce, et aI., eds.), pp. 593-614, Ocean Drilling Program,
Gass, I. G., Harris, P. G., and Holdgate, M. W. 1963. Pumice eruption in the area of the South Sandwich Islands,
Geol. Mag. 100:321-330.
Griffiths, D. H. 1963. Geophysical investigations in the Scotia Arc and Graham Land, Br. Antarct. Surv. Bull. 1:
27-32.
Griffiths, D. H., Riddihough, R. P., Cameron, H. A D., and Kennet, P. 1964. Geophysical investigations of the
Scotia Arc, Br. Antarct. Surv. Sci. Rep. 46: 43 pp.
Gripp, A E., and Gordon, R. 1. 1990. Current plate velocities relative to the hotspots incorporating the NUVEL-I
global plate motion model, Geophys. Res. Lett. 17:1I09-1112.
Hamilton, I. W. 1989. Geophysical investigations of subduction-related processes in the Scotia Sea, unpublished
Ph.D. thesis, Birmingham University, U.K.
Harland, W. B., Cox, A. V., Llewellyn, P. G., Pickton, C. A G., Smith, A G., and Walters, R. 1982. A Geologic
Time Scale, Cambridge University Press, Cambridge.
Harrington, P. K., Barker, P. F., and Griffiths, D. H. 1972. Crustal structure of the South Orkney Islands area from
seismic refraction and magnetic measurements, in Antarctic Geology and Geophysics (R. 1. Adie, ed.), pp.
27-32, Universitetsforlaget, Oslo.
Hawkes, D. D. 1962. The structure of the Scotia Arc, Geol. Mag. 99:85-91.
Hawkesworth, C. J., O'Nions, R. K., Pankhurst, R. J., and Evensen, N. M. 1977. A geochemical study of island-arc
and back-arc tholeiites from the Scotia Sea, Earth Planet. Sci. Lett. 36:253-262.
Heezen, B. C., and Johnson, G. L. 1965. The South Sandwich Trench, Deep-Sea Res. 12:185-197.
Herdman, H. F. P. 1948. Soundings taken during Discovery investigations 1932-1939, Discovery Reports 25:
39-106.
Hill, I. A, and Barker, P. F. 1980. Evidence for Miocene back-arc spreading in the central Scotia Sea, Geophys. J.
R. Astron. Soc. 63:427-440.
Holdgate, M. w., and Baker, P. E. 1979. The South Sandwich Islands. I: General description, Br. Antarct. Surv. Sci.
Rept. 91: 76 pp.
Isacks, B., and Molnar, P. 1971. Distribution of stresses in the descending lithosphere from a global survey of focal
mechanism solutions of mantle earthquakes, Rev. Geophys. Space Phys. 9:103-174.
Jarrard, R. D. 1986. Relations among subduction parameters, Rev. Geophys. 24:217-284.
Kemp, S., and Nelson, A L. 1931. The South Sandwich Islands, Discovery Reports 3:133-198.
LaBrecque, J. L., Kent, D. V., and Cande, S. C. 1977. Revised magnetic polarity time scale for late Cretaceous and
Cenozoic time, Geology 5:330-335.
Langel, R. A 1992. International geomagnetic reference field: The sixth generation, J. Geomagn. Geoelectr.
44:679-707.
Lawver, L. A., and Dick, H. J. B. 1983. The American-Antarctic Ridge. J. Geophys. Res. 88:8193-8202.
Livermore, R. A., and Woollett, R. W. 1993. Seafloor spreading in the Weddell Sea and southwest Atlantic since
the late Cretaceous, Earth Planet. Sci. Lett. 117:475-495.
Ludwig, W. J., and Rabinowitz, P. D. 1982. The collision complex of the North Scotia Ridge, J. Geophys. Res.
87:3731-40.
314 PETER F. BARKER
Luff, I. w., 1982. Petrogenesis of the island arc tholeiite series of the South Sandwich Islands, unpublished Ph.D.
thesis, University of Leeds, U.K.
Macdonald, D.1. M., Storey, B. c., and Thomson, J. W. 1987. South Georgia, p. 63, BAS GEOMAP Series, Sheet
I, 1:250,000, Geological map and supplementary text, British Antarctic Survey, Cambridge.
Mattey, D. P., Carr, R. H., Wright, I. P., and Pillinger, C. P. 1984. Carbon isotopes in submarine basalts, Earth
Matthews, D. H. 1959. Aspects of the geology of the Scotia Arc, Geol. Mag. 95:425-441.
Maurer, H., and Stocks, T. 1933. Die echolotungen des Meteor. Wiss. Ergebn. Deut. Atlantic Expedition Meteor;
1925-27, pp. 1-309.
McAdoo, D. C., and Marks, K. M. 1992. Gravity fields of the Southern Ocean from Geosat data, J. Geophys. Res.
97:3247-3260.
Menard, H. W. 1978. Fragmentation of the Farallon plate by pivoting subduction, J. Geol. 86:99-110.
Minster, J. B., and Jordan, T. H. 1978. Present-day plate motions, J. Geophys Res. 83:5331-5354.
Muenow, D. W., Liu, N. W. K., Garcia, M. 0., and Saunders, A. D. 1980. Volatiles in submarine volcanic rocks
from the spreading axis of the East Scotia Sea back-arc basin, Earth Planet. Sci. Lett. 47:272-278.
NGDC-NOAA. 1993. Global Relief CD-ROM Data Set, National Geophysical Data Center, Boulder.
Parsons, B. L., and Sclater, 1. G. 1977. An analysis of the variation of ocean floor bathymetry and heat flow with
age, J. Geophys. Res. 82:803-827.
Pearce, J. A. 1982. Trace element characteristics of lavas from destructive plate boundaries, in Andesites:
Orogenic Andesites and Related Rocks (R. S. Thorpe, ed.), pp. 525-548, Wiley, Chichester.
Pearce, J. A., Baker, P. E., Harvey, P. E., and Luff, I. W. in press. Geochemical evidence for subduction fluxes,
mantle melting and fractional crystallization beneath the South Sandwich island are, J. Petrol.
Pelayo, A. M., and Wiens D. A. 1989. Seismotectonics and relative plate motions in the Scotia Sea region, J.
Geophys. Res. 94:7293-7320.
Saunders, A. D., and Tamey, 1. 1979. The geochemistry of basalt from a back-arc spreading centre in the East
Scotia Sea, Geochim. Cosmochim. Acta 43:555-572.
Saunders, A. D., and Tamey, J. 1991. Back-arc basins, in Oceanic Basalts (P. A. Floyd, ed.), pp. 219-263, Blackie,
Van Nostrand, Reinhold, Glasgow.
Simpson, P., and Griffiths, D. H. 1982. The structure of the South Georgia continental block, in Antarctic
Geoscience (c. Craddock, ed.), pp. 185-192, University of Wisconsin Press, Madison.
Steiger, R. H., and Jaeger, E. 1977. Subcommission on geochronology: Convention on the use of decay constants
in geo- and cosmochronology, Earth Planet. Sci. Lett. 36:359-362.
Suess, E. 1909. Das Antlitz der Erde, III Band, Freytag, Leipzig.
Taylor, B. 1987. A geophysical survey of the Woodlark-Solomons region, in Marine Geology, Geophysics, and
Geochemistry of the Woodlark Basin-Solomon Islands (B. Taylor and N. F. Exon, eds.), Earth Science Ser.,
Vol. 7, pp. 25-48, Circum-Pacific Council for Energy and Mineral Resources, Earth Science Ser., Hous-
ton, TX.
Tectonic Map of the Scotia Arc. 1985. 1:3000000. BAS (Misc) 3, British Antarctic Survey, Cambridge.
Tomblin, J. F. 1979. The South Sandwich Islands. II: The geology of Candlemas Island, Br. Antarct. Surv. Sci.
Rept. 92: 33 pp.
Tyrrell, G. W. 1945. Report on rocks from West Antarctica and the Scotia Are, Disc. Rep. 23:37-102.
Vlaar, N. J., and Wortel, M. J. R. 1976. Lithospheric ageing, instability and subduction, Tectonophysics 32:
331-351.
Von Huene, R., Kulm, L. D., and Miller, 1. 1985. Structure of the frontal part of the Andean convergent margin, J.
Geophys. Res. 90:5429-5442.
Wilson, D. S. 1993. Confirmation ofthe astronomical calibration of the magnetic polarity timescale from sea-floor
spreading rates, Nature 364:788-790.
Woodhead, J., Eggins, S., and Gamble, J. 1993. High field strength and transition element systematics in island arc
and back-arc basin basalts: evidence for multi-phase melt extraction and a depleted mantle wedge, Earth
8
Bransfield Strait, Antarctic Peninsula

Active Extension behind a Dead Arc
Lawrence A. Lawver, Randall A. Keller, Martin R. Fisk,
and Jorge A. Stre/in
ABSTRACT
Bransfield Strait is a marginal basin landward of the South Shetland Trench. It lies between
the South Shetland Islands and the tip of the Antarctic Peninsula and is an example of an
extensional basin formed by rifting within a continental volcanic arc. The Antarctic Penin-
sula is the product of at least 200 m.y. of subduction with the majority of the exposed rocks
related to continental arc volcanism older than 20 Ma. Volcanism in Bransfield Strait started
by 0.3 Ma and continues today. This new volcanism maintains some of the chemical
signatures of the old arc volcanism but also has signatures transitional between arc rocks
and backarc basin rocks. On the basis of high heat flow, active volcanism, extensional
faulting, and earthquake fault plane mechanisms, Bransfield Strait is an active extensional
basin forming within the Antarctic Peninsula. Seismic refraction work in Bransfield Strait
indicates some thinning of continental crust, but the basin itself is underlain by as much
as 30 km of anomalous crustal material. The observed extension seems to be confined to
Bransfield Strait, which is bounded by the landward projections of the Hero and Shackleton
fracture zones. The present extension in Bransfield Rift started less than 4 m.y. ago, and
possibly less than 1.5 m.y. ago, following the demise of the Antarctic-Phoenix spreading
center (ANT -PHO ridge), which ceased spreading about 4 Ma. Apparent, continued sub-
duction at the South Shetland Trench after the ANT -PHO ridge stopped spreading may
occur as trench rollback. The amount of trench rollback should be comparable to the
amount of extension in Bransfield Strait.
Lawrence A. Lawver • Institute for Geophysics, University of Texas at Austin, Austin, Texas 78759-8397.
Randall A. Keller and Martin R. Fisk • College of Oceanic and Atmospheric Sciences, Oregon State
University, Corvallis, Oregon 97331-5503. Jorge A. Strelin • Departmento de Ciencias de la Tierra,
Instituto Antartico Argentino, Buenos Aires, Argentina.
Backarc Basins: Tectonics and Magmatism, edited by Brian Taylor, Plenum Press, New Yor~, 1995.
315
316 LAWRENCE A. LAWVER et al.
1. INTRODUCTION
Bransfield Strait (Fig. 8.1) is an example of a basin that formed by rifting within a
continental volcanic arc. For most of the past 200 m.y., the Antarctic Peninsula (Fig. 8.1)
has been the site of subduction of oceanic lithosphere that formed at spreading centers in the
South Pacific (Tanner et al., 1982; Barker and Dalziel, 1983; Pankhurst, 1983; Hole et aI.,
1991). Major plate motions in the vicinity of Bransfield Strait and the Antarctic Peninsula
are well constrained by magnetic anomaly patterns preserved on the Antarctic, Pacific,
Scotia, and former Phoenix plates (Barker, 1982). At about 4 Ma, the Antarctic-Phoenix
(ANT-PHO) spreading center ceased spreading and the remnant Phoenix plate was incor-
porated into the Antarctic plate. Previous authors have referred to the remnant plate as the
A1uk plate (Herron and Tucholke, 1976; Barker, 1982; Mayes et al., 1990; GRAPE Team,
1990), the Drake plate (Barker, 1982; Mayes et aI., 1990), and/or the Phoenix plate (Barker,
1982; Larter and Barker, 1991). Larson and Chase (1972) used the terms Phoenix lineations,
Phoenix model, and Phoenix plate, and related the terms to the Phoenix Islands found in the
central Pacific Ocean. Although the Phoenix lineations and the Phoenix Islands both lie on
the Pacific plate, they assigned the term Phoenix plate to the "southern" half of the Pacific-
Phoenix spreading center and traced a possible Pacific-Phoenix fracture zone from the
Phoenix lineations to the Eltanin fracture zone system's intersection with the present-day
Pacific-Antarctic Ridge. They correctly pointed out that "It is unlikely that the Phoenix
plate corresponded to the Antarctic plate until about 85 m.y.B.P.... ," because "Antarctica
and New Zealand did not begin to separate until just prior to anomaly 32." The designation
ANT
700S
FIGURE B.1. Present-day, polar stereographic map of the regional tectonic setting of the Bransfield Strait area
modified from the Tectonic Map of the Scotia Arc [1985; British Antarctic Survey Sheet (Miscellaneous) 3].
Abbreviations for the major plates are AFR = Africa, ANT = Antarctic, f PHQ = former Phoenix, SeQ = Scotia,
and SND = Sandwich plate. The South Shetland plate lies between the South Shetland Trench and the active rift
shown in Bransfield Strait. Double arrows indicate direction of strike-slip motion along the North and South Scotia
ridges. Active trenches are shown as black. The South Shetland Trench is shown shaded.
BRANSFIELD STRAIT, ANTARCTIC PENINSULA 317
Phoenix plate for the remnant plate off Bransfield Strait is therefore tectonically most
accurate and should take precedence.
Seismic reflection and refraction work, gravity, magnetics, and heat flow data, and the
history and composition of volcanism on the peninsula and in Bransfield Strait place
constraints on the processes associated with the formation of Bransfield Strait as a marginal
basin. Recent dredges from new sites in Bransfield Strait (Keller et ai., 1994) will soon be
analyzed and will give better control on active basin formation.
2. REGIONAL TECTONIC SETTING
Present-day plate boundaries are indicated in Fig. 8.1. Plate reconstructions (Barker,
1982; Mayes et ai., 1990) suggest that at least a thousand kilometers of the Phoenix plate
have been subducted beneath Bransfield Strait since 50 Ma. Between 50 and 43 Ma, a
segment of the ANT-PHO spreading center between the Tharp and Heezen fracture zones
(Fig. 8.2) was subducted (Barker, 1982). At 20 Ma, the active South Shetland Trench was
r( i I
~·-z---
•
<I /
ANT
65"8
70"$ ,
20Ma
FIGURE B.2. Reconstruction of Southeast Pacific and Antarctic Peninsula regions for 20 Ma. The reconstruction
is based on the PLATES database. Dashed fracture zones and spreading centers are taken from Larter and Barker
(1991). Plate abbreviations are as in Fig. 8.1: NAZ = Nazca plate, PHO = Phoenix, Alex. = Alexander Island, Anv.
= Anvers Island. Magnetic anomaly picks are shown as small squares. Magnetic lineations connect the small
squares. Rifting in Bransfield Strait has been closed up so that the South Shetland Islands are closer to the
Antarctic Peninsula than they are at present. The shaded region is the seafloor that was subducted since 20 Ma.
nearly twice the length it was after about 14 Ma, when the sections between the Adelaide
and Anvers fracture zones were subducted. Through time, additional segments of the ridge
were subducted to the northeast, until about 4 Ma, when the spreading center immediately
southwest of the Hero fracture zone reached the trench (Barker, 1982). The last segment of
the ANT-PHO spreading center (between the Hero and Shackleton fracture zones) was then
abandoned offshore of the trench and the last remnant of the Phoenix plate (f PHO, Fig. 8.2)
became part of the Antarctic plate.
Until 4 Ma, the subducting plate was decoupled from the overriding plate as the lower
Phoenix plate slid beneath the upper Antarctic plate. As the Phoenix-Antarctic spreading
centers were subducted, the subducted slab continued to sink, leaving in its wake a slab
window (Hole and Larter, 1993). After 4 Ma, when the last Phoenix-Antarctic spreading
center southwest of the Hero fracture zone was subducted, the remnant, unsubducted Phoenix
plate (former Phoenix plate = fPHO) was frozen in place between the Hero and Shackleton
fracture zones. Whether this resulted in the subducted Phoenix slab, to the southwest of the
Hero fracture zone, to also be frozen in place with respect to the underlying mantle or if it
continued to sink independently is a matter of speculation. If the subducted slab to the
southwest of the Hero fracture zone continued to sink but the slab beneath Bransfield Strait
was held in place by the f PHO, then there would have to be tearing or differential motion
along the subducted section of the Hero fracture zone, and there would be no slab window
beneath Bransfield Strait. On the other hand, if the slab beneath Bransfield Strait is detached
from the f PHO and sinking, then there should be deep seismicity and evidence of a slab
window in the geochemistry.
Other marginal basins along the outer edge of the Antarctic Peninsula may have
formed in response to the subduction of ridge segments to the southwest of the Hero frac-
ture zone (Gamboa and Maldonado, 1990), although none of these basins seem to have
reached the advanced state of Bransfield Strait. GEOSAT gravity data from nos to the
northeast along the western margin of the Antarctic Peninsula (Lawver et al., 1993) show a
marginal gravity low that extends from George VI Sound (700S, 70 0W) between Alexander
Island and the peninsula, to the southern edge of Anvers Island at 65°W. This gravity
signature is similar to the one seen in Bransfield Strait. Single-channel seismic reflection
lines from along this margin (Anderson et al., 1990) show features that resemble inner shelf
troughs and old forearc basins that might be interpreted as precursors to Bransfield Strait,
although these basins do not show any evidence for extension. Larter and Barker (1989)
suggested that the forearc to the south of Bransfield Strait was simply uplifted and eroded
but not extended after ridge-trench collision.
The age of the subducted Phoenix plate at the South Shetland Trench (Fig. 8.1)
increases from 14 Ma in the southwest near the Hero fracture zone to 23 Ma in the northeast
near the Shackleton fracture zone (Barker, 1982). Between the Hero and Shackleton
fracture zones, there are at least two additional fracture zones, D and E (Fig. 8.2). Their
landward extensions would cross the neovolcanic zone of Bransfield Strait at about 59°W
(FZ D) and 57°50'W (FZ E). At the trench, the age difference across FZ D is 3 m.y.; across
FZ E, 3.5 m.y. (Larter and Barker, 1991).
Presently, the subducted slab can be detected beneath the South Shetland Islands
(Grad et aI., 1993) dipping at an angle of 25° to the southeast (Fig. 8.3), although there is
virtually no seismic activity associated with the lower plate slab (Pelayo and Wiens, 1989).
This may be explained by either slow, aseismic descent of the slab, or the slab beneath the
South Shetland Islands and Bransfield Strait is frozen in place such that there is no
detectable motion of the slab relative to the asthenosphere below it. If the lower or
o Drake Passage 2.5 South Shetland Islands Bransfield Strait 42 Antarctic Peninsula
10
20
::::.:
::.:: 30
z
- 40 ~--.-
;:::
1;" 50 6.15
~ 8.2
Q 60
70
80 ~~~~~~~~~~~~~~~~~~~~~~~~J-~~
o 20 40 60 80 100 120 140 160 180 ·200 220 240 260 280 300
DISTANCE IN KM
FIGURE B.3. Seismic model of the lithosphere from the deep-ocean side of the South Shetland Islands through
Bransfield Strait and onto the Antarctic Peninsula taken from Grad et al. (1993). The symbols used are (1) sedi-
ments, vp =2.5-4.2 km s-', (2) upper crust, vp =5.4-6.3 km s-', (3) middle crust, vp =6.4-6.8 km s-', (4) lower
crust and high-velocity body in Bransfield Strait, vp > 7.0 km s-', (5) Moho boundary, vp > 8.0 km s-',
(6) reflection boundary in the lower lithosphere. Notice that the volcanic line is not centered in the middle of
Bransfield Strait.
subducted plate remains attached to the f PHO, then any extension in the upper plate must
be accompanied by trench rollback. While some may argue that trench rollback is evidence
for subduction, we make the distinction that in this case, if the lower plate is still attached to
the Antarctic plate, it cannot be subducting with respect to itself, and the independent South
Shetland plate must therefore be moving with respect to the lower plate. This in tum pro-
duces the observed rollback or seaward motion of the trench. If the slab has broken from the
former Phoenix plate or is sinking along a hinge line at the trench, then there should be deep
seismicity and there would not be evidence for deformation of the accretionary wedge.
Prior to the demise of the ANT - PHO spreading center, the Shackleton FZ to the north
of Bransfield Strait was the boundary between the Antarctic and Phoenix plates on one side
and the Scotia plate on the other (Fig. 8.1). Seafloor spreading stopped in the western Scotia
Sea immediately to the east of the Shackleton FZ at 6 Ma (Barker and Burrell, 1977). At
about 4 Ma, when the remaining Phoenix plate became a part of the Antarctic plate, the
Scotia-Phoenix plate boundary became part of the Scotia-Antarctic plate boundary. From
4 Ma to recently, the Scotia-Antarctic plate boundary to the north of Bransfield Strait has
trended NW-SE (Shackleton fracture zone), while to the east of Bransfield Strait it has
trended E-W (South Scotia Ridge). The consequence of this 110° bend in the plate bound-
ary would be compression on the inside of the bend and extension on the outside of the
bend. This should produce extension in the North Bransfield Basin, the part of Bransfield
Strait between Bridgeman Island (Fig. 8.4) and Clarence Island. Single-channel seismic
reflection profiles across the North Bransfield Basin (Lawver and Villinger, 1989) show
continentward tilted blocks, reminiscent of the continental extension models of Wernicke
FIGURE 8.4. Bathymetric map of Bransfield Strait region (modified from Klepeis and Lawver,1994). Bransfield
Strait lies between the South Shetland Islands and the Antarctic Peninsula. Contour interval is 200 m, I()()()-m
contour lines are shown as bold lines and some are labeled. The 1800-m contour southwest of King George Island
(KGI) outlines the King George Basin. Solid triangles are locations of submarine and subaerial Quaternary rift-
related volcanism that have been sampled and analyzed (Keller et al., 1992). Hollow triangles are submarine
locations where fresh, glassy basalts have been recovered (Keller et al., 1994) but not yet analyzed. BI =Bridgman
Island, CI =Clarence Island, DI =Deception Island, EI = Elephant Island, GrI =Greenwich Island, JRI =James
Ross Island, LvI = Livingston Island, MP =Melville Peak, PI =Penguin Island, and RI = Robert Island. Dashed
lines numbered I through 4 are parallel or subparallel volcanic lineations, ranging in size from tens of meters in
height to 2 lan, or more for Deception Island and the submarine volcano south of KGI.
(1981). Evidence for compression includes the extraordinarily steep SW face of Clarence
Island which varies from -1200 m to +2300 m in a distance of 10 km or less. The low
temperature, high-pressure metamorphics found on Clarence Island and on the eastern half
of Elephant Island (Dalziel, 1984) also support the idea of compression and uplift within
the bend.
3. TECTONIC SETTING OF BRANSFIELD STRAIT
The strait itself is approximately 100 km wide by 400 km long and is bounded on the
northwest by steep normal faults with as much as 4 km of down throw to the southeast
between the South Shetland Islands and the Bransfield basins (Ashcroft, 1972). Birken-
majer (1992) distinguishes between the l00-km-wide Bransfield Strait and a narrowly
defined (15-20 km wide) Bransfield rift. From field mapping on King George Island,
Birkenmajer (1992) found fossil evidence that a marine basin existed during the early
Eocene (fossiliferous glaciomarine strata from the Krakow Glaciation) and during the
Eocene part of the succeeding interglacial time. He reports that during early Oligocene the
region was covered by an ice sheet and then later flooded by a shallow sea. Regional uplift
caused the area to be above sea level until the late Oligocene when incipient rifting started
at the end of the Oligocene (26 to 22 Ma).
The basins of Bransfield Strait are thought to be mainly Quaternary features, although
the age of initiation of rifting is poorly constrained. Much of the early faulting is associated
with the arc-building period of the peninsula. Evidence for early extension consists of a
system of antithetic faults that cut the upper Oligocene and older rocks along the margin of
the rift. Arc tension continued through early Miocene when several stages of basaltic to
andesitic dyke intrusions occurred between 22 and 20 Ma and at 14 Ma (Birkenmajer,
1992). From the late Miocene to Pliocene there is no evidence of faulting or volcanism
within Bransfield Strait. The present Bransfield rift is the site of Pleistocene to Recent,
mildly alkaline to calcalkaline volcanic activity.
The present plate configuration in the Bransfield Strait region places the South Shet-
land Islands on a separate (South Shetland) microplate, bounded on the northwest by the
South Shetland Trench and on the southeast by the Bransfield rift. The southeast boundary
of Bransfield Strait rises more gently toward the Antarctic Peninsula through a series of
widely spaced normal faults (Ashcroft, 1972). The active rift appears to extend 400 km
(Gonzalez-Ferran, 1985) from Low Island (Fig. 8.4) above the landward extension of the
Hero fracture zone to Clarence Island above the landward extension of the Shackleton
fracture zone. Based on GEOSATIERS-l satellite-derived gravity data (Fig. 8.5), Brans-
field Strait can be subdivided into three basins: the south, central, and north Bransfield
basins. The south Bransfield Basin shows up as a -100-km-Iong gravity anomaly trending
southwest of Deception Island. The south Bransfield Basin approximately follows the
l000-m contour shown in Larter and Barker (1991, Fig. 8.4). The central Bransfield Basin
is a graben up to 2 km deep and roughly 40 km wide by 200 km long that is easily seen on
the bathymetric map (Fig. 8.4). It extends from Deception Island to Bridgeman Island and
contains the 20- by 50-km, flat-floored King George Basin southeast of King George Island.
The north Bransfield Basin is actually the deepest part of Bransfield Strait with a maximum
depth over 2740 m and the lowest gravity anomaly, <-30 mgal. The lowest gravity
anomaly is not coincident with the deepest part of the basin but is to the northeast toward
Clarence Island. If the positive gravity anomalies produced by Deception and Bridgeman
Islands (which are relatively young structures) are removed from the GEOSAT data shown
in Fig. 8.5, the subdivisions of Bransfield Strait lose their definition and a single major
gravity low extends from just south of Low Island to Clarence Island.
The digital bathymetric map (Fig. 8.4) of Bransfield Strait and the nearby region
(Klepeis and Lawver, 1994) shows numerous small bathymetric features that appear to be
mounds and are lineated, parallel to the basin axis defined by a line through Deception and
Bridgeman Islands. Some of these mounds can be seen in Fig. 8.4. There are at least four
parallel lines of active and incipient volcanism. The first is discussed by Birkenmajer (1992)
and includes Penguin Island, an active cinder cone just off King George Island. This
lineation is within the shelf area of the South Shetland Islands. The second line is defined in
King George Basin by at least six to eight closed contours indicating circular structures
from a few tens of meters in height to a couple of hundred meters. The third line is the main
rift axis and includes Deception and Bridgeman Islands and at least two major submarine
volcanoes that are up to 1.4 km in height above the seafloor and as much as 8 km in
FIGURE B.S. Geosat/ERS-I gravity data from Sandwell and Smith (1992) with digital bathymetric data of
Klepeis and Lawver (1994) superimposed. The gravity data bar indicates gravity values from < -40 mgal to
> + 30 mgal. Bold white bathymetric contours are shown every 500 m, while thin white contours are observed
bathymetry contoured every 100 m. Thin gray contours are interpolated bathymetric data contoured every 100 m.
A color version of this figure appears opposite page 311.
diameter. The final volcanic lineation is defined by a ridge (Fig. 8.4), the highest heat flow
in King George Basin, and a very recent extrusion at 57°W.
The observed recent faulting (Fig. 8.6) seen in the multichannel seismic data (Barker
and Austin, 1994) may define yet a fifth zone of extension and deformation, and although
diapirism is suggested, no actual intrusion is observed. Barker and Austin (1994) show an
alignment of intracrustal diapirs that trend subparallel to the other neovolcanic zones but in
fact diverge from them by about 5°. They suggest that the divergence shows a temporaU
spatial evolution of stress in Bransfield Strait, possibly related to the complex plate
interactions to the northeast.
4. EARTHQUAKES
Most of the earthquakes in Bransfield Strait are shallow events related to extension
and volcanic activity (Fig. 8.7). Deep events in the subducted plate are rare. The vast
NW Ap·3 SE NW AP-4 SE
b
3km
Vertical ExaggeratTon 3:1 (@ 2000 mls)
FIGURE 8.6. Seismic cross sections taken from Barker and Austin (1994). Locations shown in Fig. 8.8. Profile
shows fan-shaped, normal faulting pattern where the faults reverse their dip about a midpoint (at -CDP 3000 for
8.6a, -CDP 5300 for 8.6b). The fault blocks rotate away from the center of the structure, and central faults are
confined to the upper part of the section. The inferred diapiric body is unreflective.
55'W 4S'W
SCOTIA PLATE
62"S
-- NBB
On_ol ........... ' ...,.
•• an idt.aI plUJlnl tagh lont
45'W
FIGURE 8.7. Tectonic map of the area north of the Antarctic Peninsula (AP) showing interpreted geometry of
the plate boundaries between the Shackleton fracture zone and the South Scotia Ridge. EI is Elephant Island, CI is
Clarence Island, BI is Bridgeman Island, and NBB is North Bransfield Basin, SSR (inset) is South Scotia Ridge,
dashed line labeled BS axis is the volcanic rift axis of Bransfield Strait. Bold lines located north EIICI represent
faults interpreted to parallel the steep scarps of the EI platform's northern margin, bold lines south ofEIICI are the
major fault zones taken from Klepeis and Lawver (1994). GEOSAT gravity anomaly is after Sandwell (1992).
Circles are earthquake epicenters from the Tectonic Map of the Scotia Arc (1985). Focal mechanisms are for
shallow earthquakes from Pelayo and Wiens (1989). Shaded area between GEOSAT gravity anomaly lineament
and EIICI represents a broad zone of tectonic disruption where the Shackleton fracture zone transform apparently
is stepping eastward to join with a transform along the South Scotia Ridge.
324 LA WRENCE A. LAWVER et al.
majority of the teleseismic events between 1964 and 1992 are swarms of moderate-
magnitude events which lasted up to six months (International Seismological Centre
Earthquake Catalogs (ISC), 1964-1993).
The December 1967 eruption of Deception Island produced three teleseismic events,
although two of them were mislocated almost 30 kIn from Deception Island. One event
occurred in connection with the 1969 eruption. The 1970 eruption produced four events
located at depths of 45 to 49 kIn, one event at 9 kIn with no magnitude determination,
and one magnitude 4.6 event with no depth determination (ISC catalogue, 1970). Three
swarms of five to eight events not associated with volcanic eruptions occurred near
Deception Island in 1971, 1974, and 1982. In 1989 a single event at 10 kIn occurred near
Deception Island (ISC catalogue, 1989). The events in 1971 and 1982, at 15 and 12 kIn,
respectively, produced normal fault plane solutions indicative of extension (Pelayo and
Wiens, 1989).
A large swarm of 24 separate events (magnitude 4.5 to 5.3) occurred near Bridgeman
Island in the first six months of 1975. Nearly all the events were assigned depths of 33 kIn or
less, except for one at 167 kIn. [This event was undoubtedly mislocated depthwise because
of its low magnitude, 4.5, and the few stations which recorded it. Pelayo and Wiens (1989)
do not show any events at such a depth.] One event near Bridgeman Island was recorded in
1990 at a depth of 10 kIn. These earthquakes may be related to ongoing extension rather than
volcanic activity, because there is no historical evidence for volcanic activity on Bridgeman
Island.
Elsewhere in Bransfield Strait two events relocated by Pelayo and Wiens (1989) lie at
depths of 35 and 55 kIn (their events 10 and 18). Event 10 (61.4°S, 56.4°W) was a normal
faulting event that may be in the upper plate if the seismic refraction work of Guterch et al.
(1991) is correct. Event 18 (63°S, 61.9°W) at 55 km must be in the subducted slab although
the Harvard catalogue, for that event lists its depth as 42 kIn. It is located almost directly on
strike with the onshore projection of the Hero fracture zone, and may in fact be evidence
of tearing along the subducted fracture zone.
In addition to the teleseismic events mentioned, a local network on Deception Island
has operated for two months each austral summer since 1986 (Vila et aI., 1992). It has
consistently recorded 1000 events per month with a total released seismic energy of about
3.0 x 1013 ergs/day. These events seem to be aligned parallel to the main fault system which
crosses the island (E-W), and they follow the regional tectonic trend of the Bransfield rift
(Vila et aI., 1992). The 1986-1992 events were grouped into two depth ranges, those from
oto 2.5 kIn and those from 2.5 to 7.0 kIn. During the 1992-1993 austral summer this same
network recorded five deep (55-85 kIn) events of magnitude 1.6-2.4 (SEAN, 1993), but it
is unclear how such a closely spaced network could resolve deep events of such low
magnitude.
5. SEISMIC REFRACTION
Early seismic refraction work found an anomalously thick layer beneath Bransfield
Strait with a seismic velocity lower than that of normal mantle material but higher than
typical continental crust (Ashcroft, 1972). Later seismic refraction work in Bransfield Strait
(Guterch et at., 1985, 1991, 1992; Grad et aI., 1993) identified returns from the Moho that
dips from 10 kIn in the South Shetland Trench to 40 kIn below the South Shetland Islands,
and also at 40 Ian below the Antarctic Peninsula. Grad et al. (1993) recorded a seismic
boundary in the lower lithosphere that ranged in depth from 35 to 80 Ian as shown in
Fig. 8.3. Both the Moho and the lower seismic boundary dip at an angle of 25°, which was
interpreted to be the dip of the remnant Phoenix plate under the Antarctic plate. Although
published gravity coverage is not extensive, Bouguer gravity anomalies (Renner et aI.,
1985) of up to + 100 mgal correlate with the anomalous crustal structure found by Grad
et al. (1993) under the South Shetland Islands and Bransfield Strait.
The sequence stratigraphy of Bransfield Basin is discussed by Jeffers and Anderson
(1990) and was used by Grad et al. (1993) to constrain their seismic velocity model. They
found an anomalous layer under Bransfield Basin with seismic wave velocities of vp >
7.0 Ian s-'. Unfortunately this high-velocity layer obscures the normal Moho transition as
seen under the South Shetland Islands and under the Antarctic Peninsula. This anomalous
layer is overlain by layers with velocities of 2.5, 4.2, and 5.6 Ian s-'. The upper layer
consists of 0.2 to 1.0 Ian of unconsolidated to poorly consolidated young sediments with
substantial amounts of lava and tuff. The second layer consists of 1.2 to 2.5 Ian of older and
better consolidated sediments, lava, and tuff. In the 2- to 7-Ian depth range, Grad et al.
(1993) found an approximately lO-Ian-wide body with velocity vp = 6.8 Ian s-'. They
attribute this body to be the active Bransfield rift. The active rift is manifested in the basin
itself as submarine volcanoes, intrusive dykes, large magnetic anomalies, and high heat
flow.
6. SEISMIC REFLECTION
There are numerous multichannel seismic and single-channel seismic lines known to
have been collected in Bransfield Strait, including ones done by the British Antarctic
Survey (Larter, 1991; Barker et aI., 1992), German scientists as part of the GRAPE
(Geophysical Research Group for the Antarctic PEninsula) Team (1990), Polish scientists
(Guterch et aI., 1985), Brazilian scientists (GambOa and Maldonado, 1990), Spanish scien-
tists (Acosta et aI., 1992a,b), U.S. scientists (Lawver and Villinger, 1989; Nagihara and
Lawver, 1989; Jeffers and Anderson, 1990; Jeffers et al., 1991; Barker and Austin, 1994),
and other presently unpublished work from Italian, Chinese, Japanese, and Korean cruises.
Barker (1976) published the first seismic profile for Bransfield Strait, while Guterch et al.
(1985) presented the first seismic reflection lines to demonstrate the "complex crustal
structure in the trough of Bransfield Strait." Their profile crosses the southern end of the
central Bransfield Basin and clearly shows a neovolcanic intrusion zone. Jeffers and
Anderson (1990) published a number of line drawings of seismic reflection data from
Bransfield Strait, including one line that nearly duplicated the earlier Polish line. They
determined that the sedimentation is dominated by glacial-marine processes with associ-
ated lithofacies. The central basin has received terrigenous sediment from the Trinity
Peninsula and from the South Shetland Islands (Jeffers and Anderson, 1990). They discuss
basin ward transport of sediment through deeply incised troughs that form a prograding
complex of coalescing trough-sourced wedges, although when the troughs are not active,
principally during interglacial periods, the supply of terrigenous sediment to the basin is
relatively low and sedimentation is predominantly biogenic. Although they state that the
volcanic ridge acts as a barrier to sediment transport across the basin, later bathymetric
mapping of the basin revealed that the volcanic ridge is not continuous, so it would not act
as a complete barrier (Klepeis and Lawver, 1994). In places, the neovolcanic ridge is easily
defined as a linear wall less than half a kilometer across with nearly sheer sides that are at
least a few hundred meters high. It was recently dredged (Keller et at., 1994) and produced
fresh glassy basalts.
GambOa and Maldonado (1990) presented multichannel seismic reflection data col-
lected in 1987 and 1988 by the Brazilian Antarctic Program. In each oftheir crossings they
show a "spreading center" and ponded turbidites. On all multichannel seismic lines across
Bransfield Strait, there is never more than 1 s of recent ponded turbiditic sediment. The
GRAPE team (1990) published preliminary results from 1400 kIn of reflection seismic
profiles from Bransfield Strait and the adjoining peninsula region. Their profile 9 crossed
the major submarine volcano just off King George Island (also shown in Barker, 1976), and
they identified an additional "magmatically influenced" zone just to the southeast of the
volcanic edifice toward the Antarctic Peninsula. Larter (1991) questioned the designation of
certain crust in Bransfield Strait as oceanic crust based solely on the reflection character of
the top of acoustic basement, as the GRAPE team (1990) did. The unusual thickness of the
crust found from the seismic refraction data would also suggest a lack of true oceanic crust.
Acosta et at. (1992b) present line drawings of single channel seismic reflection profiles
that show the southeastern flank of the South Shetland Islands to be broken by faults into
horsts and grabens which have been intruded by dikes and plugs. Barker and Austin (1994)
display initial results of the RIV Ewing multichannel seismic reflection cruise, where they
found complex fan-faulting patterns (Fig. 8.6) along the Antarctic Peninsula side of
FIGURE B.B. Heat flow map of the King George Basin. Heat flow in mW m- 2 . Four highest values are shown
with larger type size. Bathymetric contours are shown every 100 m with 500-m contours shown as bold lines. The
inner, dashed, closed contour is the 1980-m contour showing the flat floor of the King George Basin that is
uniformly between 1980 m and 2000 m with the exception of the linear volcanic mounds found along dashed line
2. Dashed lines labeled 2, 3, and 4 correspond to parallel volcanic lineations shown in Fig. 8.4. Thick lines labeled
6a and 6b are the locations of the seismic sections shown in Fig. 8.6. Shaded area is diapir zone of recent? activity
taken from Barker and Austin (1994).
Bransfield Strait. They interpret the fanfaulting to be intracrustal diapirism that indicates
a new zone of diffuse extension. They found a new zone of active extension on the margin
of King George Basin toward the Peninsula from the four zones discussed above.
7. MAGNETICS
Aeromagnetic surveys of the Bransfield Strait region show a central positive anomaly
along the axis of the deepest part of the strait (Garrett, 1990; Gonzalez-Ferran, 1991;
Maslanyj et al., 1991). This central high has been modeled as a large, positively magnetized
igneous body associated with the inferred axis of rifting (Parra et al., 1984; Gonzalez-
Ferran, 1991). Gonzalez-Ferran (1991) concluded that reversely magnetized crust at the
flanks of the central body best fit the data, and that a total of approximately 15 km of new
crust has been created by 2 m.y. of spreading in Bransfield Strait.
Roach (1978) modeled detailed marine magnetic data from Bransfield Strait and found
that only normally magnetized crust existed beneath the axial trough. Roach later reported
(in Barker and Dalziel, 1983) that narrow strips of reversely magnetized crust were required
at the margins of the axial trough, and a total of about 30 km of spreading had occurred
since 1.3 Ma. Unfortunately his data and the details of his modeling were not published,
although his results are now widely cited in the literature as evidence for a 1-2 Ma age for
Bransfield Strait. Such a short magnetic anomaly profile is ambiguous, and models of it are
poorly constrained. It may also be inaccurate to assume for the purpose of modeling that
there is true seafloor spreading in Bransfield Strait, since seismic refraction data (see
Section 5) suggest that the crust in Bransfield Strait is too thick to be normal oceanic crust.
No spreading center has been identified, and the recent volcanism is dispersed along several
lineaments in Bransfield Strait (Klepeis and Lawver, 1994). The thick sediment cover
contains interbedded basalt flows and hides basement topography that affect the analyses of
marine magnetic anomalies in the basin and complicate any modeling, as discussed by
Lawver and Hawkins (1978).
8. HEAT FLOW
Heat flow measurements have been made in the King George Basin (Fig. 8.8)
immediately south of King George Island. In the King George Basin the seafloor is flat (1.96
to 1.99 km), well sedimented, and covers an area of roughly 10 to 18 km by 45 km. Thermal
gradients were measured at 63 locations, and in situ thermal conductivities were measured
at 25 sites. The in situ conductivity matches thermal conductivity measurements made on
piston core samples. The heat flow values range from 49 to 626 mW m -2 and are generally
high, with about 25% of the values greater than 220 mW m- 2 • There is significant local
variation in values, with the highest values in the central part of the basin and along the
southeast and northeast edges of the basin. In the western part of the basin values are
generally less than 100 mW m- 2 •
The highest value (626 mW m- 2) is near the center of the basin, although five
additional measurements within a few hundred meters of this high value vary from 150 to
250 mW m- 2 , typical of regions with active hydrothermal circulation. While this region of
very high heat flow is not in line with the volcanic islands and submarine volcanoes of
Bransfield rift, it is close to a region of recent faulting seen in the multichannel seismic
reflection data (Fig. 8.8). The second highest recorded heat flow value (411 mW m- 2) was
found on the northeast edge of the basin in an area of mounds a few meters in height and
up to a kilometer in diameter.
Evidence for hydrothermal activity in Bransfield Strait has been found in the water
column and in the sediments. In the waters of Bransfield Strait, B3He increases from less
than zero near the surface to > 7 at depth. This has been interpreted as evidence for injection
of 3He into the water by backarc rifting (Schlosser et ai., 1988). Mn concentrations in the
water also increase with depth to a maximum of almost 7 nM. Temperature changes are less
pronounced, but appear to mimic the Mn profiles (Suess et ai., 1988). Maturation of
biogenic components in the sediments (Whiticar et ai., 1985; Brault and Simoneit, 1990) is
similar to the thermogenic hydrocarbons found to be associated with hydrothermal activity
in the Gulf of California (Simoneit, 1983). Significant downcore decreases in amino acid
abundances despite relatively constant Bi3C, B15N, CIN, and percent carbonate C in the
sediments suggest hydrothermally accelerated complexation of amino acids (Silfer et ai.,
1990).
Within the area of high and variable heat flow in King George Basin, several hundred
grams of sediment were recovered in the heat flow probe. Qualitative analysis of this
sediment by an electron microprobe indicated the presence of grains of Fe sulfide, Fe-Zn
sulfide, Fe-Zn-Cu sulfide, Zn chloride, and metals or oxides of Zn and Fe, all presumed to
be products of hydrothermal activity. This indicates hydrothermal activity occurred in the
region of high heat flow (62°14'S, 57°44'W) and water-column measurements indicate
that hydrothermal circulation is still active.
10. PETROLOGY AND GEOCHEMISTRY
10.1. Regional
Volcanism near the northern end of the Antarctic Peninsula can be divided into the
main arc phase and a later phase associated with rifting of the peninSUla (Bransfield Strait)
and volcanism on the Weddell Sea side of the peninsula (James Ross Island Volcanic
Group). Volcanism is presently active on both sides of the peninsula; however, the compo-
sition of the rocks on the two sides are chemically distinct (Fig. 8.9) and are also chemically
distinct from the earlier arc volcanism.
10.1.1. Antarctic Peninsula Arc

Although the crust beneath the South Shetland Islands is assumed to be continental,
actual basement rocks are unknown. According to Birkenmajer (1992), the basement of
Bransfield Strait consists of continental-type crust modified by basic intrusions. The oldest
exposed rocks are metasediments on Livingston Island (Fig. 8.4) and are thought to be
between late Carboniferous and Triassic in age (Thomson, 1992). Similar rocks on the
Antarctic Peninsula are between Permian and Cretaceous in age (Pankhurst, 1983). The
Si02
FIGURE B.9. Weight percent Si02 versus weight percent NlIz0+K20 for samples from the South Shetland
island arc, Bransfield Strait, and the James Ross Island Volcanic Group (JRIVG). South Shetland arc data are from
Saunders et al. (1980), Smellie (1988), Smellie et al. (1984, 1988), and Keller et al. (1992). Bransfield Strait data
are from Keller and Fisk (1992) and Keller et al. (1992). Additional data from Bransfield Strait consistent with
those shown here can be found in Weaver et al. (1979). JRIVG field is from data in Table III. Classification fields
are from Le Bas et al. (1986). Dashed line is silica saturation line from Irvine and Baragar (1971).
oldest igneous rocks in the area are 155 to 185 Ma plutonic rocks on the northern Antarctic
Peninsula near James Ross Island (Saunders et aI., 1982). K-Ar dates of 351 to 384 Ma of
micas from granites from the same area have been reported (Rex, 1976) but are not
confirmed by Rb-Sr dating. Plutonic rocks along the Bransfield Strait side of the peninsula
are mostly 70 to 125 Ma, but some are as old as 130 to 145 Ma (Saunders et al., 1982). The
oldest extrusive rocks on the South Shetland Islands are late Jurassic-early Cretaceous
volcaniclastic rocks on islands at the western end of the chain (Smellie et aI., 1980). A
granodiorite intruding the volcaniclastic beds yielded a K- Ar age of 121 Ma (Smellie et aI.,
1984). Additional isotopic ages suggest relatively continuous arc volcanism in the South
Shetland Islands from late Jurassic to late Tertiary (Pankhurst and Smellie, 1983), although
dates younger than about 20 Ma are extremely rare in the subduction-related volcanic rocks
(Rex and Baker, 1973; Smellie et al., 1984; Birkenmajer et al., 1986).
Arc volcanism on the South Shetland Islands consisted mainly of calc-alkaline, high-
alumina basalts, basaltic andesites, and low-silica andesites. Dacites and rhyolites are rare
(Smellie, 1983; Smellie et aI., 1984). Exposures of intrusive rocks are less common and are
mainly gabbros, tonalites, and granodiorites interpreted to be cogenetic with the extrusive
rocks (Smellie, 1983). The chemistry and mineralogy of the pre-Quaternary South Shetland
Islands have been described in detail by Smellie et al. (1984). Islands in the southern part of
Bransfield Strait, close to the Antarctic Peninsula, are Eocene-Oligocene arc tholeiites
similar to rocks on the South Shetland Islands (Rex, 1976; Baker et aI., 1977).
Pre-Quaternary volcanic arc rocks from the South Shetland Islands and northern
Antarctic Peninsula are calc-alkaline with mild tholeiitic tendencies (Smellie et al., 1984).
Chemical variations within the arc magmas can be accounted for by fractional crystalliza-
tion and slight differences in extent of partial melting of the source (Smellie, 1983). Their
alkalinity (Fig. 8.9) and trace element abundances (Saunders et aI., 1980; Smellie et aI.,
1984) are typical of island-arc rocks; that is, the large-ion lithophile elements (e.g., Rb,
Ba, K, and Ce) are enriched relative to the high-field-strength elements (e.g., Zr, Ti, and
Nb). Normalized multielement plots show an obvious Nb depletion (Fig. 8.10).
330 LAWRENCE A. LAWVER at al.
m40 'r-~--------------------~
a:
o
~1
z FIGURE 8.10. N-MORB-nonnalized multielement plot
E for basalts from the South Shetland arc (Tertiary basalts
"GI
..!:!
only), Bransfield Strait (seamounts only), and James Ross
Island Volcanic Group. Data sources as in Fig. 8.9.
CD
E N-MORB nonnalization values are from Saunders and
o
z Tarney (1984). Increasing distance from the South Shet-
land Trench corresponds to increasing Nb and decreas-
Zr Ti Y ing Ba/Nb.
Strontium isotopic analyses are limited for the South Shetland Arc, but mainly range
from 0.7033 to 0.7043 (Smellie et at., 1984; Barbieri et aI., 1989; Jin Qingmin et at., 1991;
Keller et aI., 1992). The only Nd and Pb isotopic data available (excepting a single analysis
in Keller et at., 1992) are from the Fildes Peninsula on western King George Island and are
published as ranges only: 0.51288 to 0.51301 for 143Nd/ l44Nd, 18.503 to 18.641 for
206PbJ204Pb, and 15.588 to 15.625 for 207PbJ204Pb (Jin Qingmin et at., 1991). No 208PbJ204Pb
data were included in Jin Qingmin et at. (1991). A single South Shetland arc basalt from the
Low Head area of King George Island had 87Srf8 6Sr, 143Nd/l44Nd, and 207PbJ204Pb within the
ranges reported by Jin Qingmin et al. (1991) but had higher 206PbJ204Pb (Keller et aI., 1992) .
10.1.2. Bransfield Strait

Quaternary volcanism in Bransfield Strait occurred on Deception, Penguin, and
Bridgeman Islands (Weaver et aI., 1979), and at a few isolated locations on King George,
Livingston, and Greenwich Islands (Smellie, 1990), and on seamounts and ridges between
Deception and Bridgeman Islands (Fig. 8.4; Fisk, 1990; Keller et at., 1994). All reported
K-Ar dates from the Bransfield Strait islands and seamounts are 300 ka or less (Table I).
Deception Island is historically active (e.g., Smellie, 1990), and the most recent volcanism
on Penguin Island was probably within the last 300 years (Birkenmajer, 1980). All of the
sampled seamounts yielded fresh, glassy basalts (Fisk, 1990; Keller et aI., 1994).
The Quaternary volcanic centers in Bransfield Strait are mainly olivine basalts and
basaltic andesites, although Deception Island erupted a complete basalt-to-trachydacite
evolutionary suite (Keller et at., 1992). Bridgeman Island is composed of basaltic andesites,
while Melville Peak and Penguin Island are olivine basalts and scoria (Weaver et at., 1979;
Keller et aI., 1992). The seamounts are aphyric to olivine-plagioclase-phyric, vesicular,
tholeiitic basalts, and basaltic andesites (Fig. 8.9; Fisk, 1990; Keller and Fisk, 1992).
The Bransfield Strait samples, with the exceptions of the sodic samples from Penguin
and Deception Islands, can be classified as subalkaline basalts and basaltic andesites (Table
II). They all have higher concentrations of the alkali and alkali earth elements (e.g., Na,
K, Rb, Sr, and Ba), and less of an Fe enrichment trend than typical mid-ocean ridge
tholeiites (Fisk, 1990; Keller et at., 1992). Many of the Quaternary basalts are similar to the
older-arc rocks in silica versus total alkalies (Fig. 8.9). Trace element concentrations are
slightly to moderately enriched relative to N-MORB, and show a subtle negative Nb
anomaly when normalized to N-MORB (Fig. 8.10). Low-pressure, fractional crystallization
of olivine, spinel, and plagioclase can account for the chemical variation between samples
from an individual volcano (Keller and Fisk, 1992; Keller et aI., 1992). Sr-isotopic ratios
TABLE I
K-Ar Data for Basalts from Bransfield Strait and the James Ross Island Volcanic Group
40Arrad 40Arrad Age ± I 17

Location Sample K (wt%) (10- 9 cc/g) (%) (lQ3 yr) Ref.D
Bransfield Strait
King George Island
Melville Peak MP372 0.694 6.229 6.7 231 ± 19 3
Melville Peak MP375 0.767 2.137 2.2 72 ± 15 3
Melville Peak MP406 0.802 9.240 5.3 296 ± 27 3
Livingston Island
Gleaner Heights P.51.1 0.433 0.001 0.2 100 ± 400 4
Greenwich Island
Mt. Plymouth P.54.1 0.305 0.002 0.4 200 ± 300 4
Mt. Plymouth P.55.1 0.300 0.002 0.5 200 ± 400 4
Deception Island DI048 0.390 2.266 0.928 105 ± 46
Bridgeman Island B2 0.581 1.414 0.932 63 ± 25
Eastern Seamount 300.13 0.435 0.8779 0.2744 53 ± 36 2
Western Seamount 310.14 0.418 1.631 0.6808 103 ± 35 2
James Ross Island Volcanic Group
James Ross Island, Prince Gustav Channel islands, 300 ± 100 to 7130 ± 490 5,6
Vega Island, Humps Island, and Paulet Island
James Ross Island
Dreadnought 4MI8 0.838 216.6 26.3 6643 ± 102
Villar Fabre 5M117 1.063 162.9 13.7 3940 ± 85
Cockburn Island CK3 0.992 129.7 31.3 2781 ± 32
aReference codes: I-This paper; 2-Fisk (1990); 3-Birkenmajer and Keller (1990); 4-Smellie et al. (1984); 5-Smellie e/ al.
(1988); 6-Sykes (1988). Samples with reference codes I, 2, and 3 were analyzed at Oregon State University using techniques
described in Fisk (1990).
range from 0.70304 to 0.70386, and Nd-isotopic ratios from 0.51287 to 0.51301. Pb isotopes
range from 18.691 to 18.757 (206PbP04Pb), 15.594 to 15.624 (207PbP04Pb), and 38.441 to
38.556 (208PbP04Pb) (Keller et aI., 1992).
10.1.3. James Ross Island Volcanic Group

The James Ross Island Volcanic Group (JRIVG) crops out on the northern tip of the
Antarctic Peninsula and on numerous islands in the northwestern Weddell Sea in a backarc
setting that may be analogous to the Patagonia Plateau basalts. Many outcrops of the
JRIVG are sheer cliffs of palagonitized basaltic hyaloclastite and pillow breccia capped by
coeval basalt flows (Nelson, 1966). Basaltic dikes intrude the underlying Cretaceous
sediments and Tertiary glacial deposits. The basalts are mainly alkalic (Fig. 8.9), but
hypersthene- and quartz-normative tholeiites have also been reported (Smellie, 1987). The
oldest rocks in the JRIVG are 7 Ma basalt clasts in glacial deposits beneath the main
volcanic pile on James Ross Island (Table I). Volcanic activity continued through the
Pliocene and Quaternary, with the youngest K-Ar date being less than 0.3 Ma (Table I). The
most recent activity on James Ross Island occurred at three volcanic cones near the margin
of the ice cap. These cones are undisrupted by glacial activity and are undoubtedly of a very
young age (Strelin et aI., 1993).
The alkalic basalts ofthe JRIVG are chemically distinct from both the South Shetland
TABLE /I ~
I\l
Representative Whole Rock Major and Trace Element Concentration for Quaternary Volcanic Rocks from Bransfield Strait"
Axial Bridgeman Penguin Melville

Western seamount Eastern seamount ridge Deception Island Island Island Peak
Sample 292.18 310.07 297.Gl 309.01 DF86.32 DI048 D1082 01051 01055 D1075 DI060 Bl P174 MP405
Si02 49.5 50.4 52.5 54.3 54.3 50.9 53.8 51.6 68.9 63.9 57.8 55.8 50.6 52.8
Ti02 1.13 1.54 1.05 1.71 1.82 1.58 1.82 2.44 0.60 1.12 1.86 0.78 1.30 0.92
AIP3 14.1 16.3 16.9 15.7 15.8 17.3 16.2 15.2 14.5 15.6 15.6 18.7 17.2 16.7
FeO* 8.6 9.0 7.1 10.1 9.6 8.4 9.0 11.0 4.3 6.7 8.6 6.6 8.7 7.1
MnO 0.16 0.16 0.14 0.17 0.19 0.15 0.17 0.19 0.15 0.17 0.17 0.12 0.16 0.14
MgO 12.3 6.7 7.0 3.7 4.4 6.1 4.6 4.9 0.3 1.2 2.6 5.0 7.9 7.1
CaO 10.4 10.9 11.3 7.4 8.1 10.3 8.9 9.1 1.8 3.6 6.0 9.4 10.2 11.3
Nap 2.73 3.40 3.04 4.23 4.72 4.05 4.40 4.71 7.17 6.67 5.90 3.52 3.99 3.09
Kp 0.46 0.55 0.39 0.97 0.45 0.47 0.51 0.50 1.89 1.43 0.96 0.70 0.66 0.91
pps 0.16 0.19 0.12 0.23 0.30 0.25 0.29 0.36 0.12 0.42 0.37 0.08 0.29 0.14
Total 99.56 99.08 99.56 98.49 99.65 99.55 99.66 100.06 99.71 100.76 99.77 100.64 100.96 100.16
Sc 35 32 31 30 27 32 35 37 11 17 20 29 29 42
V 256 328 220 351 235 245 265 336 13 50 194 210 282 249
Cr 618 192 228 8 23 106 42 44 o 2 9 40 218 157
Ni 283 80 84 3 8 28 14 14 9 7 27 95 41
Cu 68 76 54 70 29 31 30 47 8 17 54 92 154 79
Zn 63 69 56 92 85 66 77 81 87 90 85 56 71 55
Ga 16 19 16 17 21 17 19 21 24 23 20 17 22 18 ~
Rb 10 10 7 17 4 7 9 5 30 21 14 17 7 17
Sr 255 266 280 330 189 465 374 350 130 254 325 357 636 598 ~
y 23 26 21 37 52 26 33 36 62 56 43 14 16 21 ~
Zr 93 104 94 154 197 143 167 171 444 355 249 72 90 112 ~
Nb 2 4 2 4 7 7 7 8 16 14 10 2 2 ;to
Ba 91 98 65 144 6 40 62 29 256 167 130 76 146 173 ~
ZrfY 4.0 4.0 4.5 4.2 3.8 5.5 5.1 4.8 7.2 6.3 5.8 5.1 5.6 5.3
aConcentrations were detennined by x-ray fluorescence at Washington State University. and are given as weight percent oxides for the major elements and parts per million for the trace elements. FeO* is total
~
J:J
Fe as FeO. Data for Western and Eastern seamounts are from Keller and Fisk (1992). Other data are from Keller et al. (1992). Data for Deception. Bridgeman. and Penguin Islands and Melville Peak have also !!l.
been published by other workers (e.g .• Weaver et al.• 1982; Smellie. 1990) and are comparable to data given here.
~
arc and from the <1 Ma Bransfield Strait basalts. The JRIVG tends to have lower Si02 (Fig.
8.9), and higher K20 and Ti0 2 (Table III). Paulet Island, especially, is high in total alkalies,
A1 20 3, and many incompatible trace elements. High concentrations of alkalies and incom-
patible trace elements in the JRIVG are reminiscent of many continental-rift basalts and
alkalic ocean island basalts. Their prominent Nb hump on the MORB-norrnalized multi-
element plot (Fig. 8.10) is a common characteristic of intraplate basalts not associated with
a subduction zone and is in marked contrast to the Nb depletions of the South Shetland arc
and Bransfield Strait backarc basalts. The 87Srf8 6Sr values range from 0.70306 to 0.70348,
and 143Ndll44 Nd from 0.51282 to 0.51293 (Keller et at., 1993), which places them in the
lower left quadrant in 87Srf8 6Sr versus 143Ndll44Nd space, displaced toward St. Helena-type
values.
10.1.4. Temporal Comparison

The hiatus between the end of arc volcanism on the South Shetland Islands and the
beginning of volcanism in Bransfield Strait was approximately 20 m.y., but the major and
trace element compositions are similar in some cases (Figs. 8.9 and 8.10). Bridgeman
Island, in particular, could be mistaken for an island-arc tholeiite (Table II). The submarine
samples, on the other hand, are more similar to backarc basin basalt (i.e., a hybrid of arc and
mid-ocean ridge basalt compositions; Fisk, 1990). The isotopic ratios of the arc and backarc
rocks are also similar, which implies that they all were derived from similar sources.
However, the lack of trace element and isotopic data from the South Shetland Islands and
the lack of samples from Bransfield Strait (partially rectified in 1993) preclude thorough
comparisons.
10.1.5. Across-Arc Comparison

The importance of the chemical contribution from a subducting slab has been shown
to decrease with distance from the trench in southern South America (Stem et at., 1990).
On the northern Antarctic Peninsula, the occurrence of volcanism at a variety of distances
from the South Shetland Trench (50-125 km for the South Shetland arc, 125-150 km for the
Bransfield Strait, and 300-400 km for the James Ross Island Volcanic Group) is an
excellent chance to constrain the spatial extent of contamination of the mantle by a
subducting slab. On the northern Antarctic Peninsula we see an obvious decrease in
contribution from the subducted slab with increasing distance from the trench. The high Ba
and low Nb of the South Shetland arc samples are typical of the subduction-zone influence
found in many volcanic arcs. The Bransfield Strait basalts are farther from the trench, have
a slight negative Nb anomaly, and are moderately enriched in the alkalies relative to
MORB. This is typical of backarc basin basalts, and suggests less of a contribution from the
subducted slab. The JRIVG basalts are farthest from the South Shetland Trench and do not
show any evidence for the influence of the slab subducting at the trench. Their high Nb is
typical of intraplate basalts and is the opposite of what would be expected if their source
had been contaminated by subduction (Keller et ai., 1993). Apparently, between 150 and
300 km is the greatest distance that the subducting slab can influence the chemistry of the
volcanism above it.
Some published comparisons of arc and backarc basalts suggest that the upwelling
mantle in backarc regions contains an ocean island basalt (OIB) component (Stem et at.,
1990; Hickey-Vargas, 1992). This OIB component can be seen in the JRIVG data but not
TABLE 11/
James Ross Island Volcanic Group Whole Rock Major and Trace Element Concentrations by X-ray Fluorescence"
James Ross Island
Hidden Lake Dreadnought Massey Conico Ventisca Kotick Villar Fabre
Lat (S) 64°02' 64°00' 63°57' 63°56' 63°57' 64°01' 64°06'

Lon (W) 58°18' 57°50' 57°58' 58°08' 58°04' 58°22' 58°22'
Si0 2 48.1 47.9 48.5 48.2 50.2 47.1 47.4 48.5 50.4 47.1 46.8 48.4 48.3 48.4 47.9 49.8 48.2
Ti0 2 1.78 2.01 2.09 1.89 1.84 1.71 1.95 2.04 1.87 1.71 1.75 1.55 1.69 1.59 1.64 1.77 1.74
AIP3 15.1 16.1 16.0 15.7 16.4 15.4 15.7 16.1 16.3 15.1 15.1 14.9 15.3 15.5 15.0 15.8 15.1
FeO* 10.5 10.0 9.8 10.4 9.5 10.9 10.2 10.6 9.7 10.7 10.5 10.9 10.6 10.8 11.1 10.4 12.0
MnO 0.16 0.17 0.16 0.16 0.15 0.17 0.16 0.17 0.16 0.16 0.16 0.17 0.15 0.16 0.17 0.15 0.18
MgO 8.9 7.2 7.6 7.9 6.7 8.1 8.6 8.4 6.8 10.3 10.0 9.2 9.1 8.9 8.6 7.8 9.1
CaO 8.2 9.9 8.9 9.2 8.3 8.4 8.4 9.5 9.2 8.0 8.5 8.6 8.0 7.9 7.8 7.9 8.0
N~O 3.34 3.56 3.39 3.37 3.83 3.34 3.64 3.57 3.58 2.58 2.96 3.11 3.27 3.16 3.43 3.71 3.27
Kp 1.20 0.93 1.13 1.01 1.59 0.83 1.44 1.19 1.37 1.13 0.89 0.72 1.01 0.92 1.05 1.28 0.95
pps 0.38 0.38 0.47 0.49 0.49 0.35 0.63 0.50 0.36 0.42 0.33 0.28 0.38 0.36 0.37 0.54 0.21
Total 97.57 98.16 98.10 98.38 98.96 96.17 97.99 100.55 99.71 97.15 96.94 97.73 97.64 97.66 97.02 99.17 98.72
Sc 19 27 28 27 23 24 22 29 30 20 26 20 22 21 19 13 25
V 176 207 191 180 169 159 169 197 195 149 167 166 146 154 151 150 148
Cr 276 178 207 256 188 223 228 239 184 313 304 272 233 231 226 199 225
Ni 166 75 94 115 96 146 157 116 86 200 180 180 177 162 168 145 170
Cu 41 39 48 50 37 58 56 49 44 48 42 52 44 51 41 47 35
Zn 92 83 85 82 90 93 87 81 87 88 90 90 101 99 99 97 90
Ga 16 17 19 15 17 16 18 15 19 21 16 18 18 18 15 17 20
Rb 17 11 13 12 19 10 17 13 25 10 13 11 10 9 13 12 10
Sr 506 510 607 611 701 453 742 618 481 686 510 359 482 446 475 653 395
y 24 25 24 26 24 24 25 26 28 22 22 25 24 22 24 21 19
Zr 154 168 191 178 203 162 226 180 164 174 134 129 142 141 142 197 139
Nb 27 29 36 40 41 38 41 39 28 28 24 17 24 25 24 38 23
James Ross Island
Cockburn
Sta Marta Dobson Ekelof Coley Marina Elba Paulet Island Island
Lat (S) 63°54' 64°02' 64°12' 64°11' 64°10' 64°15' 63°35' 64°12'
Lon (W) 57°54' 57°56' 57°14' 57°10' 57°22' 57°30' 55°46' 56°54'
Si02 47.2 48.0 47.8 48.3 46.1 48.3 48.9 48.5 48.8 48.7 48.8 48.5 48.2 47.4 46.9 47.8 48.0 48.8
Ti02 1.73 1.56 1.65 1.64 2.12 2.11 1.99 2.00 1.95 2.44 2.45 2.46 2.26 2.26 2.26 2.30 1.82 1.86
AlP3 15.4 15.5 15.2 15.4 14.6 15.7 15.4 15.3 15.3 18.2 18.4 18.3 16.9 16.9 16.7 17.0 16.0 16.4
FeO* 10.3 10.9 10.6 11.0 10.9 10.7 10.4 10.4 10.2 10.4 10.2 10.8 9.7 9.6 9.5 9.3 10.5 10.3
MnO 0.17 0.17 0.17 0.17 0.16 0.16 0.16 0.16 0.15 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16
MgO 8.5 9.5 9.3 8.7 9.4 7.6 8.9 9.0 8.9 4.7 4.8 4.6 7.7 6.6 6.9 6.7 7.0 6.7
CaO 7.8 8.7 8.7 8.6 8.2 9.5 8.0 7.9 7.8 7.6 7.4 7.3 9.3 8.9 9.1 9.2 8.6 8.7
N~O 3.74 3.11 3.27 3.37 4.03 3.50 4.10 3.99 4.35 4.89 5.32 4.93 4.21 4.16 4.30 4.33 3.69 3.59
~O 1.38 0.82 0.91 0.88 1.17 1.05 1.57 1.56 1.65 1.55 1.97 1.82 1.40 1.43 1.37 1.44 1.16 1.22
P20 S 0.52 0.32 0.37 0.36 0.96 0.38 0.63 0.62 0.64 0.75 0.84 0.84 0.81 0.75 0.75 0.86 0.40 0.43
Total 96.79 98.54 98.02 98.56 97.63 98.97 99.96 99.48 99.66 99.29 100.24 99.72 100.68 98.19 97.83 99.03 97.27 97.99
Sc 17 30 27 23 18 26 23 22 21 16 17 18 28 23 23 22 24 25
V 151 185 195 189 137 211 165 177 185 126 151 121 185 187 204 191 173 177
Cr 237 295 294 260 288 224 288 288 282 10 10 10 180 102 114 115 180 150
Ni 191 189 188 164 206 86 132 138 135 12 17 17 99 69 78 74 96 77
Cu 47 53 59 61 47 21 29 34 29 29 27 21 29 31 41 43 19 29
Zn 92 85 87 88 98 90 93 87 91 78 82 86 74 75 81 75 85 81
Ga 17 17 17 18 15 21 21 20 19 18 18 22 19 19 16 20 18 20
Rb 11 13 13 15 11 11 17 17 17 13 17 16 17 17 16 16 16 16
Sr 593 435 431 433 1383 536 728 720 732 945 911 917 951 940 943 937 536 547
Y 22 22 25 25 21 25 23 23 23 29 28 29 28 30 29 28 23 25
Zr 179 120 131 135 296 149 206 205 208 254 250 254 225 229 216 228 163 170
Nb 34 19 21 24 91 26 43 43 43 55 55 55 45 47 46 48 26 27
-Concentrations were determined by x-ray ftuorescence at Washington State University using techniques described in Keller el 01. (1992), and are given as weight percent oxides for the major elements and parts
per million for the trace elements. FeO* is total Fe as FeO. Data for Paulet Island and some of these locations on James Ross Island have also been published by other workers (Nelson, 1966; Baker el 01.,1977;
Smellie, 1987) and are comparable to the data given here.
in the Bransfield Strait data (Keller et aI., 1992). Apparently the OIB component is not
available or is not melted beneath Bransfield Strait, just as the subducted component does
not appear in the JRIVG lavas. The presence of a subducted slab beneath James Ross Island
should rule out the possibility that a rising plume or hot spot is responsible for the OIB-like
compositions of the JRIVG. Also, the "slab-window" hypothesis that has been suggested
for the origin of similar alkalic basalts found farther south on the Antarctic Peninsula at Seal
Nunataks and Alexander Island (Hole et aI., 1991) cannot apply to James Ross Island.
10.2. Intra-Bransfield Strait Comparison

All of the recent (,,;;300 ka) episode of volcanic activity in Bransfield Strait can be
classified as basalts to basaltic andesites, and on Deception Island as basalts to dacites.
These rocks appear to be the product of <5% to 15% melting of mantle that contained
0.5% to 2% of a subducted component (Keller et aI., 1992). Within these rocks, however,
we identify three contrasts in chemistry that are correlated with differences in mode or
location of eruption: (1) on-axis versus off-axis, (2) submarine versus subaerial, and (3)
northeast of 57°50'W versus southwest of 57°50'W. These differences are probably the
result of different mantle sources and processes.
10.2.1. On Axis versus off Axis

The two off-axis volcanoes, Penguin Island and Melville Peak (Fig. 8.4), have high Sr,
Ba, and K, and low Y relative to the on-axis basalts. Off-axis volcanic remnants on
Livingston and Greenwich islands, possibly contemporary with Penguin Island and
Melville Peak, also have high Sr and Ba and low Y, similar to Penguin Island and Melville
Peak (Smellie et aI., 1984). The high-Sr abundances of Penguin Island and Melville Peak
are not the result of crustal assimilation because other chemical changes that would be
associated with this assimilation are not observed (Keller et al., 1992). High Sr is also
typical of the older (>20 Ma) volcanic arc rocks of the South Shetland Islands, so the off-
axis Bransfield Strait volcanoes may have tapped the older arc mantle source. The lower
206Pbf2 04 Pb of the off-axis samples also shows that the mantle that produced the off-axis
lavas was chemically different from that which produced the on-axis basalts. The high Sr of
Penguin Island and Melville Peak excludes the possibility of residual plagioclase in the
mantle source (Keller et al., 1992). Small amounts of melting of a deeper, mineralogically
different (gamet-bearing) source is also suggested by the higher Ce/Sm (Keller et aI., 1992)
and other trace elements (Weaver et aI., 1979) of the off-axis basalts. Thus, off-axis
volcanoes appear to be the result of a smaller degree of melting of a deeper and chemically
distinct mantle than on-axis volcanoes (Weaver et al., 1979; Keller et al., 1992).
10.2.2. Submarine versus Subaerial

The seamounts have lower Zr/Y ratios (-4; Table II) than the subaerial volcanoes
(mostly >5). This difference is not a result of alteration, since both Zr and Yare relatively
immobile during weathering and most of the rocks are fresh. The degree of melting of the
mantle can affect the ZrIY ratio; however, this effect is only significant for small «5%)
amounts of melt, which would also be expressed in much larger amounts of incompatible
elements in the subaerial volcanoes than in the submarine ones. Some trace elements
(Ba, Rb, K), however, are lower in the subaerial volcanoes than in the submarine volcanoes,
indicating that chemical differences between these two groups are not related to the degree
of melting.
Most of the subaerial volcanoes are larger than the submarine volcanoes, so their
higher ZrN may represent a characteristic of a volcano as it grows; that is, continuous
partial melting of the mantle beneath the larger volcanoes could produce high ZrN, and the
smaller submarine volcanoes have not yet evolved to this chemistry. Alternatively, the
lower ZrN of the submarine basalts (which are closer to MORB values of 2.5 to 3.5) may
indicate that as the rift develops and deepens, the volcanoes become MORB-like. The rift
sample (DF86.32 in Table II) has the lowest ZrN (3.8) of any Bransfield Strait rock and
appears to be the most similar to MORB. It also has the lowest Rb, Sr, and Ba. Finally, the
larger volcanoes, which are mainly subaerial, may represent melting of deeper mantle
where garnet is present.
10.2.3. Along Axis

The chemistry of basalts from backarc rifts can vary significantly along strike of the
rift, as has been found in the Mariana Trough, Lau Basin, and Gulf of California (Saunders
and Tarney, 1984; Hawkins and Melchior, 1985; and Hawkins et aI., 1990). This is also true
of Bransfield Strait, where Na20 increases and Rb/Sr decreases abruptly from NE to SWat
approximately 57°50'W. The change occurs between Melville Peak and Penguin Island and
is evident in all of the analyzed submarine and subaerial samples. Basalts that are northeast
of this boundary also have lower total alkalies and lower 87S r /86S r at the same 143Nd/l44Nd
compared with basalts southwest of this boundary. The chemical boundary is found in the
same general region as the highest heat flow.
The abrupt change in isotopes along axis indicates there is a change in mantle sources
across the boundary, perhaps due to a change in age or composition of the subducted
oceanic crust. The boundary approximately coincides with the onshore extrapolation of
fracture zone E (Fig. 8.4), which separates younger crust to the southwest and older crust to
the northeast (Larter and Barker, 1991). Alternatively, the chemical boundary may reflect
the contrast between subducted oceanic crust created at the Antarctic-Phoenix Ridge
versus oceanic crust created at the Nazca-Phoenix Ridge. Different inputs to the mantle
wedge from the subducted slab could account for the isotopic differences as well as induce
changes in melting or mineralogy in the mantle. Elsewhere, along-arc variations in basalt
chemistry have been attributed to variations in the composition of the subducted material
(e.g., Lin et aI., 1990).
11. CONCLUSIONS
Regional tectonics of the Antarctic Peninsula support the hypothesis that little or no
active subduction is occurring beneath Bransfield Strait and the South Shetland Islands.
What little subduction that may be occurring would be equal to the amount of extension in
Bransfield Strait. Spreading ceased on the Antarctic-Phoenix spreading center at approx-
imately 4 Ma. The remnant of the Phoenix plate was pinned under the South Shetland
Islands, where it has been imaged to a depth of 40 km by seismic refraction results (Grad
et aI., 1993). Teleseismic earthquakes do not indicate any motion of a subducted slab
338 LA WRENCE A. LA WVER at al.
beneath Bransfield Strait, although one event (event 18; Pelayo and Wiens, 1989) may have
been produced along the downdip extension of the Hero fracture zone. Plate reorganization
since -4 Ma resulted in a 110° bend in the Antarctic-Scotia plate boundary that seems to
have imposed extensional stress on the Bransfield Strait region.
Bransfield Strait is a region of recent extension based on active and recent volcanism,
high heat flow, earthquake fault plane solutions, and seismic reflection data. This extension
may also be associated with trench rollback at the South Shetland Trench and defines the SE
boundary of the South Shetland Islands microplate. Seismic refraction results (Grad et aI.,
1993) indicate that the crust under Bransfield Strait is anomalously thick, up to 30 km, with
an unusual layer of crustal material with vp > 7.0 km S-I. They also found a lO-km-wide
body with seismic velocity vp = 6.8 km S-I, which they interpreted to be the active
Bransfield rift. Fault plane solutions (Pelayo and Wiens, 1989) from the Bransfield Strait
region support active extension, and some earthquake swarms can be directly correlated to
episodes of active volcanism. Barker and Austin (1994) and Lawver and Villinger (1989)
show evidence of active crustal extension from seismic reflection results, supportive of
continental crustal extension rather than normal seafloor spreading.
Active arc volcanism seems to have ceased at about 20 Ma. Recent Bransfield Strait
basalts are chemically transitional between island-arc and ocean ridge basalts, a feature
they share with many marginal basin basalts. Such chemistry agrees with the transitional
nature of the tectonics of the strait (i.e., transitional between arc and spreading regimes).
Considerable data supporting the existence of hydrothermal activity in Bransfield Strait
have already been published. The new discovery of sulfide minerals in the sediments at a
high heat flow site in Bransfield Strait is additional support for recent hydrothermal activity.
Acknowledgments
This work was supported by a number of National Science Foundation grants to the
authors and to those that have provided us with their results. Grants include, but are not
limited to, DPP-8916436 and DPP-9019247 to L. Lawver and DPP-8817126 to M. Fisk.
Field work on James Ross Island was supported by the Argentine Antarctic Institute and
Oregon State University. We are grateful to S. Porebski for donating the Cockburn Island
samples, and to E. Godoy for donating the Paulet Island samples. GPS-navigated multi-
beam bathymetry were collected on cruise 91-01 of RN Maurice Ewing and two cruises of
NOAA ship Surveyor. Heat flow data was collected on RN Polar Duke cruise PD89-IV.
We wish to thank the captains and crews from all those ships, without whose assistance this
work would not have been possible. We would particularly like to thank Rob Larter, Dallas
Abbott, and Daniel Barker for reviewing an unacceptably rough draft of this manuscript.
This is UTIG contribution number 1062.
REFERENCES
Acosta, J., Herranz, P., and Sanz, J. L. 1992a. Perfiles sismicos en el rift de Bransfield, Campafia Exantarte 90/91,
in Geologica de laAnttirtida Occidental (1. L6pez-Martinez, ed.), pp. 195-202, Simposios T 3, III Congreso
Geol6gico de Espana y VIII Congreso Latinoamericano de Geologia, Salamanca.
Acosta, J., Herranz, P., Sanz, J. L., and Uchupi, E. 1992b. Antarctic continental margin: Geological image of the
Bransfield Trough, an incipient oceanic basin, in Geological Evolution of Atlantic Continental Rises (C. W.
Poag and P. C. Graciansky, eds.), pp. 49-61, Van Nostrand Reinhold, New York.
Anderson, J. A., Pope, P. G., and Thomas, M. A. 1990. Evolution and hydrocarbon potential of the northern
Antarctica Peninsula continental shelf, in Antarctica as an Exploration Frontier: Hydrocarbon Potential.
Geology and Hazards (B. St. John. ed.). pp. 1-12. Am. Assoc. Petrol. Geol., Studies in Geology 31.
Ashcroft. W. A. 1972. Crustal structure of the South Shetland Islands and Bransfield Strait, Br. Antarc. Surv. Sci.
Rep. 66.
Baker, P. E., Buckley, E, and Rex, D. C. 1977. Cenozoic volcanism in the Antarctic, Phil. Trans. R. Soc. London
279:131-142.
Barbieri, M., Birkenmajer, K., Delitala, M. C., Francalanci, L., Narebski, W., Nicoletti, M., Peccerillo, A.,
Petrucciani, c., Todaro, M. L., Tolomeo, L., and Trudu, C. 1989. Preliminary petrological, geochemical and
Sr isotopic investigation on Mesozoic to Cainozoic magmatism of King George Island, South Shetland
Islands (West Antarctica), Mineral. Petrogr. Acta (Bologna) 32:37-49.
Barker, D. H. N., and Austin, J.A., Jr. 1994. Crustal diapirism in Bransfield Strait, West Antarctica-evidence for
distributed extension in marginal basin formation, Geology 22:657-660.
Barker, P. E 1976. The tectonic framework of Cenozoic volcanism in the Scotia Sea region, a review, in Andean
and Antarctic Volcanology Problems (0. Gonzalez-Ferran, ed.), pp. 330-346, International Association of
Volcanology and Chemistry of the Earth's Interior, Rome.
Barker, P. E 1982. The Cenozoic subduction history of the pacific margin of the Antarctic Peninsula: Ridge crest-
trench interactions, 1. Geol. Soc. London 139:787-801.
Barker, P. E, and Burrell, J. 1977. The opening of Drake Passage, Mar. Geol. 25:15-34.
Barker, P. E, and Dalziel, I. W. D. 1983. Progress in geodynamics of the Scotia Arc region, in Geodynamics of
the Eastern Pacific Region. Caribbean and Scotia Arcs (R. Cabre, ed.), Geodyn. Ser., Vol. 9, pp. \37-170,
American Geophysical Union, Washington DC.
Barker, P. E, Dalziel, I. W. D., and Storey, B.C. 1992. Tectonic development of the Scotia arc region, in Antarctic
Geology (R.I. Tingey, ed.), pp. 215-248, Oxford University Press, Oxford.
Birkenmajer, K. 1980. Age of the Penguin Island volcano, South Shetland Islands (West Antarctica) by the
lichenometric method, Bull. Acad. Pol. Sci. Ser. Sci. Terre 27:69-76.
Birkenmajer, K. 1992. Evolution of the Bransfield Basin and rift, West Antarctica, in Recent Progress in Antarctic
Earth Science (Y. Yoshida, K. Kaminuma, and K. Shiraishi, eds.), pp. 405-410, Terra, Tokyo.
Birkenmajer, K., Delitala, M. C., Narebski, w., Nicoletti, M., and Petrucciani, C.1986. Geochronology of Tertiary
island-arc volcanics and glacigenic deposits, King George Island, South Shetland Islands (West Antarctica),
Bull. Acad. Pol. Sci. Ser. Sci. Terre 34:257-273.
Birkenmajer, K., and Keller, R. A. 1990. Pleistocene age of the Melville Peak volcano, King George Island, West
Antarctica, by K-Ar dating, Bull. Polish Acad. Sci., Earth Sci. 38:17-24.
Brault, M., and Simoneit, B. R. T. 1990. Mild hydrothermal alteration of immature organic matter in sediments
from the Bransfield Strait, Antarctica, Appl. Geochem. 5:149-158.
Dalziel, I. W. D. 1984. Tectonic Evolution of a Forearc Terrane. Southern Scotia Ridge. Antarctica. Geological
Society of America Special Publication 200.
Fisk, M. R. 1990. Back-arc volcanism in the Bransfield Strait, Antarctica, 1. South Am. Earth Sci. 3:91-101.
Gamboa, L. A. P., and Maldonado, P. R. 1990. Geophysical investigations in the Bransfield Strait and in the
Bellingshausen Sea, Antarctica, in Antarctica as an Exploration Frontier: Hydrocarbon Potential, Geology
and Hazards (B. St. John, ed.), pp. 127-141, Am. Assoc. Petrol. Geol., Studies in Geology 31.
Garrett, S. W.1990. Interpretation of reconnaissance gravity and aeromagnetic surveys of the Antarctic Peninsula,
1. Geophys. Res. 95:6759-6777.
Gonzalez-Ferran, O. 1985. Volcanic and tectonic evolution of the northern Antarctic Peninsula-late Cenozoic to
recent, Tectonophysics 114:389-409.
Gonzalez-Ferran, O. 1991. The Bransfield rift and its active volcanism, in Geological Evolution of Antarctica (M.
R. A. Thomson, J. A. Crame, and J. W. Thomson, eds.), pp. 505-509, Cambridge Univ. Press, Cambridge.
Grad, M., Guterch, A., and Janik, T. 1993. Seismic structure of the lithosphere across the zone of subducted Drake
plate under the Antarctic plate, West Antarctica, Geophys. 1. Int. 115:586-600.
GRAPE Team. 1990. Preliminary results of seismic reflection investigations and associated geophysical studies in
the area of the Antarctic Peninsula, Antarctic Sci. 2:223-234.
Guterch, A., Grad, M., Janik, T., and Perchuc, E.1992. Tectonophysical models of the crust between the Antarctic
Peninsula and the South Shetland trench, in Geological Evolution of Antarctica (M. R. A. Thomson, J. A.
Crame, and J. W. Thomson, eds.), pp. 499-504, Cambridge University Press, Cambridge.
340 LA WRENCE A. LA WVER at al.
Guterch, A, Grad, M., Janik, T., Perchuc, E., and Pajchel, J. 1985. Seismic studies of the crustal structure in West
Antarctica 1979-1980-Preliminary results, Tectonophysics 114:411-429.
Guterch, A., Shimamura, H., and Polish-Japan-Argentina Research Group. 1991. An OBS-land refraction
seismological experiment in the Bransfield trough, West Antarctica, 199011991: Abstracts, Sixth International
Symposium on Antarctic Earth Sciences, 9-13 September, 1991, Japan, pp. 201-202.
Hawkins, J. w., Lonsdale, P. E, Macdougall, J. D., and Volpe, A M. 1990. Petrology of the axial ridge of the
Hawkins, J. W., and Melchior, J. T. 1985. Petrology of Mariana Trough and Lau Basin basalts, 1. Geophys. Res.
90:11431-11468.
Herron, E. M., and Tucholke, B. E. 1976. Sea-floor magnetic patterns and basement structure in the southeastern
Pacific, in Init. Repts. DSDP, 35 (C. D. Hollister and C. Craddock, et aI., eds.), pp. 263-278, U.S.
Government Printing Office, Washington, DC.
Hickey-Vargas, R. 1992. A refractory HIMU component in the sources of island-arc magma, Nature 360:57-59.
Hole, M. J., and Larter, R. D. 1993. Trench-proximal volcanism following ridge crest-trench collision along the
Antarctic Peninsula, Tectonics 12:897-910.
Hole, M. J., Rogers, G., Saunders, A. D., and Storey, M. 1991. Relation between alkalic volcanism and slab-
window formation, Geology 19:657-660.
International Seismological Centre Earthquake Catalogs. 1964-1993.
Irvine, T. N., and Baragar, W. R. A. 1971. A guide to the chemical classification of the common volcanic rocks,
Can. 1. Earth Sci. 8:523-548.
Jeffers, J. D., and Anderson, J. B. 1990. Sequence stratigraphy of the Bransfield Basin, Antarctica: Implications for
tectonic history and hydrocarbon potential, in Antarctica as an Exploration Frontier: Hydrocarbon Poten-
tial, Geology and Hazards (B. St. John, ed.), pp. 13-30, Am. Assoc. Petrol. Geol., Studies in Geology 31.
Jeffers, J. D., Anderson, J. 8., and Lawver, L. A 1991. Evolution of Bransfield basin, Antarctic Peninsula, in
Geological Evolution ofAntarctica (M. R. A. Thomson, J. A. Crame, andJ. W. Thomson, eds.), pp. 481-485,
Cambridge University Press, Cambridge.
Jin, Q., Kuang, E, Ruan, H., and Xing, G. 1991. Island arc volcanism and magmatic evolution in Fildes Peninsula,
King George Island, Antarctica, Abstracts, Sixth International Symp. Antarctic Earth Sci., Ranzan, Japan,
pp. 250-255.
Keller, R. A., and Fisk, M. R. 1992. Quaternary marginal basin volcanism in the Bransfield Strait as a modern
analogue of the southern Chilean ophiolites, in Ophiolites and Their Modern Analogues (L. M. Parson, B. J.
Murton, and P. Browning, eds.), pp. 155-170, Geol. Soc. Lond. Spec. Pub. No. 60.
Keller, R. A., Fisk, M. R., and Strelin J. A. 1993. Correlating distance from a trench with subducted component in
recent basalts from the northern Antarctic Peninsula [abs.), EOS supplement, October 26, 1993, p. 663.
Keller, R. A., Fisk, M. R., White, W. M., and Birkenmajer, K. 1992. Isotopic and trace element constraints on
mixing and melting models of marginal basin volcanism, Bransfield Strait, Antarctica, Earth Planet. Sci.
Lett. ill:287-303.
Keller, R. A, Strelin, J. A, Lawver, L. A, and Fisk, M. R. 1994. Dredging young volcanic rocks in Bransfield
Strait, Antarctic 1. U.S. XXVIII(l993 review issue):98-100.
Klepeis, K. A., and Lawver, L. A. 1994. Bathymetry of the Bransfield Strait, southeastern Shackleton fracture zone
and South Shetland Trench, Antarctic 1. U.S., XXVIII(l993 review issue): 103-104.
Larter, R. D. 1991. Debate: Preliminary results of seismic reflection investigations and associated geophysical
studies in the area of the Antarctic Peninsula, Antarctic Sci. 3:217-222.
Larter, R. D., and Barker, P. E 1989. Seismic stratigraphy of the Antarctic Peninsula Pacific margin: A record of
Pliocene-Pleistocene ice volume and paleoclimate, Geology 17:731-734.
Larter, R. D., and Barker, P. E 1991. Effects of ridge crest-trench interaction on Antarctic-Phoenix spread-
ing:forces on a young subducting plate, 1. Geophys. Res. 96:19,583-19,607.
Larson, R. L., and Chase, C. G. 1972. Late Mesozoic evolution of the western Pacific Ocean, Geoi. Soc. Am. Bull.
83:3627-3644.
Lawver, L. A., Dalziel, I. W. D., and Sandwell, D. T. 1993. Antarctic plate: tectonics from a gravity anomaly and
infrared satellite image, GSA Today 3: 117 -122.
Lawver, L. A, and Hawkins, J. W. 1978. Diffuse magnetic anomalies in marginal basins: Their possible tectonic
and petrologic significance, Tectonophysics 45:323-338.
Lawver, L. A., and Villinger, H. 1989. North Bransfield Basin: RN POLAR DUKE cruise PD VI-88, Antarctic 1.
Sci. 23:117-120.
Le Bas, M. J., La Maitre, R. W., Streckeisen, A, and Zanettin, B. 1986. A chemical classification of volcanic rocks
based on the total alkali-silica diagram, 1. Petrol. 27:45-750.
Lin, P. N., Stem, R. J., Morris, J., and Bloomer, S. H. 1990. Nd- and Sr-isotopic compositions of lavas from the
northern Mariana and southern volcano arcs: Implications for the origin of island arc melts, Contrib. Mineral.
Petrol. 105:381-392.
Maslanyj, M. P., Garret, S.w., Johnson, A. C., Renner, R G. B., and Smith, A. M. 1991. Aeromagnetic anomaly of
West Antarctica (Weddell Sea sector): BAS GEOMAP Series, Sheet 2,1:2,500,000, with supplementary text,
Cambridge, British Antarctic Survey.
Mayes, C. L., Lawver, L. A., and Sandwell, D. T. 1990. Tectonic history and new isochron chart of the South
Pacific, J. Geophys. Res. 95:8543-8568.
Nagihara, S., and Lawver, L. A. 1989. Heat flow measurements in the King George Basin, Bransfield Strait,
Antarctic J. Sci. 23:123-125.
Nelson, P. H. H. 1966. The James Ross Island Volcanic Group of north-east Graham Land, Br. Antarct. Surv. Sci.
Reports, No. 54.
Pankhurst, R 1. 1983. Rb-Sr constraints on the ages of basement rocks of the Antarctic Peninsula, in Antarctic
Earth Science (R L. Oliver, P. L. James, and J. B. Jago, eds.), pp. 367 -371, Austral. Acad. Sci., Canberra, and
Cambridge Univ. Press, Cambridge.
Pankhurst, R 1., and Smellie, 1. L. 1983. K-Ar geochronology of the South Shetland Islands, Lesser Antarctica:
Apparent lateral migration of Jurassic to Quaternary island arc volcanism, Earth Planet. Sci. Lett. 66:
214-222.
Parra, J. C., Gonzalez-Ferran, 0., and Bannister, J. 1984. Aeromagnetic survey over the South Shetland Islands,
Bransfield Strait and part of the Antarctic Peninsula, Rev. Geol. Chile 23:3-20.
Pelayo, A. M., and Wiens, D. A. 1989. Seismotectonics and relative plate motions in the Scotia Sea region, J.
Geophys. Res. 94:7293-7320.
Renner, R G. B., Sturgeon, L. J. S., and Garrett, S. W. 1985. Reconnaissance gravity and aeromagnetic surveys of
the Antarctic Peninsula, Br. Antarct. Surv. Sci. Rep. No. BO.
Rex, D. C. 1976. Geochronology in relation to the stratigraphy of the Antarctic Peninsula, Bull. Br. Antarct. Surv.
43:49-58.
Rex, D. c., and Baker, P. E. 1973. Age and petrology of Cornwallis Island granodiorite, Br. Antarct. Surv. Bull.
32:55-61.
Roach, P. J. 1978. The nature of backarc extension in Bransfield Strait [abstract], Geophys. J. R. Astron. Soc.
53:165.
Sandwell, T. T. 1992. Antarctic marine gravity field from high-density satellite altimetry, Geophys. J. Int.
109:437-448.
Sandwell, D. T., and Smith, W. H. F. 1992. Global marine gravity from ERS-I Geosat and Seasat reveals new
tectonic fabric, EOS Trans. AGU 73:133.
Saunders, A. D., and Tamey, J. 1984. Geochemical characteristics of basaltic volcanism within backarc basins, in
Marginal Basin Geology: Volcanic and Associated Sedimentary and Tectonic Processes in Modem and
Ancient Marginal Basins (B. P. Kokelaar and M. F. Howells, eds.), pp. 59-76, Geol. Soc. Sp. Pub. No. 16.
Saunders, A. D., Tamey, J., and Weaver, S. D. 1980. Tranverse geochemical variations across the Antarctic
Peninsula: Implications for the genesis of calc-alkaline magmas, Earth Planet. Sci. Lett. 46:344-360.
Saunders, A. D., Weaver, S. D., and Tamey, 1. 1982. The pattern of Antarctic Peninsula plutonism, in Antarctic
Geosciences (c. Craddock, ed.), pp. 305-314, University of Wisconsin Press, Madison.
Schlosser, P., Suess, E., Bayer, R, and Rhein, M. 1988. 3He in the Bransfield Strait waters: Indication for local
injection from backarc rifting, Deep-Sea Res. 35:1919-1935.
SEAN. 1993. Scientific Event Alert Network Bulletin for 31 March 1993, Smithsonian Institute, Washington,
DC,p.8.
Silfer, J. A., Engel, M. H., and Macko, S. A. 1990. The effect of hydrothermal processes on the distribution and
stereochemistry of amino acids in recent Antarctic sediments, Appl. Geochem. 5:159-167.
Simoneit, B. R. T. 1983. Effects of hydrothermal activity on sedimentary organic matter: Guaymas Basin, Gulf of
California-Petroleum genesis of protokerogen degradation, in Hydrothermal Processes at Seafloor Spread-
ing Centers (P. A. Rona, K. Bostrom, L. Laubier, and K. L. Smith, eds.), pp. 451-471, Plenum, New York.
Smellie, J. L. 1983. A geochemical overview of subduction-related igneous activity in the South Shetland Islands,
Lesser Antarctica, in: Antarctic Earth Science (R L. Oliver, P. L. James, and 1. B. Jago, eds.), pp. 352-356,
Austral. Acad. Sci., Canberra, and Cambridge Univ. Press, Cambridge.
Smellie, J. L. 1987. Geochemistry and tectonic setting of alkaline volcanic rocks in the Antarctic Peninsula: A
review, J. Volcanol. Geotherm. Res. 32:269-285.
Smellie, J. L. 1988. Recent observations on the volcanic history of Deception Island, South Shetland Islands, Br.
Antarct. Surv. Bull. 81:83-85.
Smellie, J. L. 1990. D. Graham Land and South Shetland Islands, in Volcanoes of the Antarctic Plate and Southern
Oceans (W. E. LeMasurier and J. W. Thomson, eds.), pp. 302-359, Antarct. Res. Ser., American Geophysical
Union, Washington, DC, 48.
Smellie, J. L., Davies, R E. S., and Thomson, M. R. A. 1980. Geology of a Mesozoic intra-arc sequence on Byers
Peninsula, Livingston Island, South Shetland Islands, Br. Antarct. Surv. Bull. 50:55-76.
Smellie, J. L., Pankhurst, R. 1., Hole, M. J., and Thomson, J. W. 1988. Age, distribution and eruptive conditions of
late Cenozoic alkaline volcanism in the Antarctic Peninsula and eastern Ellsworthland: Review, Br. Antarct.
Surv. Bull. 80:21-49.
Smellie,1. L., Pankhurst, R J., Thomson, M. R. A., and Davies, R. E. S. 1984. The Geology of the South Shetland
Islands. VI: Stratigraphy, Geochemistry and Evolution, Br. Antarct. Surv. Sci. Rep. No. 87.
Stem, C. R, Frey, F. A., Futa, K., Zartman, R. E., Peng, Z., and Kyser, T. K. 1990. Trace-element and Sr, Nd, Pb,
and 0 isotopic composition of Pliocene and Quaternary alkali basalts of the Patagonian Plateau lavas of
southernmost South America, Contrib. Min. Petrol. 104:294-308.
Strelin, J. A., Carrizo, H., Lopez, A., and Torielli, C. 1993. Actividad volcanic a holocena en la Isla James Ross.
Segundas Jornadas de Comunicaciones sobre investigaciones Antarticas, Actes pp. 335-340, Buenos Aires.
Suess, E., Fisk, M., and Kadko, D. 1988. Thermal interaction between backarc volcanism and basin sediments in
the Bransfield Strait, Antarctica, Antarctic J. U. S. 22(5):46-49.
Sykes, M. A. 1988. New K-Ar age determinations on the James Ross Island Volcanic Group, north-east Graham
Land, Antarctica, Brit. Antarct. Surv. Bull. 80:51-56.
Tanner, P. W. G., Pankhurst, R J., and Hyden, G. 1982. Radiometric evidence for the age of the subduction
complex of the South Orkney and South Shetland Islands, West Antarctica, J. Geol. Soc. Lond.139:683-690.
Thomson, M. R A. 1992. Stratigraphy and age of pre-Cenozoic stratified rocks of the South Shetland Islands:
review, in Geologia de la Antartida occidental (1. Lopez-Martinez, ed.), pp. 75-92, Simposios T3, III
Congreso Geologico de Espana y VIII Congreso Latinamericano de Geologica, Salamanca.
Vila, J., Ortiz, R., Correig, A. M., and Garcia, A. 1992. Seismic activity on Deception Island, in Geological
evolution ofAntarctica (M. R A. Thomson, 1. A. Crame, and 1. W. Thomson, eds.), pp. 449-456, Cambridge
University Press, Cambridge.
Weaver, S. D., Saunders, A. D., Pankhurst, R J., and Tamey, J. 1979. A geochemical study of magmatism
associated with the initial stages of backarc spreading: The Quaternary volcanics of the Bransfield Strait from
South Shetland Islands, Contrib. Mineral. Petrol. 68:151-169.
Weaver, S. D., Saunders, A. D., and Tamey, J. 1982. Mesozoic-Cenozoic volcanism in the South Shetland Islands
and the Antarctic Peninsula: Geochemical nature and plate tectonic significance, in Antarctic Geoscience (c.
Craddock, ed.), pp. 263-273, Univ. Wisconsin Press, Madison.
Wernicke, B. 1981. Low-angle normal faults in the Basin and Range province: Nappe tectonics in an extending
orogen, Nature 291:645-648.
Whiticar, M. J., Suess, E., and Wehner, H. 1985. Thermogenic hydrocarbons in surface sediments ofthe Bransfield
Strait, Antarctic Peninsula, Nature 314:87-90.
9
Structural and Kinematic Evolutions

of the Okinawa Trough Backarc Basin
Jean-Claude Sibuet, Shu-Kun Hsu, Chuen-Tien Shyu,
and Char-Shine Liu
ABSTRACT
Refraction data acquired in the Okinawa Trough show that the crust is of continental origin
and that its thickness increases from 10 km in the southern Okinawa Trough to 30 km in the
northern Okinawa Trough. The Okinawa Trough is still in a rifting stage. In the southern
Okinawa Trough, magnetic anomalies are linked either to limited arc volcanic intrusions
located on the northern side of the Ryukyu arc, to the axis of the main depressions where
basaltic intrusions with volcanic arc affinities are present, or to volcanic products lying at
different depths within the thinned continental crust. A link among partial melting occur-
ring just above the Benioff zone, ascent of volcanic products through extended continental
crust, and the amount of crustal extension are proposed. We also suggest that both arc and
backarc basin volcanism are derived from the same source located above the Benioff zone.
Melt preferentially rises through the fissures and faults created by extension in the conti-
nental crust. The total amount of extension across the Okinawa Trough has been estimated
from refraction and gravity data. It decreases slightly from 80 km in the southern Okinawa
Trough to 74 km in the northern Okinawa Trough. Parameters of rotation have been
established for the entire opening of the Okinawa Trough and for the three phases of
extension already identified: (1) The late Pleistocene to Recent phase of extension is
characterized by normal faults with vertical offsets of a few meters changing progressively
in direction along the Okinawa Trough. The amount of extension is about 5 km in the
middle Okinawa Trough. (2) The early Pleistocene phase of extension is characterized by
large tilted blocks which affect late Pliocene-early Pleistocene sediments. Directions of
these normal faults change progressively along the Okinawa Trough. Extension in the
northern Okinawa Trough is estimated as 25 km. The difference in azimuth of the normal
faults for these two tectonic phases increases toward the northern Okinawa Trough. (3) The
early Miocene phase of extension is poorly characterized from geological data but has been
Jean-Claude Sibuet and Shu-Kun Hsu • IFREMER, Centre de Brest, 29280 Plouzane, France. Chuen-
Tien Shyu and Char-Shine Liu • Institute of Oceanography, National Taiwan University, Taipei, Taiwan,
Republic of China.
343
344 JEAN-CLAUDE SIBUET at al.
calculated from parameters of other rotations. The corresponding pole of rotation is located
1500 kIn northeast of Kyushu; 50 kIn of extension in the northern Okinawa Trough and 75
kIn in the southern Okinawa Trough are expected. The proposed kinematic evolution of the
Okinawa Trough differs considerably from the previous reconstructions and raises nu-
merous questions, as for example, the interpretation of paleomagnetic measurements
obtained in the southeastern Ryukyu Islands. The bending of the Okinawa Trough devel-
oped since the onset of the backarc basin extension, except for the southwestern portion of
the Okinawa platelet whose pronounced bending is due to the collision in Taiwan.
1. INTRODUCTION
The island of Taiwan is located at the junction between the East China Sea continental
margin and the Luzon arc (Fig. 9.1). East of Taiwan, the Okinawa Trough (OT) is linked to
the subduction of the Philippine Sea plate beneath the Eurasian plate. South of Taiwan, the
South China Sea on the Eurasia plate is subducted beneath the Luzon arc, which belongs to
Luzon - - ---:
.. ~___ South China Sea
Ryukyu Trench
FIGURE 9.1. Block diagram of the deep ocean around Taiwan. Bathymetric data have been collected by
numerous institutions and are available through the National Geographic Data Center (NGDC). Complementary
data acquired east and south of Taiwan have been made available by the Institute of Oceanography of National
Taiwan University. After cleaning and crossover adjustments (Hsu, in press), data have been gridded and
represented with the GMT software package (Wessel and Smith, 1991). The viewing angle is chosen in order to
image the southern Okinawa Trough and its relationship with the Ryukyu subduction zone but also the transfer
of the convergent motion from the Manila Trench to the collision in Taiwan. South of Taiwan, the Philippine Sea
plate overrides the Eurasian plate, whereas northeast of Taiwan, the Eurasian plate overrides the Philippine Sea
plate. Shading from the south (azimuth 190°).
OKINAWA TROUGH BACKARC BASIN 345
the Philippine Sea plate. The main convergent motion is thus transferred from the Manila
Trench to the collision in Taiwan. In other words, south of Taiwan, the Philippine Sea plate
overrides the Eurasian plate, whereas northeast of Taiwan the Eurasian plate overrides the
Philippine Sea plate.
The Okinawa Trough is a curved backarc basin convex to the southeast and located
behind the Ryukyu Trench and Ryukyu Islands (Fig. 9.2). It extends from the Han Plain in
northern Taiwan to Kyushu Island in southern Japan. Its width increases from 100 km in the
south to more than 200 km in the north, and its curvature is mainly restricted to the south. It
abuts the narrow parallel curved belts of Taiwan, which are convex to the northwest over
approximately the same distance. Several authors have suggested that the bending of the
OT was due to the collision and the following indentation of the Philippine Sea plate within
the Eurasian plate (Jarrard and Sasajima, 1980; Letouzey and Kimura, 1986), the indenter
being the undeformed Luzon volcanic are, located on the northwestern border of the
Philippine Sea plate, against the continental part of the Eurasian plate.
[121' III"
'0'
s ""
r-----.;~ ~ Detailed maps
zo·
..
...
12'"
FIGURE 9.2. Map of the Okinawa Trough, Ryukyu are, and surrounding areas with location of seismic, 3.5-kHz,
and Sea Beam profiles shown in this study. Black squares represent Sea Beam detailed maps shown in this study
with the corresponding profile numbers.
346 JEAN-CLAUDE SIBUET et al.
The purpose of this chapter is to detennine the direction of nonnal faults that
accommodate the different phases of extension along the OT, to detennine if tensional
features are inherited from previous tensional phases, to evaluate the amount of extension
associated with each tensional phase, and to propose a coherent kinematic evolution of
the OT. We will try to establish if the pronounced bending of the southern OT occurs during
the fonnation of the backarc basin or if it is an inherited curvature since the initiation of the
OT. Basic data used in this chapter are detailed swath bathymetyric data, single-channel
seismic and 3.5-kHz data, and new compilations of all available bathymetric and magnetic
data acquired east of Taiwan.
2. TECTONIC SETTING OF THE OKINAWA TROUGH
The OT fonned by extension within the continental lithosphere, as first suggested by

Uyeda (1977). The first rifting phase is dated middle to late Miocene and is associated
with a broad uplift of the continental Ryukyu nonvo1canic arc and Taiwan-Sinzi folded
belt, followed by nonnal faulting and subsidence (Sun, 1981; Letouzey and Kimura, 1986).
This continental rifting phase (Table I) occurred after a major change in the direction of the
Philippine Sea plate moving relative to the Eurasia plate in early Miocene (Le Pichon et aI.,
1985). After 5 m.y. cessation of tectonic activity (Kimura, 1985), the second rifting phase
(Table I), which is thought to be much more important than the previous one and resulted in
the main fonnation of the OT, started about 2 m.y. ago at the Plio-Pleistocene boundary.
The initiation of this phase is associated with the recent uplift of the Ryukyu arc at the Plio-
Pleistocene boundary (Ujiie, 1980). Based on seismic correlations with drilling stratigraphy
(Tsuburaya and Sato, 1985), an upper Pleistocene phase of extension, called the third phase
of extension but much less developed than the lower Pleistocene phase, has been identified
(Furukawa et aI., 199Ib). This third phase of extension very probably extends to Recent
times, as it is clearly expressed within the sediments deposited since late Pleistocene as
nonnal faults characterized by small vertical offsets (Sibuet et aI., 1987). Consequently, the
Okinawa backarc basin was fonned through at least three tectonic phases (Table I) that
occurred within a relatively short period of time. To explain the occurrence of closely
consecutive tectonic phases in backarc basins, Sibuet et at. (1987) suggested that very slight
changes in the subduction parameters (such as convergence rate, angle of convergence,
angle of the Benioff zone) would induce major tectonic change in the backarc basin evolution.
2.1. Northern Okinawa Trough

The northern OT is considered to be in a rifting stage as revealed by refraction
experiments of Ludwig et at. (1973), Kasahara et at. (1985), and Iwasaki et at. (1990).
From a 190-km-long profile shot from Kyushu Island along the axis of the trough, Iwasaki
et at. (1990) demonstrated that the velocity structure is characteristic of continental crust
and that the Moho shallows toward the southwest from 27 to 30 km below the continental
Kyushu Island to about 23-24 km at the SW end of the profile, where the trough is 1,000 m
deep. Independently, Sibuet et at. (1987) showed that the thick sedimentary cover of the
northern OT results from a massive supply of sediments from the Yangtze and Yellow
rivers and was affected by tensional processes. Nonnal faults 120 to 150 km long and
trending N040° to N060° limit tilted fault blocks to about 10 to 30 km wide. These faults
related to the second tectonic phase and were active in late Pliocene-early Pleistocene
TABLE I
Main Phases of Extension in the Okinawa Trough
Normal faults in the OT
Southern OT Northern OT
(25°N; 124.5°E) (29.5°N; 128.5°E)
OT main
Age tensional Mean Amount of Mean Amount of
(m.y.) Period phases directiona extensionb directiona extensionb
Holocene
0.01 (Recent) 3rd phase 90° ± 10° 5 km 74° ± 15° 6 km
Upper
0.7 Pleistocene
Lower 2nd phase 80° ± 10° 8 km 55° ± 15° 25 km
Pleistocene
1.7
Pliocene
5
Cessation of
tectonic
activity
Upper Miocene 54° 75 km 58° 57 km

12 1st phase (calculated) (calculated) (calculated) (calculated)
Middle Miocene
17
Lower Miocene
aThe mean directions of normal faults are extracted from Figs. 9.18a and b.
"The amounts of extension is computed from parameters of rotation (Table II).
(Fig. 9.3). Onlapping Pleistocene sediments partially filled half-grabens but are, in tum,
normally faulted by the present-day third phase of extension. This phase of extension is
characterized by faults of a different strike, which is about 25° clockwise from the previous
trends. However, most of the deformation occurs during the early Pleistocene. All the early
Pleistocene tilted fault blocks and half-grabens are arranged in a right-lateral en echelon
pattern without evidence of transform motion between blocks, which suggests that the
amount of extension varies along the blocks. However, the amount of total extension is
approximately the same when local extensions are summed up perpendicularly to the OT
over several en echelon tilted blocks (Sibuet et aI., 1987).
2.2. Southern Okinawa Trough

The southern OT located SW of Okinawa Island, consists of a broad fiat basin with
over 2 kIn of sediments which were deposited mostly at a very high deposition rate in the
R IV JEAN CHARCOT POP 1 CRUISE

STRUCTURAL MAP
FIGURE 9.3. Geological and structural map of the Okinawa Trough from Sibuet et al. (1987): I, present-day arc
volcanism; 2, trench axis; 3, normal faults; 4, trench slopes and accretionary complex; 5, structural highs.
last 0.5 m.y. (Hilde et aI., 1984; Kimura, 1985). The trough ends at the eastern termination
of the Han Plain right above the abrupt end of the Ryukyu Benioff zone defined by the
associated cloud of deep earthquakes (Tsai, 1986). Based on the interpretation of multi-
channel seismic (MCS) lines, Kimura (1985) suggested that most of the Okinawa depres-
sion (at least a band 80 km wide) was of oceanic origin. This was first questioned by Sibuet
et al. (1987), who demonstrated that the backarc spreading phase only started very recently,
as shown by the presence of en echelon and, in some cases, overlapping depressions
oriented N070° - N080° with the presence of a few intruded volcanics (Fig. 9.3). During this
period, extension and subsidence were recorded both within the whole sedimentary column
by the downward curvature of seismic reflectors toward the center of elongated bathymetric
depressions and within the depressions by normal faulting and intrusion of volcanics
(Sibuet et aI., 1987). Magnetic anomalies up to 300 nT (Furukawa et al., 1991a) are
associated with these volcanic ridges, where vesicular basalts have been photographed and
dredged. Sibuet et al. (1987) restricted the width of the oceanic part to a maximum of 20 km
on each side of the axial depressions, and even suggest that the oceanic domain could have
been limited only to the volcanic ridges themselves. In 1988, a refraction seismic experi-
ment with a deployment of 18 ocean bottom seismometers was conducted across the
southern OT from the East China shelf to the Ryukyu nonvolcanic arc at about 124.5°E
longitude (Hirata et aI., 1991; profile I, Fig. 9.2). In the central part of the trough, the Moho
is at least 18 km below the sea surface. Even if the Moho cannot be followed outside the
trough axis, this is the first complete and comprehensive refraction data set available across
the OT. As underlined by Hirata et al. (1991), the major contribution of this refraction
profile is that the southern OT is still in a continental rifting stage and that the oceanic
domain, if any, is limited to the volcanic ridges in the axial portion of the trough. This
interpretation is confirmed by trace element analyses and isotopic ratios of basaltic samples
collected during the RN Jean Charcot POPI cruise (October 1984) on the southern OT
volcanic ridges, which show that these lavas are characteristic of arc volcanism (Sibuet
et aI., 1986; Boespflug, 1990). Analyses results are similar to those performed on samples
of the Ryukyu arc (e.g., the Iizuma arc volcano in Japan; Ishizaka et aI., 1977) which show
either that the magmatic output comes from the Benioff zone at an optimal depth for arc
magmatism generation or that the arc magmatism ascends through the crustal tensional
features of the backarc basin. Since the major part of the southern OT is underlain by at
least 10 km of continental crust, the southern OT is still in the stage of incipient rifting of a
backarc basin with an incomplete thinning of the continental crust.
2.3. Middle Okinawa Trough

This system of backarc volcanic ridges of the southern OT ends in the middle OT,
close to Okinawa Island where a series of parallel basaltic ridges oriented N075° (Fig. 9.3)
associated with linear magnetic anomalies have been identified (Sibuet et aI., 1987). These
magnetic anomalies could be interpreted either in terms of dike intrusions or emplacement
of early oceanic crust (Davagnier et aI., 1987). This specific area represents the transition
between the thinned continental crust and the first signs of possible backarc basin oceanic
crust but also the transition from present-day arc volcanism, which decreases in intensity
from Japan to north of Okinawa Island, to volcanism occurring within the backarc basin
itself. It has been named the volcanic arc rift migration phenomenon (VAMP) area (Sibuet
et aI., 1987). Refraction data collected by Nagumo et al. (1986) show that a 1O-km-thick
upper crust of continental origin overlies a lower crust (6.8 km/s) of unknown thickness.
Consequently, the crust is still of continental origin and more than 10 km thick. Lavas and
pumices were sampled during the POPI (RN Jean Charcot cruise, 1984), DELP-84
(Wakashio cruise, 1984), and Shinkai 2000 dives (Uyeda et aI., 1985a). Dredged olivine
basalts are of island-arc type (Uyeda et aI., 1985a,b; Kimura et aI., 1986; Sibuet et aI., 1986;
Boespflug, 1990) and are younger than 1 m.y. (K-Ar and fission track ages; Kimura et al.,
1986). Some dredged rhyolite, dacite, and dacitic andesite indicate a bimodal volcanism.
Thus, as for the basalts dredged in the southern OT, petrologic and geochemical evidences
do not support the existence of a typical oceanic crust in the VAMP area and the middle ~T.
In summary, the whole OT is still in a rifting stage. The minimal crustal thickness is 10
km in the deepest southern OT but increases in the northeast direction, along the axis of the
trough, to attain about 30 km in southern Japan as beneath the continental shelves of the
East China Sea and Ryukyu arc. On tensional continental margins, the thickness of thinned
crusts could reach a few kilometers, and synrift volcanism has been currently observed
(e.g., northeast Atlantic continental margins; Sibuet, 1992). Complementary continental
rifting, in order to reduce the thickness of the thinned continental crust within the OT, is still
anticipated before true oceanic crust would be emplaced. Consequently, the present-day
observed backarc volcanism is considered there as products of arc magmatism emplaced
through the system of normal faults where thinning is maximum rather than true backarc
volcanism emplaced through a fully thinned continental crust. However, in the middle OT,
as basalts have been emplaced over a width of 30 km (Sibuet et a!., 1987), a possibility
exists there that no more continental thinning would occur. In the two schematic litho-
spheric cross sections proposed in Fig. 9.4, we postulate that the magmatic supply has the
same origin for the arc and backarc basin volcanism and is issued from approximately the
same area above the Benioff zone.
3. EXTENSIONAL TECTONICS IN THE OKINAWA TROUGH
Sibuet et a!. (1987) demonstrated that normal faults that bound the tilted fault blocks
(second phase) vary in direction from N040° to N060° in the northern OT and from N075°
to N090° in the southern OT, following the direction of the northward concavity of the
trough, but with a systematic offset with respect to the mean direction of the OT (Fig. 9.3).
This is obvious on the OT structural map of Sibuet et al. (1987). These authors also noticed
that small normal faults affecting the Upper Pleistocene and Quaternary sediments (third
phase) were oriented slightly different from the trend of main normal faults related to the
previous Lower Pleistocene tectonic phase (second phase). We performed detailed analyses
of all Sea Beam data acquired during the POPI cruise in order to discriminate the two
families of normal faults.
backarc balln
dlffu . . . . '.nllon arc volcanllm
-1- ,V
EU RASIAN
PLATE
SE
FIGURE 9.4. Map of the Okinawa Trough region showing the position of the trench, the present-day active
volcanic arc, and the axes of depressions and elongated ridges of the Okinawa Trough. Depths of the Benioff zone
are from Eguchi and Uyeda (1983). Cross sections of the backarc basin and forearc across the northern and
southern OT are adapted from Sibuet et al. (1987). The magma supply has the same origin for the arc and backarc
basin volcanisms and is generated from approximately the same area above the Benioff zone.
3.1. Identification of Two Families of Faults

Figures 9.5 and 9.6 show seismic profiles shot in the northern and middle OT. These
profiles complement those already published (Lee et al., 1980; Kato et aI., 1982; Sibuet
et aI., 1987) in illustrating the shallow crustal structures. Large tilted fault blocks generally
10 to 30 km wide are clearly evident on all the seismic sections.
On profiles 42 and 43 located in the northern OT (Figs. 9.5a, 9.5b), the major tilted
block, which is the same on both sections, is about 10 km wide and presents a vertical offset
of at least 1 km. The formation of this block is late Pliocene-early Pleistocene in age and
is related to the second phase of extension. Depressions created during the second tectonic
phase are almost completely filled with post-early Pleistocene sediments. A slight tectonic
phase of early late Pleistocene could exist just after the deposition of the oldest infilled
sediments as shown by the onlapping reflectors located over the slightly deformed deep
sedimentary layer (located at 2.5 s.d.t.t. between 00H15 and OlHOO on profile 42). The late
Pliocene to Recent third phase of extension is evidenced by normal faults that affect the
upper part of the sedimentary section and by the small changes in the topography of the
seabed (thin arrows in Fig. 9.5a). Thus, seismic data tell us that the second and third
phases of extension are two different successive tectonic phases. However, no clear sign of
remobilization of the large offset normal faults belonging to the second phase exists during
the third phase.
In the middle OT, tilted blocks are especially well imaged from the arc platform to a
distance of a few kilometers to the axis of the backarc basin (Fig. 9.6, 07H45 to 12Hl5 on
profile 99). The 20-km large tilted block located on this profile between 08H1O and 09H30
is a textbook example. It displays all the expected conventional features for such a block
formed in a tensional environment: tilting of the block over two thirds of its length without
internal deformations, fast rotation of the block attested by the presence of a poorly
developed V-shaped body of synrift sediments, and internal deformation of the deeper
portion of the block between 08H1O and 08H40 by drag along the normal fault of the
adjacent eastern block. Another large tilted block (between IOH30 and llH30 on profile 99)
displays a significant brittle deformation by drag during its tilting and subsidence. Three
normal faults are observed within the southeastern half portion of the block, which allows
us to reconstruct the initial shape of the block before extension. From dredged samples
collected during the POPI cruise (30 0 04'N; 128°31'E) and the Toka 1 oil well results (Nash,
1979), tilting occurred in late Pliocene-early Pleistocene (second phase) and has affected
the whole sedimentary cover. Onlapping Upper Pleistocene to Recent sediments partly fill
the depressions of the OT axial portion where subsidence is maximum (Fig. 9.6). These
sediments are in tum normally faulted during the more recent phase of extension. Vertical
offsets of the present-day third generation of faults are considerably smaller than those of
the second generation. The recent normal faults observed on the seismic sections corre-
spond to minor deformations of the sedimentary column between 12H20 and 14H30 and to
a hummocky sea bottom between 12HOO and 12H40 on profile 99. A basaltic intrusion
appears in the deepest portion of the OT where the amount of extension is maximum. Large
normal faults developed during the second tectonic phase could have been reactivated, but,
once again, there is no evidence for that from the available seismic data.
The 3.5-kHz records corresponding to the same portions of seismic profiles 42 and 43
shown in Fig. 9.5 much more clearly reveal normal faults that developed during the last two
tectonic phases (Fig. 9.7). The vertical exaggeration of the 3.5-kHz profiles is 40. The large
normal faults at OOHIO and 02H15 limit the two tilted blocks of the second generation
NORTHERN OKINAWA TfiOUGH SCS VE 3
s.dU
Sept. 29, 19B4 PrQf!lt, 42
a ';i.d.tt
So?pt 29, 1984 Prcfiie 43
NORTHERN OKINAWA TROUGH SCS V.E.' 3

~km 20km
I I
Sf NW
23,30 00:00 00:30 00:57
Sept. 29. 1984 Prof~e 42
NW Seabeam map Sf
01:06 02,00 02:30
b S.d.t.t. Sept. 29, J984 Profde 43
FIGURE 9.5. (a) Examples of the Plio-Pleistocene tilted fault block geometry (large arrows) and Recent normal
faulting (small arrows) in the northern Okinawa Trough: IFREMER high-speed (10 knots), single-channel seismic
profiles 42 and 43. Locations of profiles are shown in Fig. 9.2. (b) Structural interpretation of seismic profiles
shown in (a). Tilted blocks formed during the second tensional phase (Table I) appear in gray. In light gray, deep
sediments slightly deformed during early Pleistocene or early late Pleistocene by tensional movements assigned to
the second phase of extension.
a
2S
~
~
:xla
MIDDLE OKINAWA TROUGH Profile 99 . _ __
~ 300
SCS VE 0 3 ]''nT §
- ~ ::t:
Oct 4 1984 - --- ------ ---------- - - " _ _ ' F - - - _ _ _ '.
-"F---------- --~-:~~-~/,~ ~-~~=_-7\:/=~-~ L ~
o -- - -+------~ --- - Seabeam Map nT
200 10
ol_---1......-..L~---L_ ~km l~--- 2
roT Sf 09.00
0- ~
~
~
5-
ti .,j f
.;
FIGURE 9.6. IFREMER high-speed (10 knots), single-channel seismic profile 99 shot in the middle Okinawa Trough from the arc platform to the axis of the trough. Location
shown in Fig. 9.2. Plio-Pleistocene tilted fault blocks are especially well imaged, except in the axial part of the trough where present-day normal faults are only marked by minor
deformations of sediments and seabed. A volcanic ridge associated with a 200-nT magnetic anomaly is emplaced in the axis of the depression, where subsidence is maximum.
Tilted blocks formed during the second tensional phase (Table I) appear in gray in the structural interpretation of the seismic profile.
~
already identified in Fig. 9.5. However, between 12H20 and 14H30, numerous normal faults
of the third generation, which are not imaged on the seismic sections because their
topographic expression does not exceed a few meters, appear on the 3.5-kHz records as
nearly vertical features. Profiles 42 and 43 differ only about 20° in azimuth with a course
change at OlH03. The same normal faults appear between OlHll and OlH20 on profile 43
and between OOH45 to OlHOO on profile 42. As profiles 42 and 43 are 2 to 3 Ian apart, this
suggests that normal faults of the third generation present a lateral continuity of at least
a few kilometers.
A section ofthe 3.5-kHz profile recorded in the axial part of the southern OT (Fig. 9.8,
profile 130) completes the information given above. Two sedimentary units are identified:
The lower folded unit corresponds to the upper part of the Pliocene sedimentary section
slightly deformed by the Lower Pleistocene extension; the upper unit corresponds to
onlapping Upper Pleistocene to Recent sediments and is still not much affected there by the
last tectonic phase. The only evidence of recent deformation is the small bulge of Upper
Pleistocene to Recent sediments which appears between OOH55 and OlHlO. The recent
FIGURE 9.7. Portions of the 3.5-kHz profiles 42 and 43 (in northern Okinawa Trough) corresponding to the
seismic sections shown in Fig. 9.5a. Location shown in Fig. 9.2. Vertical exaggeration is about 40. The Plio-
Pleistocene tilted fault block is bounded by a 30-m-offset normal fault (OOHIO on profile 42 and 02H15 on profile
43). Recent normal faults with a few-meter offsets affect Quaternary sediments.
FIGURE 9.8. Portion of the 3.5-kHz profile 130 (in

southern Okinawa Trough) located in Fig. 9.2. Vertical
exaggeration is about 40. The gently folded Plio-
Pleistocene sedimentary unit is overlain by onlapping
Upper Pleistocene Quaternary sediments slightly de-
formed between OOH55 and OlHlO. The dotted line
separates the two sedimentary units.
Profile 130 Oct. fl. 1984
defonnation seems to occur independently from the previous generation of faults and does
not seem to systematically mobilize preexisting nonnal faults or zones of weakness.
The interpretation of seismic profiles shows that the second and third phases of
extension are distinct and that nonnal faults belonging to each of these tensional phases
could be easily differentiated.
3.2. Spatial Distribution of the Two Families of Faults

Nonnal faults which belong to the two fault families described previously have been
identified from the POPI 3.5-kHz records, where they display a topographic offset larger
than 4 m, and from the seismic profiles, where the offset is greater than 20 m. Most of the
Sea Beam data acquired during this cruise have been reprocessed with a contour spacing
of 2 m and displayed with an appropriate scale in order to examine the lateral extension and
direction of the nonnal faults. The working scale is generally between 1/10,000 and
1150,000, depending on the slope of the nonnal faults and on the water depth, as the width of
the recorded band is about two thirds of the water depth. Examples are given in Figs. 9.9
to 9.14.
In the northern OT, swath bathymetric contours every 2 m are shown for a small
section of the profile 43 (Fig. 9.9) along both the seismic and 3.5-kHz sections of Figs. 9.5
and 9.7. The lateral extension of the nonnal faults is at least as wide as the width of the Sea
Beam band, which is 650 m. A nonnal fault ofthe second generation is oriented N052° and
presents a vertical offset of 38 m. Two nonnal faults of the third generation are oriented
N075° and present vertical offsets of 10 and 4 m. The difference in the orientation ofthese
two different families of faults is 20° to 25° in the northern OT. The detailed Sea Beam map
shown in Fig. 9.10 reveals seafloor topography along a small portion of the seismic profile
99 as shown in Fig. 9.6, located in the axial portion of the middle OT where present-day
sediments are defonned by the third generation of faults. The basaltic intrusion appears in
the northwest comer of the map (Fig. 9.10). Though the hummmocky character of the sea
bottom is confinned by the irregularity of bathymetric contours, the linearity of the nonnal
faults is still expressed over a minimum of I km across the profile. The three nonnal faults
and the southeastern limit of the basaltic intrusion belong to the last generation of faults.
Two faults with offsets of 20 and 22 m are oriented N065°, the other two, with larger offsets,
are oriented N045° (Fig. 9.10). The difference of 20° in the orientation of faults could be
explained by the reactivation of faults oriented N045° which belong to the previous
356 JEAN-CLAUDE SIBUET at al.
E127 43 E12743.5
N28 34. -f----fL--A-'i-<'zS4~-f-r__+---__+-N28 34.
N28 33. 5-+----+--...,..-'>'r-::~"t--'<~t_---t_N28 33.5
N28 33 .-+---+----'~-=--, - - - + - N 2 8 33.
FIGURE 9.9. Sea Beam bathymetric map of a

portion of profile 43 located in Figs. 9.2, 9.5,
and 9.7, plotted in Mercator projection and con-
toured every 2 m. Directions and vertical offset
values of the normal faults are also shown. The
azimuths of early Pleistocene and Recent nor-
N28 32.5-+-.-~-l-·r-r-.-.,...,-r"'FT~t-LT-,-. . .-,-t_N28 32.5 mal faults differ by 20° to 25° in the northern
Okinawa Trough.
generation. Due to the massive sediment supply from the China continental shelf, deposi-
tion rates up to 4 km/m.y. have been noticed for the last 0.5 m.y. (Hilde eta!., 1984; Kimura,
1985). Consequently, the subsidence at the axial portion of OT imaged on seismic profile 99
(Fig. 9.6) as well as on the swath bathymetric section (Fig. 9.10) is the expression of a
significant amount of extension. A large part of the present-day extension seems to be
absorbed by older faults.
Thus, the two families of faults may differ about 20° in azimuth locally. Small-offset
normal faults are generally associated with the present-day phase of extension, and large-
offset normal faults with the previous tectonic phase. Figures 9.11 and 9.12 (profiles 43 and
62 located in the middle OT) show that exceptions exist concerning the magnitude of
offsets. On profile 62 (Fig. 9.12), a fault with 8-m offset belongs to the second generation of
faults, while on profile 43 (Fig. 9.11), a fault with 30-m offset belongs to the third generation
of faults.
However, in the southern OT, the two families of faults may not be differentiated (as
shown in profiles 162 and 164, Figs. 9.13 and 9.14 respectively). All the hfaults, whatever is
the magnitude of their vertical offsets, present strike directions which differ only by 10° to
15°. Consequently, the difference in strike azimuth between the two families of faults
decreases from about 25° in the northern OT to about 20° in the middle OT and is not
discernible in the southern ~T.
N26 ,-+--------j--N:ZB '"
FIGURE 9.10. Sea Beam bathymetric map of a portion of profile 99 located in Figs. 9.2 and 9.6, plotted in
Mercator projection and contoured every 2 m. Directions and vertical offset values of the nonnal faults are also
shown. The azimuths of late Pleistocene and Recent nonnal faults differ by 20° in the middle Okinawa Trough.
The subsidence of the axial part of the trough is due to the present-day nonnal faulting.
These observations have several implications. As extensional faults are well expressed
in the sedimentary sections as a result of the high detritic input from mainland China, the
present-day tectonic phase of extension activates again the directions of the previous phase
of extension, as already known, but a new generation of faults also appears.
The new generation of faults progressively changes in azimuth along the OT, which
means that these faults probably record the changes in stress regime.
In the Izu-Bonin arc, extension occurs in an intraoceanic island arc. Two families of
normal faults, oriented 337° and 355°, are simultaneously active and have been identified in
the 50-km-wide Sumisu rift (Taylor et aI., 1991). This rift exhibits a zigzag pattern in plan
view, and directions of faults are the same than those of the rift borders. Taylor et ai. (1991)
have shown that faults with both trends form a geometry with an orthorhombic symmetry
which is consistent with theoretical models of fault formation in three-dimensional strain
fields (Reches, 1983). This model predicts that the maximum strain direction is perpendicu-
lar to the bisector of the acute angle between the two fault trends. For the Sumisu rift, the
extension direction is N076°. Though the difference in the directions of the two families of
faults for the OT and Sumisu rift are similar (about 20°), the tectonic contexts are different:
the two families are active simultaneously in an oceanic environment for the Sumisu rift
N27
N27
N27
N27
FIGURE 9.11. Sea Beam bathymetric map of a

N27
portion of profile 43 (middle Okinawa Trough)
located in Fig. 9.2. plotted in Mercator projec-
tion and contoured every 2 m. Directions and
vertical offset values of the normal faults are
given. Small-offset normal faults are generally
N27
associated with the present-day phase of exten-
sion. but exceptions exist. as the 30-m-offset
normal faults shows.
and belong to two consecutive tectonic phases in a thinned continental crust environment
for the OT. Following the observations of Sibuet et at. (1991) in the Bay of Biscay, another
explanation concerning the simultaneous presence of two families of faults in the Sumisu
rift would be also possible. There, the plate boundary, which was active during the
Pyrenean phase, follows previous lines of weakness as the former Bay of Biscay oceanic
rift system. When the motion along the plate boundary is tensional, rift directions of the
new oceanic crust are parallel to the plate boundaries and are not coincident with the stress
directions given by the accurately defined position of the related pole of rotation. In the
Sumisu rift, the location of the zigzag rift pattern could be also related to the presence of a
previous zone of weakness such as the intraoceanic island arc. Rift directions that develop
within the oceanic rift system would be parallel to the rift borders and independent of the
stress directions.
3.3. Quantification of the Spatial Distribution of the Two Families of Faults

Based on the previous assumptions, a systematic study of the directions of these two
families of faults has been performed. For each fault that presents a linear trend across the
surveyed bathymetric swath, the following parameters are collected: latitude, longitude,
vertical offset, azimuth, and dipping direction. As many as 510 measurements have been
performed from southern Japan to Taiwan (Fig. 9.15). Directions of faults change from
N090° east of Taiwan to N040° southwest of Kyushu (Fig. 9.15). The distribution of faults
E126 49. E126 49.5 E126 50. E126 50.5

N27 7. s-t-L..JLL.LL-L..JLL.y-..L1-L.L.L...L.l-'-4..L.1.--'---LL..L.l.--'---4-N27 7.5
~lI-'<i~rr-m-'----+------~:27 7.
-+----.::'4;.-7T;;;w.~:t:-'=~~"<I---------+-1N27 6.5
-+------t-:7"""-"7;:~~-";;:::+-'rlt:7lTr""-"----+-N27 6.
~"'--:I_A~HN27 5.5
Et26 49. E126 49.5 E126 50. E126 50.5
FIGURE 9.12. Sea Beam bathymetric map of a portion of profile 62 (middle Okinawa Trough) located in Fig.
9.2, plotted in Mercator projection and contoured every 2 m. Directions and vertical offset values of the normal
faults are given. Large-offset normal faults are generally associated with the early Pleistocene phase of extension,
but exceptions exist, as the 8-m-offset normal faults shows.
summed in 5° intervals of azimuths (Fig. 9.16) has been calculated for each of the four
geographical boxes displayed in Fig. 9.15. These diagrams show that peaks of dominant
fault directions change from 80-95° in box A, to 65-80° in box B, to 50-80° in box C,
and to 40-80° in box D.
Figure 9.17 displays the direction of the faults as a function of the longitude. The
continuous decrease of the azimuth as a function of the longitude is confirmed. The
dispersion of azimuths around the mean value is ±20° at longitude 123°E (close to Taiwan)
and ±30° at longitude 129°E (close to Japan). This observation confirms the existence
of two families of faults that diverge in the northward direction as previously suggested.
In Fig. 9.17, the directions of the OT axis and of the trench axis are also displayed as a
function of the longitude. All the directions of normal faults are located on the right-hand
side of the direction of the trough axis and also preferentially on the right-hand side of the
N25
N25 ~--------+----------·-+--------t--N25 15.
N25 -------'-------'---N25 15.5
N25 15.
N25
FIGURE 9.13. Sea Beam bathymetric map of a portion of profile 162 (southern Okinawa Trough) located in Fig.
faults are given. Normal faults belonging to the early Pleistocene and Recent phases of extension cannot be
differentiated in the southern Okinawa Trough.
direction of the trench axis. This means that both the large normal faults related to Lower
Pleistocene (as already noticed by Sibuet et al. 1987) and the late Pleistocene to Recent
small offset normal faults are systematically offset with respect to the directions of the OT
axis and continental slopes. The geodynamic context is different from the one of the Izu-
Bonin arc and of the western Bay of Biscay in the sense that several tensional phases occur
in the OT instead of a single tensional phase where rift features are parallel to the rift
borders. Because the two families of normal faults belonging to the second and third phases
of extension only differ by a maximum of 20° to 25° in direction, we have selected a few
normal faults that clearly belong to one of the two families by using the following criteria.
Seismic data are used to identify large tilted blocks formed during the second phase of
extension, bound by normal faults which present a significant vertical offset larger than a
few hundred meters, and which display a significant topographic effect larger than 30 m
(Fig. 9.18a).
Seismic data also unambiguously indicate Recent faults that affect the late Pleistocene
to Recent sedimentary layers. Among them, we have selected normal faults characterized
by small topographic vertical offsets and not related to possible deep normal faults
bounding large tilted blocks formed during the second phase of extension (Fig. 9.18b).
With these specific criteria, the two families of faults are clearly discriminated. The
mean direction of normal faults is estimated as a function of the longitude with an error of
±15° for faults of the second generation and ±100 for faults of the third generation (see mean
direction of faults in the southern and northern OT; Table I, and in the four geographical
-+...LJ-L-'-'-+LL..L.l---'-L..L...LJ+-'-'----'-LL..L.l---'-Lt...L.J---L..L.l+N25 16.
N2S 15.&--I--~ ~,.,......:AA7+ __~"t_-----+----rN25 15.5
~~"=*1Jf~~~~-+---+-N25 15.
N25 14.5+---+---==--~~~±l1l-\~~~?i-6T.:---t-1~25 14.5
FIGURE 9.14. Sea Beam bathymetric map of a portion of profile 164 (southern Okinawa Trough) located in Fig.
faults are given. Normal faults belonging to the early Pleistocene and Recent phases of extension cannot be
differentiated in the southern Okinawa Trough.
boxes; Figs. 9.15 and 9.16). The mean directions offaults related to tensional phases 2 and 3
and calculated from the selection of data shown in Fig. 9.18 correspond to the whole
spectrum of directions displayed for each of the four geographical boxes (Figs. 9.15 and
9.16), except for the secondary peak, which appears in Fig. 9.18a between 60° and 75°. In
fact, this peak corresponds to data collected west of 123°E (Fig. 9.15) on the northern OT
continental slope, which changes trend in its southwestern termination.
Consequently, as the change of direction of faults seems to be progressive, it confirms,
as previously seen, that the strain directions could correspond to the stress directions. In this
hypothesis the extensional stress regime in the backarc basin would have the minimum
principal stress <13 perpendicular to the mean directions of faults belonging to each of the
two families. However, for both tectonic phases, <13 is not parallel to the future breakup line
between the Okinawa platelet (defined by the Ryukyu arc and forearc) and the Eurasia plate
(i.e., the trends of continental margins of the OT). This means that <13 has changed through
time and that the initial tensional Miocene stress regime was different from the Lower
Pleistocene and Recent regimes. We propose to examine if the two families of normal faults
could be explained by single rotations.
121'E 122'E 123'E 124'E 12S'E 126'E 12TE 128'E 129'E 130'E 131'E 132'E
33'N 33'N
32'N 32'N
31'N 31'N
30'N EAST CHINA 3O'N
SEA c
29'N 29'N
28'N 28'N
2TN 2TN
2S'N 26'N
2S'N WEST PHILIPPINE 2S'N
BASI N
24'N 24'N
23'N -.:......=o_-===--.o=~-===>_-===-_=~-==_-===-_==-~=_-=~ 23'N

121'E 122'E 123'E 124'E 12S'E 12S'E 12TE 128'E 129'E 130'E 131'E 132'E
FIGURE 9,15, Directions of normal faults identified on large-scale Sea Beam maps contoured every 2 m, The
POPl data yielded 510 measurements (RN Jean Charcot cruise, September-October 1984) and show a progres-
sive change in direction from N090° east of Taiwan to N040° south of Japan,
4. KINEMATICS OF THE OKINAWA TROUGH OPENING
4.1. Determination of the Poles of Rotation Corresponding to the Early

Pleistocene and Recent Tectonic Phases
From previous discussion, the orientation of the normal faults probably represents the
stress regime for both the second and third tensional phases. Position of the poles of rotation
can thus be computed for these two tensional phases, For each family of faults, the position
of the pole of rotation has been computed by minimizing the error on a grid, The error (J' at
each point of the grid is equal to
I~=I (8 1 - 8 2)2
(J'=
N-l
where 81 is the observed azimuth of a measurement, 82 is the azimuth of the segment
connecting the computed point to the location of the measured point, and N is the total
"" Po
~
'"
~
fw
"" 30 120
30 150 130
30
!1
i~ '"
0>-
n
11
j 'iO
lao
'" '" 00 t<'l0 ISO
'" C flu:1'lbeT of fat~lts 1'7;)
*g
C '0
~
g"
0'
FIGURE 9.16. Distribution of fault directions by sum-

ming the faults within 50 intervals of azimuth for each of the
four geographical boxes displayed in Fig. 9.15. Azimuths of
normal faults belonging to the second and third phases of -31:'1 6iJ ~ 120 11';C
extension are from the mean curves of Figs. 9.18a and b. dir$C!!on {degfeB;
1~ t---~----~--~----~----~--~----~---T
Trench axis
120
"
"...'"" 90
8
::!" o 0
c
FIGURE 9.17. Directions of the normal 0
...
'M
60
'o8
faults as a function of the longitude (small o
open squares). Directions of the axis of the ""...
'M
'tl
'b
Okinawa Trough (black dots) and of the o

30
trench axis (black triangles) are also dis-
played as functions of the longitude. Direc-
o
tions of the normal faults are systematically
deviated from the direction of the Okinawa 122 123 124 125 126 127 128 129 130
Trough axis. longitude ('E)
150 150
A 120 B
I I
120
110 110
I" eo
30 Early
~ eo
i
" 30
Plelatocene
0 0
1:/2 123 124 125 121 127 121 1211 130 t~ 123 124 125 121 127 121 1211 130
Iongftude (. E) longitude ('E)
FIGURE 9.18. (a) Selected directions of normal faults belonging to the second phase of extension (early
Pleistocene, Table I) as a function of the longitude (small open squares) and mean value (continuous line). Black
dots represent the direction of the Okinawa Trough axis. (b) Selected directions of normal faults belonging to the
third phase of extension (late Pleistocene to Recent, Table I) as a function of the longitude (small open squares)
and mean value (continuous line). Black dots represent the direction of the Okinawa Trough axis.
number of measurements. The error has been computed every 0.2°, and the error values
have been contoured and plotted with the GMT software package (Wessel and Smith, 1991).
Poles of rotation corresponding to the second and the third generations of faults have been
derived from 56 and 63 data points, respectively. The location of the pole of the second
generation of faults (24.8°N, 121.6°E) lies in the eastern portion of the Ban Plain (Fig.
9.19a). Consequently, extension in the southern OT and Ban Plain would be weaker
compared to the northern ~T. This explains why large tilted fault blocks with significant
offsets are observed in the northern OT and not in the southern OT. The location of the pole
for the third generation offaults (24.6°N, 113.6°E) is about 800 km west of the previous one
(Fig. 9.19b). A kinematic consequence is that the extension in the Ban Plain was weak
during the early Pleistocene but more active during the Recent phase. In fact, normal faults
observed in the Ban Plain affect the Plio-Pleistocene erosional surface. However, the Tatun
and Kilung volcano groups, which are located in northern Taiwan behind the backarc basin
but are related to the Ryukyu subduction system, are mostly of Quaternary age (Ho, 1986).
High-heat-flow values (Yamano et aI., 1989) and hydrothermal manifestations (Yamano
et aI., 1986) show that N075°-oriented volcanic ridges of the VAMP area located in the
middle OT (Davagnier et aI., 1987; Sibuet et aI., 1987) are present-day to Recent features
emplaced along trends that are in the direction of the pole of rotation of the third phase. This
observation confirms the validity of the computed pole position and that present-day
volcanism is emplaced along the rift direction of the third phase.
4.2. Determination of the Total Amount of Extension in the Okinawa Trough

It has been shown that extension occurs in the continental domain along the whole OT,
from Taiwan to Japan. The bathymetric map (Fig. 9.2) shows that the OT is wider on the
Japanese side, but the maximum water depth decreases toward the northeast; thus we do not
know if the amount of total extension increases toward Japan. From available refraction and
gravity data, we have established two crustal sections for the northern and southern OT,
respectively, in order to quantify the amount of extension along these two transects.
Two techniques could be used to define the geometry of the Moho if part of the Moho
depth is known: (1) By assuming that the crust is in local isostatic equilibrium, the geometry
118'£ 119'£ 120'12 121"£ 122"E

28'111
2TN 27'111
26"111 26'111
25"111 25'111
24'N 24'N
23"111 23'111
22"111 22"111
21"N 21'111
110"" 111"E 112'E 113'!: 114"E 115"E 116"£ l1TE lIS"!: 119"E 120"E 121"E 122'E
113'£ 114'" 115"E 116"E lin: lHrE 119'" 120"E 121"£ 122'E
2S"1II
27"111 2TN
26'111 26"111
25"1'1 25"111
N"N 24'N
23"111 23"111
22"111 22"111
21'111
110'" il1'E 112"E 113"E 114"E 11S"E 116"E liTE 118"£ 119"E 120'E 121'" 122'E
FIGURE 9,19, (A) Location of the pole of rotation (star) computed by minimization of the error on a 0"2° grid
(24,8°N, 121.6°E), Data from 56 normal faults that unambiguously belong to the early Pleistocene phase of
extension have been used in this calculation, Error values have been contoured and plotted by using the GMT
software package (Wessel and Smith, 1991), (B) Location of the pole of rotation (star) computed by minimization
of the error on a 0"2° grid (24,6°N; 1I3,6°E). Data from 63 normal faults which unambiguously belong to the late
Pleistocene to Recent phase of extension have been used in this calculation. Error values have been contoured and
plotted by using the GMT software package (Wessel and Smith, 1991).
of the Moho can be deduced by using an appropriate relationship between refraction

velocities and densities. (2) If the gravity along the profile is known, the geometry of the
Moho can be laterally extended by fitting the results of a two-dimensional gravity model
with observed seismic refraction data.
We have used the first technique to obtain a model in local isostatic equilibrium, which
was in agreement with the limited available refraction constraints. Then the calculated
gravity values from this geometrical model have been compared with the observed gravity
in order to provide geometrical models across the entire OT reasonably in agreement with
both refraction and gravity data.
The seismic refraction experiment of Hirata et al. (1991) was conducted in the southern
OT along profile I (Fig. 9.2). The Moho has been identified in the axial part of the trough at
a depth of 18 kIn, with no evidence of existing oceanic crust beneath the central axis of the
trough. However, in their ray-tracing modeling, Hirata et al. (1991) assumed that both the
Moho and the boundary between the upper and lower crusts were flat outside of the trough
axis. This assumption is not valid in the context of a narrow basin because it strongly
violates the isostatic equilibrium principle. Consequently, we can only retain with caution a
depth of 18 kIn for the Moho at the axis of the trough. Lee et al. (1980) reported refraction
results along a profile parallel to the strike of the trough intersecting profile I. The P-wave
velocity distributions in the crust are similar for both studies, except Lee et al. (1980) found
that the Moho is at a depth of 15 kIn. Figure 9.20 presents models of profile I in local
isostatic equilibrium with Moho depths beneath the axis of the southern OT at 15 and 18 kIn.
Densities are obtained by using the theoretical relationship of Warner (1987) for sediments
_ S
1~r-------~------~------~--~
PROFILE I N WNW PROFILE II ESE
I~~""':"'1
i
f~
100
"I
>-
0 ::_:::_.;:.~-::-.•.- - - - - -______1
1--______-..•-...-:::__
1J5 to3 14
50 100 150 50 100 150 200 250

distance (km) distance (km)
FIGURE 9.20. Crustal models in local isostatic equilibrium established from available refraction data along
profiles I and II located in the southern and northern Okinawa Trough (Fig. 9.2). Densities are in g/cm 3. The
shallowest point of the Moho in profile I is at 18 Ian (continuous line) or at 15 Ian (continuous line with dots),
depending on the interpretation of refraction data (Lee et aI., 1980; Hirata et aI., 1991). Hatched area corresponds to
thinned continental crust. Observed gravity data (continuous line) and calculated gravitational attractions due to
the geometrical structure (crosses for the Moho at 18 Ian and dots for the Moho at 15 Ian) are also shown. Same
legends for profile II except that only one Moho position is provided.
and the velocity density relationship for crustal materials of Sibuet et al. (1990). The high
value of the mantle density (3.4 g/cm3 ) has been chosen because continental rifting is still
active and the whole lithosphere has not yet been heated. The geometrical model was
established by using a reference hydrostatic column located in the axial portion of the
trough. The deduced initial crustal thicknesses are 30.8 and 27.8 km, respectively, for axial
Moho depths at 18 and 15 km. The observed gravity data are extracted from a free-air
gravity anomaly chart of the Okinawa Trough and vicinity (Oshima et aI., 1988). The fit
between the calculated and the observed gravity data is reasonably good except on both
edges of the OT. Shallow Moho gives a better fit.
In the northern OT the refraction experiments of Iwasaki et al. (1990) in the western
and central parts and Hayes et al. (1978) in the eastern part of the trough have been
projected onto profile II (Fig. 9.2). The Moho has been identified in the western and central
parts of the trough but not in the eastern part, where refraction data only provide informa-
tion on the thickness of sediments and on the depth of the continental crust. Gravity data
and density values are derived by using the same approaches as for profile I. The geometri-
cal model is established by using a reference hydrostatic column located at 160 km where a
Moho depth estimation is available. The deduced initial crustal thickness is 31 km, a value
very close to the one obtained for profile I. The calculated gravitational attraction of this
model in local isostatic equilibrium fits fairly well the general trend of the observed gravity
data (Fig. 9.20).
If we assume that the extension is accommodated by stretching (pure shear) in the
whole lithosphere, then knowing the initial thickness of the crust, the total amount of
extension which affected the nonthinned continental lithosphere can be determined. It is the
only way to get a rough estimate of the amount of extension in the ~T. Note that this
estimate is not significantly affected by the presence of dike intrusions in the ~T. If the
shallowest Moho is 15 km and the initial crustal thickness is 27.8 km, the amount of
extension along profile I is 99 km-that is, 97 km perpendicularly to the axial direction of
the trough. With a Moho at 18 km and an initial crustal thickness of 30.8 km, the amount
of extension is 82 km-that is, 80 km perpendicularly to the direction of the trough at the
location of profile I. For profile II, the total amount of extension is 74 km for an initial
crustal thickness of 31 km.
For coherence of Moho depths along the entire OT, an amount of 80 km of extension
has been assumed for the southern OT. This value is large compared to the width of the
southern OT (100 km between the two continental slopes). However, the cross section of
profile I (Fig. 9.20) shows that the Moho topography extends laterally, outside of the
topographic expression of the OT located between 50 and 150 km. The existence of a thick
pile of sediments above the continental crust is the reason for the slow decay of Moho depth
to 31 km below the Eurasian platform. Consequently, part of the 80 km of extension is due
to the thinning of the continental crust outside of the trough sensu stricto. The amount of
extension is quite constant along the OT (Fig. 9.20), but the initial width of continental crust
extended during the process of creation of the backarc basin is 100 km in the southeastern
OT (along profile I) and 200 km in the northern OT (along profile II). This explains why the
Moho depth increases from southern to northern ~T.
Though a crude hypothesis has been made on the tensional mechanism to estimate the
total amount of extension, an important indication is that the amount of extension across the
OT does not increase from Taiwan toward Japan, as has been commonly implied in most of
the geodynamic sketches except that of Vander Zouwen (1984), who proposed a pole of
rotation located in southern Japan (Kyushu Island). The depth to Moho map of Jin et al.
(1983) grossly shows the northeastward increasing of the Moho depth and the increasing
width of the trough. The Moho is at about 30-km depth beneath the East China Sea
continental shelf. If we assume that the crustal thickness before rifting was 31 km, the total
amount of extension along the OT is roughly constant or, to be more precise, slightly
decreasing from 80 km at the location of profile I to 74 km at the location of profile II (Fig.
9.2). This means that the pole of rotation is located at about 90° from the OT location. From
a kinematic point of view, if the poles of rotation of the second and the third phases of
extension are both located west of Taiwan, the pole of rotation of the first phase of
extension, which occurred in middle Miocene, was located northeast of the OT but far from
the position given by Vander Zouwen (1984).
4.3. Determination of Parameters of the Total Rotation

Knowing the amount of extension perpendicular to the OT in two different places, we
can obtain parameters of rotation only if a third constraint is available. This constraint
comes from the shape of the continental slopes that bound the trough. At the base of slopes,
between 1000 and lS00 m, a large bathymetric gradient could be followed on both edges of
the OT (bathymetric chart of the middle and northern OT (Oshima et aI., 1988) and
bathymetric chart ofthe southern OT presented in this study). These two boundaries could
be matched remarkably well because of their arcuate shapes and the presence of conjugate
irregularities as, for example, at 26.2°N, 124.9°E and at 2S.soN, 12S.3°E. The amount of
extension has been maintained perpendicularly to the trough at 74 and 80 km in the northern
and southern OT, respectively. In such conditions, the pole of rotation has been determined
at 3SoS, SOON with a rotation angle of 0.77° (Table II). Figure 9.21 shows such a reconstruc-
tion with present-day contours and the rotated position ofthe Okinawa platelet. The rotated
1000-m isobath located north of the Ryukyu Islands fits very well with the conjugate isobath
along the northern edge of the trough, except for the southwestern edge of the OT (west of
123.S0E), which has no bathymetric counterpart on the northwestern edge, and possibly for
the southern segment located between 123.soE and 124.soE, which includes the two main
islands of lriomote Sima (24.3°N, 123.8°E) and Isigaki Sima (24.4°N, 124.2°E).
Consequently, east of 123.soE or possibly east of 124.soE, the curvature of the entire
trough was not acquired during the formation of the backarc basin but was already in
existence at the beginning of the formation of the trough and persisted through time. The
main curvature of the trough is already given by the northeastern OT slope, which has not
been significantly affected during the collision of Taiwan. The relative uniformity of the
amount of extension along the trough (Fig. 9.21) also suggests a decoupling between the OT
opening and the compressive tectonics in Taiwan, at least between profiles I and II located
TABLE /I
Okinawa Platelet Motions with Respect to Eurasia
Latitude Longitude
OT main tensional phases (ON) (DE)
1st phase Middle to late Miocene 37.6 144.0 -1.93

2nd phase Early Pleistocene 24.8 121.6 1.58
3rd phase Late Pleistocene to Recent 24.6 113.6 0.19
Total extension -35.0 50.0 0.77
12"F. 122'E 124'E 125'E 12S'E 127'E 121fE 129'E 130'E 131'E
SO'N
29'N EURASIA 29'N
(fixed)
28'N 28'N
27'/>1 27'N
26'N 2I!'N
25'N 25'1'4
24'N PHILIPPINE 24'jIj
SEA
23'jIj
121'E
-==_-==
122'1:
.....-==_-"?::=.:::
123'E 124'E
....."""'=.....'""""'==-""""="""""""'==-""""==-.....==-,.,,-'!
125'E 12S'E 128'10 129'E IS0'E 131'E
23'jIj
FIGURE 9,21. Reconstructions of the Okinawa Trough opening. Present-day positions are in continuous lines.
The Ryukyu Islands and their 2()()-m isobaths represent the Okinawa platelet, which extends as far as the trench
axis (Fig. 9.4). The continuous line with open triangles shows the position of the trench, Eurasia is fixed. The
dotted lines represent the present-day position of the Okinawa platelet at the closure of the Okinawa backarc basin
before any tensional motion in the area occurred, i.e., in middle Miocene time. The parallelism of the lOOO-m
isobath of the Eurasian continental platform with the rotated lOOO-m of the northwestern border of the Okinawa
platelet shows the quality of the fit.
in Fig, 9,2, Another consequence of this reconstruction is that the mean azimuth of the
relative motion of the Okinawa platelet with respect to Eurasia is NI43°, which is quite
similar to the mean direction of motion (NI26°) of the Philippine Sea plate with respect to
Eurasia (Seno and Maruyama, 1984), The decoupling of the tectonic evolutions of the
backarc basin and of the collision in Taiwan is maintained through time, except for the
southwestern portion of the trough, west of 123,5°E or 124,5°E, where the Okinawa platelet
and the Ryukyu Trench change significantly in direction with respect to the rest of the
OT, Numerous earthquakes occur over a maximum distance of 140 km from Taiwan in the
EW forearc Nanao basin (Tsai, 1986), Focal mechanisms indicate a compressive motion
which is related to the collisional processes in Taiwan (Kao and Chen, 1991; Cheng et ai"
1992) and could explain the ESE-WNW orientation of the southwestern extremity of the
Ryukyu forearc and trench and the narrowing of the backarc basin itself, though abnormally
high tensional motions persist in the southern OT east of the Han Plain over a distance of
170 km as confirmed by the high earthquake density, focal mechanisms (Kao and Chen,
1991; Cheng et ai" 1992), and seismic reflection profiles which display present-day normal
faults parallel to the OT axis (Sibuet, 1991),
4.4. Trial to Quantify the Amount of Extension Corresponding to the Two Last
Tensional Phases
To better define the last two phases of extension, we attempt to evaluate the horizontal
offsets of faults related to each of the two phases across two transects of the OT. The
amount of extension has been determined by assuming that the brittle surface deformation
was representative of the crustal deformation following the method of Le Pichon and Sibuet
(1981). Because pure shear is the assumed mechanism of extension, the surface extension is
identical to crustal or lithospheric extensions. Though this method has been criticized in the
past (Chenet et aI., 1983), in our opinion it is the only way to try to quantify the amount of
extension. Horizontal offsets have been determined from the dip and the vertical offset of
each fault and summed for all faults across the entire trough. Numerous seismic profiles
of the POPI cruise have been used for that purpose. Final estimates of the amount of
extension perpendicular to the trough direction are given for two complete transects of the
trough.
The Recent phase of extension is characterized by numerous small offset normal faults
and also by the reactivation of normal faults that belong to the preceding phase of
extension. Because the amount of extension cannot be estimated from faults that present a
large topographic expression, but only from those faults which offset the sedimentary
cover, this bias is eliminated by choosing only seismic sections where normal faults of the
previous phase are buried, which is mainly in the southern and middle OT. The estimates of
the cumulative extension amount on several OT cross sections are about 3 km in the
southern OT and 5 km in the middle ~T. Though there may be large errors, which could
be a factor of 2 on these estimates, the final accepted estimate is 5 km in the middle OT (at
29°N, l28.5°E) which gives a 0.2° rotation angle for the last tensional phase (Table 11).
Concerning the amount of extension that occurred in the second phase of extension
(early Pleistocene), the only locations where such estimations can be done are in the middle
or northern OT (example on profile 99, Fig. 9.6). Because of the high sedimentation rate,
the subsequent thick sediments mask the tilted blocks relative to this phase in the southern
OT. For badly defined blocks, the amount of extension has been estimated by using the
downward prolongation of both the Plio-Pleistocene surface of the blocks and the normal
faults that bound the blocks. Estimates performed in the middle and northern OT are similar
and equal to 25 ± 10 km. Though the amount of error cannot be accurately defined, we retain
25 km of extension at 29.7°N, 129°E, which gives a 1.4° rotation angle for the second
tensional phase (Table II).
4.5. Parameters of Rotations for the Three Tensional Phases

Parameters of the first phase of extension can be deduced from the total rotation and
the two last rotations (Table II). Figure 9.22 shows the relative position of the Okinawa
platelet (represented by the Ryukyu Islands and their 200-m isobaths) before extension (at
the time of closure) and at the end of each of the first, second, and third (present-day
position) phases of extension. As anticipated, the largest amount of motion corresponds to
the first phase, which in fact is the least-documented phase from geological data. The
corresponding pole of rotation is located in the northeastern prolongation of the OT, at
about 1500 km beyond the northern extremity of the OT, which gives about 50 km of
extension in the northern OT and 75 km of extension in the southern OT. From paleo-
121"10 123'!ii 124'''' 12TE 129'1::
EURASIA
(flxoo)
ItS'}!
2TN
24'N PHiLIPPINE
SEA
2S·tj
121'1: 122'E 12:1'\:
FIGURE 9.22. Reconstructions of the three extensional phases of the Okinawa Trough. Present-day positions
are in continuous lines. The Ryukyu Islands and their 200-m isobaths represent the Okinawa platelet. The
continuous line with open triangles shows the present-day position of the trench. Eurasia is fixed. The noncon-
tinuous lines represent the position of the Okinawa platelet at the closure of the Okinawa backarc basin (about
15 Ma, dotted lines), at the end of the first tensional phase (about 7 Ma, dashed lines) and at the end of the second
tensional phase (about 0.5 Ma, dotted and dashed lines).
magnetic constraints coming from the southeastern Ryukyu Islands, Miki et al. (1990)
explain the motion of these islands (at least south of the Miyako depression) with respect to
Eurasia by a clockwise rotation of 19° around a pole located in northern Taiwan (24.7°N,
121.7°E) for the portion of the Okinawa platelet located south of the Miyako depression.
The amount of extension in the OT north of Isigaki Sima Island (Fig. 9.2, 24.4 oN, 124.2°E)
is 70 km, a value close to the one proposed in this study. However, using these parameters
of rotation, the amount of extension would be 150 km near the Miyako depression and 340
km in the northern OT just south of Kyushu. Both values are unacceptable, which questions
the validity of the interpretation of these paleomagnetic measurements. The most coherent
interpretation of the 19° clockwise rotation of the southwestern portion of the Okinawa
platelet would be to link this motion to the collision of the Luzon arc in Taiwan. Paleo-
magnetic measurements were performed on samples older than 10 Ma (Miki et aI., 1990),
which do not give any information on the age of the rotation. We suggest that this rotation of
the southwestern portion of the Okinawa platelet occurred during the collision in Taiwan
(i.e., during the last 4 m.y.) and that the pole of rotation was not located in northern Taiwan
but close to or east of Iriomote Sima and Isigaki Sima Islands. This interpretation links the
curvature of the southwestern Okinawa platelet west of 123.5°E or 124.5°E to the collision
in Taiwan. The pole position of the OT total closure (Table II) also differs considerably
from that proposed by Vander Zouwen (1984), which is located in Kyushu. That pole of
rotation allows only a few kilometers of extension in the northern OT instead of several tens
of kilometers as established in this study. To test the proposed kinematic evolution of the
OT backarc basin, we examine newly compiled bathymetric and magnetic anomaly maps
of the southern OT (Fig. 9.2).
5. NEWLY COMPILED BATHYMETRIC AND MAGNETIC DATA IN THE

SOUTHERN OKINAWA TROUGH
Bathymetric and magnetic data from the National Geographic Data Center (NGDC)
and new data acquired east and south of Taiwan by the Institute of Oceanography, National
Taiwan University, have been compiled in the area east of Taiwan, in the OT, and in the
Philippine Sea Basin.
5.1. New Bathymetric Map

Bathymetric data have been acquired in 95 different cruises (Fig. 9.23). The root-
mean-squares (r.m.s.) error at the intersections of all tracklines is 119.3 m and the mean
absolute value is 55.1 m. To avoid the appearance of pseudostructures, several cruises
FIGURE 9.23. Tracklines show where bathymetric data have been acquired (95 cruises). Data have been
collected by numerous institutions and are available through the National Geographic Data Center. Complemen-
tary data acquired east and south of Taiwan were made available by the Institute of Oceanography, National
Taiwan University.
which give crossover errors larger than 200 m were removed first. The rest of the
bathymetric data were adjusted by linear interpolation (Hsu, in press) to eliminate the
crossover errors. Then data were gridded and contoured using the GMT software (Wessel
and Smith, 1991). Figure 9.24 shows a portion of the bathymetric map over the southern OT,
the direction of normal faults identified from Sea Beam data, together with offset and dip
indications. Two major depressions located at about 25.2°N, 124°E and 25.8°N, 125.5°E
in the southern OT are underlined by several segments of faults.
5.2. New Magnetic Map

Magnetic data have been acquired in 57 different cruises (Fig. 9.25). Magnetic
anomalies are computed using the International Geomagnetic Reference Field (lGRF)
1990. The r.m.s. error at the intersections of all tracklines is 55.9 nT, and the mean absolute
value is 38.8 nT. To avoid the appearance of pseudomagnetic features, we removed several
cruises that give crossover errors larger than 100 nT and applied a regional correction to
magnetic data of those cruises that do not have diurnal informations, using the general
method of Regan and Rodriguez (1981). The remaining data have been adjusted by linear
interpolation (Hsu, in press) to eliminate crossover errors. Then data were gridded and
contoured with the GMT software (Wessel and Smith, 1991). A regional magnetic anomaly
was extracted from the magnetic anomaly grid by second-order filtering and then was
substracted from the initial grid to give the residual magnetic anomaly map (Fig. 9.26).
FIGURE 9.24. Bathymetric map of the southern Okinawa Trough in Mercator projection. Contours every 500 m.
Gray scale on the right. Measurements of normal fault directions identified on Sea Beam data appear with
indications of dips. Scale of vertical offsets in the upper left corner of the map. The N142° mean direction for the
opening of the Okinawa Trough (Table II) shows the good correlation between topographic features located on
either side of the trough, except for the southeastern portion of the Okinawa platelet, west of l23.5°E or l24.5°E,
which has no morphologic counterpart on the northern continental slope of the Okinawa Trough.
FIGURE 9.25. Tracklines show where magnetic data have been acquired (57 cruises). Data have been collected
by numerous institutions and are available through the National Geographic Data Center. Complementary data
acquired east and south of Taiwan were made available by the Institute of Oceanography, National Taiwan
University.
5.3. DISCUSSION
A major feature of the residual magnetic anomaly map is the strong magnetic contrast
between the OT backarc basin and the Okinawa platelet. The southern OT, defined by the
200-m isobath, is characterized by magnetic anomalies with magnitude of a few hundred
nanoteslas and wavelengths of a few tens of kilometers. In contrast, the Okinawa platelet,
defined by the 200-m isobath on the OT side and the trench axis, is characterized by weak
magnetic anomalies, which confirm that the southern Ryukyu arc is a nonvoJcanic arc
,.... -.....
.......
~
..,..
FIGURE 9.26. Residual magnetic anomaly map of the southern Okinawa Trough in Mercator projection
obtained by subtraction of a regional field from the magnetic anomaly data. Contours every 20 nT. The N142°
direction shows boundaries between areas of similar magnetic character.
(Letouzey and Kimura, 1986). However, except for the high-amplitude anomalies located
northeast of Taiwan, which are linked to volcanics of the Taiwan Sinzi folded belt, the East
China Sea continental margin is characterized by magnetic anomalies of similar wave-
length but lower amplitude than in the ~T. Two lines of interpretations can be proposed
for the origin of the OT magnetic anomalies.
The initial continental domain, which was lately thinned and extended 80 krn, belongs
to the domain of the Taiwan Sinzi folded belt, where volcanics have been identified. As the
initial continental domain has been extended by a factor of about 2 during rifting, an
increase of the wavelength of magnetic anomalies and a reduction of the amplitude due to
the subsidence of magnetic sources were anticipated. However, these were not observed.
As a result of tensional processes, the whole OT continental domain has been thinned
and crustal and lithospheric thicknesses significantly reduced (Fig. 9.20). Products of
partial melting initiated in the lower part of the lithosphere, just above the Benioff zone,
probably come up through the lithosphere, lying as arc volcanic intrusions at different
depths within the crust, and giving rise to significant magnetic anomalies. We tend to favor
this hypothesis.
In an earlier magnetic compilation, Oshida et al. (1992) described relationships
between magnetic anomalies and geologic features. They demonstrated that the numerous
dipole-type anomalies observed along the southern side of the trough are linked to volcanic
knolls that are considered as the southwestern extension of the Ryukyu volcanic arc (Ueda,
1986; Oshida et aI., 1988). Geochemical analyses performed on basaltic samples collected
on these seamounts confirm the arc volcanic affinity of these seamounts (Sibuet et al., 1986;
Boespflug, 1990).
Because of the reduced amount of partial melting, only a small percentage of the crust
would be composed of volcanics. It seems reasonable to assume that refraction data cannot
resolve the presence of such crustal material. A major question concerns the mode of ascent
of magmatism through the crust. Does the basaltic magma go indifferently or randomly
through the crust? Or does it go preferentially along zones of weakness parallel to rift
directions or transform directions? In addition to localized dipole anomalies linked to arc
volcanism (Oshida et aI., 1992) along the southern side of the trough, a close examination
of the shape of magnetic anomalies in the OT shows a good correlation between the
location of the elongated magnetic anomalies and the axes of the two depressions already
evidenced on the bathymetric map and where basaltic elongated ridges have been mapped
and sampled (Sibuet et aI., 1987; Fig. 9.26). Oshida et al. (1992) modeled these magnetic
anomalies with a normally magnetized crust (2.5 Nm), supporting the view that seafloor
spreading has not started in the southern OT. In the southern OT only a few basaltic
intrusions have been identified in these depressions (Fig. 9.3) where surface extension and
consequent crustal extension are maximum. Thus, the residual magnetic anomaly map
clearly shows a correspondence between magnetic anomalies and rift directions related to
the last two phases of the OT opening. No correspondence seems to exist with rift directions
of the first phase, oriented N052° from the kinematic analysis. Previous existing weak
magnetic anomalies could have been obliterated by large anomalies related to the second
and third phases of extension and, in any case, better developed because extension and
partial melting increase with time.
The mean direction of the entire OT opening has been plotted on the bathymetric (Fig.
9.24) and residual magnetic anomaly (Fig. 9.26) maps. Though transform faults are not
observed in the OT (Sibuet et aI., 1987), the N142° segments limit areas of similar magnetic
character. A plausible interpretation of this observation is to relate magnetic anomalies and
associated crustal magmatic intrusions with different sections of the continental crust
experiencing different tensional factors. Extension could be different in adjacent sections.
For example, OT continental slopes are very steep on both sides of the section located
between l23.8°E and l25.4°E (Fig. 9.26). Most of the extension occurs in the deepest part
of the OT. In the adjacent southwestern section and extending to Taiwan, continental slopes
are wider and the amount of extension in the deep part of the trough seems to be reduced
compared with other sections. We consequently suggest that a link exists among partial
melting, ascent of volcanic products within the crust, and amount of crustal extension.
6. CONCLUSIONS
The major conclusions concerning the structural and kinematic evolution of the OT
are the following.
(1) The whole OT is still in a rifting stage as shown by refraction and magnetic data.
The crustal thickness increases from 10 km in the deepest part of the southern OT to 30 km
in the northern OT, close to Japan.
(2) The amount of extension across the OT has been estimated from refraction and
gravity data. It slightly decreases from 80 km in the southern OT to 74 km in the north-
ernOT.
(3) Three phases of extension are identified and their parameters of rotation are
established: (a) The late Pleistocene to Recent phase of extension is characterized by
normal faults with vertical offsets of a few meters and changing directions progressively
along the OT. The amount of extension that occurred during this phase is about 5 km in the
middle OT. (b) The early Pleistocene phase of extension is characterized by tilted blocks
that affect late Pliocene-early Pleistocene sediments and change direction progressively
along the ~T. The difference in azimuths of these two last phases increases toward the
northern OT. The amount of extension for this phase is estimated to be 25 km in the
northern OT. (c) The middle to late Miocene phase of extension, poorly characterized from
geological data, has been calculated from parameters of the total rotation and of the two last
rotations. This first phase of extension is a major tectonic phase with about 50 km of
extension in the northern OT and 75 km in the southern OT. The corresponding pole of
rotation is located 1500 km northeast of Kyushu Island.
(4) Directions of normal faults of the last two phases correspond to the stress
directions because of their progressive change along the OT axis resulting from a partial
relaxation of the motion along zones of weakness active during previous tectonic phases.
(5) The proposed kinematic evolution of the OT differs considerably from previous
reconstructions and raises questions concerning, for example, the interpretation of paleo-
magnetic measurements obtained in the southeastern Ryukyu Islands.
(6) The curved shape of the OT was acquired since the onset of the backarc basin
extension except for the southwestern portion of the Okinawa platelet, west of l23.5°E or
l24.5°E longitude, which has rotated 19° clockwise (Miki et aI., 1990) during the collision
of Taiwan around a pole located east or close to the southern Ryukyu Islands.
(7) New bathymetric and residual magnetic anomaly maps are in agreement with the
proposed kinematic evolution. Magnetic anomalies in the southern OT and geochemical
analyses performed on collected basaltic samples suggest that part of the anomalies are
linked to crustal arc volcanic intrusions emplaced close to the southern continental slope
of the OT and within topographic depressions located in the central part of the southern OT.
For the remaining magnetic anomalies, we suggest a link among partial melting, ascent of
volcanic products within the crust, and amount of crustal extension.
Acknowledgments
We thank Benoit Loubrieu, Alain Normand, and Michel Voisset for their help in
processing Sea Beam data with the IFREMER TRISMUS software package, and Daniel
Carre for drafting some figures. We acknowledge discussions with H. D. Needham about
the kinematics of the early stage of ocean opening. The GMT software package was used to
display some of the figures (Wessel and Smith, 1991). Brian Taylor gave us a lot of support
and facilities for the submission and revision of this chapter and also numerous scientific
comments on the first version of the manuscript.
REFERENCES
Boespflug, X. 1990. Evolution geodynamique et geochimique des bassins arriere-arcs. Exemples des bassins
d'Okinawa, de Lau et Nord-Fidjien, Universite de Bretagne Occidentale, Brest, France.
Chenet, P.- Y., Montadert, L., Gairaud, H., and Roberts, D. G. 1983. Extension ratio measurements on the Galicia,
Portugal and northern Biscay continental margins: Implications for evolutionary models of passive continen-
tal margins, in Studies in Continental Margin Geology (J. S. Watkins and C. L. Drake, eds.), Am. Assoc.
Petrol. Geol. Mem. 34:703-715.
Cheng, S.-N., Lee, C.-T., and Yeh, Y. T. 1992. Seismotectonics of the Ryukyu arc, in Proc. 4th Taiwan Symp.
Geophys., Taipei, pp. 507-516.
Davagnier, M., Marsset, B., Sibuet, J.-c., Letouzey, J., and Foucher, J.-P. 1987. Mecanismes actuels d'extension
dans Ie bassin d'Okinawa, Bull. Soc. Geol. France 8:525-531.
Eguchi, T., and Uyeda, S. 1983. Seismotectonics of the Okinawa Trough and Ryukyu arc, Mem. Geol. Soc. China
5:189-210.
Furukawa, M., Kondo, S., Miki, M., and Isezaki, N. 1991a. Report on DELP 1988 cruises in the Okinawa Trough.
Part V: Measurement of the three components and total intensity of the geomagnetic field in the Okinawa
Trough, Bull. Earthquake Res. Inst., U. Tokyo 66:91-150.
Furukawa, M., Tokuyama, H., Abe, S., Nishizawa, A., and Kinoshita, H. 1991b. Report on DELP 1988 cruises in
the Okinawa Trough. Part II: Seismic reflection studies in the southwestern part of the Okinawa Trough, Bull.
Earthquake Res. Inst., U. Tokyo 66:17-36.
Hayes, D. E., Houtz, R. E., Jarrard, R. D., Mrozowski, C. L., and Watanabe, T. 1978. Crustal structure, 1 sheet,
scale 1:6,442,194, in A Geophysical Atlas of the East and Southeast Asian Seas, Map Chart, MC25 (D.E.
Hayes, ed.), Geological Society of America, Boulder, CO.
Hilde, T. W. C., Lee, C.-S., and Vander Zouwen, D. E. 1984. Tectonic and sedimentation history of Okinawa
Trough: Implications for development of the East China and Yellow seas. Korea-U.S. Conference on the
Yellow Sea.
Hirata, N., Kinoshita, H., Katao, H., Baba, H., Kaiho, Y., Koresawa, S., Ono, Y., and Hayashi, K 1991. Report on
DELP 1988 cruises in the Okinawa Trough. Part III: Crustal structure of the southern Okinawa Trough, Bull.
Ho, C. S. 1986. A synthesis of the geologic evolution of Taiwan, Tectonophysics 125:1-16.
Hsu, S.-K in press. XCORR: A cross-over technique to adjust track data, Comput. Geosci.
Ishizaka, K, Yanagi, K, and Hayatsu, K 1977. A strontium study of the volcanic rocks of the Myoko volcano
group, central Japan, Contrib. Mineral. Petrol. 63:295-307.
Iwasaki, T., Hirata, N., Kanazawa, T., Melles, 1., Suyehiro, K, Urabe, T., Moller, L., Makris, J., and Shimamura,
H. 1990. Crustal and upper mantle structure in the Ryukyu island arc deduced from deep seismic sounding,
Geophys. J. Int. 102:631-651.
Iarrard, R D., and Sasajima, S. 1980. Paleomagnetic synthesis for Southeast Asia: constraints on plate motions, in
The Tectonic and Geologic Evolution of Southeast Asian Seas and Islands (D. E. Hayes, ed.), Geophys.
Monogr. Ser., Vol. 23, pp. 293-317, American Geophysical Union, Washington, DC.
lin, X., Yu, P., Lin, M., Li, c., and Wang, H. 1983. Preliminary study on the characteristics of crustal structure in
the Okinawa Trough (in Chinese with English abstract), Oceanol. Limnol. Sin. 14:105-116.
Kao, H., and Chen, W.-P. 1991. Earthquakes along the Ryukyu-Kyushu arc: Strain segmentation, lateral compres-
sion, and thermomechanical state of the plate interface, J. Geophys. Res. 96:21,443-21,485.
Kasahara, I., Nagumo, S., Koresawa, S., Ouchi, T., and Kinosita, H.1985. Seismic features in the central Okinawa
Trough-an active incipient rifting, Abstract for the 23rd General Assembly of IASPEI, Tokyo 1:279.
Kato, S., Katsura, T., and Hirano, K. 1982. Submarine geology off Okinawa Island, Rep. Hydrogr. Res. 17:31-70.
Kimura, M. 1985. Back-arc rifting in the Okinawa Trough, Mar. Petrol. Geol. 2:222-240.
Kimura, M., Kaneoka, I., Kato, Y., Yamamoto, S., Kushiro, I., Tokuyama, H., Kinoshita, H., Isezaki, N., Masaki,
H., Oshida, A., Uyeda, S., and Hilde, T. W. C. 1986. Report on DELP 1984 cruises in the middle Okinawa
Trough. Part V: Topography and geology of the central grabens and their vicinity, Bull. Earthquake Res. Inst.,
U. Tokyo 61:269-310.
Le Pichon, X., Huchon, P., and Barrier, E. 1985. Geoid and the evolution of the western margin of the Pacific
Ocean, in Formation of Active Ocean Margins (N. Nasu, ed.), pp. 3-42, Tokyo, Terrapub.
Le Pichon, X., and Sibuet, J.-C. 1981. Passive margins: A model of formation, J. Geophys. Res. 86:3708-3710.
Lee, C. S., Shor, G. G., Ir., Bibee, L. D., Lu, R S., and Hilde, T. W. C. 1980. Okinawa Trough: Origin of a back-arc
basin, Mar. Geol. 35:219-241.
Letouzey, I., and Kimura, M. 1986, The Okinawa Trough: Genesis of a back-arc basin developing along a
continental margin, Tectonophysics 125:209-230.
Ludwig, w., Murauchi, S., Den, N., Bull, P., Hotta, H., Ewing, M., Asanuma, T., Yoshii, T., and Sakajiri, N. 1973.
Structure of the East China Sea-West Philippine Sea margin of southern Kyushu, J. Geophys. Res. 78:2526-
2536.
Miki, M., Matsuda, T., and Otofuji, Y. 1990. Opening mode ofthe Okinawa Trough: Paleomagnetic evidence from
the South Ryukyu arc, Tectonophysics 175:335-347.
Nagumo, S., Hinoshita, H., Kasahara, I., Ouchi, T., Tokuyama, H., Asamuma, T., Koresawa, S., and Akiyoshi, H.
1986. Report on DELP 1984 cruises in the Middle Okinawa Trough, Part II: Seismic structural studies, Bull.
Nash, D. E 1979. The geological development ofthe north Okinawa Trough area from Neogene times to Recent, J.
Jpn. Assoc. Petrol. Technol. 44:121-133.
Oshida, A., Midorikawa, Y., Kawabata, K., Kanazawa, J., Kimura, M., and Kato, Y. 1988. Submarine acoustic and
geomagnetic surveys in the north of Yaeyama-Gunto, Ryukyu arc (RN86 and RN87 cruises), Bull. Coli.
Sci., U. Ryukyus 46:123-138.
Oshida, A., Tamaki, K., and Kimura, M. 1992. Origin of the magnetic anomalies in the southern Okinawa Trough,
J. Geomagn. Geoelectr. 44:345-359.
Oshima, S., Takanashi, M., Kato, S., Uchida, M., Okazaki, I., Kasuga, S., Kawashiri, C., Kaneko, Y., Ogawa, M.,
Kawai, K., Seta, H., and Kato, Y. 1988. Geological and geophysical survey in the Okinawa Trough and the
adjoining seas of Nansei Syoto, Rep. Hydrogr. Res. 24.
Reches, Z. 1983. Faulting of rocks in three-dimensional strain fields. II: theoretical analysis, Tectonophysics
95:133-156.
Regan, R. D., and Rodriguez, P. 1981. An overview of the external field with regard to magnetic surveys, Geophys.
Surv. 4:255-296.
Seno, T., and Maruyama, S. 1984. Paleogeographic reconstruction and origin of the Philippine Sea, Tectono-
physics 102:53-54.
Sibuet, J.-C.1991. The southern Okinawa Trough, in Taicrust Workshop Proc., June 10-12,1991, National Taiwan
University, Taipei, Taiwan, RO.C., pp. 117-126.
Sibuet, J.-C. 1992. New constraints on the formation of non-volcanic continental Galicia-Flemish Cap conjugate
margins, J. Geol. Soc. London 149:829-840.
Sibuet, I.-C., Dyment, I., Bois, C., Pinet, B., and Ondrt!as, H. 1990. Crustal structure of the Celtic Sea and western
approaches from gravity data and deep seismic profiles: Constraints on the formation of continental basins,
J. Geophys. Res. 95:10,999-11,020.
Sibuet, I.-C., Letouzey, I., Barbier, E, Charvet, I., Foucher, I.-P., Hilde, T. W. C., Kimura, M., Ling-Yun, C.,
Marsset, B., Miiller, C., and Stephan, I.-E 1987. Backarc extension in the Okinawa Trough, J. Geophys. Res.
92:14,041-14,063.
Sibuet, J.-c., Letouzey, J., Marsset, B., Davagnier, M., Foucher, J.-P., Bougault, H., Dosso, L., Maury, R., and
Joron, J.-L. 1986. Tectonic evolution and volcanism of Okinawa Trough (abstract), Am. Assoc. Petrol. Geol.
Bull. 70:934.
Sibuet, J.-C., Monti, S., Rehault, J.-P., Durand, C., Gueguen, E., and Louvel, V. 1991. Quantification de I'extension
liee a la phase pyreneenne et geometrie de la fronliere de plaques dans la partie ouest du golfe de Gascogne,
C.R. Acad. Sci. Paris 317:1207-1214.
Sun, S. C. 1981. The Tertiary basins of off-shore Taiwan, ASCOPE, Manila.
Taylor, B., Klaus, A., Brown, G. R., and Moore, G. F. 1991. Structural development of Sumisu rift, Izu-Bonin arc,
l. Geophys. Res. 96:16,1l3-16,129.
Tsai, Y.-B. 1986. Seismotectonics of Taiwan, Tectonophysics 125:17-37.
Tsuburaya, H., and Sato, T. 1985. Petroleum exploration well Miyakojima-Oki, l. lpn. Assoc. Petrol. Techno!.
50:25-53.
Ueda, Y. 1986. Geomagnetic anomalies around the Nansei Soto (Ryukyu Islands) and their tectonic implications,
Bull. Volcanol. Soc. lpn. 31:177-192.
Ujiie, H. 1980. Significance of "500 m deep island shelf" surrounding the southern Ryukyu Island arc for its
Quaternary geological history, Quat. Res. 18:209-219.
Uyeda, S. 1977. Some basic problems in trench-arc-back-arc-system, in Island Arcs, Deep Sea Trenches and Back-
Arc Basins (M. Talwani and W. C. Pitman III, eds.), Maurice Ewing Ser., Vol. I, pp. 1-14, American
Geophysical Union, Washington, DC.
Uyeda, S., Kimura, M., Tanaka, T., Kaneoka, I., Kato, Y., and Kushiro, I. 1985a. Spreading center ofthe Okinawa
Trough, Tech. Rep. lAMSTEC, pp. 123-142.
Uyeda, S., Nagumo, S., and Hilde, T. W. C.1985b, Okinawa Trough-An early stage of continental margin rifting,
in 1985 Geodynamics Symposium on Intraplate Deformation: Characteristics, Processes, and Causes, Texas
A&M University, College Station, TX.
Vander Zouwen, D. E. 1984. Structure and evolution of Southern Okinawa Trough, Master's thesis, Texas A&M
University, College Station.
Warner, M. R. 1987. Seismic reflections from the Moho: The effect of isostasy, Geophys. l. R. Astron. Soc. London
88:425-435.
Wessel, P., and Smith, W. M. F. 1991. Free software helps map and display data, EOS, Trans. AGU 72:441-446.
Yarnano, M., Uyeda, S., Foucher, I.-P., and Sibuet, I.-C. 1989. Heat flow anomaly in the middle Okinawa Trough,
Yarnano, M., Uyeda, A., Kinoshita, H., and Hilde, T. W. C. 1986. Report on DELP 1984 cruises in the Middle
Okinawa Trough. Part IV: Heat flow measurements, Bull. Earthquake Res. Inst., U. Tokyo 61:251-267.
10
Shikoku Basin and Its Margins

Kazuo Kobayashi, Shigeru Kasuga, and Kyoko Okino
ABSTRACT
The Shikoku Basin is an inactive backarc basin located south of the southwest Japan arc in
the northwestern Pacific margin. Its characteristic features are summarized on the basis of
updated geophysical and geological data, including swath bathymetry, gravity, magnetics,
and seismic reflection profiling records as well as results from the DSDP/ODP drilling
holes. It has been proposed that the Shikoku Basin was born as a rift separating the N-S
trending paleo-Kyushu-Palau Ridge at its northern end. The rifting rapidly propagated
southward. It was succeeded by seafloor spreading that fonned a narrow triangular trough
bounded by steep scarps on both east and west margins. Magnetic and bathymetric data
indicate that the spreading center has changed its trend at least twice, first at 23 Ma and then
at 19 Ma. Widespread off-ridge volcanism occurred after extinction of spreading at 15 Ma.
The Kinan seamount chain was fonned at this stage. Most of the rocks constituting the
basin are tholeiite similar to MORB, whereas some of the off-ridge magmas are alkali
basalt. Igneous basement is overlain mostly by hemipelagic sediments containing dispersed
detrital clays and interbedded tephra layers. The Shikoku Basin lithosphere is subducting at
the Nankai Trough beneath southwest Japan in a NNW direction. Deep-focus earthquakes
are not observed at depths greater than 80 km, implying that the subducted young
lithosphere loses its rigidity below such depths. It seems plausible to presume that the
Shikoku Basin was subducting at the Nankai Trough at 15-12 Ma while the basin floor
underwent extensive volcanic intrusions. Emplacement of intennediate to felsic rocks at
the outer zone and Setouchi Province of southwest Japan probably originated from this
unusual circumstance.
1. INTRODUCTION
The Shikoku Basin is a N-S elongated fan-shaped basin surrounded by the Kyushu-
Palau Ridge on its west and by the Nishi-Shichito Ridge (westernmost ridge of the Izu-
Bonin arc) on its east. Its northern margin is bounded by the Nankai Trough at which the
Kazuo Kobayashi • Japan Marine Science and Technology Center, Yokosuka 237, Japan. Shigeru
Kasuga and Kyoko Okino • Hydrographic Department, Maritime Safety Agency, Chuo-ku, Tokyo
104, Japan.
381
382 KAZUO KOBAYASHI et al.
135E
Japan
40N Sea 40N
Pacific
SHIKOKU Plate 30N
.296 .443
442. .444
\ BASIN
\0~ '"\. FIGURE 10.1. Index map of the Shikoku
25N \ Philippine Sea
Plate
25N Basin and its margins. Positions of relevant
DSDP and ODP sites are shown by numeri-
130E l40E
cal figures beside solid circles. Arrows de-
135E
note directions of relative plate motions.
basin is subducted under the southwest Japan arc (Figs. 10.1 and 10.2). On the south it
merges with the Parece Vela Basin (or West Mariana Basin in Japanese maps), with which it
shares a similar evolution.
The Shikoku Basin is a backarc basin that formed to the west of the Izu-Bonin arc-
trench system, which is active at present. Opening of the basin is, however, no longer
occurring. The fundamental nature of the Shikoku Basin as an inactive backarc basin was
first pointed out by Karig (1971).
Bathymetry, magnetic and gravity anomalies, and subbottom structure of the Shikoku
Basin have been surveyed comprehensively by SN Takuyo of the Hydrographic Depart-
ment, Maritime Safety Agency (HD-MSA) of Japan (Kasuga et aZ., 1987, 1992; HD-MSA,
1990). The majority of survey tracklines are along an E-W direction with spacing of 5
nautical miles (approx. 9 km). Bathymetric coverage of the area by this swath survey is
roughly 40%. Denser tracks were chosen for several selected areas having complex
topography. The north-central margin of the basin close to the N ankai Trough was surveyed
in detail with nearly 100% swath coverage for bathymetric mapping by the French ship
Jean Charcot as part of the French-Japanese cooperative KAIKO Project (KAIKO I
Research Group, 1986; Le Pichon et aI., 1987).
The crustal structure of this area was investigated by two-ship refraction studies as
early as the 1960s (Murauchi et aI., 1968), and typical oceanic characteristics of the crust
under the basin were clearly demonstrated. The distribution of sediment cover was sur-
veyed by single-channel seismic reflection profiling along the HD-MSA's 5-mile spaced
EW tracks cited above, as well as by several multichannel reflection lines carried out by
SHIKOKU BASIN AND ITS MARGINS 383
HOE
FIGURE 10.2. Major topographic trends in the area surrounding the Shikoku Basin. The contour interval is 1 kIn.
Seamount names are shown by letters as follows: A, Koshu; B, Daiichi-Kinan; C, Daini-Kinan; D, Hakuho; E,
Komahashi-Daisan; F, Komahashi-Daini; G, Komahashi; H, Kita-Koho, I, Minami-Koho. Lines (a) and (b)
locate seismic profiles shown in Fig. 10.4.
various investigators (e.g., IPOD-Japan, 1977; HD-MSA, 1989, 1990). Most topographic
features near the margins of the Shikoku Basin are masked by thick clastic wedges
overlying local topographic lows filled by sedimentation in an early stage of basin forma-
tion (White et al., 1980; Klein and Kobayashi, 1981).
A Deep Sea Drilling Project (DSDP) hole was first drilled in the northern margin of the
basin (at Site 297) during Leg 31, although its penetration was insufficient to reveal the
structure and age of basement. On the other hand, hole 296 on the Kyushu-Palau Ridge
penetrated the igneous crust and provided crucial information on the origin and subsidence
history of the remnant ridge (Ingle et aI., 1975). A set of holes was drilled in the inner and
axial zones of the Shikoku Basin during Leg 58 (Klein et al., 1980; Klein and Kobayashi,
384 KAZUO KOBAYASHI at al.
1981). Holes at Site 442 west of the axis of the basin penetrated the pillow lava layer
overlain by sediments intercalated with massive intrusive volcanic flows. Sites 443 and
444 in the eastern zone of the basin provided only the minimum age and sedimentary
history of the area, because sediments immediately above the basement pillow were not
recovered. Drilling of the east and west margins of the basin has not been attempted.
Hole 298 of Leg 31 and holes 582 and 583 of Leg 87 (Kagami et al., 1987) were drilled
in the landward slope of the Nankai Trough and provided information only on the basin
sediment. Holes at Site 808 in Ocean Drilling Program (ODP) Leg 131 penetrated the entire
accretionary wedge of the western Nankai Trough south of Cape Muroto on Shikoku Island
into the underlying oceanic basement rocks, which were subducted from the Shikoku Basin
(Taira et al., 1991; Hill et aI., 1993) so that the recovered cores yielded information on the
northern margin of the Shikoku Basin, which is now overlain by the landward wedge. The
present article is based upon these results combined with other survey data.
2. TOPOGRAPHY OF THE BASIN AND ITS MARGINS
2.1. General Bathymetric Features and Basement Topography

Fig. 10.3 shows a bathymetric map of this area with 200-m contour intervals. Water
depths gradually increase southwestward, and the maximum depth of over 5000 m is
observed in the western margin close to the Kyushu-Palau Ridge. A sharp contrast in
topography is seen between the axial zone and both wings of the basin floor. The axial zone
is composed of small ridges and troughs trending N400W with several seamounts (the
Kinan seamount chain, Fig. 10.2), whereas the wings show linearity nearly parallel to the
trend of the whole basin, implying an episodic change in the opening pattern. A triangle
bordered by longitude 135°E, northern Kyushu-Palau Ridge and western Nankai Trough
has flat topography caused by a thick cover of sediments. The east wing is relatively flat
under the wedge of clastic deposits supplied from the Izu-Bonin arc. In contrast, the west
wing south of 300N is particularly rugged. It consists of small ridges and troughs with
lengths of 30 to 50 km, widths of 5 to 10 km, and relative heights of 500 to 1000 m, trending
along NNW-SSE.
Figure 10.4 represents a composite bathymetric profile of the Shikoku Basin in which
13 parallel profiles trending N67.5°E from 25°N to 33°N are stacked relative to the Kinan
seamount chain (Park et al., 1990). It shows that water depths of the Shikoku Basin are
quite asymmetric with a general tendency of westward deepening. Figure 10.5 shows
multichannel seismic reflection profiles traversing the Shikoku Basin at about 29°N and at
25°N. Sediment cover in the axial and inner zones is about 300 m except for the north
margin close to the Nankai Trough, whereas both west and east wings north of 27°N are
covered by horizontally layered sediments thicker than 1000 m. Sediments in the east wing
are particularly thick, as seen in the right-hand side of the profiles. General westward-
deepening topography of the eastern limb may partly be explained by the thick sediment
covers.
The western margin of the basin is distinguished by steep and stepwise scarps along
the Kyushu-Palau Ridge as indicated in the left-hand comer of the seismic profile in Fig.
1O.5(b). The boundary in the eastern margin along the en echelon Nishi-Shichito Ridge is
not so distinct as in the western scarps. Instead a distinct steep westward-dipping scarp
trending in a direction of NlOoW is recognized in the east wing of the basin floor around the
135 140E
km
,o 100
, 200
,
FIGURE 10.3. Bathymetric map of the Shikoku Basin and its margins (from Kasuga et aI., 1992), 200-m
contours. Numerical figures beside thick contour lines denote depths in kID. Lambert Projection.
longitude of 137°30'E north of 27°N. The linear Kinan escarpment has a maximum
elevation of 800 m and a length of 500 km at 29°30'N, 137°30'E. Detailed analysis of swath
bathymetric maps shows that this scarp is composed of segments of en echelon cliffs each
trending along NNW-SSE with lengths of about 30 km in average. Dives of the submersible
Shinkai 6500 (dive nos. 176 and 177 in 1993) indicated that the surface of the cliffs and
taluses is completely covered by Mn oxides (Olcino, 1993). This observation implies that
the faulting that forms the scarps is relatively old, perhaps as old as the age of the basin
SHICHITO-IWOJIMA RIDGE
KYUSHU-PALAU RIDGE CHAIN
6
(km) (km)
300
_ _C::====:J_ _0 V_E = 30.3
FIGURE 1004. Stacked profiles of water depths of the Shikoku Basin between 25°-33°N along N67 .5°E trending
lines (from Park et ai., 1990).
generation (19 Ma). East of the escarpment, arrays of small cliffs cutting the upper layer of
sediment are recognized. They appear to have been formed by recent tectonic activity.
2.2. The Southern Border of the Basin

The southern margin of the Shikoku Basin merges with the N-S-elongated Parece Vela
Basin. Although no distinctly lineated ridge or trough separating the Shikoku Basin and
Parece Vela Basin has been found, the Shikoku Basin can be distinguished from the
southern basin by a line trending in a direction of N400E at latitude of about 23 to 25°N. The
following viewpoints support this subdivision:
1. Width of the basin is narrowest along this border: the maximum width of the fan-shaped
Shikoku Basin is roughly 900 km along its north margin (on the Nankai Trough), and that of
the Parece Vela Basin exceeds 1200 km along a latitude of l7°N, whereas the width along this
presumed border is only 550 km.
2. The predominant topographic trend changes north and south of this border. In the north it is
generally parallel to the Kyushu-Palau Ridge, whereas it trends N400E in the south.
3. A chain of seamounts exist in the north. In contrast, there are deep troughs along the axis of
the south basin.
4. The east side of the Shikoku Basin is fringed by the Izu-Bonin arc with embryonic rifting,
whereas the south adjoins the West Mariana Ridge and the well-developed Mariana Trough
to its east.
2.3. Axial Seamount Chain

Along the axis of the northern part of the basin, a chain of seamounts has long been
known and named the Kinan seamount chain, which is composed of Koshu, Daiichi-
Kinan, Daini-Kinan, and Hakuho seamount from north to south (see Fig. 10.2). The
northernmost seamount consists of a linear 80-km-Iong chain with multiple crests trending
N700W (Shino et al., 1991). Water depth of the shallowest peak is 2060 m. Two independent
seamounts with crestal depths of 3130 m and 1990 m are situated just west of the northern
end of Koshu seamount. One conical seamount exists in the east wing of the basin and is
named Komahashi Daisan seamount. Its 1770-m-deep crest is nearly 150 km ENE of the
peak of Koshu seamount.
.., ...c
>. ~ 1:1 ~
>. ....
[ -111 Sp r eading NE- SW Spreading ~~ f!-w 8
:lit: (,1 Spreading
! i!
I~~~ SWW
j a aI .:l e
~
~
»
~
~
§
~
D
I
SDk.
•
~
. ~ ..~.. :'.:'~~:,,:.~: .. ::.....:~.I:"06~"; ......: _.... ;t:.::!~:.;.:

'.;'" '"
..........
t: ; '-'.
".: ..... ~ •• ;..';.. :.~':.'.. ..:!~!..-~~' . ," •....•• ,
•. ~.~.. ....~. or
sec
D SDk. Nish l-
Sh lchl l o
Ridge
FIGURE 10.5. Multichannel seismic reflection profiles traversing the Shikoku Basin (migrated time sections). For positions see Fig. 10.2. (a) IPOD-Japan (1977)
section (220 km long) in the central portion of the basin. Shot-point spacing =50 m, 24-fold stack, deconvolution and time variant filtering. (b) HD-MSA section in a
EW direction at 25°N (HD-MSA, 1989). Shot-point spacing = 50 m, 48-fold stack, deconvolution and time variant filtering. Numbers on the profile denote trace
numbers corresponding to one every 25 m. ~
'-I
Two relatively large features, called Daiichi- and Daini-Kinan seamounts, are lo-
cated on the axial zone roughly 120 km south of Koshu seamount. The twin peaks of each
seamount, with water depths as shallow as 700 m, are aligned N40oW, parallel to linearity
of the basin in the axial zone. Each seamount body has an elongated shape that also parallels
the general trend. Hakuho seamount located at latitude of 28°oo'N is also elongated in a
direction parallel to the trend of the axial zone (Furuta et aI., 1980; Kobayashi, 1984;
Kobayashi and Fujiwara, 1993).
2.4. Nishi-Shichito Ridge- The Eastern Border of the Basin

The Izu-Bonin arc, the east-bordering arc-trench system of the Shikoku Basin,
appears to be composed of at least three linear N-S trending ridges: the Nishi-Shichito
Ridge, Shichito-Iwojima Ridge, and Ogasawara (Bonin) Islands. The Nishi-Shichito
Ridge consists of an en echelon alignment of ridges trending N600E (Fig. to.2). Seafloor
ruggedness is particularly distinct in the region north of 30 oN. Separated crests of sea-
mounts are recognizable in each ridge. The size of seamounts generally increases south-
westward (toward the Shikoku Basin). The northernmost ridge, called Zenisu Ridge, is the
most prominent in morphology and is the longest. Its northeastern end extends toward the
line of islands Kozushima, Toshima, and Shikineshima.
In the region between 300N and 26°N the trend of the en echelon alignment is N45°E.
Seamounts are nonexistent in the Nishi-Shichito Ridge south of 26°N, although NE-SW
trending troughs or canyons such as Iwo Canyon are still distinguishable there. To the east,
a series of rifts, including the Hachijo, Aogashima, Sumisu, Torishima, Sofu Gan, and
Nishino-shima rifts (from 33°N to 28°N) segregates the Nishi-Shichito Ridge from the
Shichito-Iwojima Ridge, which comprises a chain of active volcanic islands and sub-
marine volcanoes in the Izu-Bonin arc (Brown and Taylor, 1988; Taylor, 1992).
2.5. Kyushu-Palau Ridge-The Western Border of the Basin

The Kyushu-Palau Ridge is a 3OOO-km-long submarine ridge that connects Kyushu,
Japan, at 32°N and the Palau Islands located at 8°N. It divides the floor of the whole
Philippine Sea into east and west. North of 24 oN it trends NNW-SSE, whereas it runs along
a NNE-SSW direction further south. The northern segment of the Kyushu-Palau Ridge is
composed of an elongated ridge morphology overlain by seamounts. Saddles between
adjacent seamounts are still nearly 2000 m higher than adjacent basins. Particularly large
seamounts are named Komahashi-Daini [30 0N, CD (crestal depth) = 289 m], Komahashi
(28°N, CD = 440 m), Kita-Koho (26°45'N, CD = 329 m) and Minami-Koho (26°tO'N,
CD = 367 m).
On each seamount more than two crests are aligned NE-SW, oblique to the ridge axis.
The seamount bodies on the NE side are usually larger than those to the SW. Okino-
tori shima at 20025'N, 136°03'E is the only emerged island (uplifted atoll) in the Kyushu-
Palau Ridge and is situated close to its junction with the E-W trending Oki-Daito Ridge,
which borders the north of the West Philippine Basin (Fig. to.2). The southern portion of
the Kyushu-Palau Ridge beyond this emergent island changes its morphology to an
elongated ridge with fewer seamounts.
The boundary of the Kyushu-Palau Ridge with the Shikoku Basin is marked by sharp,
steep scarps. The escarpment is supposed to have been formed during the initial rifting of
the Shikoku Basin and subsequent fast subsidence of the basin. In some places the east-
dipping scarps are stepwise with sedimented depressions between them as shown in Fig.
1O.5b. The west margin of the Kyushu-Palau Ridge, on the other hand, gradually increases
its water depth to reach relatively flat basins covered by thick sediments such as the Kikai,
north Daito, and south Daito basins.
3. GRAVITY ANOMALIES AND ISOSTATIC COMPENSATION OF THE BASIN

AND ITS MARGINS
The gravity field of the northwestern Pacific, including the Shikoku Basin, has long
been measured by shipboard gravity meters (e.g., Tomoda et al., 1968), and several maps of
free-air anomalies have been published (Tomoda, 1974; Watts, 1976; Kasuga et al., 1987,
1992). Free-air anomalies in the Shikoku Basin, as well as those of the West Philippine
Basin, are near 0 mgal except for seamounts and knolls, indicating isostatic compensation
of the basin crust. Bouguer gravity anomalies in the Shikoku Basin gradually increase
southwestward in harmony with thinning of the sediment cover. In the eastern margin of the
basin close to the Nishi-Shichito Ridge the Bouguer anomalies decrease, implying thicker
crust there.
Along the Kyushu-Palau Ridge free-air anomalies amounting to 50-100 mgal are
observed on each individual topographic high (Fig. 10.6). No continuous anomaly belt
exists along the ridge, whereas distinct Bouguer anomalies are seen in the portion north
of 26°N. The results seem to imply that the crust of the ridge is thicker than that of the basin
floor but that it is isostatically compensated.
Free-air anomalies of + 100 to + 120 mgal are aligned en echelon along the Nishi-
Shichito Ridge in a manner similar to topography. In contrast, such an en echelon align-
ment is not seen with Bouguer anomalies, suggesting that topographic features are caused
only by rugged shallow crust but are not related to the deeper structure and that they are
regionally supported.
4. STRATIGRAPHY OF THE SHIKOKU BASIN
The structure and geological history of bottom sediments were revealed by seismic
reflection profiling and by ocean drilling. The accretionary wedge sediments recovered by
drilling at DSDP/ODP Sites 298, 582, 583, and 808 consist of a succession of turbidites and
hemipelagic clays (Ingle et aI., 1975; Kagami et aI., 1987; Taira et al., 1991; Pickering et aI.,
1993). Most turbidites recovered from these sites appear to have been accreted from the
Nankai Trough to which they were supplied from river and land shelves northeast of the
basin via channels along its axis, given that the thickness of turbidite layers revealed
by seismic reflection profiling decreases westward (Taira and Niitsuma, 1986; Le Pichon
et al., 1987).
Most of the cores from Site 297, located at the northwestern comer of the Shikoku
Basin south of the axis of Nankai Trough, are composed of hemipelagic sediments. Only
the cores in a limited range between 3 and 5 Ma contain many turbidites. Karig (1975)
postulated that the trough topography, which now traps turbidites transported from the
northern islands, disappeared in that period. An alternative explanation is fill-up and
FIGURE 10.B. Free-air gravity anomalies of the Shikoku Basin area (from Kasuga et ai., 1992). The contour
interval is 20 mgal. Solid lines denote positive and broken lines negative anomalies.
overflow of turbidites at the trough axis occurring when a large amount of turbidites was
transported from the highly elevated central mountains in Honshu. The Shikoku Basin
facies of sediments in the lower 620-m column (between 620 m and 1240 m below seafloor)
at Site 808 consists of hemipelagic sediments (Underwood et al., 1993a).
Drilled cores from the inner wings of the basin (442, 443, and 444) contain hemi-
pelagic sediments interbedded with a number of volcanic tephra layers and dispersed
detrital clays (Fig. 10.7). Ages of each horizon in the cores were determined by micro-
paleontological correlation using nannofossils (Okada, 1980) and, in less frequent cases,
foraminifera and radiolarians. Ages of the sediment interbed recovered from a level
between massive intrusives and underlying pillow basalt at Site 442 are correlated to be
18 to 21 Ma (Discoaster drugii subzone of nannofossils), whereas sediment immediately
overlying the upper intrusive sill basalt at the same site is 15 to 17 Ma (Helicosphaera
ampliaptera zone of nannofossils). The oldest sediment age here is consistent with the
magnetic isochron 6 (see the next section), suggesting a seafloor spreading origin for the
underlying pillow layer. The age of the red-colored mudstone interbedded with basaltic
intrusives at Site 808 was determined to be 15 Ma by using nannofossils (Olafsson, 1993).
This age is consistent with that of the final stage of opening of the Shikoku Basin, as Site
808 is situated immediately west of the presumed spreading axis of the subducted basin.
FIGURE 10.7. Stratigraphic correlation of DSDP Sites 442. 443. 444. and 297 in the Shikoku Basin (Klein and
Kobayashi. 1981) and ODP Site 808 in the Nankai Trough wedge (Pickering et al.• 1993).
The age of the oldest sediment overlying massive basalts at Sites 443 and 444 situated
east of the basin axis is 14 to 15 Ma (Spheonolithus heteromorphus zone). There is a
discrepancy between this age and the magnetic isochron ages (6A according to Watts and
Weissel, 1975, and Kobayashi and Nakada, 1978, or 5D-5C identified by Shih, 1980). The
lower strata were not recovered at these sites owing to technological difficulty and time
limitation. The occurrence of postspreading off-ridge basaltic intrusions widespread in the
basin (Klein et aI., 1978) obscured the paleontological age relations of Sites 443 and 444.
More than 100 layers oftephras have been recognized in the cores from Sites 442, 443,
and 444. Volcanic glass shards are well preserved, since they are overlain by fine-grain
sediments with low water permeability. Refractive index and chemical composition of glass
shards revealed a sharp contrast between Site 442 in the west wing and the other two in the
east wing. Site 442 contains rhyolitic and dacitic tephras throughout the cores exclusively,
whereas distinct layers of basaltic tephras occur in middle to early Miocene sediment at
Sites 443 and 444 situated east of the axis (Furuta and Arai, 1980). This result seem to be
explained by the prevailing northwesterly wind in this region. In Pliocene and Pleistocene
cores collected at these sites, tephras display felsic to intermediate compositions, indicating
predominance of such volcanism in southwestern Japan. Plentiful rhyolitic tephras and
volcaniclastics have been found at the base of a Site 808 sedimentary section immediately
above the basaltic basement. Source of the rhyolitic volcanics is presumed to be the
episodic forearc volcanism in southwestern Japan taking place in middle Miocene (13-15
Ma). They were most probably transported to Site 808 by wind and/or by turbidity currents.
As Site 808 is now situated about 10 Ian landward from the axis of the trough, it was located
roughly 200 Ian oceanward side of the trough at 13 Ma, if the rates of subduction at the
Nankai Trough are assumed to be constant (-2 cm/yr) throughout the geological period. Its
distance from the source region (forearc volcanoes) seems to be still close for rhyolitic
volcanics to reach the site.
FIGURE 10.B. Chronological variations in clay sedi-

Poorly mentation in the Shikoku Basin since its opening based
drained soils on DSDP 297, 442, 443, and 444 cores (Charnley,
1980).
Chamley (1980) reported changes in content of clay minerals in the cores drilled from
the Shikoku Basin. As shown in Fig. 10.8, sedimentation of brown pelagic clay dominated
by in situ smectite formation began immediately after the basin formed in the early-middle
Miocene (15-14 Ma). Latest Miocene to Pleistocene clay sedimentation was progressively
influenced by continental climatic factors. Smectite came from volcanic islands such as the
Izu-Bonin arc. Illite and chlorite were supplied from southwest Japanese Islands, while
kaolinite was transported from the southwest by the Kuroshio current. The amount of
detritus has changed with time, depending on vertical tectonic movements of Izu-Bonin
and Kyushu-Palau ridges as barriers against ocean circulation. Vermiculite, attapulgite,
and most of the irregular mixed-layer clays were blown by wind from the Asiatic continent
(Fig. 10.8). A different pattern of detrital clay content was reported at Site 808 (Underwood
et aI., 1993b). One plausible interpretation of these differences may be proximity of the
site to the extinct spreading axis at which hydrothermal circulation to the seafloor was
active for some time after spreading ceased.
The present bottom of the Shikoku Basin is below the calcium carbonate compensa-
tion depth (CCD, nearly 4500 m at present) and the carbonate lysocline (generally 1000 m
shallower than CCD), since CaC0 3 content in bottom surface sediments is very low. The
DSDP cores from Sites 442-443 indicated that the bottom of this basin in Miocene time
was also below the CCD, as evidenced by badly dissolved foraminiferal tests of that age
(Echols, 1980). On the other hand, occurrence of a 40-cm-thick limestone layer imme-
diately overlying basalt at hole 442A and chalk at the bottom of hole 443 cores implies that
the depth of the Shikoku Basin was shallower than the CCD and lysocline when the basin
was just born. This result seems to imply two possibilities: that the young Shikoku Basin
was much shallower than now, or the CCD and lysocline at about 20 Ma were much deeper
than at present. Van Andel et al. (1975) showed that the CCD in the Miocene was nearly
500 m shallower than at present in the Pacific Ocean. If this is valid for the Shikoku Basin
at 20 Ma, the occurrence of carbonates indicates that the basin probably was much
shallower than at present. Pillow basalts recovered from Site 442 are extremely vesicular,
whereas sills contain fewer vesicules. Dick (1980) interpreted the origin of high vesicularity
to be caused by high volatile contents of backarc basin magma, whereas Kobayashi (1984)
postulated shallow water depths at the time of eruption of these rocks.
5. PETROLOGY AND AGES OF THE BASIN AND MARGINS
Igneous rocks of the Shikoku Basin "basement" were collected by the DSDP and
ODP. Pillow basalts recovered from Site 442 are supposed to have erupted on the seafloor at
the spreading center. Their petrology and texture are similar to those of the mid-ocean ridge
basalt (MORB; Dick, 1982; Dick et al., 1980). Major and trace elements analyses indicate
that they are tholeiite with Mg number between 0.6 and 0.7, similar to MORB. Many of the
Site 442 massive units are diabases with chemical composition of subalkaline tholeiite
resembling N-type MORB (Marsh et aI., 1980; Wood et aI., 1980). Most other massive
flows and sills collected at Sites 443 and 444 have similar MORB composition, except for
one sill of alkali olivine basalt at Site 444.
Igneous rocks were reached at a depth of 1289.9 m below the seafloor under an
accretionary wedge, and 37.1 m in total of basalts were penetrated (Taira et aI., 1991). The
upper unit overlain by middle Miocene sediment is composed of sill, and the lower 10 m is
pillow basalt, both analogous to the normal MORB similar to Site 442 (Siena et al., 1993).
Isotope ages of some basaltic rocks were determined. The 40Arf3 9Ar ages of two
basalts and one dolerite from Site 443 (total fusion ages) are IS.6 ± 1.9,8.3 ± 7.8, and 10.9 ±
3.S Ma, respectively (Ozima et aI., 1977; 1980). The KlAr ages of a pillow basalt from Site
443 and an intrusive sill from site 444 are 17.2 ± 3.2 and 14.7 ± 2.1 Ma (McKee and Klock,
1980).
A boulder of basalt was collected by dredge haul from a slope of the Hakuho seamount
in the Kinan seamount chain (28°Ol'N, 137°27'E, D = 2700-2200 m) (Furuta et aI.,
1980). Petrography of the sample appears to be intermediate between MORB and BABB
(backarc basin basalt), although exact judgment is difficult because of alteration. Several
boulders of igneous rocks coated by Mn oxides were collected from other seamounts in
the Kinan seamount chain. Among them a piece of basalt from Koshu seamount (31°32'N,
13so36'E, D = 2400 m) was dated by the KI Ar method to be 7.91 ± 0.19 Ma (Katsura et al.,
1994).
In contrast to basaltic properties of rocks from the Shikoku Basin floor and axial
seamounts, igneous rocks recovered from marginal ridges have petrographic and chemical
affinities to island-arc magmatism. Volcaniclastic materials with composition of two-
pyroxene andesite were recovered at DSDP Site 296 on a saddle point between seamounts
along the Kyushu-Palau Ridge (29°20'N, 133°32'E, D = 2920 m). Their 4oArf3 9 Ar age is
47.S Ma (Ozima et aI., 1977). A number of granodioritic fragments coated by thick Mn
oxides have been collected since 1973 by dredge hauls from the slopes of Komahashi-
Daini (29°S6'N, 133°19'E, D = 22S0 m; Kobayashi, 1984), Kita-Koho (26°46'N, 13so27'E,
D = 1830 m; Koyama et aI., 1986), and Minami-Koho seamounts (26°06'N, 13so49'E, D =
S03 m; Mizuno et aI., 1977). The KlAr ages of these samples are plotted in Fig. 10.9,
indicating that the Kyushu-Palau Ridge is much older than the Shikoku Basin.
Rocks from Nishi-Shichito Ridge are generally andesitic. One KlAr age (6 Ma)
reported with a sample from a crest at 29°17'N, 138°38'E (D = 903 m) (Hayashida et aI.,
1989) is much younger than those of either the Shikoku Basin and Kyushu-Palau Ridge,
indicating that this ridge was active at periods after the basin ceased to open. A rock sample
from Komahashi-Daisan seamount situated in the northeastern portion of the basin
(31°38'N, 137°16'E, D = 1900 m) has andesitic composition and age of 11.3 ± 0.4 Ma (Shino
et al., 1991), suggesting that this seamount has affinities to the Nishi-Shichito Ridge
activities regardless of its position in the east wing of the basin.
132E 134E 136E 138E 140E 142E 144E
32N'
I
30N~
~1
.~
,
FIGURE 10.9. Isotope ages (in Ma) of rocks collected from the Shikoku Basin and its margins (from various
sources mentioned in the text). Double circles denote ages of basement rocks from DSDP holes. For solid triangles
see Kobayashi (1983) and for solid circles see Shino et al. (1991).
6. MAGNETIC ANOMALIES AND SPREADING HISTORY

OF THE SHIKOKU BASIN
The existence of linear magnetic anomalies in the Shikoku Basin was first reported by
Tomoda et al. (1968). Watts and Weissel (1975) published magnetic anomaly profiles and
their correlation with the standard polarity reversal chronology. They proposed correlatable
linear anomaly patterns in the western limb of the basin, while they suggested the possi-
bility of either symmetric or single-limb spreading for the eastern portion. Tomoda et al.
(1975) and Kobayashi and Isezaki (1976) postulated a symmetric spreading model for the
Shikoku Basin. Kobayashi and Nakada (1978) and Shih (1980) added survey data and
extended interpretation.
Figure 10.10 represents a magnetic anomaly map based upon the HD-MSA's survey
data. In this figure the axial zone can be clearly distinguished from the west and east wings
of the basin in latitudes north of 26°N. The western wing appears to be divided into the
marginal and inner zones, although the eastern margin is overprinted by recent activity
of the Nishi-Shichito Ridge.
Good control of magnetic anomaly ages was provided by DSDP site 442, which
penetrated the pillow basalt basement underlying a sequence of intrusive sill and sediment
(see Section 5). The micropaleontological age of the oldest sediment-pillow basalt contact
is 18-21 Ma, which is consistent with the magnetic isochron age of anomaly 6 first
400
133~: 135E Kyushu-Palau

Ridge
FIGURE 10.10. Magnetic anomalies of the Shikoku Basin and its margins (from Kasuga et aI., 1992). The
contour interval is 100 nT. Areas with negative anomalies exceeding -100 nT are shaded. Dipole anomalies
without seamount topography are indicated by solid arrows. Steep scarps in the eastern margin of Kyushu-Palau
Ridge and topographic boundary of Nishi-Shichito Ridge are shown by 4000-m and 3000-m depth contours,
respectively.
proposed by Watts and Weissel (1975). Sites 443 and 444 drilled in the eastern part of the
basin led to ambiguous crustal ages, because these holes failed to date pillow basalt
basement. Nevertheless, the single-limb model seems to be untenable as the hole bottom
ages are 17 and 15 Ma for Sites 443 and 444, respectively, which are far older than predicted
by such a model.
Okino et al. (1994) postulated an updated identification of magnetic isochrons (Fig.
10.11). They defined three distinct fracture zones in both west and east wings and attempted
to correlate both wings by the fracture zones assuming symmetric spreading of the basin.
Five episodic stages were thus recognized by their analysis: rifting, N700E opening, E-W
opening, and NE-SW opening, followed by off-ridge volcanism and tectonic deformation.
Taylor (1992) postulated a revised identification of Shikoku Basin lineations, suggest-
ing existence of a failed rift at subchron 5D in the west wing of the basin. He proposed the
initial age of spreading to be 22 Ma. However, we prefer an earlier age of the initiation
of spreading on the basis of evidence and rationale mentioned below.
Chamot-Rooke et al. (1987) interpreted isochrons in the northern margin ofthe basin
~14r°;:..E.....,~ 30M
26M
FIGURE 10.11. Magnetic isochrons and fracture zones identified by Okino et al. (1994). In the axial zone fracture
zones have S-shape. Positions of magnetic dipole anomalies are shown by asterisks. Solid circles with numerical
figures indicate positions of DSDP/ODP Sites.
based upon finding three fracture zones during the KAIKO survey. They proposed NlOoW
trending offsets associated with N-S opening in the axial zone close to the Nankai Trough.
Okino et al. (1994) questioned such offsets and suggested existence of NNW-SSE trending
grabens associated with intense defonnation of the basement. They proposed that the
rotation of spreading axis resulted from tectonic defonnation of the northern margin of
the basin caused by the southward drift of Japanese Islands during opening of the Japan Sea
at about 15-20 Ma (Tamaki et al., 1992).
Okino et al. (1994) found pairs of magnetic dipole anomalies unaccompanied by
basement highs both in the flat terrace in the eastern slope of the Kyushu-Palau Ridge
and along the trough east of the Nishi-Shichito Ridge. As chains of these anomalies appear
to be symmetric relative to the basin axis, it seems plausible to assume that they fonned
a single line at a time prior to the opening of the Shikoku Basin and separated afterwards
to the marginal ridges, although no age constraints have been obtained for these anomalies.
In the following (see Fig. 10.12), the history of the Shikoku Basin is interpreted based
on Okino et al. (1994) and other earlier considerations.
6.1. Rifting of the Shikoku Basin

The oldest magnetic lineation identified in the northwestern comer of the basin is
inferred to be the isochron 7 (26 Ma using the magnetic time scale of Cande and Kent,
1992). Prior to that the Kyushu-Palau Ridge and Izu-Bonin arc were joined to fonn a
single active island arc. The rifting was initiated at the northern edge of the arc and
propagated southward. It was then succeeded by seafloor spreading of the floor to fonn a
fan-shaped basin. The exact age of the beginning of rifting is unknown, because the oldest
FIGURE 10.12. Five stages of the Shikoku Basin evolution based upon magnetic and related results (Okino
et al., 1994). (A) Magmatic injection and rifting, (B) southward propagation of rift and fan-shaped spreading,
(C) E-W spreading with transform offsets, (D) NE-SW spreading with en echelon centers, (E) postspreading
volcanism. Scale of figure is the same in all stages except for (A) showing schematic illustration of events only.
part of the basin has been subducted along the Nankai Trough. Considering timing of early
stages of subsidence of the Kyushu-Palau Ridge described in the next section, the rifting
at the latitude of site 296 (29.3°N) started most probably by 30 Ma. Onset of the Shikoku
Basin rifting revealed by ODP Sites 787, 792, and 793 was 30 Ma, in harmony with the
western margin data (Taylor, 1992).
6.2. Early Opening of the Shikoku Basin

Opening of the basin proceeded at a rate of 2.3-4.6 cm/year in the early stage.
Chamot-Rooke et al. (1987) pointed out that isochron 6C (triplet anomalies) and 6B abut
successively with the Kyushu-Palau margin so that the rate of southward propagation of
rifting can be estimated to be about 10-30 cm/year, which is one order of magnitude greater
than the spreading rate. At about 23 Ma the rifting and initial stage of opening reached the
southern boundary zone between the Shikoku and Parece Vela basins. Mrozowski and
Hayes (1979) postulated that the Parece Vela Basin began opening at anomaly 10 (29 Ma)
from the central latitudes and propagated north and south. Complicated but deep topography
and diffuse magnetic anomalies formed at the narrow transition between the two basins.
6.3. Spreading Phase from Anomaly 6B to Anomaly 6

After 23 Ma the spreading center changed its trend from NNW-SSE to NlOoW. The
center was cut by at least three transform faults and several offsets, but spreading proceeded
nearly constantly to form parallel lineations, although rates of spreading decreased from
4.7 to 2.2 cm/year at anomaly 6A (21 Ma). Although magnetic lineations are obscure in the
eastern wing, three fracture zones apparently symmetric to those in the west wing were
identified there.
6.4. The Last Stage of Spreading and Eruption of the Kinan Seamounts
At anomaly 6 or slightly later (20 to 19 Ma) the spreading center reoriented to NW-SE.
The adjustment appears to have been gradual, since analysis by Okino et al. (1994) revealed
S-shaped fracture zones as shown in Fig. 10.11. The linear anomalies and topography are
aligned en echelon along the NNW-SSE trending axis of the basin. This type of readjust-
ment of the spreading centers seems to have been caused by balancing of areas in the rigid
plate surrounded by remnant and active arcs. Similar en echelon alignment in the extinct
axis has been found at the central basin fault in the axis of the West Philippine Basin
(Andrews, 1978).
Opening of the Shikoku Basin finally ceased at a time of approximately 15 Ma, but
eruption of the Kinan seamounts continued until about 12 Ma. Magnetic dipole anomalies
associated with the Kinan seamount chain are relatively small compared to their appreci-
able sizes and somehow masked by the linear anomalies of the surrounding basin, as is seen
with many seamounts existing in the areas close to the spreading ridge. This may be caused
by mixed normal and reversed polarities of igneous rocks composing each seamount body.
6.5. Postspreading Off-Ridge Volcanism in the Shikoku Basin

DSDP and ODP holes have revealed evidence of widespread off-ridge volcanism
throughout the Shikoku Basin (Klein et aI., 1978). Some of the volcanism occurred 4-7
Ma at a flank of the spreading ridge after cessation of spreading (most frequently at 12-15
Ma). Such postspreading activity was discovered in many other areas (e.g., Reykjanes
Ridge; Luyendyk et aI., 1979). In any case it seems likely that the amplitudes of linear
magnetic anomalies formed at the spreading center were reduced by overprinting of
remanant magnetization of off-ridge intrusions, since magnetic polarity opposite that of the
observed anomaly was found with these intrusive rocks (Faller et aI., 1979; Klein et aI.,
1980).
7. UPLIFT AND SUBSIDENCE OF THE BASIN AND ITS MARGINS
Cores from DSDP Hole 296 display a history of vertical movements of the Kyushu-
Palau Ridge in a period prior to and during the rifting and initial stage of opening of the
Shikoku Basin. Shoal fossils of roughly 30 Ma were obtained from the volcaniclastic
sediment overlying the 47.5 Ma rocks contained in the bottom 650 m of the hole. The hiatus
between 30 and 47.5 Ma and occurrence of shoal fossils indicate that the ridge crest was
upheaved and emerged until 30 Ma.
The Kyushu-Palau Ridge was uplifted by at least 2 km above the present elevation
before 30 Ma. Crests of seamounts in the Kyushu-Palau Ridge were emerged above sea
level as evidenced by a late Oligocene reefal limestone sample dredged from the upper
slope (D = 1350 m) of Komahashi Seamount (28°08'N, 134°40'E, crestal depth = 944 m).
The upheaval of the ridge was caused by upwelling and, in some spots, by eruption and
intrusion of magma from the wedge mantle overlying the inclined subducting slab. Mag-
netic dipoles found along both Kyushu-Palau and Nishi-Shichito ridges seem to be
correlatable to this magmatism. Petrography of the volcaniclastics at Site 296 (andesitic)
and seamount rocks (granodiorite) in the Kyushu-Palau Ridge show that the magma has
the chemical composition of arc volcanism.
It may be reasonable to presume that the paleo-Kyushu-Palau Ridge began rifting
at 30 Ma along a chain of the magma reservoirs, because the crust of the ridge became thin
and soft by heating from below in a similar manner to the process of the initial rifting of the
Atlantic Ocean preceded by a chain of hot-spot activities (Burke, 1976). Since no holes
have been drilled at the margins of the basin, we cannot determine the exact age and depth
of the rift valley in the initial stage of rifting. Nevertheless, it seems likely that the rifting
was associated with the normal faults which caused the linear and steep eastward-dipping
scarps in the eastern margin of the Kyushu-Palau Ridge amounting to 2000 m in total
offset. Extensional force in a direction roughly perpendicular to the ridge axis may have
originated from eastward retreat of the Izu-Bonin-Mariana trenches together with the
westward drift of the West Philippine Basin plate associated with initiation of subduction
at the Ryukyu and Philippine trenches.
The data from Site 296 also revealed subsidence of the western half of the paleo-ridge
as rapidly as 325 mlm.y. to form the present Kyushu-Palau Ridge until the hole bottom
reached a depth of 650 m at 28 Ma (Fig. lO.13). Such a fast rate of subsidence was
associated with stretching and normal faulting of the crust.
Lithology of hole 442 implies that the active center of the Shikoku Basin was still as
shallow as 3000 m (the lysocline) in the rifting and initial spreading stages. The crests of the
Kinan seamounts situated at the axis of the basin emerged above sea level and emitted
anhydrous mafic tephras found in the DSDP cores in the Shikoku Basin (Furuta and Arai,
1980; Kobayashi, 1984). The bottom of the basin subsided synchronously with or even
tl S IYUSHU-PALAU RIDGE
KM OYASHIO Paleoland 448 _-_....
438-439 ... __ _-.,::::---- : 296
o
DEPTH
1
2
4
5J.----
FIGURE 10.13. Postulated subsidence curves of two sites in the Kyushu-Palau Ridge (north and south) with
those for sites 438-439. the normal oceans and backarc basins after Kobayashi (1984). Subsidence of remnant arc
and backarc basin is at first very fast following a curve defined by DSDP Site 296 and later approaches the vt
formula of the "normal" backarc basins (Park et al.• 1990).
faster than subsidence of the Kyushu-Palau Ridge after the entire activity ended. The
Shikoku Basin has finally subsided following the VI-law of the ocean basin subsidence by
lithosphere cooling (Kobayashi, 1984; Park et ai., 1990). The difference in the present
elevations between the Kyushu-Palau Ridge and the Shikoku Basin is isostatically con-
trolled by crustal thickness which is greater beneath the ridge than the basin.
8. INTERACTION OF THE SHIKOKU BASIN WITH SOUTHWEST JAPAN

AND NANKAI TROUGH
The Shikoku Basin has been subducted beneath southwest Japan at the Nankai
Trough. The present convergence rates in the Nankai Trough are 2 to 4 cm/year in a NNW
direction according to the NUVEL-l (DeMets et aI., 1990). The landward slope is now very
compressional, and much of the sedimentary section is deformed and accreted to the
upper wedge.
The length of the Wadati-Benioff zone recognized by deep-focus earthquakes along
the Nankai Trough is so short that the time duration since its initiation appears to be roughly
3 Ma (Kanamori and Tsumura, 1971), in harmony with Karig's postulation of a cessation
period of 3-5 Ma (Karig. 1975). However, Sugi and Uyeda (1984) and Shiono and Sugi
(1985) pointed out that oceanic lithosphere younger than 15 Ma easily loses the rigidity
required to generate deep-focus earthquakes. The shape of the Wadati-Benioff zone
beneath southwest Japan (Fig. 10.14) seems to be consistent with their conclusion, as the
zone is shortest at the axis of the Shikoku Basin and longer in both wings. Assuming that
subduction along the Nankai Trough has continued at the present rate since about 15 Ma, at
least 300 km of the northern margin of the Shikoku Basin has been lost under the trough.
FIGURE 10.14. Schematic illustration of deep-focus

earthquakes plane (Wadati-Benioff zone) showing the
Shikoku Basin plate subducting at the Nankai Trough
(modified from Shiono and Sugi, 1985). Double line de-
notes inactive spreading center of the Shikoku Basin.
Depths of epicenters at points Y and Z are 70 km and 90 km,
respectively. Length of slab from trough axis to point Y is
roughly 150 km. Aware of the fan-shaped edge at which the
subducted lithosphere loses rigidity for earthquake genera-
tion at shallower depths at its spreading center than older
margins.
It seems very plausible that rates and direction of subduction at the Nankai Trough before
15 Ma were substantially different from those at present, because southwest Japan was
rotated clockwise in this period in association with the opening of the Japan Sea.
Whatever the history of the Nankai Trough subduction was, interaction between the
Shikoku Basin and southwest Japan has influenced both sides of the trough. Widespread
emplacement of felsic to intermediate magmas at 12-15 Ma in the outer zone and Setouchi
province of southwest Japan (e.g., Nakada and Takahashi, 1979; Kobayashi, 1983) can be
explained by subduction of a hot lithosphere in which widespread off-ridge magmatism
occurred immediately before and even after subduction at the Nankai Trough (Klein et al.,
1978). Abnormally high heat flow in the Nankai Trough region (Yamano et aI., 1984) may
be caused by this origin, although fluid circulation in the accretionary wedge may also
contribute to heat transport (Yamano et aI., 1992). North-south compression of the northern
portion of the Shikoku Basin may have formed the late-stage faulted structures as revealed
by KAIKO and later surveys, either by jumps of spreading centers (Chamot-Rooke et al.,
1987) or by tectonic deformation after the spreading (Okino et al., 1994).
9. SUMMARY, DISCUSSION, AND FURTHER PROBLEMS
Analyses of geophysical and geological data, including those from the DSDP/ ODP
cores, have revealed the processes of tectonic evolution of the Shikoku Basin, a backarc
basin surrounded by island arcs. In the period prior to 30 Ma, the present Kyushu-Palau
Ridge and the lzu-Bonin arc were joined to constitute a large NNW-SSE trending arc-
trench system with the crests of local highs above sea level. The arc was then rifted,
beginning in the north and propagating southward at a speed of 10 cm/year. The rift valley
was bounded by steep scarps on both sides. The western ridge quickly subsided after the rift
was formed. Rifting was succeeded by spreading of the basin. The whole Shikoku Basin
was rifted by 27 Ma and formed a triangular trough during embryonic spreading which
proceeded until 23 Ma.
At 23 Ma the spreading center changed its trend to NlOoW and continued to open the
basin nearly symmetrically until 20-19 Ma. Major portions of the flat basin were formed
during this stage. Rocks constituting the basement are tholeiite with chemical composition
similar to MORB. The center was offset by at least three transform faults, which are
recognized in the present basin as fracture zones. The western limb of the spreading center
remains in the west wing of the basin, whereas the eastern one underwent volcanic
activities of the Nishi-Shichito Ridge and lost its original morphology and linear magnetic
anomalies.
Widespread intrusion of magma occurred roughly 5 m.y. after the end of opening of
the basin. Some of the magma was much differentiated or contaminated with surrounding
rocks to generate alkali-basalt, as recovered from a sill at Site 444. Such off-ridge volca-
nism has been known at a number of locations in the world oceans. It may modify the water
depth and complicate magnetic patterns in the area and heat flow values, although we have
not yet discovered any remnant volcanic vents supplying the intrusive lavas.
It is recognized that the Shikoku Basin is one of the type localities for study of ancient
backarc processes. Comparative investigation of this region with active backarc basins and
rifts such as the Mariana Trough, Sumisu Rift, and the Okinawa Trough will provide much
crucial information on backarc processes and their driving mechanisms.
We still need more data in the Shikoku Basin itself, particularly on the basement
morphology and ages of the marginal wings which are overlain by thick sediment. Deep
drilling through the sediment layers to the pillow basement in the east and west margins
of the Shikoku basin will be required. Drilling of the magnetic dipoles in Kyushu-Palau
and Nishi-Shichito ridges for the purpose of determining their ages and petrography will
be extraordinarily important for elucidation of the role of arc volcanism in the initiation of
the rifting. Further comprehensive surveys of the slopes of the Nankai Trough will improve
understanding of the physical and chemical processes associated with young basin sub-
duction.
Acknowledgments
The authors would like to thank our colleagues in the Hydrographic Department,
Ocean Research Institute, University of Tokyo, and JAMSTEC for their help and encour-
agement to us to complete this chapter.
REFERENCES
Andrews, J. E. 1978. Morphological evidence for the evolution of the central basin fault spreading center in the
west Philippine basin, GSA Annual Meeting, Toronto, abstract.
Brown, G., and Taylor, B. 1988. Sea-floor mapping of the Sumisu Rift, Izu-Ogasawara (Bonin) island arc, Bull.
Geol. Surv. lpn. 39:23-38.
Burke, K. 1976. Development of graben associated with the initial rupture of the Atlantic Ocean, Tectonophysics
36:93-112.
Cande, S. c., and Kent, D. V. 1992. A new geomagnetic time scale for the Late Cretaceous and Cenozoic, l.
Geophys. Res. 97(BlO):13,917-13,952.
Charnley, H. 1980. Clay sedimentation and paleoenvironment in the Shikoku Basin since the middle Miocene
(Deep Sea Drilling Project Leg 58, north Philippine Sea), in Init. Repts. DSDP, 58 (G. deY. Klein and K.
Kobayashi, eds.), pp. 669-682, U.S. Govt. Printing Office, Washington, DC.
Chamot-Rooke, N., Renard, Y., and Le Pichon, X. 1987. Magnetic anomalies in the Shikoku Basin: a new
interpretation, Earth Planet. Sci. Lett. 83:214-218.
DeMets, C., Gordon, R. G., Argus, D. E, and Stein, S. 1990. Current plate motions, Geophys. J. Int. 101:425-478.
Dick, H. J. B. 1980. Vesicularity of Shikoku Basin basalt: a possible correlation with the anomalous depth of
backarc basins, in Init. Repts. DSDP, 58 (G. deY. Klein and K. Kobayashi, eds.), pp. 895-904, U.S. Govt.
Printing Office, Washington, DC.
Dick, H. J. B. 1982. The petrology of basaltic rocks of the northern Philippine Sea, Am. J. Sci. 282:644-700.
Dick, H. J. B., Marsh, N. G., and Bullen, T. D. 1980. Deep Sea Drilling Project Leg 58 abyssal basalts from the
Shikoku Basin: Their petrology and major-element geochemistry, in [nit. Repts. DSDP, 58 (G. deY. Klein
and K. Kobayashi, eds.), pp. 843-872, U.S. Govt. Printing Office, Washington, DC.
Echols, D. J. 1980. Foraminifer biostratigraphy, north Philippine Sea, Deep Sea Drilling Project Leg 58, in [nit.
Repts. DSDP, 58 (G. deY. Klein and K. Kobayashi, eds.), pp. 567-586, U.S. Govt. Printing Office,
Washington, DC.
Faller, A. M., Steiner, M., and Kobayashi, K. 1979. Paleomagnetism of basalts and interlayered sediments drilled
during DSDP Leg 49 (N-S transect of the northern mid-Atlantic Ridge), in [nit. Repts. DSDP, 49 (B. P.
Luyendyk, J. R. Cann, eds.), pp. 769-780, U.S. Govt. Printing Office, Washington, DC.
Furuta, T. and Arai, E 1980. Petrographic and geochemical properties of tephras in Deep Sea Drilling Project cores
from the North Philippine Sea, in Init. Repts. DSDP, 58 (G. deY. Klein and K. Kobayashi, eds.), pp. 617-627,
U.S. Govt. Printing Office, Washington, DC.
Furuta, T., Tonouchi S., and Nakada, M. 1980. Magnetic properties of pillow basalt from the Kinan seamount
chain, the Shikoku Basin, J. Geomagn. Geoelectr. 32:567-573.
Hayashida, M., Takanashi, M., Ikeda, K., Kaneko, Y., Kato, Y., Ogawa, M., and Kasuga, S. 1989. Preliminary
report of continental shelf surveys of "Middle part of Shikoku Basin" quadrangle (in Japanese), Tech. Bull.
Hydrogr. 8:85-91.
Hill, I. A., Taira, A., Firth, J. Y., et al. 1993. Proc. ODP. Sci. Results, 131, College Station TX.
Hydrographic Department, Maritime Safety Agency of Japan. 1989. Data Report of Hydrographic Observation,
Series of Continental Shelf Survey, HD-MSA, Tokyo 6.
Ingle, J. C., Karig, D. E., et al. 1975. [nit. Repts DSDP, 31, U.S. Govt. Printing Office, Washington, DC, 31.
IPOD-Japan. 1977. Multichannel seismic reflection records of the Shikoku Basin and the Daito Ridges, 1976,
Part I, IPOD-Japan Data Series no. I, Ocean Res. Inst., Univ. of Tokyo.
Kagami, H., Karig, D. E., Coulbourn, W. T. et al. 1987. [nit. Repts. DSDP, 31, U.S. Govt. Printing Office,
Washington, DC.
KAIKO I Research Group. 1986. Topography and Structure of Trenches around Japan, Univ. of Tokyo Press.
Kanamori, H., and Tsumura, K. 1971. Spatial distribution of earthquakes in the Kii peninsula, Japan, south of
the median tectonic line, Tectonophysics 12:327-342.
Karig, D. E. 1971. Origin and development of marginal basins in the western Pacific, J. Geophys. Res. 76:2542-
2561.
Karig, D. E. 1975. Basin genesis in the Philippine Sea, in [nit. Repts. DSDP, 31 (1. C. Ingle, D. E. Karig et aI., eds.),
pp. 847-879, U.S. Govt. Printing Office, Washington, DC.
Kasuga, S., Iwabuchi, H., and Kato, S.1987. Results of ocean bottom surveys injunction of the Shikoku Basin and
West Mariana Basin (in Japanese with English Abstr.), Rept. Hydrogr. Res. 22:13-134.
Kasuga, S., Kato, Y., Kimura, S., and Okino, K. 1992. Present and former members of the Continental Shelf
Surveys Office, characteristics of arc-trench systems and backarc basins in the southern waters of Japan-
Outline of the geophysical survey by the Hydrographic Department of Japan (in Japanese with English
Abstr), Rept. Hydrogr. Res. 28:9-45.
Katsura, T., Shimamura, K., and Collaborators in Continental Shelf Surveys Office. 1994. Geological, geochemi-
cal research on bottom samples, from continental shelf surveys, H.D. Japan (part I)-Preliminary study for
ocean floor on the Japanese continental shelves, Rept. Hydrogr. Res. 30:345-381.
Klein, G. deY., and Kobayashi, K. 1981. Geological summary of the Shikoku Basin and northwestern Philippine
Sea, Leg 58, DSDPIIPOD drilling results, Oceanol. Acta, Spec. No. 181-192.
Klein, G. deY., Kobayashi, K., Chamley, H., Curtis, D. M., Dick, H. J. B., Echols, D. J., Fountain, D. M.,
Kinoshita, H., Marsh, N. G., Mizuno, A., Nisterenko, G. V., Okada, H., Sloan, 1. R., Waples, D. M., and
White, S. M. 1978. Off-ridge volcanism and seafloor spreading in the Shikoku Basin, Nature 273:746-748.
Klein, G. deV., Kobayashi, K. et al. 1980. Init. Repts. DSDP, 58, U.S. Govt. Printing Office, Washington, DC, pp.
567-586.
Kobayashi, K. 1983. Cycles of subduction and Cenozoic arc activity in the northwestern Pacific margin, in
Geodynamics of the Western Pacific-Indonesian Region (T. W. C. Hilde and S. Uyeda, eds.), Geodyn. Ser.,
Vol. 11, pp. 287-302, American Geophysical Union/Geological Society of America, Washington, DC.
Kobayashi, K. 1984. Subsidence of the Shikoku backarc basin, Tectonophysics 102:105-117.
Kobayashi, K., and Fujiwara, T. 1993. Survey ofHAKUHO Seamount, in The Shikoku Basin O. Segawa, ed.), pp.
243-246, Preliminary Rept. Hakuho Maru Cruise KH92-1, OR! Univ. of Tokyo.
Kobayashi, K., and Isezaki, N. 1976. Magnetic anomalies in the Sea of Japan and the Shikoku Basin: Possible
tectonic implications, in The Geophysics of the Pacific Ocean Basin and Its Margin (G. H. Sutton, M. H.
Manghnani, and R. Moberly, eds.), Geophys. Monogr. Ser., Vol. 19, pp. 239-251, American Geophysical
Union, Washington, DC.
Kobayashi, K., and Nakada, M. 1978. Magnetic anomalies and tectonic evolution of the Shikoku inter-arc basin,
l. Phys. Earth 26:S391-S402.
Koyama, K., Katsura, T., Ikeda, K., Uchida, M., Kasuga, S., Nagano, M., and Hayashida, M. 1986. Preliminary
report of continental shelf survey of Minami-Koho Seamount and adjacent areas (in Japanese), Tech. Bull.
Hydrogr. 4:39-46.
Le Pichon, X., Iiyama, T., Charnley, H., Charvet, J., Faure, M., Fujimoto, H., Furuta, T., Ida, Y., Kagami, H.,
Lallemant, S., Leggett, 1., Murata, A., Okada, H., Rangin, C., Renard, v., Taira, A., and Tokuyama, H. 1987.
Nankai Trough and the fossil Shikoku Ridge: Results of Box 6 Kaiko survey, Earth Planet. Sci. Lett. 83:
186-198.
Luyendyk, B. P., Cann, J. R. et al. 1979. Init. Repts, DSDP, 49, U.S. Govt. Printing Office, Washington, DC.
Marsh, N. G., Sanders, A. D., Tamey, J. and Dick, J. B. 1980. Geochemistry of basalts from the Shikoku and Daito
Basins, Deep Sea Drilling Leg 58, in Init. Repts. DSDP, 58 (G. DeV. Klein, K. Kobayashi et aI., eds.), pp.
805-842, U.S. Govt. Printing Office, Washington, DC.
McKee, E. H., and Klock, P. 1980. K-Ar ages of basalt sills from Deep Sea Drilling Project Sites 444 and 446,
Shikoku Basin and Daito Basin, Philippine Sea, in Init. Repts. DSDP, 58 (G. DeV. Klein, K. Kobayashi et aI.,
eds.), pp. 921-922, U.S. Govt. Printing Office, Washington, DC.
Mizuno, A., Shibata, K., Uchiumi, S., Yuasa, M., Okuda, Y., Nohara, M., and Kinoshita, Y. 1977. Granodiorite
from the Minami-Koho Seamount on the Kyushu-Palau Ridge and its KfAr age, Bull. Geol. Surv. lpn.
28:507-511.
Mrozowski, C. L., and Hayes, D. E. 1979. The evolution of the Parece Vela Basin, eastern Philippine Sea, Earth
Murauchi, S., Den, N., Asano, S., Hotta, H., Yoshii, T., Asanuma, T., Hagiwara, K., Ichikawa, K., Sato, T.,
Ludwig, W. J., Ewing, J., Edgar, N. T., and Houtz, R. E. 1968. Crustal structure of the Philippine Sea, l.
Geophys. Res. 73:3143-3171.
Nakada, S., and Takahashi, M. 1979. Regional variation in chemistry of the Miocene intermediate to felsic
magmas in the outer zone and the Setouchi province of southwest Japan, l. Geol. Soc. lpn. 85:571-582.
Okada, H. 1980. Calcareous nannofossils from Deep Sea Drilling Project Sites 442 through 446, Philippine Sea, in
Init. Repts. DSDP, 58 (G. DeV. Klein, K. Kobayashi et aI., eds.), pp. 549-565, U.S. Govt. Printing Office,
Washington, DC.
Okino, K. 1993. Formation processes of a backarc basin-Shikoku Basin (in Japanese), Abstr. Papers presented at
the 10th Shinkai Symp., JAMSTEC, pp. 23-25.
Okino, K., Shimakawa, Y., and Nagaoka, S. 1994. Evolution of the Shikoku Basin, l. Geomagn. Geoelectr. 46,
Spec. Issue: 463-479.
Olafsson, G. 1993. Calcareous nannofossil biostratigraphy of the Nankai Trough, in Proc. ODP, Sci. Results, 131
(I. A. Hill, A. Taira, 1. V. Firth, eds.), pp. 15-34, Ocean Drilling Program, College Station, TX.
Ozima, M., Kaneoka, I., and Ujiie, H. 1977. 40Ar- 39Ar age of rocks and development mode of the Philippine Sea,
Nature 267:816-818.
Ozima, M., Takigami, Y., and Kaneoka, I. 1980. 4OAr_ 39 Ar geochronological studies on rocks of Deep Sea Drilling
Project Sites 443, 445, and 446, in Init. Repts. DSDP, 58 (G. DeV. Klein, K. Kobayashi, eds.), pp. 917-920,
U.S. Govt. Printing Office, Washington, DC.
Park, C.-H., Tamaki, K., and Kobayashi, K. 1990. Age-depth correlation of the Philippine Sea backarc basins and
other marginal basins in the world, Tectonophysics 181:351-371.
Pickering, K. T., Underwood, M. B, and Taira, A 1993. Stratigraphic synthesis of the DSDP-ODP sites in the
Shikoku Basin, Nankai Trough, and accretionary prism, in Proc. ODP, Sci. Results, 131 (I. A. Hill, A Taira,
and 1. V. Firth, eds.), pp. 313-330, Ocean Drilling Program, College Station, TX.
Shih, T. C. 1980. Magnetic lineations in the Shikoku Basin, in [nit. Repts, DSDP, 58 (G. DeV. Klein and K.
Kobayashi, eds.), pp. 783-788, U.S. Govt. Printing Office, Washington, DC.
Shino, M., Ikeda, K., Tozaki, T., Nagaoka, S., Kato, Y., Shimakawa, Y., and Seta, H. 1991. Preliminary report of
continental shelf surveys of "Kosyu Seamount" quadrangle (in Japanese), Tech. Bull. Hydrogr. 9:
26-31.
Shiono, S., and Sugi, N. 1985. Life of an oceanic plate: Cooling time and assimilation time, Tectonophysics
112:35-50.
Siena, E, Coltorti, M., Saccani, E., and Vaccaro, C. 1993. Petrology of the basaltic rocks of the Nankai Trough
basement, in Proc. ODp, Sci. Results, 131 (I. A. Hill, A Taira, and J. Y. Firth, eds.), pp. 197-210, Ocean
Sugi, N., and Uyeda, S. 1984. Subduction of young oceanic plates without deep focus earthquakes, Bull. Soc. Geol.
France 26:245-254.
Taira, A., Hill, I. A., Firth, J. Y., et al. 1991. Proc. ODp, [nit. Repts., 131, Ocean Drilling Program, College Sta-
tion, TX.
Taira, A, and Niitsuma, N. 1986. Turbidite sedimentation in the Nankai Trough as interpreted from magnetic
fabric, grain size, and detrital modal analyses, in [nit. Repts. DSDP, 87 (H. Kagami, D. E. Karig, and W. T.
Coulbourn, eds.), pp. 611-632, U.S. Govt. Printing Office, Washington, DC.
Tamaki, K., Suyehiro, K., Allan, 1., McWilliams, M., et al. 1992. Proc. ODP, Sci. Results, 1271128, Pt. 2, Ocean
Drilling Program, College Station, TX, pp. 779-1478.
Taylor, B. 1992. Rifting and the volcanic-tectonic evolution of the Izu-Bonin-Mariana arc, in Proc. ODp, Sci.
Results, 128 (B. Taylor and K. Fujioka, eds.), pp. 627-651, College Station, TX.
Tomoda, Y. 1974. Reference Bookfor Gravity, Magnetic and Bathymetric Data of the Pacific Ocean and Adjacent
Seas, 1963-71, Univ. of Tokyo Press.
Tomoda, Y., Kobayashi, K., Segawa, J., Nomura, M., Kimura, K., and Saki, T. 1975. Linear magnetic anomalies in
the Shikoku Basin, northeastern Philippine Sea, J. Geomagn. Geoelectr. 28:47-56.
Tomoda, Y., Ozawa, K., and Segawa, J. 1968. Measurement of gravity and magnetic field on board a cruising
vessel, Bull. Ocean Res. [nst. Univ. of Tokyo 3: 1-170.
Underwood, M. B., Orr, R., Pickering, K. T., and Taira, A. 1993a. Province and dispersal patterns of sediments in
the turbidite wedge of Nankai Trough, in Proc. ODP, Sci. Results, 131 (I. A. Hill, A Taira, and J. Y. Firth,
eds.), pp. 15-34, Ocean Drilling Program, College Station, TX.
Underwood, M. B., Pickering, K., Gieskes, J. M., Kastner, M., and Orr, R. 1993b. Sediment geochemistry, clay
mineralogy, and diagenesis: A synthesis of data from leg 131, Nankai Trough, in Proc. ODp, Sci. Results, 131
(I. A. Hill, A. Taira, and J. Y. Firth, eds.), pp. 343-364, Ocean Drilling Program, College Station, TX.
Van Andel, T. H., Heath, G. R., and Moore, T. c., Jr. 1975. Cenozoic history and paleooceanography of the central
equatorial Pacific Ocean, Geol. Soc. Am., Mem. 143.
Watts, A. B. 1976. Gravity field of the northwest Pacific Ocean basin and its margin: Philippine Sea, Geol. Soc.
Am. MC-12.
Watts, A. B., and Weissel, J. 1975. Tectonic history of the Shikoku marginal basin: Philippine Sea, Earth Planet.
Sci. Leu. 25:315-328.
White, S. M., Charnley, H., Curtis, D., Klein, G. DeY., and Mizuno, A, with a special contribution on physical
properties by D. M. Fountain. 1980. Sediment synthesis: Deep Sea Drilling Project Leg 58, Philippine Sea, in
[nit. Repts. DSDP, 58 (G. DeY. Klein, and K. Kobayashi, eds.), pp. 963-1014, U.S. Govt. Printing Office,
Washington, DC.
Wood, D. A., Joron, J. L., Marsh, N. G., Tamey, J., and Treuil, M. M. 1980. Major- and trace-element variations in
basalts from the north Philippine Sea drilled during Deep Sea Drilling Project Leg 58: A comparative study of
backarc basin basalts with lava series from Japan and mid-oceanic ridges, in [nit. Repts. DSDP, 58 (G. DeV.
Klein, and K. Kobayashi, eds.), pp. 873-894, U.S. Govt. Printing Office, Washington, DC.
Yamano, M., Honda, S., and Uyeda, S. 1984. Nankai Trough: A hot trench? Mar. Geophys. Res. 6:187-203.
Yamano, M., Foucher, 1.-P., Kinoshita, M., Fisher, A., Hyndman, R. D., and ODP Leg 131 Shipboard Scientific
Party. 1992. Heat flow and fluid flow regime in the western Nankai accretionary prism, Earth Planet. Sci. Leu.
109:451-462.
11
Opening Tectonics of the Japan Sea

Kensaku Tamaki
ABSTRACT
The examination of the crustal structure and ODP deep-sea drilling results introduced the
following opening model of the Japan Sea. The opening of the Japan Sea was initiated by
the extension and thinning of the proto-Japan island arc, which was situated on the margin
of the Eurasian continent at ca. 30 Ma. During the extension process a large strike-slip fault
was generated at the eastern margin of the proto-Japan Sea. The strike-slip movement cut
through the entire lithosphere and caused it to split, triggering development of a seafloor
spreading system at ca. 28 Ma. The seafloor spreading system propagated to the WSW
direction into the thinned crustal zone and formed the eastern part of the Japan Basin that is
presently underlain by oceanic crust. In the meantime, the southwestern part of the Japan
Sea suffered continuing extension and thinning of the arc crust and formed basins and rises.
The basins are composed of thinned lower arc crust overlain by volcanic layers, and the
rises are composed of fragmented upper arc crusts. The propagation ceased at ca. 18 Ma,
leaving the contrasting topography of the present northern and the southern Japan Sea. The
extension and thinning of arc or continental crusts and subsequent development of a
propagating spreading system, which is initiated at a basin margin strike-slip zone, are
common and fundamental processes of most backarc basins.
1. INTRODUCTION
The backarc basins are divided into two types in terms of their occurrence. One group
is the oceanic type that develops at the backarc side of oceanic island arcs such as the
Marina Trough of the Mariana arc and the Lau Basin of the Tonga arc. The other group is
the continental type that develops at the backarc side of continental arc. The Japan Sea is
one of the most typical backarc basins of the continental type; another may be the South
China Sea. A 200-m contour line surrounding the Japan Sea shows a closed depressional
feature (Fig. 11.1). The Japan Sea is characterized by a complicated seafloor topography
with many ridges, rises, banks, and seamounts. The Japan Sea is the most topographically
complicated marginal basin of the world.
Kensaku Tamaki • Ocean Research Institute, University of Tokyo, Nakano-ku, Tokyo 164, Japan.
407
408 KENSAKU TAMAKI
FIGURE 11.1. Topography of the Japan Sea. Solid circles are DSDP and ODP drilling sites with site number.
The cause of the formation of backarc basins has not been answered, in spite of
advances in our understanding of seafloor spreading theory and plate tectonics. General
observations of backarc basin geology and geophysics suggest that backarc basins are
formed under an extensional tectonic condition at the continental margin. However, one
major enigma on the formation of backarc basins is why the extensional basins are formed
at the plate convergent margin. Another major enigma is why the backarc basins are mostly
distributed in the western Pacific and why they are not associated with all island arcs. The
origin and the opening processes of the backarc basins are better understood because of the
recent data accumulation by Ocean Drilling Program (ODP) deep-sea drilling in several
western Pacific backarc basins. In this chapter I summarize principal aspects of geology
and geophysics of the Japan Sea and discuss the opening process of the Japan Sea by
introducing a model for the opening of the Japan Sea. I further discuss the fundamental
aspects of the introduced model by applying it to other backarc basins.
2. TOPOGRAPHY
The physiographic contrast is striking between the southern and northern parts of the
Japan Sea (Figs. 11.1 and 11.2). The southern part is characterized by a complicated
topography with several rises and intervening basins. The northern part shows a rather
simple morphology with a wide, deep-sea basin. This striking contrast is closely related to
the opening tectonics of the Japan Sea.
OPENING TECTONICS OF THE JAPAN SEA 409
130' 135' 140'
FIGURE 11.2. Shaded relief topography of the Japan Sea. Minor topographic features are discriminated.
The wide, deep basin in the northern part of the Japan Sea is called the Japan Basin.
The eastern part of the Japan Basin is an abyssal plain that is the result of turbidite
deposition; the sediment thickness there reaches 2500 m. In contrast, the western part ofthe
Japan Basin has shallower water depth and slightly rougher seafloor topography (Fig. 11.2);
the sediment thickness there is accordingly thinner, 1000-1500 m. The contrast between the
western and eastern parts of the Japan Basin is significant to our understanding of its origin.
There are many topographic rises in the southern part of the Japan Sea. Larger ones are
the Yamato Rise in the center of the Japan Sea and the Korea Plateau at the east of the Korea
peninsula. Abundant granitic rocks recovered from these and surrounding topographic
highs suggest that these rises comprise continental crust (e.g., Lelikov and Bersenev, 1973;
Tamaki, 1988). The oldest rock is the Archean gneiss recovered from the Korea Plateau
with the absolute age of 2729 Ma (Tamaki, 1988). Seismic refraction results also suggest
the existence of a layer with seismic velocity of 6.0 km/s, suggesting continental crust
(Ludwig et al., 1975). Granitic rocks were recovered from most other topographic highs
in the southern part of the Japan Sea.
Two sedimentary basins are distributed between these topographic highs. The Yamato
Basin is between Honshu Island and the Yamato Rise. The Tsushima (Ullung) Basin is at
the southwestern corner and surrounded by the Korea Plateau, the Korea Peninsula, and
Honshu Island. The water depth of the Yamato Basin is 3000 m, which is 500 m shallower
than that of the Japan Basin. The sediment thickness of the Yamato Basin is 500 m thinner
than that of the Japan Basin. Accordingly, the basement depth of the Yamato Basin is 1000-
410 KENSAKU TAMAKI
1500 m shallower than that of the Japan Basin. Although the data are not sufficient in the
Tsushima Basin, the sediment thickness of the Tshushima Basin is estimated to be over
2000 m and the basement depth appears to be comparable with that of the Yamato Basin.
3. THE JAPAN AND YAMATO BASINS
The topography is closely related to the crustal structure of the Japan Sea. The crustal
structures present information critical to understanding the basin formation tectonics. The
crust of the Japan Sea has been intensively studied by seismic refraction methods since the
1950s. The volume of crustal data is the largest for any of the backarc basins of the world;
hence the Japan Sea presents us the best opportunity to study basin formation tectonics. In
the last several years, techniques for using long arrays of ocean bottom seismometers were
introduced for studies of detailed crustal structure in the basin areas (e.g., Hirata et al., 1987,
1992). During ODP Leg 128, a borehole seismometer was used for the first time for actual
crustal study of an ocean bottom hole (Shinohara et aI., 1992).
Figure 11.3 summarizes principal results of these studies in the Japan Sea. The figure
clearly displays the differences between the crustal structure of the eastern part of the Japan
Basin and that of the Yamato Basin. The crustal thickness of the Japan Basin is 6-6.5 km,
while that of the Yamato Basin is 10-13 km, twice as thick as the Japan Basin. Regardless
of the difference of crustal thickness, the seismic propagation velocity of each layer is
identical between the two basins. Below the sedimentary layer the crustal structure with
seismic velocity shows that the crust of the eastern part of the Japan Basin is a typical
oceanic crust. Magnetic anomaly lineations are also identifiable in the eastern part of
the Japan Sea, and the tentatively identified age is consistent with the age of the crust
estimated on the basis of heat flow data (Tamaki et aI., 1990). Thus, it is concluded that the
eastern part of the Japan Basin was formed by the normal seafloor spreading system.
Yarnato Basin Japan Basin Japan Arc

Southern Northern (Honshu)
Lum~;laL Murauchi Hirata etal.
(1972) (1991)
1.5 1.5
~1.7
~3.0
10
:-:6,0:':
10 10
8.1 8.2
14 14
18 7.8 18
(km) (krn/s)
(km)
Site 794
I D Water ~ Sediments • Layer 2 6J Layer 3 .................. , Moho I
FIGURE 11.3. Comparison of the crustal structure of the eastern Japan Basin and the Yamato Basin.
In contrast, the origin of the thicker crust of the Yamato Basin is not simple. One
possible explanation is that it is oceanic crust with thickness twice that of normal oceanic
crust (Hirata et at., 1989). The thickness of oceanic crust positively correlates to the
temperature of upwelling upper mantle beneath the spreading center (McKenzie and
Bickle, 1988). Thick oceanic crust is usually caused by the contamination of the spreading
center and hot spots associated with high-temperature mantle. However, it is difficult to
assume hot-spot interaction for the formation of the Japan Sea for the following reasons.
The oceanic crust, observed at the Yamato Basin, is twice as thick as the normal
oceanic crust and is comparable to those of oceanic plateaus (Schubert and Sandwell, 1989)
and suggests activity of a large hot spot. Present heat flow in the Yamato Basin, however,
does not show any appreciable anomaly indicating previous hot-spot activity (Langseth and
Tamaki, 1992). The average heat flow value of the Yamato Basin is actually 10 mW/m2
lower than that of the Japan Basin, which has normal oceanic crust. It is implausible to
introduce hot-spot interaction for the formation of the Yamato Basin in this regard.
The other idea for the origin of the Yamato Basin is that it is extended continental
crust. The seismic velocity, 6.2-7.2 km/s, of the lower crust of the Yamato Basin appears to
be identical to that of the lower crust of Honshu Island of the Japanese island arc in Fig.
11.3. The thickness of the lower crust of Honshu Island is twice that of the Yamato Basin.
The hypothesis of extended continental crusts as the origin of the Yamato Basin crust
assumes that the lower crust of the Yamato Basin is identical to that of the Japanese island
arcs. According to the hypothesis, the lower crust of Honshu Island, with a thickness of
20 km, was extended and thinned to 10 km during the opening process of the Japan Sea to
form the present Yamato Basin. The initial thickness of the lower crust of the proto-
Japanese island arcs may have been thicker than the present 20 km, and then, in this case,
the ratio of extension needs to be increased.
In contrast to the case of the lower crust, the seismic velocities of the upper crust are
quite different for the Yamato Basin and for Honshu Island. The upper crust velocity of
Honshu Island is 6.0 km/s, a typical seismic velocity of the upper crust of a continent, while
that of the Yamato Basin is less than 6.0 km/s. The upper crust of the Yamato Basin has a
velocity of 3.6-5.4 km/s, nearly identical to the upper crust of the Japan Basin. If we
assume the extensional origin of the Yamato Basin, the above problem on the origin of
upper crust of the Yamato Basin should be answered consistently with the extension
tectonics of the lower crust. To approach this problem, we introduce a whole Japan Sea
crustal distribution map in the next section.
4. EXTENSION OF ISLAND-ARC CRUST
Fig. 11.4 shows a compilation of crustal structures in the Japan Sea based on seismic
refraction studies, basement topography, and bottom rock sampling (Tamaki et at., 1992).
Abundant results of seismic refraction studies by Russian scientists were included in the
compilation. Crusts of the Japan Sea are classified into four groups in Fig. 11.4. They are
continental crust, rifted continental crust, extended continental crust, and oceanic crust. The
Yamato Basin-type crust was denoted as extended continental crust in the compiled map.
Topographic highs such as the Yamato Rise, which were not appreciably thinned by
extension, are classified as continental crusts. Rifted continental crust represents the
intermediate condition between extended crust and nonextended crust. Rifted continental
412 KENSAKU TAMAKI
FIGURE 11.4. Crustal structure of the Japan Sea. The discrimination of the crustal structure is based on seismic
reflection/refraction survey data, bottom sampling data, geomagnetic data, and basement depth and topography.
crust, which is distributed surrounding the continental crusts, is characterized by the

occurrence of many normal faults. The region of rifted continental crust represents deeper
water depths than that of continental crust, suggesting submergence by some extension and
thinning of the crust.
Characteristics of the distribution of the variable crusts in the Japan Sea are summa-
rized as follows.
1. Distribution of the oceanic crusts is restricted to the eastern part of the Japan Basin. The
region of the oceanic crust widens to the east and narrows to the west in the western Japan
Basin.
2. Extended continental crust is widely distributed in the Japan Sea, in contrast to the limited
distribution of the oceanic crust. Distribution in the southern part of the Japan Sea is
predominant.
3. Blocks of continental crust are distributed mostly in the southern part of the Japan Sea.
Restricted distribution of the oceanic crust in the Japan Sea is one of the most
outstanding features of Fig. 11.4, which is different from the previously accepted idea that
the oceanic crust is widely distributed in the Japan Sea. The predominant distribution of the
extended continental crust suggests that the extension and thinning of the continental crust
have an important role for the formation of the Japan Sea.
The southern part of the Japan Sea is characterized by the basins that are formed by the
extended continental crust and by the rises that are the continental blocks. These charac-
teristics introduce the following hypothesis: proto-Japanese island arcs, which were at the
continental margin of the Eurasian continent, suffered a differential deformation between
the upper crust and the lower crust under extensional tectonics. The lower crust was
extended and thinned by ductile deformation, while the upper crust was collapsed by brittle
deformation. The fragmented upper crusts were left as rises in the basins, while the
extended lower crust formed basins and lacked upper crusts. The upper mantle was
upheaved by the thinning of the crust and produced magma by partial melting of the upper
mantle. Seismic refraction studies show that the upper crust of the Yamato Basin has a
velocity of oceanic layer 2 (basaltic effusive and intrusive rocks) and basinal ubiquitous
distribution of this layer. The upper crust of the Yamato Basin is hypothesized to be formed
by basinal volcanism during extension tectonics. The extended lower crust in the basins that
lack upper crust was thus covered by volcanic rocks. A cross-sectional reconstruction of the
southern Japan Sea in Fig. 11.5 shows this model. We need more detailed crustal data,
especially beneath the rises, to improve the above hypothesis.
5. FORMATION AGE
ODP Legs 127 and 128 successfully recovered basaltic rocks beneath the sediments
and provided reliable information on the formation age of the Japan Sea (Tamaki et ai.,
1992). The basement volcanic rocks were recovered at ODP Site 795 in the northern Japan
Basin, at Site 794 in the northern Yamato Basin, and at Site 797 in the southern part of the
Yamato Basin (Fig. 11.1). Two sets of data are available on age information; one is 40Ar_39Ar
radiometric age determination measured by Kaneoka et al. (1992), and the other is the age
of sediments overlying the uppermost part of the basement volcanic rocks. The latter age is
determined by using microfossils such as foraminifera, diatoms, and pollen. Tada and
- -
FIGURE 11.5. Cross section of the evolution of the southern Japan Sea (left), and northern Japan Sea (right).
414 KENSAKU TAMAKI
Tamaki (1992) compiled and evaluated the analyzed results of microfossils and determined
the fossil age of the uppermost volcanic rocks. Table I summarizes the radiometric ages and
fossil ages of the uppermost volcanic rocks at the three drilling sites.
Table I shows that the fossil age is always younger than the radiometric age. The
difference between the radiometric age and fossil age at Site 795 is interpreted as due to
an unconformity between the sediments and underlying volcanic basement that is observed
in seismic reflection profiles (Tamaki et al., 1990). The difference between radiometric and
fossil ages at site 794, which forms a slightly shallower topography than the major part
of the Yamato Basin, is also explained as an unconformity (Tamaki et al., 1992). The
radiometric and fossil ages are consistent at Site 797 where most volcanic layers are sills
intruded into the sediments of concurrent deposition (Tamaki et ai., 1990). In conclusion,
radiometric ages of volcanic rocks are considered to represent the age of eruption of ba-
salts in the basins.
The oldest age among three sites is 24 Ma in the Japan Basin. The age of 24 Ma is
consistent with the preliminary identification of magnetic anomalies in the eastern part
of the Japan Basin. An older radiometric age of 32 Ma is also reported from the basaltic
rocks recovered at the outcrops along the Okushiri Ridge in the eastern margin of the Japan
BaSin (Kaneoka, personal comm.); this information suggests that the age of Japan Basin
volcanism is older than that of Yamato Basin volcanism. The radiometric age at Site 797
in the western Yamato Basin is I to 3 m.y. younger than that at Site 794 in the eastern
Yamato Basin.
6. MODEL OF OPENING OF THE JAPAN SEA
The existence of oceanic crust in the eastern Japan Basin indicates that the mid-
oceanic-type seafloor spreading tectonics actually happened and that the history of seafloor
spreading tectonics should have been preserved and observed as magnetic anomaly linea-
tions. Preliminary identification of magnetic anomalies in the eastern Japan Basin proposes
the age range of the crust of the Japan Basin as 28-18 Ma (Tamaki et aI., 1992). The age
range of 28-18 Ma includes a part that is older than that of the Yamato Basin.
Critical information comes from the configuration of magnetic anomaly lineations in
the eastern Japan Basin. A clear oblique displacement is observed in the magnetic anomaly
lineations (Tamaki et ai., 1992). The trend of displacement is NW-SE, and the overall
feature suggests WSW propagation of the spreading system in the eastern Japan Basin.
TABLE I
Age of Basement Volcanic Rocks Recovered by ODP
Radiometric age" Fossil ageb

ODP Sites (Ma) (Ma)
Site 795 (Japan Basin) 17-24 14-15.5

Site 794 (Yamato Basin) 20-21 16-18
Site 797 (Yamato Basin) 18-19 17-19
aKaneoka et al. (1992).
"Tada and Tamaki (1992).
The WSW propagation of the spreading system is consistent with the westward narrowing
of the oceanic crust in the Japan Basin as shown in Fig. 11.4.
Jolivet and Tamaki (1992) stressed the importance of the described propagation
tectonics and introduced an idea that the oceanic crust of the Japan Basin was initiated from
the eastern margin of the Japan Basin and propagated to the westward (Fig. 11.6). A N-S
trending large strike-slip fault was assumed at the eastern margin of the Japan Basin. The
strike-slip fault was generated in the margin of the basin to accommodate the strike-slip
movement in accordance with the pull-apart-type opening of the Japan Sea. Because strike-
slip faults form the weakest line of the lithosphere (Mueller and Phillips, 1991), the initial
split of lithosphere was assumed to be formed along the strike-slip fault. The split of
lithosphere evolved to seafloor spreading, and the seafloor spreading system propagated
into the extended and thinned arc crusts of the southwestern Japan Sea.
The hypothesized opening process of the Japan Sea is summarized as follows.
1. The Japan Sea began forming by the extension and thinning of the proto-Japan arc that was at
the margin of the Eurasian continent. The timing of initiation of extension is estimated to be
before 30 Ma, based on the reconstruction of the basin margin's subsidence history of the
Japan Sea (Ingle, 1992).
2. A split of the lithosphere was initiated along the strike-slip fault at the eastern margin of the
proto-Japan Sea. Seafloor spreading was initiated from the split and started propagating to the
west into the thinned arc crustal region. The initiation of the seafloor spreading is estimated
to be at 28 Ma, based on the oldest magnetic anomaly lineations in the Japan Basin.
3. The propagation of the seafloor spreading system lasted till 18 Ma and formed the oceanic
crust in the western Japan Basin. In contrast, the southwestern part of the Japan Sea was
opened in association with the extension and thinning of the arc crust, and the resultant
topography with rises and basins topography was formed during this period.
Thus, the contrasting topography in the northern and southern Japan Sea and the
characteristic crustal distribution of the Japan Sea are well explained by our opening model
28 -18 Ma
FIGURE 11.6. Simplified cartoon of opening of the Ja-

pan Sea (modified from Jolivet and Tamaki, 1992). Crus-
tal thinning and extension prevailed at the initial stage.
Seafloor spreading was triggered by the breakup of the
lithosphere along the strike-slip margin at the eastern side
of the Japan Sea. The spreading center propagated south-
westward to increase the area of the oceanic crust. In the
meantime the crust of the southern Japan Sea was being
extended and thinned.
416 KENSAKU TAMAKI
of the Japan Sea. Fig. 11.5 shows contrasting tectonics in the southern and northern Japan
Sea. The principal processes of this model are the subsequent development of the extension
and thinning of the arc crust, a split of the lithosphere along the strike-slip fault in the basin
margin, initiation of seafloor spreading, and propagation of the spreading system into the
thinned crustal region. This process appears to be fundamental for the backarc basin
formation as discussed in the next section.
7. COMMON PROCESS OF BACKARC BASIN OPENING
The opening process of the Japan Sea introduced in the previous chapters appears to be
applicable to the Lau, North Fiji, South China Sea, and Shikoku basins, too (Fig. 11.7).
ODP Leg 135 drilled seven sites along a transect in the Lau Basin (Leg 135 Scientific
Party, 1992). The Lau Basin was considered, before ODP drilling, to be formed by a simple
seafloor spreading system. Drilling results, however, showed that the western part of the
basin is composed not of oceanic crust but of extended and thinned island-arc crust. A
detailed geophysical survey for ODP drilling revealed a southward propagating spreading
center in the eastern part of the Lau Basin (Fig. 11.7a). There is a large strike-slip fault at the
northern margin of the Lau Basin. The fault is a boundary between the Pacific plate and the
Lau Basin (the Australia plate). The area of the oceanic crust widens to the north,
approaching the strike-slip fault as shown in Fig.ll.7a. The distribution of the oceanic crust
and the extended crust shows a pattern similar to that of the Japan Sea. It is assumed that
(a) Lau Basin
20
24
(d) South China Sea

180" 172"W
(c) North Fiji Basin
20"
~ Oceanic crust
110"
FIGURE 11.7. Opening process of the Lau Basin (a), Shikoku Basin (b), North Fiji Basin (c), and South China
Sea (d).
the Lau Basin was initially opened by extension of island-arc crust, and seafloor spreading
was triggered at the strike-slip fault in the northern basin margin. The seafloor spreading
subsequently propagated southward into the thinned arc crust region.
The present configuration of the Shikoku Basin widens to the north (Fig. 1l.7b).
Magnetic anomaly lineations are clearly identified in the Shikoku Basin (Okino et aI.,
1994). The oldest magnetic anomalies along the western margin of the Shikoku Basin show
a systematic variation, older to the north and younger to the south. This variation is the
unequivocal result of the southward propagation of the spreading system of the Shikoku
Basin. The Nankai Trough subduction zone bounds the northern margin of the Shikoku
Basin. As the Nankai Trough is supposed to have been originally a strike-slip fault,
initiation of a spreading system at the strike-slip fault is again introduced. The extension
of the Bonin arc during the opening of the Shikoku Basin is well documented by seismic
reflection profiling and ODP drilling (Taylor, 1992). The Bonin arc widens and deepens to
the south, suggesting more intensive extension in the southern Bonin arc than in the
northern Bonin arc. This variation is consistent with the southward propagation of the
Shikoku Basin seafloor spreading system. When the seafloor spreading intruded into
the area, the extension of arc crust is degraded and ceases. The southern part of the Bonin
arc underwent a longer period of extension tectonics than the northern part and resulted in
the present along-axis variation of water depths and widths. Thus, the Shikoku Basin and
the Bonin arc also show a combination of propagation of seafloor spreading apart from a
strike-slip fault, and extension of island-arc crust.
The accumulation of marine geophysical data is outstanding in the North Fiji Basin
during the past several years, especially by the cooperative France-Japan project (e.g.,
Tanahashi et aI., 1991). Tanahashi et al. (1991) detected northward propagation of the
currently active spreading center by a swath mapping survey (Fig. 11.7c). The magnetic
anomaly lineations associated with the spreading system widen to the south, suggesting
broader distribution of oceanic crust generated by the spreading activity. A large strike-slip
fault, the Hunter fracture zone, bounds the southern margin of the spreading system. The
configuration above also suggests an initiation of seafloor spreading at the strike-slip fault
zone and subsequent propagation to the north. The crustal structure of the area surrounding
the presently active spreading system, however, appears to have been formed by another
event of seafloor spreading (Auzende et al., 1988). Then, in the case of the North Fiji Basin,
the current propagation is considered to be propagating into the previously formed oceanic
crust (Fig. 11.7c).
The last example is the South China Sea. The South China Sea shows a configuration
similar to the Japan Sea. The topography and the distribution of magnetic anomaly
lineations suggest that the zone of oceanic crust widens to the east (Fig. 11.7d). The Manila
Trench bounds the eastern margin of the zone of oceanic crust. As the trench subduction
zone is preferentially formed along the strike-slip fault (Mueller and Phillips, 1991), it is
assumed that the eastern margin of the South China Sea was initially bounded by a strike-
slip fault. Detailed magnetic anomaly data show a westward propagation of a past spread-
ing system (Ishihara et aI., 1994). Abundant banks are distributed in the southwestern part
of the South China Sea. Since the South China Sea is an intracontinental extensional basin,
the zone of banks is supposed to be the result of the extension of continental crust; the
mechanism is similar to that of the southern Japan Sea. This situation, again, begins the
basin opening by thinning of continental crust and initiation and propagation of the
spreading system at a strike-slip margin.
418 KENSAKU TAMAKI
Among the foregoing cases, the Lau Basin, the Shikoku Basin, the South China Sea,
and the Japan Sea show a common process of opening. Opening of the backarc basin is
started by the extension and thinning of island-arc or continental crust. A strike-slip fault is
generated at a basin margin during the extension tectonics. The lithosphere splits at the
strike-slip fault and a seafloor spreading system appears, which propagates into the thinned
crustal region to generate oceanic crustal region. Starting the propagation at a margin
strike-slip fault is also well documented in the North Fiji Basin, suggesting the common
setting of the initiation of propagation in backarc basins.
The same scenario may be applicable to other backarc basins. The configuration of the
Mariana Trough suggests an ongoing northward propagation. The Okinawa Trough appears
to have started splitting the thinned lithosphere at its southwestern margin (Oshida et aI.,
1992).
8. CONCLUSIONS
The opening of the Japan Sea and the other backarc basins is described in the
following simplified model (Fig. 11.8).
I. Backarc basins commence opening by an extension and thinning of island arc crust
(Fig. 11.8a).
2. A large strike-slip fault develops at the margin of the basin and triggers a split of the
lithosphere and the subsequent initiation of seafloor spreading (Fig. 11.8b).
3. Seafloor spreading propagates into the thinned crustal region and forms an oceanic crustal
region (Fig. 11.8c).
This model recognizes that the extension and thinning of island-arc crust are ubiqui-
tous phenomena during the opening of backarc basins. This means that the island-arc crust
is easily extended and thinned because of its hot and weak condition. The extension,
however, does not develop to seafloor spreading without the help of a strike-slip fault. The
strike-slip fault forms the weakest lithospheric zone and initiates the seafloor spreading
process by a lithosphere split. Once the seafloor spreading process commences, the spread-
ing system propagates easily into the thinned and weakened crusts. The resultant configura-
tion of backarc basins is a composite structure of oceanic crust, extended arc crust, and
continental crust. The Japan Sea is most typically formed by these processes, and its
extensive data base of crustal structure and geology will provide an opportunity to further
improve the proposed model.
~ ~
C1I C1I
.lI:
~
~ 0
CD u..
a b c
FIGURE 11.8. A common process of opening of backarc basins: (a) crustal extension, (b) breakup oflithospere at
margin strike-slip fault, (c) propagation of seafloor spreading.
Acknowledgments
I wish to thank B. Taylor of SOEST, University of Hawaii, for his encouragement and
review of this work and P. Jarvis of Ocean Research Institute, University of Tokyo, for his
comments on the manuscript.
REFERENCES
Auzende, J., Lafoy, Y., and Marsset, B. 1988. Recent geodynamic evolution of the north Fiji basin (southwest
Hirata, N., Karp, B. Ya., Yamaguchi, T., Kanazawa, T., Suyehiro, K., Kasahara, J., Shiobara, H., Shinohara, M.,
and Kinoshita, H. 1991. Seismic structure of the Japan Basin in the Japan Sea by the 1990 Japan-USSR
expedition, Abstract, IUGG 20th Symp. U11.
Hirata, N., Karp, B. Ya., Yamaguchi, T., Kanazawa, T., Suyehiro, K., Kasahara, 1., Shiobara, H., Shinohara, M.,
and Kinoshita, H. 1992. Oceanic crust in the Japan Basin of the Japan Sea by the 1990 Japan-USSR
expedition, Geophys. Res. Lett. 19:2027-2030.
Hirata, N., Kinoshita, H., Suyehiro, K., Suyemasu, M., Matsuda, N., Ouchi, T., Katao, H., Koresawa, S., and
Nagumo, S. 1987. Report on DELP 1985 Cruises in the Japan Sea, Part II: seismic refraction experiment
conducted in the Yamato Basin, Southeast Japan Sea, Bull. Earthquake Res. Inst. Univ. Tokyo 62:347-365.
Hirata, N., Tokuyama, H., and Chung, T. W. 1989. An anomalously thick layering of the crust of the Yamato Basin,
southeastern Sea of Japan: the final stage of backarc spreading, Tectonophysics 165:303-314.
Ingle, J. C. Jr. 1992. Subsidence of the Japan Sea: evidence from ODP sites and onshore sequence, in Proc. ODp,
Sci. Results, 1271128, (K. Tamaki, K. Suyehiro, J. Allan, M. McWilliams et al., eds.), pp. 1197-1218, Ocean
Ishihara, T., et al. 1994. Magnetic anomaly map of East Asia, Magnetic Anomaly Map, Geological Survey of
Japan, 2.
Jolivet, L., and Tamaki, K. 1992. Neogene kinematics in the Japan Sea region and volcanic activity of the NE-
Japan arc, in Proc. ODp, Sci. Results, 127/128, Pt. 2 (K. Tamaki, K. Suyehiro, J. Allan, M. McWilliams et aI.,
Kaneoka, I., Takigami, Y., Takaoka, N., Yamashita, S., and Tamaki, K. 1992. 4OAr_39Ar analysis of volcanic rocks
recovered from the Japan Sea floor: Constraints on the formation age of the Japan Sea, in Proc. ODP, Sci.
Results, 1271128, Pt. 2 (K. Tamaki, K. Suyehiro, 1. Allan, M. McWilliams et al., eds.), pp. 819-836, Ocean
Langseth, M., and Tamaki, K. 1992. Geothermal measurements: Thermal evolution of the Japan Sea basins and
sediments, in Proc. ODP, Sci. Results, 127/128, Pt. 2 (K. Tamaki, K. Suyehiro, J. Allan, M. McWilliams et al.,
Leg 135 Scientific Party. 1992. Evolution of backarc basins: ODP Leg 135, Lau Basin, EOS, Trans. AGU. 73:
241-247.
Lelikov, Yeo P., and Bersenev, I. I. 1973. Early Proterozoic gneiss-magmatite complex in the southwestern part of
the Sea of Japan, Dokl. Akad. Nauk SSSR 223:74-76.
Ludwig, W. L., Murauchi, S., and Houtz, R. E. 1975. Sediments and structure of the Japan Sea, Geol. Soc. Am.
Bull. 86:651-664.
McKenzie, D., and Bickle, M. J. 1988. The volume and composition of melt generated by extension of the
lithosphere, J. Petrol. 29:625-679.
Mueller, S., and Phillips, R. J. 1991. On the initiation of subduction, J. Geophys. Res. 96:651-665.
Murauchi, S. 1972. Crustal structure of the Japan Sea, Kagaku (Science) 38:367-375. (in Japanese)
Okino, K., Shimakawa, Y., and Nagaoka, S. 1994. Evolution of the Shikoku Basin, J. Geomagn. Geoelectr.
46:463-479.
Oshida, A., Tamaki, K., and Kimura, M. 1992. Origin of the magnetic anomalies in the southern Okinawa Trough,
J. Geomagn. Geoelectr. 44:345-359.
Schubert, G., and Sandwell, D.1989. Crustal volumes of the continents and of oceanic and continental submarine
plateaus, Earth Planet. Sci. Lett. 92:234-246.
Shinohara, M., Hirata, N., Nambu, H., Suyehiro, K., Kanazawa, T., and Kinoshita, H. 1992. Detailed crustal
420 KENSAKU TAMAKI
structure of northern Yamato Basin, in Proc. OD?, Sci. Results, 127/128, Pt. 2 (K. Tamaki, K. Suyehiro, 1.
Allan, M. McWilliams et al., eds.), pp. 1075-1106, Ocean Drilling Program, College Station, TX.
Tada, R., and Tamaki, K. 1992. Scientific results ofODP Japan Sea legs, and their implications for stratigraphy, J.
Jpn. Assoc. Petrol. Technol. 57:103-111.
Tamaki, K. 1988. Geological structure of the Japan Sea and its tectonic implications, Bull. Geol. Surv. Jpn 39:
269-365.
Tamaki, K., Pisciotto, K., Allan, J., et al. 1990. Proc. OD?, [nit. Repts., 127, Ocean Drilling Program, College
Station, TX.
Tamaki, K., Suyehiro, K., Allan, J., Ingle, J. C., and Pisciotto, K. A. 1992. Tectonic synthesis and implications of
Japan Sea ODP drilling, in Proc. OD?, Sci. Results, 127/128, Pt. 2 (K. Tamaki, K. Suyehiro, J. Allan, M.
McWilliams et aI., eds.), pp. 1333-1348, Ocean Drilling Program, College Station, TX.
Tanahashi, M., Kisimoto, K., Joshima, M., Lafoy, Y., Honza, E., and Auzende, 1. M. 1991. Geological structure of
the central spreading system, North Fiji Basin, Mar. Geol. 98:187-200.
Taylor, B. 1992. Rifting and the volcanic-tectonic evolution of the Izu-Bonin-Mariana arc, in Proc. OD?, Sci.
Results, 126, (B. Taylor, K. Fujioka et al., eds.), pp. 627-651, Ocean Drilling Program, College Station, TX.
Yoshii, T. 1979. Crustal Structure of the Japan Sea, Tokyo, Univ. Tokyo Press. (in Japanese)
12
Kuril (South Okhotsk) Backarc Basin

Helios s. Gnibidenko, Thomas W C. Hilde, Elena V. Gretskaya,
and Andrey A. Andreyev
The researches of many commentators have already thrown much darkness on
this subject and it is possible that. if they continue. we shall soon know nothing
at all about it.
Mark Twain
ABSTRACT
The Kuril backarc basin is a 3300-m-deep basin located behind the Kuril island arc and
subduction zone. It is underlain by oceanic crust, the top of which is 7-8 km deep, and it
contains 4 to 5 + km of tectonically undisturbed sedimentary fill. The oldest sedimentary
rocks sampled from the basin margin are Miocene/Oligocene. Although no seafloor spread-
ing magnetic lineations exist, we conclude that the basin has formed by backarc extension
and spreading. Elongated in the NE-SW direction, parallel to the Kuril arc, the basin is
-250 km wide at its SW end, narrowing to closure at its NE termination at the southern tip
of Kamchatka. Opening of the basin was apparently related to a relative motion pole at or
near the south end of Kamchatka, consistent with Miocene/Oligocene lateral fault trends
between Sakhalin and Hokkaido and the strike of transverse basement ridges extending
across the basin. The age of the basin may be as young as Miocene or as old as late
Cretaceous. Ocean drilling is needed to establish its precise age.
1. INTRODUCTION
The Kuril (South Okhotsk) backarc basin lies behind the Kuril island arc in the posi-
tion of a typical subduction related backarc basin (Fig. 12.1). However, the origin of this
basin, long inactive, has remained unclear. The Kuril backarc basin structural evolution has
been interpreted (Belousov, 1982; Belousov and Pavlenkova, 1986; Ermakov, 1991) as a
result of alteration of continental crust to oceanic crust by a process of basification or
oceanization, and, alternatively (Karig, 1971; Matsuda and Uyeda, 1971; Hilde et aI., 1977;
Bogdanov, 1988) as a result of crustal extension and seafloor spreading behind the Kuril
island arc, related to subduction of the Pacific oceanic plate. Taking into account data that
Helios S. Gnibidenko • Shirshov Institute of Oceanology, Russian Academy of Sciences, Moscow 117851,
Russia. Present address: Geodynamics Research Institute, Texas A&M University, College Station, Texas
77843. Thomas w.e. Hilde • Geodynamics Research Institute, Texas A&M University, College Station,
Texas 77843. Elena II. Gretskaya and Andrey A. Andreyev • Institute of Marine Geology and Geo-
physics, Russian Academy of Sciences, Yuzhno-Sakhalinsk, Sakhalin 693002, Russia.
421
422 HELlOS S. GNiBIDENKO at al.
FIGURE 12.1. General bathymetry of the Okhotsk Sea. Framed area: the Kuril (South Okhotsk) backarc basin.
indicate the basin is inactive and possibly has existed since Cretaceous time, it has also been
suggested to be a relict structural element (Vassilkovsky, 1967; Snegovskoy, 1974; Sychev
and Snegovskoy, 1976; Gnibidenko and Svarichevsky, 1984) entrapped by Kuril island-arc
formation.
Early models of the geodynamic evolution of the Okhotsk sea plate (Chapman and
Solomon, 1976; Savostin et ai., 1983) invoked counterclockwise rotation of the Okhotsk
plate converging with the Eurasian and/or Amurian plates along Sakhalin with subsequent
subduction in the Kuril-Kamchatka Trench. Two recent geodynamics models for the
Okhotsk plate, including the Kuril backarc basin, suggest additional recent relative plate
motions:
a. The Okhotsk plate lithospheric block is under compression between the North American and
Eurasian plates and is being extruded to the southwest (Riegel et al., 1993). In such a case
recent extension would be unlikely in the Kuril (South Okhotsk) backarc basin.
KURIL BACKARC BASIN 423
b. There is no separate Okhotsk plate. It is part of the North American plate with southerly
motion (DeMets, 1992a,b). Again, no recent extension would be likely in the Kuril backarc
basin.
We examine the bathymetric, structural, and geophysical characteristics of the

Okhotsk Sea and particularly the Kuril (South Okhotsk) backarc basin in an attempt to
better understand the origin of this backarc basin.
2. BATHYMETRY
The bathymetry of the Kuril backarc basin (Fig. 12.2) reveals an unbroken, relatively
flat deep-sea plain bounded by the Okhotsk continental slope on the northwest and the steep
Kuril backarc slope on the southeast. The basin extends NE-SW, parallel to the Kuril arc,
being 200-250 km wide at its southwest end and narrowing toward Kamchatka. The nearly
horizontal floor of the basin lies mainly at a depth of 3200-3300 m with isolated 3360+-
m-deep northwest trending subbasins. The basin floor morphology suggests depositional
origin with some underlying structural control. Seamounts, rising up to 1000-1500 m above
the basin floor, are found only along the slope of the Kuril island arc. No axial rift or
structural trends are observed.
The high density of bathymetric data allows three subbasins to be distinguished (Fig.
12.2). From the northeast, these are Atlasov, Urup, and Iturup basins, with depths greater
than 3360 m. They are separated by subtle northwest trending swells, in line with the Bussol
FIGURE 12.2. Bathymetric map of the Kuril backarc basin based on an echo-sounding profiling grid of 1.5-3.0
miles between profiles (Institute of Marine Geology & Geophysics, Russian Academy of Sciences). Location of
refraction (Fig. 12.16) and reflection (Figs. 12.17 and 12.18) profiles are shown.
and Friza straits of the Kuril arc, and are underlain by basement structures of the same
trend.
All the slopes bounding the basin are relatively steep with submarine canyons, slumps,
and slides. The Kuril arc slope of the basin is steepest, ranging from 10° to 25° (Fig. 12.3).
This slope exhibits a pattern of en echelon ridges striking into the basin at angles of about
25° to perpendicular to the general trend of the Kuril arc. Some of these ridges are volcanic
edifices, others are faulted arc basement. Flat terraces are found at various depths: from
shelf level at about 100-170 m and at 950-1200 m (Svarichevsky and Svarichevskaya,
1982). These terraces indicate substantial Pliocene-Holocene subsidence of the north-
western inner slope of the Kuril arc.
The Sakhalin-Hokkaido slope of the basin has a well-defined shelf break between
140 m in the north and 190 m in the south (Figs. 12.4-12.6). The slopes dip basinward at
10°-15°, steepening to the plane of the basin. Canyons, slumps, and deep sea terraces of
probable structure origin characterize the western slope of the basin. Clearly, Terpeniya
Ridge and the steep slopes below 500 m in this region (Figs. 12.4-12.6) are southward
extensions of structural features of Sakhalin.
The northern slope of the basin is distinctly different, with a slope break near 1400 to
1600 m and constituting the structural boundary with the Academy of Sciences Rise. It dips
from this depth to the basin plain at angles between 3° and 10° (Fig. 12. 7). It is also dissected
by canyons, exhibits slides and a sharp angle boundary with the basin caused by basin
turbidite sediment onlap.
Raykoke I.
I.,.~
. ~/II
3° S· 10"'5"
Iturup I.
VI
I.
VII
VIII
;.:.x_------ _____ _
FIGURE 12.3. Bathymetry profiles across the Kuril arc slope of the Kuril backarc basin.
I Tulenly I.
"=;======~~~~~~~----------
,~II~==~====~==~==
FIGURE 12.4. Bathymetry profiles across the northern (Terpenyia bay area) Sakhalin-Hokkaido slope of the
Kuril backarc basin.
~~
I\\\~ .
w'" ,. I- l.r
- -- II"
II..
--
IItID-
IV·
VI·
FIGURE 12.5. Bathymetry profiles across the central (offshore Tonino-Aniva peninsula) Sakhalin-Hokkaido
slope of the Kuril backarc basin.
426 HELlOS S. GNIBIDENKO et al.
r~·=i"·
----1
I
J_Ii_ G
I~__ II'
1111
--I
~:
S3IIhal~~
I ,• IV
Im~
'''~''
I ~: II
FIGURE 12.6. Bathymetry profiles across the southern Sakhalin-Hokkaido slope of the Kuril backarc basin.
FIGURE 12.7. Bathymetry profiles across the northern slope (south slope of the Academy of Sciences of the
USSR) of the Kuril backarc basin.
3. GEOPHYSICAL FIELDS
3.1. Magnetic Anomalies

The magnetic anomaly map presented here is based on combining previously com-
piled magnetic anomalies of the Okhotsk Sea (Tereschenkov et al., 1988) with the latest
marine magnetic surveys (Fig. 12.8). It is necessary to point out that much of the available
data for the Okhotsk Sea backarc basin were obtained along NE-SW ship tracks. Neverthe-
less, the data is sufficient to allow us to conclude that there are no well-pronounced NE-SW
linear magnetic anomalies oriented along the strike of the long axis of the basin (Krasny,
1990; Andreyev and Vorobyev, 1991).
Generally, negative magnetic values are associated with the SW portion of the basin
and the southward offshore extension of Sakhalin, while more positive values are associ-
ated with the remaining area of the Okhotsk Sea. A rather subtle approximately N-S set of
broad anomalies is associated with the SW portion of the South Okhotsk Basin that appear
to be marine extensions of Kuril are, Hokkaido, and Sakhalin structural trends. An abrupt
change in magnetic character of the basin is observed along a NW-SE strike between 1500
and 151°E.
FIGURE 12.8. Magnetic anomalies map of the Kuril backarc basin. Positive anomalies are shown by solid lines,
negative by dashed lines, dash-dotted lines mark zero anomaly contour; the unit of measurement is xlOO nT.
428 HELlOS S. GNiBIDENKO et al.
3.2. Gravimetry
A 5 x 5 minute interpolated grid of satellite- (SEASAT) derived gravity data were used
for preparation of the free-air gravity anomaly map of the Okhotsk Sea including the South
Okhotsk backarc basin (Fig. 12.9). The contours generally follow structurallbathymetric
boundaries. Positive anomalies are associated with basement highs, while negative values
outline deep sedimentary basins. A broad negative anomaly along the northern and western
parts of the backarc basin suggests a thick sediment fill in this portion of the basin between
the Mesozoic (and probably Paleozoic) Academy of Sciences Rise and the Kuril arc, and
along the Sakhalin-Hokkaido island-arc system. The anomaly pattern indicates that the
greatest sediment thicknesses beneath the basin are along the base of the slope of the
Academy of Sciences Rise. On the basis of marine free-air anomalies, Kogan (1975) has
proposed a reduced upper mantle density beneath the Kuril backarc basin due to elevated
upper mantle temperatures.
3.3. Heat Flow

The Kuril backarc basin is characterized by high heat flow (Fig. 12.10; Sychev et al.,
1983; Tuezov et aI., 1984) of up to twice the global average. According to models ofU, Th,
and K content in the crust (Veselov and Volkova, 1981; Volkova, 1982) the main part of heat
flow is contributed from the upper mantle (80-90%) in the Kuril backarc basin, whereas in
thick crustal continental margins this is not more than 30%. The sources of such anomalies
are located at depths of 15-20 km.
Heat flow distribution in the Kuril backarc basin (Fig. 12.11) as a whole is charac-
145'
. I .
/. \
!(~)
\.<;!..(
I: .
l C. \
",'\J
\0--' POSITIVE, MGaL
.\0-'''''' NEGATIVE, MGaL

0, MGaL
_0/
45'
150- 155'
FIGURE 12.9. Free-air gravity anomalies map of the Kuril backarc basin based on the satellite altimetry data.
FIGURE 12.10. Heat flow map of the Kuril backarc basin. I-measured heat flow value (mW/m2); 2-heat flow
isolines in mW/m2; 3-isobaths in m.
terized by a heat flow of greater than 70 mW/m2 . All three subbasins (Atlasov, Urup, and
Iturup) have high heat flow values of 90-110 mW/m2 • Heat flow is also particularly high
along the southern Kuril arc.
Estimations of the crustal temperature distribution (Fig. 12.11) gives temperatures of
about 300-S00°C for layers 2 and 3, respectively, and upper mantle temperatures beneath
the basin of ll00-1200°C at depths of 20-40 km. Given these heat flow values and
temperature/crustal relationships, it is possible that any formerly magnetized oceanic crust
beneath the basin could be partially demagnetized, especially in the lower crust.
3.4. Seismicity and Stress Field

The Kuril arc-Kuril backarc basin region has intensive earthquake seismic activity
(Tarakanov and Kim, 1983). Twelve earthquakes with M>S.S and more than 100 with
M>4.S have occurred in the basin area since 1986 (Far East, 1986-1989; Seismological
Bull., 1990-93). A few earthquakes have occurred at relatively shallow depths directly
beneath the basin (0-90 km, M<4.S), but most are beneath the Kuril arc and associated
with the subducting Pacific plate (Figs. 12.12 and 12.13).
Two zones of seismic activity are defined by the distribution of seismicity of the
subduction zone (Fig. 12.13): a more compact upper zone extending to a depth of 170 km
NW SE
Tartar Stran Sakhalin Kurll baaln ocean

91 51 46 52 71 105 108 42 46 50
- - - - - 300~ 20
------400·__
30 _ - - - 500-·_ 30
km~ ________ ~~~ __ ~~~~~ __________________ ~~~~LL ________________ ~40km
o____
~, ~ _ _ _ _~_ _ _ _~_ _~~_ _~I 500 km
FIGURE 12.11. Temperature distribution pattern along the crustal-upper mantle section crossing Kuril backarc
basin (Modified from Sychev et al., 1983). I-heat flow stations in mW/m2; 2-isotherms in °C; 3-generalized
boundaries according to refraction studies and P-wave velocities in kmls; 4-upper mantle-crust boundary
(Moho).
.. 0
0'
,
.a ..
0'
o~ •
••
1
l~klll ;z::j •
-'0
FIGURE 12.12. Earthquake locations from 1986 to February 1993 in the Kuril backarc basin. Earthquake
magnitudes: I: <4.55, 2: 4.5<5.5, 3: 5.5<6.5. Focal depths: 4: <30 km, 5: 30<90 km, 6: 90<300 km,
7: 300<500 km, 8: >500 km. 10: cross-sectional line (see Fig. 12.13).
NW
100 2•• 3•• ... 5'
500 km
~o ~)
0
I ••
og~
0
2••
0
0c9
3 •• 0 0
0)0
COo c9
...
0
0
o 0
00
0
FIGURE 12.13. Cross section of hypocenters from o 0 c9 01
1986 to February 1993 events. The cross section scatter 5•• 0 02
is 100 km wide (see location in Fig. 12.12). Earth- 'm 0 03
quakes magnitudes: 1: <4.5, 2: 4.5<5.5, 3: 5.5<6.5.
and a lower more diffuse zone from 240 km to about 500 km. A 100 km zone of relative
seismic quiescence separates these zones (Tarakanov and Leviy, 1968).
It has been shown (Simbireva et al., 1976) that there is a remarkable change in the
stress field of the subduction zone at the depths between 100 and 300 km, beneath the basin.
Subhorizontal tension characterizes this zone (north of Busso1 Strait) beneath the Atlasov
subbasin, while subhorizontal compression is dominant for this zone beneath the southwest
part of the basin (Urup and Iturup subbasins).
4. TECTONICS
4.1. Crustal and Lithospheric Structure

The morpho structural depression of the Kuri1 backarc basin coincides with a de-
creased crustal thickness, generally outlined by the 20-km crustal isopach (Figs. 12.14
and 12.15).
Three principal layers are distinguished in the crustal structure of the basin (Figs.
12.15-12.18): layer 1 (V = 1.7-4.3 kmls) is a sedimentary cover that consists of mainly
layers of clay, silt, claystone, and siltstone with an average thickness up to 5 km (Fig.
12.19); layer 2 (V = 4.8-5.2-kmls) is volcanic and volcanoclastic in composition with a
thickness of 0.5-2.0 km in the basin; and layer 3 (V = 6.4-7.2 kmls) is probably gabbro-
basaltic in composition with average thickness about 5-7 km.
Starshinova (1980) suggests that the upper mantle beneath the basin down to 30 km
deep has a sandwich-like structure consisting of alternating 8- and 7-kmls layers. Deeper
discontinuities in the upper mantle are noted at depths about 80-85, 330-335, and 400 km
(Zhang and Lay, 1993) with the first probably being the regional base of the lithosphere.
The roof of the crustal layer 2 serves as a basement (Fig. 12.20) for the sedimentary
cover (layer 1) in the basin and it can be traced from the basin to its slopes through the
seismic profiling (Figs. 12.17 and 12.18).
FIGURE 12.14. Crustal thickness map of the Okhotsk

Sea plate including Kuril backarc basin based on the data
(Galperin and Kosminskaya, 1964; Zverev and Tulina,
1971; Popov and Anosov, 1978; Bikkenina et al., 1987;
and Zlobin and Zlobina, 1991). Crustal cross sections
shown in Fig. 15.
30km
o lOOkm
..........
SAKHALIN
FOlDED KURIL BACKAAC BASIN
D SYSTEM
---L::'::::::::':" E
30km
o lOOkm
'----'
OKHOTSK ARCH
FIGURE 12.15. Crustal cross sec-

tions of the Okhotsk Sea plate includ-
ing the Kuril backarc basin. Locations
in Fig. 15. I-terrigenous sedimen-
tary layer; 2-volcanic sedimentary
complex; 3-oceanic crust; 4-upper
continental crust; 5-lower continen-
tal crust; 6-upper mantle; 7 -faults.
TERPENIVA SHE L F
I m:!~OV I mornvinov basin I

ICDNTINENTAl SLOPE I K URI L BACK ARC BA SIN KURIL ARC
2.3 2.0
SO 1.7 4.1
---,-.- 3.0
43 •. 25 ~5 :=::u:--
4.3
4.0 4.2
72 64
66
6.B
6B
S> ~5,2
6.8 ~s~
7.0 6.1
Id- I
10
VELOCITIES. KMjS
/ I TOP OF OCeANIC BASEMENT .m
FIGURE 12.16. Upper crustal cross section of the Kuril backarc basin based on refraction studies (Bikkenina
et aI., 1987). Location in Fig. 12.2.
FIGURE 12.17. Multichannel seismic profile illustrating the structure of the sedimentary cover and its relation-
ship with basement. Location in Fig. 12.2: profile 1.
PIIOnl.ll 2
o 10 2,0 1,0 50· n:
1iBi!lm!'O .0
I~iillllililliliii~~:-SE 1.0
2..0
3.0
FIGURE 12.18. Single-channel (profile 2) and multichannel (profile 3) seismic profiles (profile 2) illustrating the
structure of the sedimentary cover and its relationship with basement. Location in Fig. 12.2.
FIGURE 12.19. Isopach map of the Kuril backarc basin. i-basement outcrops according to seismic data; 2-
isopachs of sedimentary cover.
FIGURE 12.20. Structure-contour map of the Kuril backarc basin basement. I-basement outcrops according to
seismic profiling data; 2-basement depth contours, in km; 3-faults. List of structural elements: I-Abashiri
Trough; 2-Tokoro Basin; 3-Kitami-Yamato Uplift; 4-Kitami-Yamato Basin; 5-Aniva horst; 6-Aniva
Basin; 7-La Perouse Uplift; 8-Taranay Basin; 9-Mordvinov Basin; lO-Muravyev Basin; ll-Makarov
Uplift; 12-Makarov Basin; 13-Terpeniya Uplift; 14-Nevsky Uplift; 15-Terpeniya Trough; 16-Terpeniya
horst; 17-Peski Basin; 18-Polevoy Uplift; 19-Pegas Uplift; 20-Pegas Trough; 21-Schmidt Trough; 22-
Petro Uplift; 23-Academy of Sciences of the USSR Uplift; 24-Atlasov Subbasin; 25-Bussol Uplift; 26-
Urup Subbasin; 27-Prostor Uplift; 28-Iturup Subbasin.
4.2. Basement
The data from dredging (Gnibidenko and Ilyev, 1992) of basement outcrops on the
slopes around the Kuril backarc basin are combined here (Fig. 12.21) for only those stations
for which either radiometric or fossils ages have been determined.
4.3. Composition
Among the dredged samples are extrusive, intrusive, and metamorphic rocks. Extru-
sive rocks consist of basalts to rhyolites with an abundance of widespread outcrops of
intermediate and basic composition. Many of the extrusive rocks exhibit varying degrees of
greenstone alteration (Vasilyev et aI., 1990), metabasalts, and meta-andesites. There are
also relatively fresh olivine basalts, bipyroxene basalts, and andesites. The presence of
transparent plagioc1ases and nonaltered volcanic glass characterize these rocks. Fresh
hyaloc1astites are also found. The nonaltered rocks were dredged from the Kuril arc slope
of the basin only.
Extrusive rocks with moderate-to-weak greenstone alteration are the most widespread
FIGURE 12.21. Dredge sites (only sites with ages determinated are presented) around the Kuril backarc basin.
l-isobaths, 2-magmatic rocks (K-Ar), 3-sedimentary rocks (biostratigraphy).
on the South Okhotsk backarc basin slopes. This group includes basalts, andesite-basalts,
andesites, andesite-dacites, and rhyolites. Fresh non altered minerals exist within the matrix
of these rocks in spite of clear indications of autometamorphic processes.
Intrusive rocks are mainly represented by granitoids, diorites, monzonites, and sye-
nites. Gabbroids are found rarely as single blocks. A fragment of lherzolite has been
exposed eastward of the Academy of Sciences Rise.
Granodiorites and quartz-diorites are the most abundant of all basement magmatic
rocks. These rocks are light gray and gray, usually medium and coarse grain, sometimes
porphyry rocks composed of acidic and medium plagioclase, quartz, potassic feldspar, and
biotite. Autometamorphic alteration is slightly developed in these rocks.
Fine-grain, medium-grain, and coarse-grain, sometimes porphyry biotite-hornblende
granites, plagiogranites, and quartz monzonites have been dredged from the Academy of
Sciences Rise and on the lower slope of the central part of Kuril arc. Cataclastic texture is
sometimes found.
Metamorphic rocks are mainly represented by hornfels. Relict porphyroblast textures
are typical for many of them, which testifies to a volcanogenic origin. Angular blocks of
granite-gneisses were also dredged from the Academy of the Sciences Rise. These leu co-
cratic rocks consist of acidic plagioclase, quartz, muscovite, biotite, and gamet.
The chemical composition of the dredged rocks shows that the extrusives belong
mainly to the calc-alkali series or close to it (Fig. 12.22). The swarm of intrusive rock
Na,O+K,O. weight 'l(,
•
10
FIGURE 12.22. Relationship between al-

kali and Si02 (Kuno diagram) in effusive
rocks from basement outcrops around the 5
Kuril backarc basin. I -tholeiite (alkalic
basalt) series area; II-calc-alkalic series
area; III -calc-basaltic series area. 1-3 =
extrusive rocks of: I-Academy of Sci- +
ences Rise, 2-Kuril arc slope, 3-Hok- 40 50 60 70
SIO .. weight 'l(,
kaido-Sakhalin slope. ., 02 +3
compositions coincide with the calc-alkali series field (Fig. 12.23) and testifies to the close
chemical relationship of these rock groups. All magmatic rocks of the Kuril backarc basin
slopes belong to the sodium type where N~O>K20.
4.4. Age
K-Ar ages of the magmatic rocks from 40 samples of basement outcrops along the
basin margins vary between 219 Ma and 30 Ma (Figs. 12.21 and 12.24). Most of the dated
magmatic rocks (75%) are Cretaceous. Keeping in mind that all radiometric ages are of
magmatic rocks, the host strata should be older, likely Jurassic in age. Granodiorites of ages
146 Ma and 219 Ma from basement outcrops on the Kuril arc slope of the basin, north of
Simushir Island, and other similar age samples from the Kuril arc slope generally match the
ages of rocks from the north side of the basin. Since the Kuril backarc basin clearly has a
thin crustal structure and velocities characteristic of oceanic crust, it is likely that its
development postdates and separates the Cretaceous and Jurassic basement rocks described
above.
Because of the high latitudes for the Okhotsk Sea and the impact of Neogene and
Pleistocene glaciation on the region, ice rafting must be considered as a process for
distributing rocks over broad areas of the seafloor and especially for those rocks that seem
Na,O, weight %
FIGURE 12.23. A diagram of alkalinity of intrusive

rocks around the Kuril backarc basin. 1-3 =intrusive rocks
of: I-Academy of Science Uplift, 2-Kuril arc slope,
4 6 8 K2 0. weight 11
3-Hokkaido-Sakhalin slope, 4-contour of extrusive
rocks of differentiated calc-alkalic series. • 1 0 2 + 3 ',_ 4
..... . ..=
3 ....
III
.. III 0111
n
.. III
=:!
0::l III
8... EPOCH AGE N.Y.
,,"..,111
IBi~
III oJ ..
DoO
t:I !;~ .. ...

oJ
~~
Do ~1IIi!:
. e
0
III
0 ..
1IIi!:
,...---
•
24,6
cnatt~an
Oligocene Rupelian
•
u 38 ,...---
c PriabOnIan 42
u Bartonian
II' Eocene
0 Lutetian
!..
Do Paleocene
YDreaian
Thanetlan
50,5
54,9
60,2
,...---
Danian
MaastrIchtian 65
73 ~
• • A ...
Callpanian
••
83
.
Santonian 87,5
II Late Coniacian
:I 88,5
0 Turonian
...,..
u
li:i:-
91
u Ceno.anian 97,5
Al.blan
u Aptian
113
ll9 ...~
Early
Barr... ian
Hauterivian
125 ••••
•
Valanginian 131
138 •
....
Berriaaaian
Maim
Titnonian
Ki_eridgian
144
150 ~
158
..
OXfordian 163 !!........-
CallOVIan 169
u Bathonian
Dogger 175
-•
Bajocian
.,".
II 181
Aalenian 188
:I Toarc1&n 190
Pleinsbachian
••
Lia. Sine.urian 200
206
Hettangian
.4
21.3
• 1 .2 ... 3
FIGURE 12.24. Radiometric age of the basement outcrops rocks around the Kuril backarc basin. The geo-
chronology scale is according to Harland et al. (1985). I-granitoids, 2-intrusive rocks of intermediate and basic
composition, 3-rocks of intermediate and basic composition, 4-extrusive rocks of acid composition.
to be exotic to the tectonic setting from which they were dredged. This was considered in
the reporting of the dredge samples described in this report. The dredge sites for these
samples were all from topographic highs that were mostly clear basement exposures. The
analyzed and reported samples were all angular, and several exhibited fresh surfaces. Any
rounded or striated samples were excluded from the study. Further, those dredge samples
from the Okhotsk Sea that have been identified as ice-rafted debris are typically Upper
Cretaceous to Paleocene acidic volcanic and granitic rocks and were also excluded. Their
source area is from onshore exposures north of the Okhotsk Sea. These rocks are distinctly
different from the rocks reported here from around the Kuril backarc basin. While the
possibility remains that ice rafting could be responsible for distributing the rocks described
in the report, we feel that it is more likely that they represent the basement of the southern
Okhotsk Sea region.
4,5, Structural Features

The acoustic basement of the basin is characterized by seismic velocities of 5.1-5.2
km/s (Fig. 12.16) and is overlain by 4 to 5 km of sediment throughout most of the basin (Fig.
12.19). Basement highs are observed which appear to be associated with either arc struc-
tures or basement at the edge of the basin (Figs. 12.17 and 12.18). They outcrop at the sea
bottom only near and along the slopes of the basin. These features have previously been
described as elevated portions of basement, volcanoes, or outcrops of layer 2 (Soloviev
et at., 1977). We interpret them as a parts of the arc basement or the basement of the
northern margin of the basin.
Distinct differences exist in the velocity structure beneath the Kuril backarc basin and
shallower portions of the Okhotsk Sea to the north and the northwest (Fig. 12.16). Beneath
the basin there is a well-defined layered structure with the top of the 5.1-5.2 km/s layer
correlating with acoustic basement of the seismic reflection data. The boundary of the basin
and the continental slope (Fig. 12.16) marks a dramatic change in the velocity structure.
To the north of the basin, this boundary is characterized by greater structural vari-
ability in relief and velocities. These structures are representative of the Academy of
Sciences Uplift and the Sakhalin-Hokkaido fold belt and can be correlated with Mesozoic
geosynclinal and ophiolitic (Cretaceous and older) formations. The overlying sedimentary
cover displays less folding and faulting.
Samples of sedimentary outcrops from around the margin of the Kuril backarc basin
(Fig. 12.21) are all Miocene in age except for one Oligocene sample. They appear to
correlate with the lower or middle parts of the Kuril basin sedimentary section as revealed
in the seismic reflection profile (Fig. 12.17). Assuming that these rocks represent early
deposition into the Kuril backarc basin, their age then places a minimum age on the
formation of the basin.
5. BASIN SEDIMENTARY FILL
5.1. Seismic Stratigraphy

The most distinctive feature of the sedimentary section of the Kuril backarc basin is
that from seafloor to basement, or near basement, it is an undisturbed, horizontally stratified
section (Figs. 12.17 and 12.18). No evidence for extension exists within the sedimentary
section. Gentle folds are observed at the base of the Kuril arc (Fig. 12.17) which could be
evidence of relatively recent minor compression at the base of the arc. A minor unconfor-
mity exists between the uppermost sediments and the gentle folds at 4.7-s depth near the SE
end of profile l. These features are, however, indications of only very minor deformation.
Basement is clearly distinguished all across the basin (Fig. 12.17) as a rough, high-
amplitude, highly reflective surface, typical of oceanic basement. This surface corresponds
to the top of the 5.2-km/s layer (Fig. 12.16) detected in seismic refraction studies (Popov
and Anosov, 1978; Bikkenina et aI., 1987). The sediments immediately overlying basement
display the greatest variability, in places horizontally onlapping basement (SE portion of
profile 1 and profile 3) and displaying dip toward the basin beneath both margins.
The relatively transparent lowest deposits overlying basement in the central portion of
profile 1, displaying apparent drape, are possibly associated with the basement high that
extends perpendicular to the arc, across the basin (Fig. 12.2). Onlap onto the basement high
displayed in profile 2 (Fig. 12.18) also shows tectonic quiescence during deposition and
another basement feature that is perpendicular to the arc (Fig. 12.2). In general, the
sedimentary section and its basement relationships indicate that the Kuril backarc basin has
been inactive throughout the basin's depositional history.
440 HELlOS S. GNIBIDENKO at al.
1\vo additional general features of the basin sedimentary section are worth noting. As
a whole, approximately the upper half of the section is more reflective than the deeper
sediments. On the basis of this and and other characteristics, we divide the basin fill into
upper and lower sedimentary sequences. The boundary between these sequences corre-
sponds to the boundary between the 2.4-3.0-km/s unit and the underlying 4.0-4.3-km/s
unit (Fig. 12.16). This boundary is observed in the reflection profiles as a high-amplitude
reflector and subtle unconformity at about 5.5-s depth across profile 1 (Fig. 12.17),4.6-4.9 s
in profile 2 and at 5.9-6.0 s in profile 3 (Fig. 12.18). The upper sequence clearly onlaps the
margin slopes of the basin, while the lower sequence dips basin ward at the margins and is
exposed on the margin slopes (Fig. 12.20). These relationships suggest whole basin
subsidence during deposition of the lower sequence.
5.2. Thickness and Distribution

A generally greater than 4-km-thick sedimentary section fills the Kuril backarc basin,
with an axial extent closely matching the deep topographic basin defined by the 3300-m
isobath (Figs. 12.2 and 12.19). While the basin as a whole has an arc-parallel long axis, it is
divided by transverse basement highs or ridges into three subbasins: Atlasov, Urup, and
Iturup. Sediment thicknesses exceed 5 km in the central Urup and the SW Iturup subbasins
(Fig. 12.19). Basement relief has clearly controlled the distribution of the primarily turbidite
deposits.
5.3. Composition and Age

Seismic reflection profiles and cores (see following section) indicate that the upper
sedimentary sequence of the Kuril backarc basin is primarily of turbidite origin, consisting
of turbidites, biogenic, and ash layers. Likewise the lower sedimentary sequence, while
somewhat less reflective, also generally displays numerous closely spaced horizontal
reflectors in the seismic reflection profiles that characterize turbidites (Figs. 12.17 and
12.18). The reflection profiles and dredge samples along the margin slopes (Gnibidenko and
Ilyev, 1992; Fig. 12.21) indicate that the lower sequence is also primarily turbidites with
alternating ash layers. These are the deposits that are exposed at the basin margins, that dip
toward the basin, and correlate with the deeper sediments in the basin.
The dredged samples along the basin margin are Miocene and Oligocene in age (Fig.
12.21). Assuming that these samples represent the lower sequence within the basin as we
have correlated them to the reflection profiles, the basement age of basin could then be as
young as Miocene/Oligocene in age.
6. HOLOCENE SEDIMENTATION
Data obtained for the uppermost part of the sedimentary cover of the South Okhotsk
backarc basin by coring (Bezrukov, 1955, 1960; Zhuze, 1957; Saidova, 1960; Ilyev et aI.,
1979, and Gretskaya, 1990) provide us with Holocene sedimentation characteristics which
may be used to consider limits on the depositional history for the older deposits.
Bottom sediments of the basin are the mainly fine-grain greenish gray, soft and plastic
sediments up to a depth of 2.2 m. They are dominantly silts and clays and are generally
poorly sorted with concentrations of silt 25-40%; clay 50-75%; water content 60-70%;
and bulk density 1.2-14 g/cm3. In the cores located in the central part of the basin (sites 62
and 63, Fig. 12.25) the fine clay fraction (0.001 mm) is considerably increased (70-80%
Fig. 12.26) with water content up to 80% (Fig. 12.27).
A fine ooze is abundant in the central part of the basin. Silts and sands exist closer to
the margins. Distribution of the terrigenous material can be traced throughout the basin in
the upper horizon (0-15 cm) and deeper layers, suggesting powerful turbidity flows.
The distance of transportation of terrigenous material traced according to variations in
grainsize was studied from the cores of Sites 60-63. In Core 61 (Fig. 12.26) silt-rich
sediments form the upper 0-15 cm and three layers within the interval of 100-180 cm. The
saturation with silt in this part of the basin reflects both the periodicity of the turbidity flows
(four cycles) and the area of their unloading. Two turbidity cycles are distinguished in Core
62, while in core 63 in the central part of the basin no turbidites are detected (Fig. 12.26).
All Holocene basin sediments are rich in biogenic silica, 23-35% weight, and have a
pronounced biogenic-detrital texture. Diatom plankton is the source of silica. The areal
zoning of biogenic silica corresponds, in general, to the circumbasin pattern of the sedimen-
tary material distribution. All sediments deposited closer to the base of the Kuril arc slope
contain greater quantities of silica.
Thus, there are two main sources of Holocene sedimentary material: terrigenous and
biogenic. The terrigenous material is mainly supplied by the turbidity flows, whereas the
I
- 7
8·
~, ~. ~, -
'.
[OJ. U' cz::::J •

, ,. , "XI m
.... ,
FIGURE 12.25. Core locations and composition in the Kuril backarc basin. l-isobaths, 2-core sites, 3-ooze,
4-silty clay, 5-silty ooze, 6-ash layer, 7-faunal residuals, 8-dispersed sand and gravel material.
S, 61 S. 62
11 FRACTION <0,001 mrn 11 FRACTION <0,001 mm
20 40 10 10 20 40 10 10
O~--~----~---L----T---
>11
I,
50 50
...
~
in
100 100 >11
,,,II1II
,
·••,
.• ,
I
150 150
20
CII
>11
I
10 40 10 I I 7 I 5
. •,
IEDIAN (.)
, • ,
I I 7 5
IEDIAN (.)
FIGURE 12.26. Grain size distribution of sediments for the central part of the Kuril backarc basin. For location
of sites see Fig. 12.25.
CLAY
100"
oo"L----------.:1L--------~100"
SILT 50 <0,001 "'"'
FIGURE 12.27. Log and measured grain-size parameters of sediments at Sites 61 and 62.
biogenic component is provided by high bioproductivity, resulting in a relatively rich

organic content in the sediments.
Age and Sedimentation Rate

The upper 2.0-2.5 m of sediments are dated as Holocene. Sedimentation rates have
been calculated as 20-25 cm/l000 year (200-250 m1Ma; Zhuze, 1957; Bezrukov, 1960;
Saidova, 1960). More recent studies (Gretskaya, 1990) estimate 30-40 cm/l000 year (300-
400 m1Ma). A high degree of water content and abundant <O.OOI-mm-size fraction in the
sediments are cited as causing slow consolidation and preservation of initial thicknesses for
long periods of time.
On the whole, Holocene sedimentation of the South Okhotsk backarc basin occurs as a
result of the turbidity flows into a closed basin along with high diatom productivity. If we
assume a rate of turbidite sedimentation of about 300 m/m.y. has existed for the late
Cenozoic, the onset of intense turbidite deposition would have been in the middle Miocene.
7. DISCUSSION: STRUCTURAL EVOLUTION AND GENESIS
During the last decade the geological and geophysical characteristics of Okhotsk Sea
plate, including the Kuril backarc basin and the Sakhalin-Hokkaido-Kuril-Kamchatka
island-arc system, have been extensively studied. These studies allow us to distinguish the
main features of the basement structure and age, and the distribution, age, and nature of the
sedimentary basins. They reveal a Sakhalin-Hokkaido-Kuril-Kamchatka island-arc sys-
tem which began its development in early Mesozoic time surrounding a cratonic Okhotsk
Sea Plate. Extensive fold and thrust deformation began in late Cretaceous time and
continues to present in the current island-arc system. The Kuril backarc basin developed
within this complex sometime between late Cretaceous and Miocene time.
While the Kuril backarc basin is underlain by oceanic crust the remainder of the
Okhotsk Sea region is underlain by continental crust of predominantly Mesozoic age
(Gnibidenko, 1985). Crustal thickness varies between 25 km in the central Okhotsk plate
up to 35 km in the surrounding folded zones. The crust of the Okhotsk plate is highly block-
faulted with Cretaceous and Cenozoic sediment-filled basins organized in NW-SE and E-W
tectonic patterns. Normal-faulted Cenozoic sedimentary basin development is inherited
from these Mesozoic structural trends.
The Okhotsk Sea is bounded on the west and north by the Hokkaido-Sakhalin and
Okhotsk-Chukotka Cretaceous and Cenozoic volcanic-plutonic belts, which are products
of convergence and subduction of the Okhotsk plate beneath the Eurasian plate during
Mesozoic time (Jolivet et aI., 1992). Collision of the continental crust of the Okhotsk Plate
along this zone is evident by the late Mesozoic highly deformed subduction complexes of
Sakhalin. This zone includes ophiolite and molasse formations of Mesozoic and Cenozoic
age. Paleozoic rocks have been found in the lower sections (Melnikov, 1988). The zone is
composed of en echelon anticlinoria, structurally connected between Sakhalin and Hok-
kaido across La Perouse Strait (Gnibidenko and Snegovskoy, 1975). All geological forma-
tions are highly folded and thrust faulted with vergence mainly to the west. During the
Cenozoic a dextral wrench-faulting system was superimposed. This zone is still tectoni-
cally active with a N-S zone of seismicity all along Sakhalin that extends southward and
links with the seismicity of Japan (Okada, 1982; Kiminami and Kontani, 1983; Rozhdest-
vensky, 1986; Jolivet et aI., 1992).
The southern and eastern boundary of the Okhotsk Sea consists of the Kuril-
Kamchatka subduction zone/arc system where the Pacific plate is subducting northwest-
ward beneath the Okhotsk Sea plate. Here, also, along the entire Kuril-Kamchatka zone,
geological evidence exists for a late Mesozoic convergence/subduction history. The Kuril-
Kamchatka fold system is composed of horst-anticlinorial uplifts of Mesozoic formations,
separated by Cenozoic sediment-filled graben-synclinorial troughs which strike northeast-
ward, parallel with the general trend of the arc. The average crustal thickness of the region
is about 30 km. Crustal and upper mantle inhomogeneities are recognized with a resolution
of a few kilometers (Averynova, 1975; Balakina, 1979, 1981; Gnibidenko et at., 1983).
To the southwest, the Kuril fold system is structurally connected with eastern Hok-
kaido structural elements (Geological Map, 1978; Honza, 1978; Gnibidenko et at., 1983). To
the northeast this zone extends into Kamchatka, where the axial parts of uplifts are traced
by outcrops of Mesozoic ophiolites (Gnibidenko et aI., 1974) within horst-anticlinorial
uplifts on the continental slope. These formations consist of deformed geosynclinal vol-
canogenic and sedimentary rocks which are intruded by gabbroides, granitoides, and grano-
diorites. Graben-synclinal troughs are filled with Neogene sedimentary deposits, which in
some basins exceed 3 km in thickness.
The crustal and tectonic evolution of the Okhotsk Sea region has involved a Mesozoic
(and possible older) cratonic block colliding with the eastern Eurasian plate margin during
late Mesozoic to early Cenozoic time (Den and Hotta, 1973; Savostin et at., 1983; Taka-
nashi, 1983; Watson and Fujita, 1985; Kimura, 1986; Jolivet et at., 1989; Maeda, 1990). This
was preceded and accompanied by construction of the Hokkaido-Sakhalin-Chutkotka
volcanic/deformation belts discussed above.
Recent and present-day plate configurations within the region are deduced from zones
of seismicity and focal mechanism solutions (Chapman and Solomon, 1976; Zonenshain
and Savostin, 1979; Tarakanov and Kim 1983; Lundgren and Giardini, 1990; DeMets,
1992a,b; Riegel et at., 1993). Bounded along the Kuril-Kamchatka Trench by the subduct-
ing Pacific plate and by a shallow zone of seismicity extending from Japan through
Sakhalin into the northeast of Russia, the Okhotsk Sea region may presently be part of the
North American plate.
Placing the origin and development of the Kuril backarc basin within the broader
geological framework and tectonic evolution of the Okhotsk Sea region must take into
account both the data from the basin and the surrounding structures. The basin is appro-
priately located behind the Kuril island arc where a backarc spreading produced basin
may be expected, and it has oceanic crustal thicknesses. However, the magnetic anomalies
are not seafloor spreading-type anomalies, so it is not possible to date the basement on the
basis of reversal time-scale modeling. To the extent that the anomalies are linear, they strike
transverse to the NE-SW long axis of the basin and are associated with basement highs.
Unfortunately, there are no samples of the deep oceanic crust for direct dating.
The thick sedimentary section filling the Kuril backarc basin is essentially tectonically
undisturbed, exhibiting no evidence for either significant extension or compression. Forma-
tion of the basin was apparently completed before most or all of the sedimentary fill was
deposited. Only vertical motion is suggested (subsidence) based on sedimentary structural
features at the basin margins.
Three lines of evidence are available to use in attempting to determine the age of
formation for the Kuril backarc basin: the ages of the dredged rock, sedimentation rates,
and age/depth values for oceanic crust. As we have noted, the Miocene/Oligocene age
dredge samples place a likely minimum age on the basin. Likewise, the Mesozoic age
dredge samples from the arc and the northwestern margin of the basin place a maximum
possible age on the basin of late Cretaceous. The Mesozoic rock samples represent the older
basement of the arc and the cratonic portion of the Okhotsk Sea, as found by many
investigators and discussed above, and the Kuril backarc basin certainly postdates the age
of these rocks. Based solely on the dredge samples, these are, however, rather broad limits
for the basin's age.
Sedimentation rates in the Kuril backarc basin are known only from the sampling of
the Holocene deposits, and estimates range from 200 to 400 mlMa from these studies.
Because these rates are only for the uppermost few meters of the basin deposits and have a
great range, it is highly speculative to use them alone for estimating rates for the entire
sediment section. In nearby backarc settings, the Japan Sea Basin and the Kamchatka-
Komandorsky Basin, where there has been ODP and DSDP drilling, deposition rates are
known for most or all of the sediment sections there and may be representative of
sedimentation rates for the Kuril backarc basin. Rates in the Japan Sea Basin rates range
from 9 to 77 m/m.y. (Burckle et at., 1992). In the Kamchatka-Komandorsky Basin, the
range is 70 to 200 m/m.y. (Scholl and Creager, 1973). Therefore, if we use a minimum of
30 m/m.y. and a maximum of 400 m/m.y., these values encompass those reported for
Miocene to present rates from the nearby backarc basins for Holocene sedimentation in the
Kuril basin. At 30 m/m.y. for an average 4500-m thickness of sediments the basement age
would be 150 Ma. At 400 m/m.y. the basement age would be 11.25 Ma. Again, these
extreme values are not very instructive. We note that for all except one of the Japan Sea and
Kamchatka-Komandorsky basin sites the average range of depositional rates is between
30 and 90 mlMa. If, on this basis, an average rate of 60 mlMa is applied to the Kuril Basin,
the basement age would be 75 Ma. The objective conclusion of this depositional rates
exercise is that ocean drilling is needed in the Kuril backarc basin.
Basement depth for the deeper portion of the Kuril backarc basin is between 7 and 8
km with 4 to 5 km of overlying sediments. Taking into account that backarc basins are
generally on the order of 1 km deeper than normal ocean crust for a given age and the
loading of the sedimentary deposits, these basement depths are still deeper than that
expected for a young backarc basin. Basement depths and sediment thicknesses more
closely match those of the Aleutian Basin known to be underlain by Mesozoic crust
(Marlow et at., 1990) than those for the Japan Sea Basin where the sediment thickness is
generally not as great and basement is 1 to 2 s shallower. These observations and compari-
sons would suggest the Kuril Basin is older than the minimum Miocene-Oligocene age
limit based on the dredged sedimentary rocks.
The lack of reversal-type magnetic lineations in the Kuril backarc basin can be
explained in two ways: The crust has been demagnetized by heating, or the basin was
completely formed during a period of single magnetic polarity. The only reasonable time
period for long single-magnetic polarities is the late Cretaceous.
We conclude that the Kuril backarc basin is floored by oceanic crust and that it was
formed by backarc spreading within the southwestern margin of the Mesozoic continental
structure of the Okhotsk Sea plate, behind the Kuril arc in response to Pacific plate
subduction. Based on the broadly distributed, rather uniformly thick and undisturbed
sediment fill, it apparently opened rapidly and has since remained inactive. The greater
width of the basin at its SW end, narrowing to closure at the south tip of Kamchatka,
suggests that the basin opened about a rotation pole very near the south end of Kamchatka.
The transverse shallow basement ridges which strike across the basin and the NNW striking
basement structures which extend between Sakhalin and Hokkaido at the SW end of the
basin are believed to be shearllateral fault zones that defined the opening direction.
Displacement on the onland portion of these faults is documented during Oliogene(?)-
Miocene age (Jolivet et aI., 1992). The Mesozoic age of the transverse basement ridges
indicates that these features are shear/rafted remnants of the pre basin Kuril arc-Okhotsk
Sea plate basement.
While we would like to be more definitive about the age of opening for the basin, the
only objective conclusion we can reach is that the basin opened over a short period,
sometime between late Cretaceous and Miocene time. The higher deposition rates and the
known age of lateral faulting from Sakhalin to Hokkaido support a Miocene age, while the
lower deposition rates, the depth of basement, and the lack of magnetic reversal lineations
support a late Cretaceous age.
Acknowledgments
The authors thank Dave Scholl for his many useful comments, Larisa Khankishieva
for her help in drafting illustrations, and Sandy Dunham for typing the manuscript. This is
Geodynamics Research Institute contribution no. 98.
REFERENCES
Andreyev, A. A., and Vorobyev, V. M. 1991. On the tectonics of the Okhotsk Sea region in terms of geomagnetic
data, Pac. Geol. 1:27-33. (in Russian)
Averynova, V. N. 1975. Deep Seismotectonics of Island Arcs (Nonhwest Pacific), Moscow, Nauka. (in Russian)
Balakina, L. M. 1979. Orientation and ruptures and movements in the sources of strong earthquakes of the north
and northwest parts of the Pacific ocean, Phys. Earth 4:43-52. (in Russian)
Balakina, L. M. 1981. Mechanism of intermediate earthquakes in the Kuril-Kamchatka focal zone, Phys. Eanh
8:3-24. (in Russian)
Belousov, V. V. 1982. Transition Zones between Continent and Ocean, Nedra, Moscow. (in Russian)
Belousov, V. V., and Pavlenkova, N. I. 1986. Interaction of crust and upper mantle, Geotectonics 6:8-20. (in
Russian)
Bezrukov, P. L. 1955. On the distribution and accumulation rate of siliceous sediments in the Okhotsk Sea, Proc.
Acad. Sci. USSR 103:433-476. (in Russian)
Bezrakov, P. L. 1960. Bottom sediments of the Okhotsk Sea, Proc. Inst. Oceanol. USSR 32:15-95. (in Russian)
Bikkenina, S. K., Anosov, O. I., Argentov, V. v., and Sergeyev, K. F. 1987. Crustal Structure of the Okhotsk Sea
Southern Part According to Seismic Data, Moscow, Nauka. (in Russian)
Bogdanov, N. A. 1988. Tectonics of Marginal Sea Deep-Sea Basins, Moscow, Nedra. (in Russian)
Burckle, L. H., Brunner, C. A., Alexandrovich, J., DeMenocal, P., Briscoe, J., Hamano, Y., Heusser, L., Ingle, J. c.,
Jr., Kheradyar, T., Koizumi, I., Krumsdiek, K. A. 0., Ling, H.-Y., Muza, J. P., Rahamn, A., Sturz, A.,
Vigliotti, L., White, L. D., Wippem, J. J. M., and Yamanoi, T. 1992. Biostratigraphic and biochronologic
synthesis of Legs 127 and 128: Sea of Japan, in Proc. ODp, Sci. Results, 1271128, Pt. 2 (K. Tamaki, K.
Suyehiro,1. Allan, M. McWilliams et aI., eds.), p. 1228, Ocean Drilling Program, College Station, TX.
Chapman, M. E., and Solomon, S. C. 1976. North American-Eurasian plate boundary in north-east Asia, 1.
Geophys. Res. 81:921-930.
DeMets, C. 1992a. Oblique convergence and deformation along the Kuril and Japan trenches, 1. Geophys. Res.
97:17,615-17,617, 17,625.
DeMets, C. 1992b. A test of present-day plate geometries for north-east Asia and Japan, J. Geophys. Res.
97:17,627-17,635.
Den, N., and Hotta, H. 1973. Seismic refraction and reflection evidence supporting plate tectonics in Hokkaido,
Pap. Meteorol. Geophys. 24(1):31-54.
Ermakov, V. A. 1991. The origin of the Kuril deep-sea basin, Pac. Geol. 1:34-49. (in Russian)
Far East Seismological Bulletin. 1986-1989. Institute of Marine Geology & Geophysics Publisher, Russian
Academy of Sciences, Yuzhno-Sakhalinsk. (in Russian)
Galperin, E.I., and Kosminskaya, I. P. (Eds.).1964. Structure of the Earth's Crust in the Transition Zonefrom the
Asian Continent to the Pacific Ocean, Moscow, Nauka. (in Russian)
Geological Map of the Japan and Kuril Trenches and the Adjacent Areas. 1978. Marine Geology Map Ser., 11,
Geol. Soc. Japan.
Gnibidenko, H. S., Gorbachev, S.Z., Lebedev, M. M., and Marakhanov, V. I. 1974. Geology and deep structure of
Kamchatka peninsula, Pac. Geol. 7:1-32.
Gnibidenko, H. S., and Snegovskoy, S. S. 1975. Structural relations between Sakhalin and Hokkaido, Proc. Acad.
Sci. USSR 224:1391-1394. (in Russian)
Gnibidenko, H. S., Bykova, T. G., Veselov, O. V., Vorobiev, V. M., and Svarichevsky, A. S. 1983. The tectonics of
the Kuril-Kamchatka deep-sea trench, in Geodynamics of the Western Pacific-Indonesian Region (T. W. C.
Hilde and S. Uyeda, eds.), pp. 249-285, Geodyn. Ser., Vol. 10, American Geophysical Union, Washing-
ton, DC.
Gnibidenko, H. S. 1985. The Sea of Okhotsk-Kuril islands ridge and Kuril-Kamchatka trench, in The Ocean
Basins and Margins, Vol. 7 A, The Pacific Ocean (A. E. S. Nairn, E G. Stehli, and S. Uyeda, eds.), pp. 377-
418, Plenum Press, New York.
Gnibidenko, H. S., and Ilyev, A. Ya. (Eds.). 1992. Catalogue of Dredging Stations in the Okhotsk Sea, Yuzhno-
Sakhalinsk, Institute of Marine Geology and Geophysics. (in Russian)
Gnibidenko, H. S., and Svarichevsky, A. S. 1984. Tectonics of the South Okhotsk deep-sea basin, Tectonophysics
102:225-244.
Gretskaya, E. V. 1990. Initial Oil and Gas Parent Potential of Organic Matter of Sediments with Reference to the
Okhotsk Sea Basins, Vladivostok, Nauka. (in Russian)
Harland, W. B. (Ed.). 1985. A Geological Time Scale, Moscow, Mir. (in Russian)
Hilde, T. W. c., Uyeda, S., and Kroenke, L. W. 1977. Evolution of the western Pacific and its margins,
Honza, E. (Ed.). 1978. Geological Investigation of the Okhotsk and Japan Sea off Hokkaido June-July 1978
(GH77-3 Cruise), Hisamoto Geological Survey of Japan.
Ilyev, A. Ya., Voronova, V. A., Zakharova, M. A., Nesterova, O. N., Tarakanova, L. I., Sheremetieva, G. N., and
Shustov, L. N. 1979. Bottom Sediments of the Okhotsk Sea Southern Part, Moscow, Nauka. (in Russian)
Jolivet, L., Fournier, M., Huchon, P., Pozhdestvenskiy, V., Sergeyev, K. E, and Oscorbin, L. S. 1992. Cenozoic
intracontinental dextral motion in the Okhotsk-Japan Sea region, Tectonics 11:968-977.
Jolivet, L., Huchon, E, and Rangin, C. 1989. Tectonic setting of Western Pacific marginal basins, Tectonophysics
160:23-47.
Karig, D. E. 1971. Origin and developments of marginal basins in the Western Pacific., J. Geophys. Res. 76:2542-
2561.
Kiminami, K., and Kontani, Y. 1983. Mesozoic arc trench system in Hokkaido, Japan, in Accretion Tectonics in the
Circum-Pacific Regions (M. Hashimoto and S. Uyeda, eds.), pp. 107-122, Terrapub, Tokyo.
Kimura, G. 1986. Oblique subduction and collision: forearc tectonics of the Kuril arc, Geology 14:404-407.
Kogan, M. G. 1975. Gravity field of the Kuril-Kamchatka arc its relation to the thermal regime of the lithosphere,
J. Geophys. Res. 80:1381-1390.
Krasny, M. L. 1990. Geophysical Fields and Deep Structure of the Okhotsk-Kuril Region, Vladivostok, Nauka. (in
Russian)
Lundgren, P. R., and Giardini, D. 1990. Lateral structure of the subducting Pacific plate beneath the Hokkaido
comer from intermediate and deep earthquakes, Pure Appl. Geophys. 134:385-404.
Maeda, J. 1990. Opening of the Kuril basin deduced from magmatic history of central Hokkaido, North Japan,
Marlow, M. S., Cooper, H. K., Dodisman, S. v., Geist, E. L., and Carlson, P. R. 1990. Bowers swell: Evidence for a
zone of compressive deformation concentric with Bowers Ridge, Bering Sea, Mar. Petrol. Geol. 7:398-409.
Matsuda, T., and Uyeda, S. 1971. On the Pacific-type orogeny and its model-extension of the paired belts concept
and possible origin of marginal sea, Tectonophysics 11:5-27.
Melnikov, O. A. 1988. Geology of Sakhalin-Hokkaido Folded Belt, Vladivostok, Nauka. (in Russian)
Popov, A. A., and Anosov, G. I. 1978. New data on crustal structure of the Kuril Basin, Proc. Acad. Sci. USSR
240: 166-168. (in Russian)
Okada, H. 1982. Geological evolution of Hokkaido, Japan: An example of collision orogenesis, Proc. Geol. Assoc.
93:201-212.
Riegel, S. A., Fujita, K, Kozmin, B. M., Imaev, V. S., and Cook, D. B. 1993. Extrusion tectonics of the Okhotsk
plate, north-east Asia, Geophys. Res. Leu. 20:607-610.
Rozhdestvensky, V. S. 1986. Evolution of the Sakhalin fold system, Tectonophysics 127:331-339.
Saidova, H. M. 1960. Regulations of foraminiferal distribution in the Okhotsk Sea bottom sediments, Proc. Inst.
Oceanol. USSR 32:96-159. (in Russian)
Savostin, L., Zonenshain, L., and Baranov, B. 1983 Geology and plate tectonics of the Sea of Okhotsk, in
Geodynamics of the Western Pacific-Indonesian Region (T. W. C. Hilde and S. Uyeda, eds.), pp. 189-221,
Geodyn. Ser., Vol. 11, American Geophysical Union, Washington, DC.
Scholl, D. w., and Creager, 1. S. 1973. Geologic synthesis of Leg 19 (DSDP) results. Far North Pacific, and
Aleutian Ridge, and Bering Sea, in Init. Repts. DSDP, 19 (1. S. Creager, D. W. Scholl et al., eds.), pp. 897-
913, U.S. Government Printing Office, Washington, DC.
Seismological Bulletin. 1990-1993. Russian Academy of Sciences, Obninsk. (in Russian)
Simbireva, I. G., Fedotov, S. A., and Feofilaktov, V. D. 1976. Heterogeneities of the stress field in the Kuril-
Kamchatka arc as derived from seismological data, Geophys. Geol. 1:70-86. (in Russian)
Snegvskoy, S. S. 1974. Seismic Reflection Studies and Tectonics of the Okhotsk Sea Southern Part and the
Adjacent Pacific Margin, Nauka, Novosibirsk. (in Russian)
Soloviev, V. V., I. K. Tuezov, I. K, and Vasilyev, B.1.1977. The structure and origin of the Okhotsk and Japan Sea
abyssal depressions according to new geophysical and geological data, Tectonophysics 37:153-166.
Starshinova, E. A. 1980. Crustal and upper mantle structure inhomogeneity of the Okhotsk Sea, Proc. Acad. Sci.
USSR 255:1349-1343.
Svarichevsky, A. S., and Svarichevskaya, L. V. 1982. The Okhotsk Sea slope relief in the southern part of the Main
Kuril Ridge, in Relief and Volcanism of KurilIsland Arc System (B. V. Ezhov, ed.), pp. 54-62, Vladivostok,
Nauka. (in Russian)
Sychev, P. M., and Snegovskoy, S. S. 1976. Abyssal depression of the Okhotsk, Japan and Bering Seas, Pac. Geol.
11:57-80.
Sychev, P. M., Soinov, V. v., Veselov, O. v., and Volkova, N. A. 1983. Heat flow and geodynamics of the transition
zone from Asia to the North Pacific, in Geodynamics of the Western Pacific-Indonesian Region (T. W. C.
Hilde and S. Uyeda, eds.), pp. 237-247, Geodyn. Ser., Vol. 11, American Geophysical Union, Washing-
ton, DC.
Takanashi, M. 1983. Space-time distribution of late Mesozoic to early Cenozoic magmatism in East Asia and its
tectonic implications, in Accretion Tectonics in the Circum-Pacific Region. (M. Hashimoto and S. Uyeda,
eds.), pp. 69-87, Terrapub, Tokyo.
Tarakanov, R. Z. and Kim, Chung Un. 1983. Seismofocal zones and geodynamics of the Kuril-Japan region, in
Geodynamics of the Western Pacific-Indonesian Region (T. W. C. Hilde and S. Uyeda, eds.), pp. 223-236,
Geodyn. Ser., Vol. 11, American Geophysical Union, Washington, DC.
Tarakanov, R. Z., and Leviy, N. V. 1968. A model for the upper mantle with several channels oflower velocity and
strength, in The Crust and Upper Mantle of the Pacific Area (G. H. Sutton, M. H. Manghnani, and R.
Moberly, eds.), pp. 43-50, Geophys. Monogr., American Geophysical Union, Washington, DC.
Tereschenkov, A. A., Baboshina, V. A., Tuezov, I. K, and Kharahinov, V. V. 1988. Structure of the Okhotsk Sea
region anomalous magnetic fields, Geophys. Geodyn. Res. 10:10-19. (in Russian)
Tuezov, I. K., Veselov, O. v., and Lipina, E. N. 1984. Heat Flow of Asia, Australia and Western Pacific,
Vladivostok, Nauka. (in Russian)
Vassilkovsky, N. P. 1967. On the geological nature of the Pacific mobile belts, Tectonophysics 4:583-593.
Veselov, O. v., and Volkova, N. A. 1981. Radioactivity of rocks in the Okhotsk Sea region, in Geophysical Fields of
the Transition Zone of Pacific Type (M. L. Krasny, ed.), pp. 51-709, Vladivostok, Nauka. (in Russian)
Volkova, N. A. 1982. The therrnoconductivity model of the crust of the Okhotsk Sea region, Geol. Geophys. 5:
92-97. (in Russian)
Watson, B. E, and Fujita, K 1985. Tectonic evolution of Kamchatka and the Sea of Okhotsk and implications for
the Pacific basin, in Tectonostratigraphic Terranes of the Circum-Pacific Region (D. G. Howell, ed.), pp.
333-336, Circum-Pacific Council Energy and Mineral Resources, Houston, TX.
Zhang Zhi, and Lay, T. 1993. Investigation of upper mantle discontinuities near northwestern Pacific subduction
zones using precursors to SSH, J. Geophys. Res. 98:4389-4405.
Zhuze, A. P. 1957. Diatomic in the surface layer of the Okhotsk Sea sediments, Proc. Inst. Oceano/. USSR 22:164-
220. (in Russian)
Zlobin, T. K., and Zlobina, L. M. 1991. Crustal structure of the Kuril Island arc system, Pac. Geol. 6:24-35. (in
Russian)
Zonenshain, L. P., and Savostin, L. A. 1979. Introduction to Geodynamics, Moscow, Nedra. (in Russian)
Zverev, S. M., and Tulina, Yu. V. (Eds.). 1971. Deep Seismic Sounding of the Earth's Crust in the Sakhalin-
Hokkaido-Primorye Zone, Moscow, Nauka. (in Russian)
13
Hydrothermal Activity Related

to Arc-Backarc Magmatism
in the Western Pacific
Jun-ichiro Ishibashi and Tetsuro Urabe
ABSTRACT
A compilation of 27 sites of hydrothermal mineralization in the western Pacific was made

in terms of their tectonic settings, mineral commodity, and fluid chemistry. These sites
constitute about 20% of the known seafloor hydrothermal sites, which are dominated by
those occurring at mid-ocean ridges (MORs). High-temperature hydrothermal activities in
the western Pacific exclusively occur in association with submarine volcanism found either
on backarc spreading centers, backarc rifts, or volcanic fronts in arc-backarc systems.
Magmatic contribution to the hydrothermal systems in arc-backarc settings is more obvious
than those in MORs, and several lines of evidence suggest direct interactions between the
magma chamber and hydrothermal fluids.
Hydrothermal activity at backarc spreading centers shares several common chemical,
mineralogical, and morphological features with those in MORs. This is understandable
because of small differences in parameters, such as magma signature and composition of
oceanic crust, between those two systems. Frontal-arc volcanoes are proved to be the sites of
vigorous high-temperature hydrothermal activity. Several investigated submarine calderas
in the Izu-Bonin (Ogasawara) and Mariana arcs have signs of present or recent activity,
indicating Ubiquitous hydrothermal mineralization regardless of their magma composition.
Backarc rifting centers are particularly interesting because of their analogous tectonic
settings with the Japan arc of Miocene age where many Kuroko-type volcanic massive sulfide
deposits (VMSDs) formed in association with bimodal volcanisms. Some hydrothermal sites
in the Okinawa Trough backarc basin represent modem examples of such mineralization.
The wide variety in modes of occurrence of hydrothermal systems in the western
Pacific region, in contrast to those in the MOR setting, reflects diversity in magma
composition, contribution from heterogeneous island-arc crust rather than homogeneous
oceanic crust, and the unique tectonic setting of the arc-backarc system. These systems
Jun-ichiro Ishibashi • Laboratory for Earthquake Chemistry, Faculty of Science, University of Tokyo,
Bunkyo-ku, Tokyo 113, Japan. Tetsuro Urabe • Geological Survey of Japan, Tsukuba 305, Japan.
451
452 JUN-ICHIRO ISHIBASHI AND TETSURO URABE
provide the best field to investigate analogs to VMSDs, because most ancient VMSDs are
hosted by the rock types in a subduction-related volcanic suite or an extension-related
bimodal suite representative of island-arc- or continental-rift-related tectonic settings.
1. INTRODUCTION
Five or six years after the discovery of hydrothermal mineralization at 21°N East
Pacific Rise (EPR; e.g., Francheteau et al., 1979; Spiess et al., 1980), reports of similar
seafloor hydrothermal activity were noted in the western Pacific (e.g., Both et aI., 1986;
Hawkins, 1986; Hawkins and Helu, 1986; Craig et al., 1987a,b; Urabe et al., 1987).
Continued exploration efforts in the last decade have shown that the western Pacific region
has as much seafloor hydrothermal mineralization associated with it as with the mid-ocean
ridges (MORs). Hydrothermal mineralization in the western Pacific appears to have wider
variations in tectonic setting, mode of occurrence, and chemistry of fluid and resultant
deposits than those of MORs.
Prior to 1970, the number of known seafloor hydrothermal mineral deposits was only
one, that discovered in the Red Sea in 1966 (Miller et aI., 1966). But it had increased to
50 by 1983 (Rona, 1983), 80 by 1988 (Rona, 1988), and to about 140 in the latest compilation
(Rona and Scott, 1993). By contrast, the number of the arc-backarc-related hydrothermal
occurrences in the western Pacific compiled in this chapter is 27, thus comprising about
20% of all known seafloor hydrothermal occurrences. It is notable that this estimation
would be biased from the actual distribution, because only a small portion of the seafloor
has been investigated. Indeed, in recent years, the Japanese deep submergence research
vessel (DSRV) Shinkai 2000 and Shinkai 6500, the American DSRV Alvin, the French
DSRV Nautile, and the Russian DSRV Mir have been deployed energetically and widely in
the western Pacific, away from their home countries, and have contributed to successive
discoveries of hydrothermal occurrence.
This chapter summarizes contemporary hydrothermal mineralization in arc-backarc
settings in the western Pacific region. Several occurrences of hot-spring activity around
shallow submarine volcanoes and volcanic islands, such as the Esmeralda Bank of the
Mariana arc (Sttiben et aI., 1992) and the Wakamiko caldera in the Kagoshima Bay in
southwest Japan (Nedachi et aI., 1991), low-temperature hydrothermal manganese oxides
on frontal volcanoes (e.g., Moorby et al., 1984; Usui et al., 1986), and cold seepage from
accretionary prism of the Japan Trench and Nankai Trough (e.g., Taira and Pickering, 1991)
are excluded because of their less important nature in terms of metal concentrations.
However, cold venting found at serpentinite seamounts of the Mariana forearc (Fryer et al.,
1987b, 1990) is discussed because of the formation of aragonite chimneys.
Relevant comparative data from MOR hydrothermal systems are not cited here;
readers are referred to excellent review papers by Rona (1983, 1984, 1988), Backer and
Lange (1987), von Damm (1990), and Rona and Scott (1993).
2. TECTONIC SETTING AND ASSOCIATED VOLCANISM
Seafloor hydrothermal mineralization has occurred throughout the history of the

Earth. The oldest known example is at Isua, Greenland, where copper sulfide layers occur
HYDROTHERMAL ACTIVITY IN THE PACIFIC 453
concordantly within a banded iron formation (Appel, 1979) in a supracrustal belt of about
3.8 Ga (Compston etal., 1986). More than 500 VMSDs are known on land. The VMSDs are
found in association with volcanism of virtually every lithologic type: felsic to ultramafic
rocks in predominantly volcanic or mixed volcanic-sedimentary environments.
Since the 1970s, many attempts have been made to relate the geologic setting of
VMSDs to certain tectonic environments such as MORs, volcanic arc-subduction zones,
and marginal basins (e.g., Sawkins, 1972; Sillitoe, 1972; Mitchell and Bell, 1973; Baranov
and Levin, 1993). There is unanimous agreement that most of the major Phanerozoic
VMSDs formed in island-arc and/or backarc settings rather than at MORs (e.g., Sawkins,
1972; Franklin et aI., 1981; Scott, 1985). Rona (1988) indicated that basalt-hosted
VMSDs that might occur at spreading centers of MORs or backarc basins constitute only
17% of the 508 deposits of his compilation, and that the majority of deposits (56%) are
hosted by rhyolite, which is considered to form in a subduction-related volcanic suite or
to an extension-related bimodal basalt-rhyolite suite within an island arc or continental
rift. Because of the range of acidic to basic submarine volcanism existing in diverse
tectonic settings, the western Pacific provides the best field test as to whether or not
present-day VMSDs are forming in similar settings to those formed during the Phaner-
ozoic.
In the western Pacific, many backarc basins have been accreted and preserved along
the margin of the Pacific plate since late Cretaceous time (Kroenke, 1984; Honza, 1991;
Tamaki and Honza, 1992). About 75% of backarc basins worldwide occur in the western
Pacific region and six of them have active spreading centers: Mariana Trough, Andaman
Sea, Manus Basin, Woodlark Basin, North Fiji Basin, and Lau-Havre Basin (cf. Tamaki
and Honza, 1992). These backarc spreading centers are the most important locations of
submarine volcanism comparable to that at MORs. In contrast, the majority of present-day
arc-trench systems in the western Pacific lack an associated backarc basin (Taylor and
Kamer, 1983). Volcanism in such arc-trench systems peaks on the volcanic front, which is
located on the trench-side border of the active volcanic sites. Other tectonic units where
sporadic volcanoes are observed include the forearcs, intra-arc rifts, and backarc rifts. Since
the mechanism of this rifting activity is one of the least understood aspects of the evolution
of arc-backarc systems, it is in practice difficult to categorize without ambiguity every
known volcano into a simple tectonic regime such as forearc, volcanic front, backarc rift, or
backarc spreading center.
Regardless of their tectonic category, it is quite reasonable to expect vigorous hydro-
thermal activity at most ofthe active volcanic centers in arc-backarc systems. Rifting of an
island-arc or continental crust is likely to provide a favorable setting for the circulation of
seawater eventually evolving as a hydrothermal fluid. This, combined with the fact that a
submarine volcano has a shallow magma chamber inside the edifice will provide an ideal
structure for the hydrothermal system to focus high-temperature venting at the summit
(Scott, 1985, 1987). Backarc spreading centers show several similarities to MOR systems
where hydrothermal activity is widely distributed. This variability in possible location
makes investigation of hydrothermal activity in the western Pacific more complex but
potentially more interesting than in those arguably simple tectonic settings, such as MORs
and hot spots.
This chapter reviews some major sites of hydrothermal mineralization associated with
three major tectonic settings: backarc spreading center, backarc rift, and volcanic front.
These, and other miscellaneous examples, are briefly summarized in Table I together with
~
TABLE I
Hydrothermal Mineral Occurrences in Arc-Backarc Systems in the Western Pacific
Location Water
No. location depth
(Fig. 13.1) (m) Geologic structure Type of deposit Mineralogy
I. Backarc spreading center

I North Fiji Basin 1980 Axial graben at topographic high of Active (T = 290°C) anhydrite chim- Anhydrite, amorphous silica in dead
"Station 4" N.-central segment near triple junc- neys standing on dead sulfide chimneys; pyrite, marcasite
(l6°59'S, 173°55'E) tion. Sheet lava floor. mound. Forest of dead sulfide chalcopy., sphal., wurtz., goethite.
chimneys.
2 North Fiji Basin 2720 Collapsed lava lake on flat rise crest Warm (T = 5.2°C) fluid discharge None.
"Station 14" of fast spreading S-central seg- through mussel bed. No hydrother-
(l8°50'S, 173°30'E) ment. No sediment cover. mal minerals. Site of megaplume ~
(Nojiri et al., 1989).
3 Fiji transform fault 1860- Short spreading ridge axis which dis- Hydrothermal sulfide impregnation in Magnetite, pyrrhotite, chalcopyrite
"Extensional Relay Zone A" 2335 places Fiji transform fault as inter- MORB-Iike basalt dredged from and opal on fracture surface. ~
(l6°IO'S, 177°'25E) preted by Jarvis et al. (1994) axial valley.
4 Central Manus Basin 2500 2-km-wide axial rift graben of the Sulfide chimneys up to 20 m high are Sphalerite, wurtzite, pyrite, marcas,
~
"Vienna Woods" northeast spreading center. Mostly venting clear, milky and black chalcopy, galena, am. silica, barite.
~
(3°IO'S, 1500 17'E) massive pillow lava floor. fluids. Sulfate smokers are also Sulfate chimney; anhy., silica, bar- ~
):.
present. ite.
5 Eastern Manus Basin 2000 Caldera of basaltlbasaltic andesite at Sulfide ores were not recovered. Ferruginous oxide deposits. pyrite ~
"DESMOS cauldron" an intersection of a spreading cen- Megaplume-like methane anoma- and native sulfur disseminated in rr1
(3°42'S, 151°52'E) ter and a transform fault? lies in water column over the cal- basaltic andesite. Ci1
dera.
6 Eastern Manus Basin 1650 Crest of a prominent ridge of dacite 4-m-high sulfide chimney venting Cohesive anhydrite, chalcopy., bor- ~
"PACMANUS field" flows and domes called Pual Ridge "smoke" (no temperature informa- nite, tennantite, and sphalerite. c::
(3°42'S, 15 P42.6'E) which exists in pull-apart basin. tion). Only one ore chip (I cm) ~
was recovered. m
7 Valu Fa Ridge, Lau Basin 1700 4OO-m-Iong zone along a normal fault Black and white smokers venting up Pyrite, chalcopy., marcasite, sphal., ::x:
"Vai Lili site" on a ridge crest. Basaltic andesite to 400°C fluid. Height up to 17 m. barite, tennantite, galena, gratonite, t§
(22°18'S, 176°35'W) to rhyodacite lava. Massive Zn- and Cu-sulfide ores native gold. II
0
are present.
8 Valu Fa Ridge, Lau Basin 1900 On south Valu Fa Ridge crest domi- Low-temperature Mn-oxide crust is d-Mn02, birnessite, todorokite, amor-
:t
"Hine Hina site" nated by andesite and dacite. covering high-temperature fossil phous Fe-oxide, goethite, sphal- ~
(22°33'S, 176°43'W) Strongly bleached volcanic rocks. sulfide deposits. Diffusive fluid (T erite, barite, pyrite, chalcopy. ~
,....
= 40°C). ),.
C)
9 Valu Fa Ridge, Lau Basin 1946- At the top and along the flanks of 1O-15-m-high inactive barite-sulfide Barite and sphalerite are dominant ::!
"White Church site" 1966 Northern Valu Fa Ridge. Brecci- chimney field extends over 300 m with lesser galena, tennantite, :s-t
(21 °55'S, 176°32'W) ated and pillow lava, and pumice along normal faults. chalcopy, gold, and pyrrhotite. -<
field. ~
10 Northeastern Lau Basin 2100 Axial region of northeasterly trending Dredged black smoker chimney sam- Wurtzite, pyrite, chalcopyrite, barite,
"Papatua expedition site" active spreading ridge of the north- pies. and amorphous silica. Thin film of
:tIII
(15°17'S, 174°45'W) em Lau Basin. Mn-oxyhydroxide. ~
II Peggy Ridge, Lau Basin 1664- At a fracture zone offsetting north- Dredged hyaloclastite basalt frag- Opal, barite, montmorillonite, phillip- S}
(J6055'S, 176°49.5'W) 1900 westerly trending active spreading ments which are cemented by opal site (oxygen-isotope study suggests :n
C)
center of the northern Lau Basin. and euhedral barite crystals. temp. of 117-134°C).
12 Western Woodlark Basin 2143- Westernmost propagating tip of Spires and mounds of Fe-Mn-Si ox- Inactive barite silica chimneys con-
"Franklin Seamount" 2366 spreading center. Basaltic andesite ide up to several meters thick and tain up to 21 ppm Au. Si-bearing
(9°55'S, l5l °50'E) and inferred sodic rhyolite. 200 m in extent. Venting 20-30°C Fe oxyhydroxide.
clear solution.
13 Central Mariana Trough 3600- Hank of an axial volcano of basaltic Active barite-rich sulfate chimneys, I Barite chimneys. Sulfides; abundant
"Alice Spring field" and 3700 andesite on backarc spreading cen- m high, venting clear fluid (T = sphal with lesser amounts of gal-
other ter. Volcanics are fractionated. 287°C). ena, chalcopy, pyrite.
(18°12'E, 143°30'E)
14 Central Mariana Trough 3675 Crest of an axial ridge where the re- Several active chimneys surrounded Sphalerite, barite, amorphous silica.
(18°02'N, 144°45'E) lief is greatest (ca. 800 m). A low by a low mound of hydrothermal
mound, 20-30 m in diameter. precipitates.
(continued)
~
~
TABLE I
( Continued)
Location Water
No. location depth
(Fig. 13.1) (m) Geologic structure Type of deposit Mineralogy
2 Backarc rift
15 Sumisu rift 1530- On the flank of a rhyolite dome of en Several tens of fossil barite-silica Barite, amorphous silica, opal-CT
Izu-Ogasawara (Bonin) arc 1600 echelon ridge of bimodal (rhyolite- chimneys with thin manganese (oxygen isotope study suggests
(31 °06'N, 139°54'E) BABB) volcanism. oxyhydroxide coating. temp. of formation <150°C).
16 Northern Okinawa Trough 690- Volcanic depression and hills where Several anhydrite chimneys venting Anhydrite chimneys. Sphal., wurtz.,
"Minami-Ensei Knoll" 705 pumice, diorite and granite frag- clear fluid (T = 267-278°C) are tetrahedrite, chalcopy., galena, py-
(28°23.5'N, 127°38.5'E) ments are recovered. observed on barite-sulfide mound. rite, barite, bornite, coveIlite
17 Middle Okinawa Trough 1400 Vescicular pillow basalt ridge located Carbonate-rich hydrothermal precipi- Calcite, rhodocrosite, anhydrite,
"Iheya Ridge (CLAM site)" on the axial part of backarc rift of tates with disseminated sulfides. wurtzite, pyrrhotite, galena, iso-
(27°33'N, 126°58'E) the Okinawa Trough. Fluid emanation (T = 100-220°C) cubanite, chalcopyrite, argentite.
from fissures. ~
3 Volcanic front
18 Middle Okinawa Trough 1300- Northeastern slope of a sediment-rich Active black smoker (T = 320°C), Sphalerite, tetrahedrite, galena, barite,
"Izena Cauldron (JADE 1550 rectangular depression with rhyoli- white smokers and sulfide crusts. chalcopyrite, pyrite, anhydrite, ~
site)" tic lava domes. Liquid-C0 2-hydrate bubbling are stibnite and other sulfides.
(27°I6'N, 127°05'E) also observed. ~
19 Izu-Ogasawara (Bonin) arc 1300- A submarine caldera with diameter of Submersible dive failed to locate ac- Barite, pyrite, chalcopyrite, sphalerite, ~
"Kita-Bayonnaise Caldera" 1400 5-7 km and has dacitic central tive discharge but box coring re- tetrahedrite-tennantite, epidote, car- ~
cone. covered barite-sulfide fragments ):.
(32°06'N, 139°51 'E) bonate, smectite.
from the floor. ~
20 Izu-Ogasawara (Bonin) arc 1000- Partly emerged submarine caldera Massive barite and sulfide samples Barite, pyrite, and other minor sul- ~
"Myojinsho Caldera" 1100 with flat-topped central cone of were recovered from the caldera fides. ~
(31°53'N, 139°59'E) dacite. floor sediments.
21 Izu-Ogasawara (Bonin) arc 1370 Calc-alkaline dacite seamount of low- High-temperature (T= 230-311°C) Chalcopyrite, sphalerite, anhydrite,
~
o
"Suiyo seamount" K series. Dimension of the caldera fluid discharged from chimneys barite. High Au content (up to 71 c::
(28°34'N, 1400 39'E) is 1.5 x 0.8 km. «1m) in a field of 150 x 300 m ppm). ~
on the caldera floor. ~
22 Izu-Ogasawa (Bonin) arc 930 Andesitic caldera with basaltic central Veinlets of sulfides in altered an- Pyrite, sphalerite, Cu-Fe-sulfide (fill- ::r:
"Kaikata seamount" cone called KC caldera. Low-temp. desite. Water in caldera is contami- ing temp. of fluid inclusions are ~
(26°42'N, 141°05'E) activity on the cone. nated with iron hydroxide. 290°C).
Basalt and basaltic andesite volcano Warm fluid (T = 39°C) rich in carbon Elemental sulfur, Fe- and Mn-oxide,
~
23 Northern Mariana arc 402
"Kasuga 2 seamount" at the intersection of Mariana vol- dioxide and sulfur dioxide is vent- nontronite.
~
(21°35'N,143°37'E) canic arc and a fracture. ing from sulfur-encrusted volcanic SJ
sand. ~
r-
24 Southern Mariana Ridge 1470 At the summit of a seamount of White smoker chimneys emanating Anhydrite, barite, with lesser amounts ):.
()
"Forecast Vent Field" island-arc basalt located about 17 clear fluid (T = 202°C). Height of of sulfide minerals. :::!
(13°24'N, 143°55'E) km east of spreading axis. chimneys are about 1 m. S
4 Other seamounts ~
25 Derugin Depression, east of 1460- A knoll in volcano-tectonic depres- Fragments of travertine-like barite ag- Barite (834S = +20 per mil, which ~
Sakhalin Island, Okhotsk 1480 sion. Submarine caldera? Strong gregates were dredged. No tem- suggests its seawater sulfate ori-
hydrogen sulfide odor on sample. perature measurements. gin).
~
Sea I'll
(54°02'N, 146°16'E) }!
26 Komandorsky Basin A volcanic cone of fresh dacite to the Emission of volcanic gases (methane, Nontronite, Fe- and Mn-hydroxide Q
"Piip volcano" rear of a transform fault of Aleu- hydrogen) and precipitation of non- found in dredged sample. !!
()
(55°23'N, 167°27'E) tian arc. Backarc rift? tronite and ferro- and manganese
hydroxide.
5 Forearc Area
27 Mariana forearc region 1500 Mud volcano or solid intrusion of Cold seepage of dilute fluid from Aragonite, calcite and amorphous
Conical seamount serpentinite in forearc region. Mg-silicate chimneys. Aragonite Mg-silicate, Mg-silicate, ferro-
(19°33'N, 146°39'E) chimneys are also common. manganese oxide
Major reference of each site: 1,2. Bendel et al. (1993); Gracia et al. (1994); 3. von Stackelberg et al. (1990); 4. Tufar (1990); 5. Garno et al. (l993c); 6. Binns and Scott (1993); 7-9. Fouquet et al. (1993);
Lisitsyn et al. (1992); 10. Hawkins and Helu (1986); II. Bertine and Keene (1975); 12. Lisitsyn et al. (1991); Binns et al. (1993); 13. Craig et al. (l987b); 14. Hawkins et al. (1990); IS. Urabe and Kusakabe
(1990); 16. Nedachi et al. (1992); 17. Gamo et al. (199lb); 18. Halbach et al. (1989); 19. lizasa et al. (1993); 20.lizasa et al. (1992); 21. Watanabe and Kajimura (1993); 22. Urabe et al. (1987); 23. McMurtry
et al. (1993); 24 Garno et al. (1993a); 25. Astakbova (1992); 26 Baranov et al. (1991); 27. Fryer et al. (1990).
~
100' 120'
, 40'
IIt
I'
II
1 20 ,
20'
o· 0'
i'
'I
i
I,
20' 20
40· 40
=
120' 140· 160' 180' 160'
FIGURE 13.1. Hydrothermal mineral occurrences related to arc-backarc systems of the western Pacific. Num-
bers are listed in Table 1.
their mineralogy and mode of occurrence. The locations of the study fields are plotted in
Fig. 13.1.
3. CASE STUDIES
3.1. North Fiji Basin

The North Fiji Basin is an active, mature marginal basin in the southwestern Pacific
(Fig. 13.2), which was investigated during the Japan-France-SOPAC cooperative project
STARMER (e.g., Auzende et ai., 1989, 1990, 1991, 1992; Tanahashi et ai., 1991; Auzende
and Urabe, 1994; Auzende et al., Chapter 4 this volume). The 800-kIn-long linear backarc
spreading center is comparable in size to a MOR spreading system (Tanahashi et ai., 1994).
The Wadati-Benioff zone does not extend beneath the spreading center. Petrological
studies of basalt samples have revealed features similar to N-MORB with some influence of
coexisting oceanic island basalts (OIB) components (Eissen et at., 1991, 1994; Nohara et at.,
1994). This spreading ridge is divided into four first-order segments: northern (NI600E),
northern-central (NI5°E), southern-central (N5°E), and southern offset (NS) segments
(Tanahashi et ai., 1994).
.,
Q
o Q
10°5
o
20°
1700W
FIGURE 13.2. Location and tectonic setting of North Fiji and Lau basins. The solid circles mark the hydrother-
mal fields listed in Table I. The thick lines indicate backarc spreading centers, wheras the thin lines delinate
contours of 2000 m water depth.
Hydrothermal activity along the North Fiji Basin spreading ridge has been denoted by
chemical anomalies detected in hydrothermal plumes during the PAPATUA expedition of
1986 (Craig et al., 1987a), the SEAPSO III cruise of 1986 (Auzende et aI., 1988), and the
RN Moana Wave cruise of 1987 (Sedwick et aI., 1990). Systematic investigation of the
spreading ridge sites during the STARMER project, using a deep-tow TV camera system in
conjunction with dives made by the DSRV Nautile and DSRV Shinkai 6500, confirmed two
active hydrothermal sites (Honza et al., 1989; Auzende et aI., 1991).
An active hydrothermal chimney known as "White Lady" occurs in the 2-km-wide
axial graben on a bathymetric rise (location 1 on Fig. 13.2). This site is located at the
northern tip of the northern-central (N15E) segment and adjacent to a ridge-ridge-fracture
zone triple junction (Lafoy et aI., 1991). The White Lady consists exclusively of anhydrite
and is developed on a 2-m-high sulfide mound. Other anhydrite chimneys were found at the
STARMER II site about 150 m southwest of the White Lady chimney. The broad hydro-
thermal field named "Station 4" is distributed over a thinly sedimented fresh sheet lava
flow (Bendel et al., 1993). Dense hydrothermally supported fauna of balanoid, black
gastropods, hairy gastropods, and mussels thrive around the chimneys (Desbruyeres et aI.,
1994). These anhydrite chimneys were venting clear hydrothermal fluid of 230 to 290°C.
Their fluid chemistry is characterized by low chlorinity (47-49% of seawater) and deple-
tion in metal elements such as iron and manganese. This chemistry indicates segregation
between brine and vapor beneath the seafloor and that the latter phase has condensed to
form the hydrothermal fluid that was sampled (Grimaud et al., 1991; Ishibashi et al., 1994a).
The hydrothermal activity at White Lady persisted for two years between the two diving
cruises from 1989 to 1991 (Auzende et aI., 1992), with little change in the fluid chemistry
(Ishibashi et aI., 1994a). In contrast to active sulfate mineral occurrences, no active sulfide
chimneys have been observed. At the northern side of the White Lady site, a 2-km-wide,
fossil hydrothennal area was discovered. The height of the sulfide chimneys in this "Pere
Lachaise" area reaches IS m, implying that intense black smoker hydrothennal activity
previously existed in this same graben (Auzende et al., 1991; Bendel et a!., 1993). The
mineral assemblages of the sulfide mounds and of the fossil chimneys are very simple and
resemble those found at mid-ocean ridges.
In the southern-central (NSOE) segment, diving revealed collapsed lava lakes on
sediment-free axial depressions a few tens of meters wide and a few meters deep (Auzende
et a!., 1992). Biological communities of mussels, galatheids, and synaptic holothurians
exist in the collapsed lava lake at a depth about 2720 m (Auzende et a!., 1991). At this site,
known as "Station 14" (location 2 on Fig. 13.2), a massive release of fluid as a megaplume
was identified from a water column anomaly during the KAIYO 87 cruise (Nojiri et al.,
1989). However, only warm fluid emanations (S.2°C) were observed during the Shinkai
6500 dives of 1993 (Gracia et al., 1994). The low-temperature fluid samples showed slight
chemical anomalies in Si02, Mn, Li, and CH4 contents and helium isotopic composition,
suggesting that the high-temperature fluid was diluted by a high degree of mixing with
ambient seawater (Ishibashi et al., 1994c).
3.2. Manus Basin

The Manus Basin, located in the Bismarck Sea, is a backarc basin situated to the north
of the New Britain island arc-trench system (Fig. 13.3). Backarc spreading in this area is
characterized by an exceptionally high spreading rate, greater than 10 crn/yr full rate
(Taylor, 1979). Spreading centers in the western, central, and eastern basins are connected
by transfonn faults.
The first discovery of hydrothennal deposits in the Manus Basin was made in 1986 via
seafloor camera photographs (Both et a!., 1986). The photographs showed several inactive
FIGURE 13.3. Location and tectonic setting of Manus and Woodlark basins. The solid circles mark the
hydrothermal fields listed in Table I. The solid lines indicate backarc spreading centers and associated transform
faults.
chimneys about 2 m in height with associated fauna composed of gastropods, galatheid

crabs, and horny corals. This area, subsequently named "Red Star," is located on a small
axial volcanic swell at a depth of about 2500 m inside the rift valley of the spreading center
in the central basin. In 1990, the OLGA II expedition located a number of hydrothermal
fields including four major fields, in this spreading ridge and retrieved several mineral
samples (Tufar, 1990). At the same time as the OLGA II cruise, the DSRV Mir investigated
some of these hydrothermal fields (Lisitsyn et ai., 1993; Shadlun et aI., 1993).
The active hydrothermal site named "Vienna Woods" (3°9.9'S, 150016.8'E, depth
2490-2500 m; location 4 on Fig 13.3), extending for about 1000 m along strike of a N600E
direction on the Red Star swell, rests on pillow basalt lava without any sediment cover.
Active and inactive massive sulfide chimneys occur close to one another, resembling tree
trunks (Lisitsyn et aI., 1993). Several active fluid emissions were observed as colorless,
milky, and, in places, black smoke. The highest-fluid-temperature measurement at one of
the vent orifices was 276°C (Lisitsyn et aI., 1993). A fluid sample was substantially diluted
with ambient bottom seawater (14% fluid vs. 86% ambient seawater), although the chemi-
cal composition was considered to be comparable to hydrothermal fluids from the hydro-
thermal systems in MaRs. Near the area of high-temperature activity, sulfide precipitates
and stockwork mineralization into oxides and hydroxides were noted. Mineral samples
from the Vienna Woods and adjacent fields are characterized by sulfides dominantly rich in
zinc, with lesser copper and locally lead, and by the occurrence of barite. Along the same
spreading ridge where Vienna Woods site, two other active hydrothermal fields (3°6.7'S,
150021.8'E, depth 2600 m and 3°22.2'S, 15002.2'E, depth 2185 m) were located. These
hydrothermal fields are restricted to the central graben of the ridge, and their linear
distribution along a N600E direction is attributed to fissures paralleling the spreading axis
(Tufar, 1990).
Hydrothermal activity was also observed in some sites along the spreading ridge in the
eastern basin. Water-column anomalies of CH4 and helium-isotopic composition attributed
to hydrothermal activity were first detected during the PAPATUA Expedition of 1986
(Craig et aI., 1987a). Garno et al. (1993c) confirmed a large hydrothermal plume from CH4 ,
Mn, and Al anomalies during the AQUARIUS expedition in 1990. The plume was centered
over an elongated caldera structure, named DESMOS cauldron, which has a relative depth
of 200 m and a diameter of 1-2 km (location 5 on Fig. 13.3). A deep-tow camera and TV
survey identified vent-associated biological colonies of squat lobsters, clams, and tube
worms and bottom-temperature anomalies up to 0.26°C at the northern terrace. The
hydrothermal dredged minerals were only native sulfur and disseminated pyrite attached to
basaltic rock samples.
Another active hydrothermal site (PACMANUS field; location 6 on Fig. 13.3) in the
eastern basin was identified about 20 km west of the DES MaS cauldron by dredging,
sediment coring, and bottom photography during the PACMANUS I expedition of 1991
(Scott and Binns, 1992; Wheller et at., 1992; Binns and Scott, 1993). Discontinuous
hydrothermal deposits were observed throughout a 3-km by 800-m zone on the crest of Pual
Ridge, where lavas range in chemical composition from andesite through dacite to rhyo-
dacite. are fragments were recovered by the camera video sled, which had collided with a
chimney. These fragments included anhydrite, chalcopyrite, and other sulfide minerals.
The spreading center of the western Manus Basin was surveyed during the RN
Hakurei-Maru No. 2 cruise in 1992. Large amounts of iron oxyhydroxide were dredged
from the axial part of the ridge, but no high-temperature vent fields were discovered
(J. Date, pers. comm.). Recently, discovery of submarine hydrothermal venting was re-
ported in the New Ireland forearc basin, which is in a rifted forearc setting behind Manus-
Kilinailau trench (Herzig et at., 1994). At volcanic cones on the southern flank of Lihir
Island, distinct hydrothermal fields of vein mineralization contain clays, silica and dissemi-
nated sulfides, and clam colonies were identified.
3.3. Lau Basin

The Lau Basin is a typical example of an active backarc basin accreted between a
remnant arc of Lau Ridge and the volcanic front of Tofua arc (Fig. 13.2). Its backarc
spreading center is propagating southward as suggested by the triangular shape of the basin
(von Stackelberg et aI., 1985, 1988; Hawkins, Chapter 3 this volume). The Valu Fa Ridge is
150 km long in the southern Lau Basin and considered to have started its spreading activity
at 1 Ma with 6 crn/yr full spreading rate. This ridge is oriented asymmetrically within the
basin, situated only 20 to 40 km west of the active Tofua volcanic arc. Small nontransform
offsets and an overlapping spreading center divide the ridge into three major segments: the
southern, central, and northern Valu Fa ridges (von Stackelberg et aI., 1988). Petrological
studies show highly fractionated, diverse volcanic rock suites of backarc basin lava, i.e.,
basaltic andesite, andesite, and rhyodacite (e.g., Jenner et at., 1987). These studies also
show an increasing effect of the subducted slab component from north to south as the
distance between the backarc spreading center and the Tofua volcanic arc narrows. Highly
vesicular and autobrecciated lava dominate samples from this area, strongly suggesting that
there is a high volatile content in the backarc magma.
Indications of hydrothermal activity were observed at the Valu Fa Ridge during the
SO-35 cruise of 1984 and the SO-48 cruise of 1987 (von Stackelberg et aI., 1985,1988;
Herzig et at., 1990). Subsequent dives with the DSRV Nautile in 1989 located four
hydrothermal sites (Fouquet et aI., 1990, 1991a,b, 1993). Fouquet et al. (1993) discussed a
range of hydrothermal activity in these sites, which are controlled by the relative interplay
of volcanism and tectonism. Early, low-temperature activity forms manganese crusts (the
Hine Hina site) in the volcanic stage, while sulfide deposition starts to develop as the
influence of tectonism improves fluid pathways (the Vai Lili site), and finally hydrothermal
activity is completely controlled by fault systems (the White Church site).
The high-temperature fluid (up to 342°C) venting in the Vai Lili site (location 7 on Fig.
13.2) on the central ridge was the first discovery of black smokers in backarc basins of the
western Pacific (Fouquet et at., 1990). Besides numerous black smokers and white smokers
both venting high-temperature fluid, diffuse discharges from the highly porous brecciated
andesite causes temperature anomalies within the surrounding bottom seawater up to 30°C.
This vigorously active hydrothermal field has dimensions of 400 m by 100 m and lies along
a normal fault at a depth of 1720 m (Fouquet et aI., 1993). Compared to MOR sulfide
deposits, the surface deposits of the Vai Lili site are characterized by abundant barite,
sphalerite, tennantite, and galena with rare pyrrhotite and no pyrite. The chemical composi-
tion of sulfides samples shows an enrichment in Ba, Zn, As, Pb, Ag, Au, and Hg and
depletion in Mo, Se, and Co. Primary grains of native gold were identified in the outer layer
of Ba-Zn chimneys, which was the first documented occurrence of gold in seafloor sulfides
(Herzig et aI., 1993). The fluid chemistry of the Vai Lili vents is characterized by very low
pH and high dissolved base metals contents (Zn, Pb, Cu, and Cd). Herzig et al. (1993)
pointed out that the chemistry of the hydrothermal fluid is significantly modified by
subseafloor mixing and cooling processes that probably induced sulfide precipitation and
oxidation.
Inactive barite-sulfide chimneys at the White Church site (location 8 on Fig. 13.2)
were observed along the top and on the flanks of the northern Valu Fa Ridge segment, at
depths between 1960 and 1800 m (Fouquet et at., 1993). This hydrothermal field consists of
chimneys up to 15 m in height and barite boulders mixed with massive sulfides. Hundreds
of small manganese oxide chimneys up to 50 cm high and low-temperature hydrothermal
discharges of 25°C were found along the faults.
On the southern Valu Fa Ridge segment (location 9 on Fig. 13.2), occurrences of
manganese oxide crusts covering sulfide deposits, strongly altered and bleached surface
rocks, and widespread discharge of low-temperature (40°C) fluid characterize a third site
called Hine Hina which located at a depth of 1900 m (Fouquet et at., 1993).
In the northern Lau Basin, the occurrence of hydrothermal deposits at some sites have
been reported with less information: sulfides from the northeastern ridge (Hawkins and
Helu, 1986; location 10 on Fig. 13.2) and barite from the Peggy Ridge (Bertine and Keene,
1975; location 11 on Fig. 13.2). In addition, extensive inactive sulfide deposits with andesitic
lava at 15°23'S, and with basalt at 18°36'S, were recently discovered by DSRV Mir and are
cited in Fouquet et at. (1993).
3.4. Mariana Trough

The Mariana Trough is an actively spreading backarc basin that accreted between the
remnant volcanic arc of the West Mariana Ridge and the volcanic front of the Mariana arc
(Fig. 13.4). Its spreading rate is considered to have been 3.0-3.4 cm/yr as a full rate for the
past 3 m.y. (Hussong and Uyeda, 1981). The age of the Mariana Trough is less than 7 Ma,
and the Wadati-Benioff zone does not lie beneath the backarc spreading axis (Hawkins et
at., 1990; Fryer, Chapter 6 this volume).
In 1987, DSRV Alvin dive programs confirmed present-day hydrothermal activity in
the Mariana Trough (Craig et al., 1987b), which had been evidenced by hydrothermal
plume identification during the CEPHEUS cruise in 1982 (Horibe et at., 1986). Together
with observations by the DSRV Shinkai 6500 in 1992 and 1993, hydrothermal occurrences
are known in various tectonic settings across the trench-arc-backarc systems-that is, (1)
axial volcanoes of backarc spreading center between 18°13'N and 18°02'N, (2) frontal
volcanoes at 21°N in the northern Mariana arc (McMurtry et al., 1993) and at 13°24'N in the
southern Mariana Trough (Gamo et at., 1993a; Johnson et al., 1993), (3) a forearc serpen-
tinite mud volcano (Fryer et at., 1987b, 1990), and (4) off-axis hydrothermal mounds at
18°02'N (Leinen et al., 1987; Wheat and McDuff, 1987,1994).
Active hydrothermal vents were first discovered on the flanks of axial volcanoes on
the spreading center at the depth of 3600-3700 m (Craig et al., 1987b; location 13 on Fig.
13.4). During the DSRV Alvin dive programs of 1987, four separate hydrothermal fields
were discovered along the crestline of the trough (Hessler and Lonsdale, 1991). The Alice
Spring site (18°12.6'N, 144°42.4'E, depth 3640 m) and another high-temperature venting
site (18°12.8'N, 144°42.4'E, depth 3595 m) are located on the south flank of the central
peak. Many sulfide-sulfate chimneys rise from large mounds of amorphous iron oxy-
hydroxide and amorphous silica (Hawkins et at., 1990). Chimney samples collected from
this area are characterized by barite deposition with the absence of anhydrite as the matrix
to the chimney walls (Kastner et al., 1987; Sttiben et at., 1994). The mineral composition of
145°E
25°
I
\ ,
( \
I I
'"' \ \ (\ \.
)23,',,-
\,
c-I ./ "- ,
i'
I \
\
" "-
"- ,-
(\ I
'D
I
\ \ \ \1 20°
\ aI \
ie '5c
>:a', /
\
I I!!
1\1
~~ 13, \\ , I-
~ (~I 14 I oj.
\ 1\1
C
1\1
GI
,1\B I ,:F' ·Ii
~ ;:EI ~ a: I :E
i' e
;:) I
~ 1 !II I
c:: I
II,
I ) ~ /
~ f! 15°
".f! ':ll
}
"~
( (/
/ I
)':"24 'r~
{,-:.. I, e IJ
L// ,.,
L-____ ~ ____________ ~ __________ ~100N
FIGURE 13.4. Location and tectonic setting of Mariana Trough. The solid circles mark the hydrothermal fields
listed in Table I. The broken lines delinate contours of 3000-m water depth.
the sulfides is dominantly sphalerite, abundant galena, and lesser chalcopyrite, which
clearly contrasts with those from MOR hydrothermal systems (Kastner et al., 1987;
Kusakabe et aI., 1990; Moore and Stakes, 1990). Some chimneys are venting a clear fluid
with temperatures up to 287°C (Craig et aI., 1987b). The fluid chemistry, especially the
chemical and isotopic composition of boron, indicated evidence for fluid interaction with
oceanic crust in a backarc setting that is enriched in slab-derived boron (Campbell et aI.,
1987; Palmer, 1991). Hydrothermal vent communities are represented by hairy gastropods,
bresilid shrimp, and brachyuran crabs (Hessler et aI., 1987, 1988; Okutani and Ohta, 1988).
In 1992, the Alice Spring site was revisited by the DSRV Shinkai 6500 (Gamo et al.,
1993a,b). Negligible changes occurred in the five years since the DSRV Alvin dives, both in
the fluid chemistry and biological communities. On the southeastern flank of the south
peak, at 18°1O'N, "Anemone Heaven" site is a milky, low-temperature (23°C) fluid
discharging from an open crack. This field has been colonized by large anemones and
galatheid crabs, although little mineralization was observed (Hessler et al., 1987; Hawkins
et al., 1990). A high-temperature fluid (250-285°C) venting field, the "Snail Pits" site
(18°10.95'N, 144°43.2'E, depth 3660 m), was also found on the south peak. At 20 km south
of these sites, another hydrothermal field was located at 18°02.8'N, 144°45.2'E in a water
depth of 3676 m (location No. 14 on Fig. 13.4), where sulfide-sulfate chimneys, including
active high-temperature fluid discharges and large irregularly shaped mounds of soft
ochreous oxide mud composed mainly of amorphous iron oxyhydroxide and some amor-
phous silica, were observed (Hawkins et aI., 1990).
The Kasuga seamounts in the northern Mariana arc are located along a N-S trending
cross-arc fault zone (location 23 on Fig. 13.4) which probably represents the tectonic region
of egress of arc magma leaking into the actively spreading Mariana backarc (Fryer et aI.,
1987a). At the summit areas of seamounts Kasuga 2 (depth 400 m) and Kasuga 3 (depth
1140 m), emanations of warm fluid and hydrothermal deposits were confirmed by DSRV
Alvin dives (McMurtry et al., 1993). A sample of 39°C fluid collected from the Kasuga 2
seamount has very unusual characteristics, which are attributed to significant subseafloor
addition of volatile components (C0 2 and S02) directly from magma into the hydrothermal
fluids (McMurtry et aI., 1993). Enrichment in CO 2 results in "chemical weathering"
reactions, whereby igneous minerals and/or alteration phases are strongly altered, adding
magnesium and other cations to the hydrothermal fluid. Hydrolysis of S02 to sulfate and
native sulfur is responsible for the high sulfate content, with a corresponding light 834S
signature. Hydrothermal deposits consisting of elemental sulfur, Fe- and Mn-oxides, and
nontronite are consistent with this fluid chemistry. The magmatic contribution caused by
the reduced solubility of the volatile components at a shallow depth controls both the major
elements composition of the fluid and the hydrothermal mineralization (McMurtry et al.,
1993).
On the southern Mariana Ridge, Gamo et al. (1993a) discovered the hydrothermal site
named "Forecast Vent Field" (location 24 on Fig. 13.4) at the summit of the seamount
Peak B (13°24'N, 143°55'E, depth 1470 m). Around this region, the distance between the
Mariana Ridge and the Mariana Trough narrows and both island-arc volcanism and backarc
spreading activity appear along the fault zone, which crosscuts the arc-backarc system
(Fryer, 1993; Masuda et al., 1993). Basalt samples recovered from the Peak B seamount
have low K20 content consistent with the Mariana backarc basalts, but their MgO-Ti02
relationship would classify them as an island-arc basalt (Masuda et aI., 1993). Hydrother-
mal fluid up to 202°C emanates from anhydrite chimneys (Johnson et al., 1993).
The Conical seamount is a conical edifice 20 km in diameter and 1500 m in height,
located in the Mariana forearc, 80 km west of the Mariana Trench and 120 km east of the
Mariana Ridge (location 27 on Fig. 13.4). The seamount is composed of unconsolidated
clay to sand-sized serpentinite with suspended clasts of variously serpentinized harz-
burgite, dunite, and rare metabasalt (Fryer et aI., 1990). DSRV Alvin dives found white
chimneys of two types, carbonate and silicate, near the summit knoll at a depth of about
3150 m (Fryer et al., 1987b). The carbonate chimneys consist of aragonite, calcite, and
minor amounts of amorphous gel-like Mg silicate. The other chimneys are Mg silicate with
trace amounts of carbonate and a thin coating of ferromanganese oxide (Haggerty, 1987a;
Fryer et aI., 1990). Age determinations by the 2IOPb method show that the chimneys are very
young, <150 years old. A combined carbon- and oxygen-isotopic study suggests that the
carbonates originated from the oxidation of methane of probable thermogenic origin and
from ambient, low-temperature seawater (Haggerty, 1987b; Fryer et aI., 1990). Cold fluid
was noted to seep from orifices of the silicate chimneys when they were scraped by the
Alvin manipulator. Fluid temperatures were slightly lower (0.03°C) than that of ambient
seawater, and the chemistry has several unique features, such as high pH and enrichments in
silica, carbonate, sulfate, and H 2S. Pore fluids sampled by ODP drilling at the Conical
seamount (Sites 778-780) confirmed the existence of this peculiar fluid (Mottl, 1992) with
chemical characteristics (e.g., pH = 12) less mixed with seawater than the chimney fluid.
Mottl (1992) attributed the fluid chemistry to a combination of deep-seated serpentinization
processes and seawater interactions with crustal rocks including the surficial serpentinite.
He proposed that the serpentinite mud volcanoes such as the Conical seamount in forearc
regions act as important conduits through which volatile components (H20, CO2, and other
Sand C species) are released from the subducting slab and return to the ocean.
3.5. Izu-Bonin (Ogasawara) Arc

The Izu-Bonin (Ogasawara) arc is a well-developed intraoceanic island arc associated
with the Izu-Ogasawara Trench that includes a volcanic chain called Sichito-Iwojima
Ridge and discontinuous backarc depressions (Fig. 13.5). The backarc depressions just
behind the volcanic front are considered to be fault-bounded backarc rift grabens (Honza
and Tamaki, 1985). The chain of active volcanic islands and submarine volcanoes about
1200 km long extends from the Izu Peninsula of Japan to the northern end of the Mariana
arc (Yuasa, 1985). This trench-arc-backarc system is situated on the eastern rim of the
Philippine Sea plate, where the Pacific plate is being subducted toward the northwest.
In 1984-1988, the Geological Survey of Japan conducted a research program to
evaluate the possibility of submarine hydrothermal mineralization in this region (Nakao et
aI., 1990). Several hydrothermal occurrences were found within submarine calderas of the
frontal volcanoes. Subsequent dive programs by the DSRV Shinkai 2000 confirmed the
present-day hydrothermal activity. Moreover, in 1990, the first example of active high-
temperature venting (up to 311°C) and accompanying mineralization located on the sub-
marine frontal volcanoes was discovered (Kasuga and Kato, 1992; Watanabe and Kajimura,
1993,1994). Together with frontal volcanoes, the rift grabens were considered to be an ideal
tectonic setting for hydrothermal activity. Fujioka (1983) pointed out several topographic
140· 145 E
0
35°N
FIGURE 13.5. Location and tectonic setting of

Izu-Bonin arc. The solid circles mark the hydro-
thermal fields listed in Table I. The broken lines
delinate contours of 3000-m water depth.
similarities between the lzu-Bonin arc system and the restored Japan arc of Miocene age
where major Kuroko-type VMSDs are known to exist.
The Kaikata seamount is one of the active frontal volcanoes and has four peaks: KM,
KN, KS, and KC (location 22 on Fig. 13.5). The KC peak has a submarine caldera of 2-km
radius in which the central cone is composed of lava and volcanic sand with a rim of
andesitic lava (Yuasa et ai., 1988; Naka et ai., 1989a). From the caldera floor, altered
andesite, intensely disseminated and veined with sulfide, was dredged during the GH85-1
cruise of 1985 (Urabe et ai., 1987). Sulfide minerals present are dominated by pyrite with
minor amounts of sphalerite. High homogenization temperatures of fluid inclusions in vein
quartz of 290°C and enrichment of gold in vein materials are indicative of vigorous fossil
hydrothermal activity within this caldera (Urabe et ai., 1987). During a subsequent Shinkai
2000 dive, Yuasa et ai. (1988) noted an alteration zone along the lower part of the caldera
wall more than 150 m high. Warm, shimmering fluid about 10°C higher than ambient
seawater, associated biological communities dominated by white blind crab, and volcanic
sands cemented by marcasite and minor amounts of pyrite at the shimmering site were
confirmed in both the central cone at a depth about 900 m and the caldera wall (Naka et al.,
1989a; Fujikura et ai., 1993). Contrary to this evidence, present-day high-temperature
hydrothermal activity was not seen.
On the northwestern slope of the KM peak of the Kaikata seamount, at depths between
900 and 1200 m, hydrothermal manganese oxide deposits cover the volcanic sands in the
NE-SW trending zone over an area 10 km by 2 km (Usui et ai., 1986, 1989; Usui and
Nishimura, 1992). The manganese deposits are characterized by high MnlFe ratios and very
low transition element contents. Usui and Nishimura (1992) concluded that low-temperature
hydrothermal solutions ascending through fractures or faults in permeable volcanic sands
are responsible for precipitation of the several-centimeters-thick manganese layers.
The Suiyo seamount is one of the Sichito seamounts chain that is a part of the volcanic
front in the middle part of the Izu-Bonin arc (location 21 on Fig. 13.5). This seamount is
composed of dacitic rocks that belong to the low-potassium calc-alkaline rock series
(Nagaoka et ai., 1991, 1992; Watanabe and Kajimura, 1993). In 1991, dives of the DSRV
Shinkai 2000 discovered sulfide chimneys venting clear, high-temperature fluid (230°C) at
the slope of the caldera (Kasuga and Kato, 1992). During the next year, vigorous hydrother-
mal activity comparable to that of MORs with respect to their temperature (260-31l°C)
was located on the caldera floor at a depth of 1370 m (Watanabe and Kajimura, 1993, 1994;
Watanabe et ai., 1994). Active smokers and a hydrothermal alteration zone are distributed
within an area of about 300 m by 150 m in a NNW-SSE direction. Mineralogical studies
show abundant chalcopyrite, especially inside chimneys walls. Columnar structures of
chalcopyrite crystals are enclosed by colloform sphalerite, anhydrite, and barite. High Au
contents (71 ppm maximum) and high AulAg ratios (average = 0.142) in the samples were
also noted (Watanabe and Kajimura, 1994). The fluid chemistry of the Suiyo seamount is
characterized by a higher CI content than seawater, coupled with Ca enrichment and Na
depletion (Tsunogai et ai., 1994; Ishibashi et ai., 1994b).
The Mokuyo seamount (28°14'N, 1400 34'E) is located about 30 km south of the Suiyo
seamount and is composed of basaltic rocks (Nagaoka et al., 1991). Within the caldera
wall, at a depth of 1210 m, low-temperature fluid emanation (40°C maximum) is associated
with clam-dominated biological communities (Nagaoka et ai., 1992). Although mineraliza-
tion was not observed, helium isotopic compositions of fluid samples suggest the existence
of a hydrothermal end-member fluid originating from magmatic activity (Tsunogai et ai.,
1993).
The Kita-Bayonnaise caldera (Myojin Knoll) is located in the northern Izu-Bonin

arc, which has an outer rim of 5-7 km (location 19 on Fig. 13.5). The caldera rim and a
central cone are composed of dacitic rock, while significant amounts of pumice are
observed in the upper caldera wall (Yuasa et aI., 1991; Iizasa et aI., 1993). From the caldera
floor at depths around 1300-1400 m, abundant sulfides and barite, comprising about 1.0-3.5
wt% of the sediment, were recovered by box-and-gravity coring (lizasa, 1993a). The
mineral assemblage of barite-pyrite-chalcopyrite-sphalerite, with or without galena and
tetrahedrite, was identified in a single grain sample, indicating hydrothermal mineralization
took place and was subsequently eroded by mass wasting (Iizasa, 1993a). During a 1992
diving expedition with Shinkai 2000, mineralized tuff consisting of sulfides and barite was
sampled from the eastern part of the caldera floor (lizasa et al., 1993a). The chemical
composition of this tuff is characterized by enrichments in Zn, Fe, Cu, Ba, Au (up to 19
ppm), and Ag (up to 272 ppm). Fluid inclusions in barite indicating formation temperature
of 190-215°C and those in sphalerite indicating 205-225°C support the hydrothermal
mineralization (Iizasa, 1993a).
The Myojinsho submarine caldera is located about 30 km south of the Kita-Bayonnaise
caldera (location 20 on Fig. 13.5). It has an outer rim of 5-6 km and a flat-topped central
cone composed of dacitic rocks. From the caldera floor at depths of 1000-1100 m, several
massive hydrothermal barite and sulfide samples, together with rocks containing dissemi-
nated sulfides, were recovered. On the basis of a textural study, Iizasa et al. (1992) showed
there are three sulfide mineralization sequences at the Myojinsho caldera: pyritization,
quartz and sulfide vein let-type, and massive sulfides. They also reported chemical composi-
tions of sulfide samples characterized by enrichments in Zn, Pb, Cu, Au (up to 2.7 ppm),
and Ag (up to 350 ppm). Factor analysis on the chemical compositions of sediment samples
collected from the whole caldera floor suggests that the mineralization took place in the
western part of the Myojinsho caldera (lizasa, 1993b).
The Sumisu rift is one of the rift grabens in the northern Izu-Bonin arc (location 15 of
Fig. 13.5). Taylor et al. (1990) conducted a DSRV Alvin-Sea Beam study and found a fossil
hydrothermal site on an axial volcano. Chimneys, veins, and crusts are composed of silica,
barite, and iron oxide. They occur primarily on the crown of a rhyolite lava dome at a depth
of 1600 to 1530 m (Urabe and Kusakabe, 1990). The 8 180 values of barite and the
occurrence of amorphous silica suggest low-temperature «150°C) formation of the de-
posits, a possible analog to the hydrothermal hematic chert associated with the Kuroko
deposition (Urabe and Kusakabe, 1990). Surface sediments of the Sumisu rift basin show
manganese contents commonly in excess of 1%, suggesting a hydrothermal input (Nishi-
mura and Murakami, 1988), although lower manganese content «0.3%) of the pelagic and
hemipelagic sediments sampled by ODP Leg 126 indicate this hydrothermal influence has
been negligible back to 1.1 Ma (Nishimura et aI., 1992a,b). Combined with high heat flow
values in the basin (Yamazaki, 1988), the evidence given above would suggest that the
recent hydrothermal activity that formed the dead barite-silica chimneys could have been a
source of the surficial manganese, derived from manganese-rich plumes spread all over the
rift basin (Urabe and Kusakabe, 1990).
3.6. Okinawa Trough

The Okinawa Trough is considered to be a nascent backarc basin within the continen-
tal margin (Japanese DELP Research Group on backarc basins, 1986; 1991). Several
geophysical and geological investigations (e.g., Sibuet et al., 1987, Chapter 9 this volume)
have confirmed rifting activities in the middle Okinawa Trough, which is characterized by
the development of normal faulting in brittle continental crust (Fig. 13.6). Kimura et al.
(1986) have demonstrated recent bimodal volcanism in the middle Okinawa Trough.
Ishizuka et al. (1990) concluded that the magmatic sources in this area are in transition from
island-arc type to backarc type, based on petrological studies. Extremely high and localized
heat flow anomalies support very recent magmatic intrusions in this region (Yamano et al.,
1989).
In 1984 and 1986, dive programs with the DSRV Shinkai 2000 first located low-
temperature hydrothermal activity in the middle Okinawa Trough (Kimura et al., 1988). A
large hydrothermal field with extensive occurrence of sulfides was discovered during the
SO-56 cruise of 1988 (Halback et aI., 1989). Subsequent Shinkai 2000 dives confirmed the
presence of high-temperature fluid venting up to 320°C (Nakamura et al., 1990; Sakai et al.,
1990a). The middle Okinawa Trough mineralization is considered to be a modern analog
for Kuroko-type mineralization (Halbach et al., 1993). In 1988, a deep-tow TV camera
survey conducted by RN Kaiyo located hydrothermal activity at other sites in the Okinawa
Trough (Momma et al., 1989). These hydrothermal fields have been visited every year since
their discovery (recently, however, the Safety Evaluation Committee of JAMSTEC prohib-
ited Shinkai 2000 diving at one of these sites, because the installation of submerged buoys
by fisheries groups makes further dive studies at this site unsafe). All of the hydrothermal
FIGURE 13.6. Location and tectonic setting of Okinawa Trough. The solid circles mark the hydrothermal fields
listed in Table I. The broken lines show contours of lOOO-m water depth. The triangles indicate Quaternary
volcanoes.
systems in the Okinawa Trough formed in environments with extensive terrigenous sedi-
ment, supplied from the adjacent Asian continent.
The Izena cauldron is a rectangular (6 km by 3 km) depression formed by strike-slip
strain accompanying the rifting (Halbach et al., 1989, 1993; location 18 on Fig. 13.6). It is
situated approximately in the southwestern continuation of a chain of Quaternary volcanoes
on the Ryukyu arc; however, this location is also just in front of the eastern end of the Aguni
rift graben segment. It is difficult to judge whether the Izena magmatic activity should be
ascribed to volcanic front or backarc rifting. From the outer wall of the cauldron, occur-
rence of rhyolite, dacite, poorly vesciculated pumice, mudstone pumice, and tuff breccia
have been reported, with twin rhyolite lava domes located in its center part (Nakamura
etai., 1989; Tanaka etai., 1990). The hydrothermal field known as "Jade site" lies within an
elongated region of 1000 m by 200 m on the northeastern slope at depths between 1300 and
1550 m. At the center of the hydrothermal field, the "Black smoker" chimney is vigorously
discharging 320°C hydrothermal fluid. Surrounding this chimney, several groups of clear
smoker chimneys are emitting transparent fluids with temperatures up to 220°C (Nakamura
et al., 1990). The fluid chemistry can be explained by the interaction with acidic to
intermediate volcanic rocks and by fluid-sediment interaction (Sakai et ai., 1990a). One
characteristic of the Izena hydrothermal activity is the abundance of CO2, reflecting a
magmatic source (Ishibashi et ai., 1990). At the outskirts of the hydrothermal field, liquid-
CO2 bubble emanation was observed (Sakai et ai., 1990b) Mineral samples from this area
contain sphalerite, tetrahedrite, galena, chalcopyrite, pyrite, barite, and anhydrite (Aoki and
Nakamura, 1989; Urabe, 1989; Halbach et ai., 1993). High Au (up to 24 ppm) and Ag (up to
1.1%) contents have also been noted.
At a small depression along the Iheya Ridge, which lies in the axial part of the Iheya
rift graben, another active hydrothermal field has been discovered (location 17 on Fig. 13.6).
This field is called the "Clam site," after an associated biological community that consists
dominantly of clams, scalpellids, and slender vestimentiferan tube worms (Ohta, 1990).
The Ibeya Ridge consists of highly vesicular pillow basalt; the gassy lava surfaces are
suggestive of recent eruptions (Naka et ai., 1989b). Hydrothermal activity is limited to
narrow area at a depth of 1400 m. Mineralization at the Iheya site is characterized by
carbonate precipitates, which occur in a form of cemented multilayered eaves covering
large portions of the seafloor (Gamo et ai., 1991b). The sampled carbonates consist of
manganoan calcite, rhodochrosite, anhydrite, and amorphous silica, with only trace
amounts of sulfides including wurtzite, pyrrhotite, and galena present as disseminations
(Tanaka et ai., 1989; Izawa et ai., 1991). Native sulfur was also recovered from the outer
region of the field. Moderate emanations of fluid from edges and fissures of the cemented
carbonates were observed in several places (Sakai et al., 1990a; Gamo et ai., 1991a,b). Fluid
temperatures showed both temporal and spatial variation, with temperature ranging from
100 to 220°C. Nakashima et ai. (1993) estimated the fluid-rock interaction temperature at
between 250 and 300°C, based on fluid inclusion homogenization temperatures in calcite.
Carbon-isotopic compositions for carbonates are very close to those of the venting fluids,
providing evidence that the carbonate precipitated directly from the CO2-rich fluids
(Kimura et ai., 1989). Gamo et ai. (1991a) demonstrated some unique features of the fluid
chemistry, such as anomalously high ammonium content, high alkalinity, and 34S-enriched
sulfate, all interpreted as a result of organic matter decomposition during subseafloor
mixing with penetrated seawater. Gamo et ai. (1991a) proposed a model of multiple
secondary hydrothermal circulation cells within sediment layers which are in turn associ-
ated with high-temperature (around 300°C) primary circulation.
At the summit area of the Natsushima 84-1 knoll (water depth of 1540 m), which is
located about 15 km east of the Iheya Ridge, low-temperature hydrothermal activity was
first discovered in 1986 in the middle Okinawa Trough (Kimura et aI., 1988). This small
knoll is composed of dacite and rhyolitic rocks (Kimura et al., 1986) and situated on the
south side of a small depression, where Yamano et al. (1986) reported extremely high-heat-
flow measurements. Shimmering water from one of the mounds has a temperature 2-3°C
higher than that of the ambient seawater, while methane and manganese anomalies were
detected (Gamo et aI., 1987). The mound was covered with a yellowish brown deposit of
Fe-rich smectite and Fe-Mn oxyhydroxide (Masuda et al., 1987). Oxygen isotopes of these
deposits indicate a temperature of precipitation of about 50°C, which confirms the low-
temperature activity at this site (Masuda et al., 1987).
The Minami-Ensei knoll is located at a rifting center in the northern Okinawa Trough
(location 16 on Fig. 13.6). Its topography is characterized by small hills and depressions
within a rim of 2-km radius and is attributed to a volcanic complex (Aoki et al., 1993).
Poorly vescicu1ated pumice, diorite, and granite samples were recovered from this knoll.
Living and/or dead biological communities dominated by mussels were observed in several
places in the small depressions (Hashimoto et aI., 1990, 1993). The strongest hydrothermal
activity is in "C depression," located on the westem slope of the volcanic complex. Several
chimneys are aligned along a N-S direction at depths between 690 and 705 m. Most of the
chimneys are discharging transparent fluids with temperatures of 267 - 278°C (Chiba et aI.,
1993). Nedachi et al. (1992) described the hydrothermal samples as siliceous ores, sulfate
chimneys, massive sulfide ores, and clastic ores. Siliceous samples contain high Si02
contents (up to 83%) and some sulfides and barite. All of the active hydrothermal chimney
samples are composed mainly of anhydrite, with only fine-grain sulfide minerals precipi-
tated on the anhydrite surfaces. Some Zn-rich massive sulfides were found at the foot of the
chimneys. The clastic ore samples are characterized by enrichment in both Zn and Pb and a
high average Au!Ag ratio (around 0.08). Moreover, components of the clastic material do
not coincide with the samples observed on the seafloor. Nedachi et al. (1992) suggest that
the clastic ore samples were formed beneath the seafloor and ejected on the seafloor during
an explosive boiling stage. A similar scenario has been proposed for the sulfides from the
Jade site (Urabe, 1987). The chemical composition of the sampled hydrothermal fluid is
characterized by a depletion in both metal elements and H 2S and by a chlorinity only 10%
less than that of seawater, which can be explained by a model of phase separation just below
the seafloor which induces subseafloor sulfide precipitation (Chiba et aI., 1993). Nakashima
et al. (1993) reported fluid inclusion homogenization temperatures are similar to the
measured venting fluid temperature. Hydrothermal activity at the Minami-Ensei knoll has
been shown to be very close to P-T conditions for the phase separation.
4. DISCUSSION
4.1. Volcanism and Hydrothermal Activity

About three quarters of terrestrial volcanic rocks have been produced at divergent
plate boundaries during the Cenozoic era (McBirney, 1984). The oceanic hot spots and arc-
backarc systems contributed 7-10% and 10-15% of the total volume of volcanic rocks,
respectively. Although submarine arc-backarc volcanism is less important volumetrically,
it is very significant for its lithologic diversity.
All the known high-temperature hydrothermal activity in the western Pacific region is
located at the loci of such submarine volcanism. This is understandable because the
volcanism provides favorable crustal environments for hydrothermal mineralization: ex-
tremely high heat flow and fracture- or rift-induced permeability to facilitate voluminous
hydrothermal circulation, magma chambers of favorable geometry (e.g., Rona, 1988),
supply of volatile and other components from the subduction zone (e.g., Stolper and
Newman, 1992), the volatile-rich nature of arc and backarc magmas (e.g., Hochstaedter
et aI., 1990), and tectonic events conducive to intense hydrothermal activity. The hydrother-
mal activity at hot spots seems to be the exception for these favorable features. Fluids from
some hot spots in the central Pacific, the Loihi seamount in the Hawaii volcanic chain
(Sedwick et al., 1992,1994), the Teahitia seamount at the southern extension ofthe Society
Islands (Michard et aI., 1993), and the Macdonald seamount in the Austral Island volcanic
chain (Cheminee et al., 1991) are highly enriched in volatile components of magmatic
origin such as CO 2 and He. However, the hydrothermal deposits in these submarine
volcanoes are mainly composed of iron hydroxides, and no sulfide mineralization has been
identified (De Carlo et aI., 1983; Puteanus et aI., 1991).
Rona (1988) pointed out, on the basis of a compilation of ancient massive sulfide
deposits, that basalt-hosted massive sulfide deposits that may have formed at spreading
centers constitute 17% of the known 508 deposits. Most of them are less than 30 x 106
tonnes in size. On the other hand, the most numerous (56%) and largest (up to 231 X 106
tonnes) massive sulfide deposits are hosted in rhyolite. Hydrothermal activity associated
with dacitic to rhyolitic volcanism is known at ten sites (6, 7,8, 12, 15, 16, 18, 19, 20, and 21,
Table I) in the western Pacific, and eight (except 12 and 15) of them have sulfide mineraliza-
tion. This number is comparable to the nine sulfide mineralized sites (1, 4, 9, 10, 13, 14, 17,
22, and 24, Table I) that are associated with basaltic to andesitic volcanism. Since rhyo-
dacite is extremely rare in MOR systems (Christie and Sinton, 1981), this statistic supports
the idea that the ancient arc-backarc setting is the exclusive site for the formation of
rhyolite-hosted massive sulfide deposits. It is still premature to discuss the size and ton-
nage of the present-day mineralization in the western Pacific, but relatively large deposits
(e.g., 7, 8, 18, and 21) tend to be associated with acidic volcanism.
4.2. Time Scale of Volcanic and Hydrothermal Events

Oshima island, the northernmost frontal volcano of the Izu-Bonin arc, displays a
remarkably regular periodic volcanic activity. The stratovolcano of arc tholeiite basalt
erupted at average time intervals of 135 ± 50 years in the past 1330 years and of 100 years in
the past 10,000 years (Aramaki et aI., 1992). Although there are insufficient data to prove
such periodicity in other frontal volcanoes of the arc, it is reasonable to deduce that the
replenishing of arc magma happens periodically at interval of order 100 years. The
associated hydrothermal activity is, therefore, likely to periodically reactivate and cease in
accordance with the volcanic activity.
Estimation of the time scale of hydrothermal activity at submarine arc volcanoes
seems to be even more difficult. However, repeated Shinkai 2000 dives on the Kaikata Sea-
mount provide some clues. In November 1986, Yuasa et al. (1988) observed reddish brown
flocculates of Fe-oxyhydroxide drifting within the caldera and found weak fluid shimmer-
ing from hydrothermally altered caldera wall. A revisit in May, 1988 (1.5 years later) by one
of the authors (T. U.) showed that all the Fe-oxyhydroxide had been completely swept away.
An intensive search for the shimmering failed, probably because hydrothermal activity had
ceased. In contrast, activity at the summit of the central cone was confirmed by later dives
in the same month. Shimmering of warm fluid (max. temperature 22°C) marcasite-
cemented sediments (T. Urabe; unpublished data) and associated biological community of
mussel, deep-sea sole, and brachyuran crab characterize the latter site (Mitsuzawa et aI.,
1989). Furthermore, the shimmering became hardly identifiable in 1990. These lines of
evidence suggest that the lifetime of some hydrothermal activity at frontal volcanoes is
likely to be short, on the order of years, and possibly much shorter than that of MORs.
4.3. Hydrothermal Mineralization in Relation to Ancient Analogs

Sawkins (1976) proposed a fourfold classification of massive sulfide deposits based on
their composition and tectonic settings: Kuroko, Cyprus, Besshi, and Sullivan types. The
hydrothermal mineralization in the western Pacific provides present-day examples of the
first three deposit types, but there are no observed present-day examples for the sediment-
hosted massive sulfide deposits of the last type. The Kuroko deposits in Japan are located in
a fault-bounded backarc trough of the northeast Japan arc of Miocene age (Ohmoto et at.,
1983). Rifting with 1-3 km of vertical displacement, intense bimodal volcanism, and
associated Kuroko mineralization occurred between 16 and 15 Ma. Several lines of evi-
dence, including volcano-tectonic setting, geology, and isotopic composition, all suggest
their magmatic hydrothermal origin (Urabe, 1987; Urabe and Marumo, 1991). A Besshi
type deposit, named after the Besshi deposit in Japan, is defined as a cupriferous-pyrite
massive sulfide deposit hosted in metamorphosed, submarine tholeiitic basalt (Kanehira
and Tatsumi, 1970). These deposits differ significantly from the Cyprus-type by an abun-
dance of epiclastic rocks and the lack of an ophiolite suite. The Kuroko and Besshi types are
considered to have formed in arc-backarc environments rather than MORs, which may
have produced the Cyprus type deposits (Sawkins, 1976).
Hydrothermal mineralization in the Okinawa Trough, the Sumisu rift, and the Lau
Basin occurs in association with rhyodacite volcanism at backarc rifting-spreading centers,
which develop by splitting preexisting island-arc crust and are regarded as the best
examples of contemporaneous Kuroko-type mineralization (Halbach et at., 1989, Urabe
and Kusakabe, 1990, Fouquet et at., 1993). The basalt/andesite-hosted deposits such as
those in the central Mariana Trough and the Vienna Woods site of the Manus Basin seem to
provide the modem analog of the Besshi-type deposits. The present-day deposit will be
covered by epiclastic sediment like the latter deposits after the termination of the minerali-
zation: Recent submersible dives at various volcanic centers in the western Pacific provided
unequivocal evidence that most of the flanks of the volcanic edifices are thickly covered by
sand to pebble-size sediments of volcanic rock and/or pumice origin. This is in contrast
with sediment-starved MORs, where we observed almost no coarse-grain sediments except
scarce glassy fragments. Such volcaniclastic sediments may be derived either from erup-
tion of arc-backarc volcanism or by mass wasting on the flank of the volcano. Later
accretion of the deposits and their host rocks during subduction may give rise to metamor-
phism and intense deformation (Karig, 1985) which are commonly observed in the Besshi-
type deposits in the world (e.g., Franklin et aI., 1981).
The deposit in the North Fiji Basin that is located at the spreading center of a mature
backarc basin may be a candidate for the Cyprus-type deposit. This type of deposit may
also be formed at the MORs.
4.4. Mineralogy, Geochemistry, and Sulfur Isotopes of the Deposits
The hydrothennal deposits of the Western Pacific (Table I) are characterized by their
low pyrite/marcasite content and absence of pyrrhotite (except at a few sites such as Lau
Basin and the Fiji transfonn fault). This is partly a result of the low iron content in
hydrothennal fluids in the western Pacific and the higher 102 condition in the backarc
settings, since pyrite is more stable than pyrrhotite at such conditions, as discussed by
Herzig et al. (1993). They attributed high gold content of the western Pacific deposits to
higher 102 and IS 2 conditions.
Mineralogical zoning observed in chimneys and mounds is essentially the same as that
of MORs; Cu-rich sulfides in the center and Zn-rich sulfides in the outer edges (e.g., Bendel
et aI., 1993). Galena and tetrahedrite-tennantite (and realger, orpiment, and sulfosalts
minerals in the case of the Okinawa Trough), which are unique to the arc-backarc systems,
are enriched in the outer zone, sometimes together with barite.
The chemical analyses on the sulfide samples from the western Pacific region com-
monly show a wide range of values even in a single hydrothennal field. This is due to the
gross mineral zoning observed in these ores and to their complex growth history. However,
the average chemical composition of the sulfides shows certain characteristics that differ
from those of the MORs (Table II).
Figure 13.7 shows the variation in representative element contents, which are nor-
malized against average EPR sulfide deposits. The average composition of sulfide samples
of the North Fiji Basin spreading center (Fig. 13.7 A), which is in the most advanced stage of
backarc spreading, strongly resembles that of the EPR. It is also obvious that the composi-
tion is almost identical to the average Besshi-type ores, which are believed to have fonned
in backarc spreading centers of the Mesozoic to Paleozoic eras. On the other hand, samples
of the Manus Basin (PACMANUS site) show slight enrichment in gold and arsenic, which
are characteristic elements in subaerial epithennal gold deposits (e.g. Hedenquist et aI.,
1990) and high-temperature hot springs observed on the acidic to intennediate volcanoes
(Karpov and Naboko, 1990). This is rather remarkable because the crustal abundance of
these elements is quite small and the enrichment factors should be around several thousand.
Such an extreme enrichment is unlikely to occur simply by leaching from the host rocks.
One of the most striking occurrences of present-day gold mineralization is observed at
the shore of a crater lake of Osorezan volcano, northern Japan (Aoki and Thompson, 1990;
Izawa and Aoki, 1991). Extremely enriched gold (6510 ppm Au) and arsenic (3650 ppm) are
precipitating near the surface together with Hg, Pb, Zn, and Sb from neutral pH, chloride-
rich hot-spring water with high H 2S content (Aoki, 1992a). Oxygen- and hydrogen-isotope
data suggested a significant contribution of the fluid from cooling dacite magma of the
Osorezan (Aoki, 1992b). Dacitic volcanism observed at the PACMANUS site is the likely
source of such "epithennal elements" in the seafloor deposits, by an analogy of the
terrestrial example.
The composition of sulfide samples from the Lau Basin, Mariana Trough (Fig. 13.7B),
and Minami-Ensei (Fig. 13.7C) matches closely that of the average black ore (Zn-Pb-Ba
massive ore) of the Kuroko deposits in Japan, which strongly indicates their common
mechanism of fonnation. In fact, the Kuroko and these three present-day examples all
fonned at centers of rifting and early stages of backarc spreading.
The composition of the sulfide samples from frontal volcanoes (Suiyo seamount,
Okinawa Trough) shows an unusual variation (Fig. 13.7C): The iron-poor nature of the
Okinawa Trough ore makes all the elements more enriched than that of the EPR. The sulfide
samples from the Suiyo seamount are enriched in copper but depleted in lead and barium,
compared with other backarc examples.
The significant contribution of magmatic components to the hydrothermal fluids is
manifested by the presence of volatile elements such as sulfur and carbon. Sulfur-isotopic
values of the known sulfide samples are around 8%0 at the Izena cauldron (H. Sakai; pers.
comm.), 2.1-3.1%0 at the Mariana Trough (Kusakabe et aI., 1990), 0.3-2.2%0 at the
Minami-Ensei knoll (Nedachi et ai., 1992), 2.1-2.8%0 at the Kita-Bayonnaise caldera
(Iizasa, 1993a), and 0.9-1.2%0 at the Kaikata caldera (vein part) (Urabe, unpubl. data). This
value range from 0 and 8%0 corresponds to the compositional range of magmatic sulfur.
4.5. Chemical Composition of the Hydrothermal Fluid

Geochemical studies of venting hydrothermal fluid are considered to be important and
instructive, because fluid chemistry evolves from original seawater and is affected by
chemical interactions with various materials encountered during its circulation beneath the
seafloor. Active venting of high-temperature fluid (temperature around 300°C) was ob-
served in several sites in the western Pacific. At six sites, enough fluid samples were
collected and analyzed to estimate chemical composition of each hydrothermal end mem-
ber. Table III summarizes chemical data of these hydrothermal fluids together with those of
the EPR 21oN, OBS vent for comparison with the MOR hydrothermal system. There are
some reports of chemical data of fluid samples collected at other sites. Samples from
Vienna Woods in the Manus Basin (Lisitsyn et aI., 1993) and Franklin seamount in the
Woodlark basin (Binns et aI., 1993) were not listed in this table, because those samples were
significantly diluted with ambient seawater, which made end-member estimation ambig-
uous. Several fluid samples were collected and studied from low-temperature vents: the
Clam site (lheya site) in the Okinawa Trough (Gamo et aI., 1987, 1991a) and the Kasuga 2
seamount in the northern Mariana arc (McMurtry et aI., 1993). However, these data are
excluded from the following discussion, because the significant temperature difference
makes it difficult to compare the chemistry.
The major element composition of hydrothermal fluids is considered to be essentially
controlled by chemical equilibria with various mineral phases. In the reaction zone just
above the magma chamber, the hydrothermal fluid is saturated with respect to all of the
primary and secondary (alteration) minerals at hydrothermal conditions. This model has
been confirmed via geochemical studies by several works (e.g., Seyfried, 1987; Campbell
et aI., 1988; Bowers, 1989). As compiled in Table III, the high-temperature fluids from the
western Pacific have some characteristics in common, such as the depletion of Mg and
SO4' the enrichment in cations relative to seawater, and low pH and reducing conditions,
regardless of the type of hydrothermal system or the tectonic setting, including MOR. The
spectrum of major element compositions does not diverge more than an order of magnitude.
This similarity of major element composition is understandable, because both mineral
assemblages and physical conditions (,1'> 300°C, P> 100 bars) where hydrothermal fluid-
mineral interactions occur are similar in these systems.
Among the chemical equilibria at hydrothermal conditions, the control of dissolved
SiOz concentration, via quartz dissolution, has been well studied (e.g., Von Damm et ai.,
1991). In Fig. 13.8, the SiOz concentration of some fluids are plotted together with quartz
solubility data for isobars from 100 to 500 bars. Most data for the EPR systems (open
....
til
TABLE /I
Chemical Compositions of Hydrothennal Sulfide Samples in the Western Pacific
Location no. 4 7 8 9 7-9 13

N. Fiji N. Fiji Manus Lau Lau Lau Lau Mariana
(Bendel et al., (T. Urabe, (Binns and Vai Lili Hine Hina White Church (Fouquet et al., (Hannington,
1993) unpubl. data) Scott, 1993) (Fouquet et aI., (Fouquet et aI., (Fouquet et al., 1993) 1989)
1993) 1993) 1993)
Fe (%) 30.05 30.96 30 10.9 30.4 7.23 17.4 2.39

Cu (%) 7.45 2.76 30.3 8.18 2.77 4.19 4.56 1.15 c...
Pb (%) 0.057 0.Q25 0.035 0.27 0.88 0.4 0.33 7.4
Zn (%) 6.64 5.49 1.4 26.4 15.4 12.2 16.1 9.96
~
Mn (ppm) 761 65 522 744 542 175
Co (ppm) 238 90 20 6.9 3 2 i
0
Ni (ppm) 0.45 6
S (%) 36.67 38.28 30.2 29.05 42.4 20 30.12 17.8 ~
As (ppm) 182 170 2440 1575 4629 445 2213 126 ~
Sb (ppm) 30 21 320 35 89 51 190
Se (ppm) 168 230 56 21 8 10
~
).
Cd (ppm) 260 150 73 780 350 482 465
Sn (ppm) 13 4
~
W (ppm) ~
Mo (ppm) 270 80 360 27 20 56 32 5 Cil
Au (ppm) 1.077 1.2 10 1.21 2.51 2.42 1.4 0.784 §i
Ag (ppm) 47.7 42 147 680 118
0
151 256 184
c:
Hg (ppm) 22
Ba (%) 0.82 0.043 0.75 11.76 2.75 25.8 11.56 33.33
~
III
III
Location no. 13 16 18 18 21 :::t:
Mariana Okinawa Okinawa Okinawa Izu-Bonin EPR average Black Ore Besshi ti
(T. Urabe, Minami-Ensei Izena Izena Suiyo (Fouquet et at., average average lJ
0
unpubl. data) (Nedachi et al., (Halbach et al., 1989) (Urabe, 1989; (Watanabe and 1993) (n = 160) (n = 32)
1992) Aoki and Kajimura, 1994) (ANREa, (ANREa,
:r!
Nakamura, 1989) 1993) 1993)
~
~
r-
Fe (%) 4.57 3.29 7.33 4.84 16.7 25.93 9.62 36.38 :to.
Cu (%) 1.4 1.46 1.77 9.42 14.5 2.10 ()
1.67 5.95
14.27 15.8
::!
Pb (%) 0.29 5.01 0.15 0.06 6.76 0.09 :s-t
Zn (%) 8.83 12.8 22 31.3 11.6 11.92 13.58 1.55
1567 1236
-<
Mn (ppm) 300 165 961 301 348 ~
Co (ppm) 13 336 23.19 1116
Ni (ppm) 2.5 16.94 12.50 27.93 :r!rn
S (%) 36.92 25.16 36.56
~
As (ppm) 110 417 27537 15300 892 168 873 112 ()
Sb (ppm) 495 14.12 489 36.20 !i
()
Se (ppm) 0.92 112 1.73 17.92
Cd (ppm) 950 925 938 434 860 48.33
Sn (ppm) 2.56 12.70 6.94
W (ppm) 15.00 178 67.52
Mo (ppm) 120 52.62 113 80.28
Au (ppm) 0.65 2.43 4.6 3.6 19.7 0.20 2.02 0.84
Ag (ppm) 55.5 90.3 2100 1445 113 51.07 285 33.92
Hg (ppm) 3.68 42.31 3.31
Ba (%) 18.1 1.11 2.76 0.089 0.08 10.33 0.03
aAgency of Natural Resources and Energy.
.".
:::j
_ 1,000
e
0
A.
a: 100 6. Manus
a..
w
~ 10
~
;
-a.
E
c?l 0.1
e
0
0.Q1
Cu Zn Ag As Au Pb Ba
_ 1,000
e
0
B.
a: 100
a..
w
CD
01
I!! 10
~
CD
-a. 13. Mariana
E
c?l 0.1
e
0
0.01
Cu Zn Ag As Au Pb Ba
1,000
e c.
0
a: 100
a..
w
CD FIGURE 13.7. Variation in composition of hydrother-
If
.-,
10
mal mineral samples from the weatem Pacific. Metal
~ contents are nomalized with the average East Pacific
; Rise ore (listed in Table II). Average assay of Besshi-type
-a.
E ore and Balk ore (sphalerite-galena-barite ore) of the
c?l 0.1
e Kuroko deposit are also superimopsed (data after
0
0.01 Agency of Natural Resources and Energy, 1993). Num-
Cu Zn Ag As Au Pb Ba bers in the figure correspond to the location in Table I.
squares) are plotted on isobars deeper than the seafloor, which is considered evidence for
deeper hydrothermal reaction zones. In contrast, Si02 data for arc-related fluids, such as
those at Lau Basin, Suiyo seamount, Minami-Ensei knoll, and Izena cauldron, suggest
hydrothermal interaction at pressures and temperatures very close to the seafloor. This
could be attributed to very shallow hydrothermal circulation cells, caused probably by a
near surface magmatic body. For example, at the Izena site in the Okinawa Trough, heat
flow measurements suggest convection cells at about I-Ian depth (Kinoshita, 1990). An
alternative explanation is that hydrothermal fluid interacts with conduit rock during its
ascent because quartz dissolution occurs very quickly. Highly permeable volcaniclastic
rocks resulting from volatile-rich volcanic activity are commonly observed around the
above four sites, which is consistent with both explanations. For the Mariana Trough and
North Fiji Basin hydrothermal fluids, substantial cooling near the surface might explain the
reason why silica dissolution equilibrium is not attained in these systems.
The process of subseafloor phase separation, which appears more commonly in arc-
TABLE 11/
Chemical Compositions of Hydrothermal Fluids in the Western Pacific
Location no. 7 13 16 18 21
N. Fiji Lau Mariana Okinawa Okinawa Izu-Bonin EPR
White Lady Vai Lili Alice Spring Minami-Ensei Izena Suiyo OBS
Temperature 285 334 285 278 320 311 350

(0C)
pH 4.7 2.0 4.4 5.0 4.7 3.7 3.4
Li (fLM) 200 623 5800 1860 600 891
Na (mM) 210 590 431 430 446 432
K(mM) 10.5 79.0 31.0 50.9 73.7 29.7 23.2
Rb (fLM) 8.8 68 30 360 28
NH4 (mM) 0 4.70 5.32 <0.1
Mg (mM) 0 0 0 0 0 0 0
Ca (mM) 6.5 41.3 23.6 22.1 23.2 89.0 15.6
Sr (fLM) 30 20 73 227 110 300 81
Ba (fLM) 5.9 >39 55 60 100 8
Mn (fLM) 12 7100 94 370 587 960
Fe (fLM) 13 2500 21 435 1660
Cu (fLM) 34 0.003 35
Zn (fLM) 3000 7.6 106
Pb (nM) 3900 36 308
S04 (mM) 0 0 0 0 0 0 0
Cl (mM) 255 790 544 527 550 658 489
Br (fLM) 306 1140 866 1045 802
Alkalinity 0.12 0.43 3.51 0.88 -0.20 -0.40
(meq)
B (mM) 0.47 0.83 0.81 4.0 3.41 1.43 0.51
Al (fLM) 6.0 4.9 17.0 5.2
Si (mM) 14.0 14.5 14.0 10.8 12.5 13.2 17.6
H2S (mM) 2.0 <5 2.6 2.44 13.1 1.6 7.3
CO2 (mM) 14.4 15 96 200 40 8
87Srf8 6Sr 0.7046 0.7044 0.7033 0.7100 0.7089 0.7031
834S (H 2S) 3.6-4.8 3.6 7.4-7.7 1.3-5.5
(%0)
Data sources:
I North Fiji Basin, White Lady (Grimaud et aI., 1991; Ishibashi et aI., 1994a).
7 Lau Basin, Vai Lili Field, VL-3 (Fouquet et al., 1993).
13 Mariana Trough, Alice Spring (Campbell et al., 1987; Campbell and Edmond, 1989; Palmer and Edmond, 1989; Palmer, 1991).
16 Minami-Ensei knoll (Chiba et al., 1993).
18 Izena cauldron, (Sakai et aI., 1990a; Garno et ai., 1991a; Ishibashi, 1991; Chiba et aI., 1993; You et ai., 1994; Shitashima, 1994).
21 Suiyo seamount (Ishibashi et ai., I994b).
EPR East Pacific Rise 21°N, OBS vent (Von Damm et ai., 1985a).
related hydrothermal systems, is one of the most important factors that control fluid
chemistry and mineral occurrence. Boiling drastically changes the chemical composition of
the hydrothermal fluid during its ascent from the reaction zone. Phase segregation and
condensation between the vapor and the brine components result in a wide divergence in
the concentrations of the major elements (Butterfield et aI., 1990). Moreover, boiling
induces subseafloor precipitation of ore-forming metal elements (e.g., Drummond and
Ohmoto, 1985), because of the sudden increase in pH, mainly caused by CO 2 partitioning
into the vapor phase. The White Lady site in the North Fiji Basin was the first example
30
• Mariana, N.Fiji
'" Lau, Izu-Bonin
• Okinawa Trough
o Easl Pacific Rise
20
C\I 15
o
(j)
10
FIGURE 13.8. Relationship between Si0 2 concentration

5 and equilibrium condition with quartz. Isobars are drawn
according to the calculated data and plots of Von Damm et al.
O'----'----'--~ .........--'-~--I.---' (1991). Si02 concentrations of EPR 21°N fluids are from Von
200 250 300 350 400 450 500
Damm et at. (1985a) and Campbell et at. (1988), and those of
temperature others are from Table III.
described in the western Pacific (Grimaud et ai., 1991) of a phase-separated hydrothermal

system and the second in the world after the ASHES vent field, Juan de Fuca Ridge,
reported by Massoth et al. (1989). Since the venting fluids at the White Lady site are the
vapor-rich fraction of the phase-separated hydrothermal fluid, they have low concentrations
in all of the major elements (see Table III). The fluids discharging from these seafloor
orifices do not form sulfide deposits, only sulfate chimneys. Figure 13.9 shows the relation-
ship between temperature and depth (pressure) of the venting hydrothermal fluids in the
western Pacific region. While the fluids at East Pacific Rise (EPR), Mid-Atlantic Ridge
(MAR), and Mariana Trough plot in the one phase region, the Lau Basin, Suiyo seamount,
Minami-Ensei knoll, and Izena cauldron plot close to the two-phase boundary. For North
Fiji hydrothermal fluid, significant cooling of the condensed vapor-rich fluid near the
surface of the seafloor is responsible for the shift of the plot far from the two-phase
boundary region, although the fluid certainly experienced boiling (Ishibashi et ai., 1994a).
temperature ("C)
200 300 400 500
0
e~ .• Mariana. N.Fiji
•.
100 Lau. Izu-Bonin
•
0
Okinawa Trough
EPR,MAR
f!!
::I
200
•
I 300
0
400
• 0
500
FIGURE 13.9. Relationship between temperatures and depth of hydrothermal fluids at the seafloor. Two-phase
boundary curve for 3.2 wt% NaCl in H20 system (equivalent to seawater) is drawn from the data of Bischoff and
Rosenbauuer (1985).
Many examples of arc-related hydrothermal systems are located at shallower water depths
than those at MOR, which gives rise to P-T conditions close to the two-phase boundary.
Besides, the contribution of volatiles from arc magmas into the hydrothermal fluid would
further promote the initiation of boiling. In several deposit samples from within the
Okinawa Trough, there is evidence for explosions that are attributed to subseafloor boiling.
In contrast to the major elements, trace elements do not form separate secondary
(alteration) minerals. Their distribution between the hydrothermal fluid and the alteration
minerals is likely to be controlled by ion exchange with major elements. For example, the
Sr/Ca ratio of hydrothermal fluid has common values among the various hydrothermal
systems, which is attributed to a single-step control, probably by recrystallization of
plagioclase (Berndt et aI., 1988). The "soluble elements," such as lithium and boron, are
considered to be rarely involved in secondary (alteration) minerals during hydrothermal
interaction (Seyfried et ai., 1984). In other words, they are only leached from primary
minerals of rocks during early stages of fluid circulation. Thus, if we assume similar water/
rock ratios during hydrothermal interactions among the various tectonic settings, lithium
and boron concentrations of the hydrothermal fluid should reflect their contents in the
underlying rocks. This relationship could be used as an indicator of the rock type that the
fluid encountered. Palmer (1991) showed that the Mariana Trough hydrothermal fluid has a
higher boron content and lower 8 11 B isotopic composition than the EPR fluids, which is
attributed to the abundance of boron in backarc basalt magma having been influenced by
the subducting slab. The Lau Basin hydrothermal fluid has boron systematics quite similar
to the Mariana Trough hydrothermal fluid. Together with enrichment in other trace ele-
ments compared to the MOR fluid, the trace element composition of the Lau hydrothermal
fluid could be interpreted by a similar explanation (Fouquet et aI., 1991a,b). The high boron
content of the Suiyo seamount fluid would be consistent with the underlying rhyolitic
magma, since the latter is, in general, enriched in boron compared to basaltic magma.
Palmer (1991) also showed that higher boron contents occur in hydrothermal fluids of
sediment-hosted systems (e.g., 1.5-1.7 mM at Guaymas Basin and 1.7-2.1 mM at Escanaba
Trough). The boron abundance in the middle Okinawa Trough fluids could be attributed to
extensive fluid-sediment interaction (You et aI., 1994).
4.6. Gas Geochemistry of the Hydrothermal Fluid

Gas species are basically unrelated to fluid-mineral interactions, but they have been
shown to be good tracers for the source of the hydrothermal fluid during subseafloor
circulation. For sediment-starved hydrothermal systems on ridge crests, both carbon- and
helium-isotopic compositions are very close to those of MORB, indicating that they are
dominantly derived from a magma (e.g., Welhan and Craig, 1983). In contrast, in sediment-
hosted environments, fluid interaction with organic matter during hydrothermal circulation
results in the significant increase of CO 2 and CH4 with their 13C-depleted signature (e.g.,
Welhan and Lupton, 1987). Helium-isotopic compositions are known to be an unequivocal
geochemical indicator of the mantle component in hydrothermal fluids, because helium has
less fractionation (e.g., Lupton et ai., 1980). Compilation of the helium-isotopic composi-
tion of various fluids from the western Pacific (Fig. 13.10) provides information on the
isotopic signature of their magmatic sources beneath the hydrothermal system. The Mari-
ana Trough hydrothermal fluid has helium-isotopic composition encompassed by the
MORB range, in spite of the subtle difference in fluid chemistry. The North Fiji hydrother-
3HeJ4He (RIRalmj
4 6 8 10
East Pacific Rise
Mid-Atlantic Ridge
North Fiji Basin III
Mariana Trough
Izu-Bonin Arc
Okinawa Trough
• '"
FIGURE 13.10. Comparison of helium-isotopic compositions of various hydrothermal fluids in the western
Pacific. The hatched area represents the averaged helium-isotopic composition of MORB (Hilton et aI., 1993).
Data sources: East Pacific Rise: 21°N (Welhan and Craig, 1983), l3°N (Merlivat et aI., 1987), Mid-Atlantic Ridge:
Snake Pit site (Jean-Baptiste et aI., 1991), North Fiji Basin: Station 4 (Ishibashi et aI., 1994c), Mariana Trough:
central peak (Craig et aI., 1987b), Izu-Bonin arc: Suiyo seamount (Tsunogai et al., 1994), Okinawa Trough: Izena
cauldron, Iheya Ridge (Sakai et aI., 1990a), Minami-Ensei knoll (Chiba et aI., 1992).
mal fluid has a higher 3He/4 He ratio, which is likely to reflect the influence of the ocean
island basalt (alB) component into the magma source (Ishibashi et ai., 1994c). The
Okinawa Trough hydrothermal fluids have lower helium-isotopic ratios than MORB; these
ratios are commonly found in subaerial island-arc volcanic gases and geothermal systems.
It is interesting to note, therefore, that the hydrothermal fluid of the Suiyo seamount has a
MORB-like isotopic ratio in spite of its arc-related setting. Tsunogai et ai. (1994) attributed
this to minor contamination by crustal radiogenic helium from the thin continental crust of
the lzu-Bonin intraoceanic island arc.
As shown in Table III, CO 2 concentrations in the hydrothermal fluids of arc-related
systems range from 15 to 200 mM, spanning more than an order of magnitude. This range is
significantly higher than that of the MaR hydrothermal fluids which are typically 10 mM.
This abundance of CO 2 should be attributed to the volatile-rich nature of arc magmas
compared to MORB magmas, since the gas species are dominantly derived from magmatic
source. Although organic-derived CO 2 are also, at least partially, involved in the Okinawa
Trough hydrothermal systems, it is worthy to note that the CO 2 concentrations of 200 mM
correspond well with CO 2 contents observed in fluid inclusions (50-350 mM) contained
within stockwork ores from Kuroko deposits (Pisutha-Amond and Ohmoto, 1983). More-
over, the wide variance of CO 2 concentration in hydrothermal fluids in different tectonic
settings is notable, since the major element compositions are rather restricted, being
controlled by fluid-mineral equilibria, as discussed before. As an interesting demonstration
of the abundance of CO 2 in the Okinawa Trough, the discharge of liquid-C0 2 bubbles
forming CO 2 hydrate has been observed at two sites, the Izena hydrothermal field (Sakai et
ai., 1990b) and the Minami-Ensei knoll (Chiba et ai., 1992). In both cases, the bubbles are
dominated by 85-90% CO 2 , and their carbon-isotopic compositions coincide with that of
the dissolved CO 2 in high-temperature fluids. This suggests that these bubbles originated
from a magmatic source, although the mechanism of transportation (direct degassing,
separated fluids due to boiling) is not clear. Sakai et ai. (1990b) proposed that the CO 2-rich
fluids accumulate under a gas-hydrate seal barrier within a porous sediment layer and that
the bubble emanation leaks through a crack of this barrier, probably caused by a change of
hydrothermal activity.
4.7. Relationship of Fluid Chemistry to Seafloor Mineralization

The contents of dissolved ore-forming metal elements are controlled by various
parameters, such as temperature, pH, redox condition, and chlorine concentration. For
hydrothermal systems on MORs, the pH and redox condition of the hydrothermal fluid is
controlled only by fluid-rock mineral equilibria (e.g., Bowers et aI., 1988). In contrast, in
sediment-hosted systems, organic-derived species play an important role in regulating
these parameters. The Okinawa Trough hydrothermal systems are examples of sediment-
hosted systems located in a continental margin, where hydrothermal fluid has interacted
with volcanic, terrigenous, and organic material (Sakai et al., 1990a; Garno et aI., 1991a,b).
Hydrothermal fluid interactions within the sediment layer would be another important
source for the various fluid components, including gas species (e.g., Von Damm et aI.,
1985b). In particular, the thermal decomposition of organic matter causes significant
enrichment in CO 2, CH4 , and NH3 . Involvement of these species in the hydrothermal fluid
would control various parameters of the fluid chemistry, in particular pH, as NH3 has a high
pH buffering potential because of its strong dissolution capacity at high temperatures
(Ishibashi, 1991). Such pH regulation may be advantageous in continuous ore precipitation.
Chlorine concentration is one of the important parameters, since many metal elements are
transported as species of chloride complexes in the hydrothermal fluid. Although chlorine
is not related to any major alteration mineral phases, many of the chlorine concentrations in
hydrothermal fluids (Campbell and Edmond, 1989), including the Suiyo and Lau fluids,
deviate from that of ambient seawater. To explain the chlorine enrichment, several mecha-
nisms have been proposed including rock hydration, hydrothermal fluid phase separation,
mixing of the fluid with deep brine formed during previous phase separation (Von Damm
and Bischoff, 1987), dissolution of a retrograde soluble phase (Campbell and Edmond,
1989), and trapping of mantle-derived chlorine-rich fluids, but they are still controversial.
In summary, the fluid chemistry of hydrothermal systems located in the western
Pacific region is variable, reflecting their diverse geological environment and tectonic
setting. Among them, a magmatic contribution of volatile components plays an important
role in regulating some parameters of the fluid chemistry, such as the boiling point, pH, and
redox condition, which in tum control the solubility of ore-forming metal elements. The
variety of fluid chemistries provides a good opportunity to examine the relationship of fluid
chemistry and mode of occurrence of seafloor mineralization.
5. SUMMARY
A compilation of 27 sites of hydrothermal mineralization in the western Pacific is

given in terms of their tectonic settings, mode of mineral occurrence, and fluid chemistry.
These sites comprise about 20% of the known seafloor hydrothermal sites which are
dominated by those occurring at MORs. All the high-temperature hydrothermal sites
discussed above occur in association with submarine volcanism found either in backarc
spreading centers (14 sites), backarc rifts (3 sites), or volcanic fronts (7 sites) within arc-
backarc systems in the western Pacific region. No sulfide mineralization has been identified
in hot-spot volcanoes.
The lavas of the backarc spreading and backarc rift centers span a compositional range
that include slightly to moderately fractionated basalt, andesite, and rhyodacite. Hydrother-
mal activity associated with dacitic to rhyolitic volcanism is known at ten sites in the
western Pacific and eight of them accompany sulfide mineralization. Those include the
PACMANUS site in eastern Manus Basin, the southern Valu Fa Ridge sites in the Lau
Basin, three sites in the Okinawa Trough, and Suiyo seamount in the Izu-Ogasawara arc.
Since rhyodacite is extremely rare in MOR systems, ancient rhyolite-hosted deposits
(Kuroko type) that constitute 56% of the known massive sulfide deposits now on land were
likely to have formed in arclbackarc setting.
There are nine high-temperature hydrothermal sites with sulfide mineralization in
association with basaltic to andesitic volcanism, such as central Manus Basin, central
Mariana Trough, and northern Valu Fa Ridge. These deposits seem to provide the modern
analog of the Besshi-type deposits, which is defined as a cupriferous-pyrite massive sulfide
deposit hosted in metamorphosed, submarine tholeiitic basalt. The deposits differ signifi-
cantly from the Cyprus type by an abundance of epiclastic rocks and the lack of ophiolite
suite. The ubiquitous occurrence of sand- to pebble-size epiclastic sediments, which are
absent in MOR, on the flank of arclbackarc volcanoes suggests that these volcanoes were
the sites for the Besshi-type mineralization. The deposit in the North Fiji Basin, located at a
spreading center of the matured backarc basin, may be a candidate for the Cyprus type.
It is reasonable to deduce that the replenishing of arc magma happens periodically at
an interval of 102-103 years. The associated hydrothermal activity is, therefore, likely to
periodically reactivate and cease in accordance with the volcanic activity. There is not
enough data to estimate the time scale of hydrothermal activity at submarine arc volcanoes.
However, repeated Shinkai 2000 dives on the Kaikata seamount suggest that the lifetime of
the hydrothermal activity at frontal volcano is likely to be less than 10 years.
The hydrothermal deposits of the western Pacific are characterized by their high gold
content, low pyrite/marcasite content, and lack of pyrrhotite except at a few sites. This is
due to low iron content in hydrothermal fluids and higherj02 andjS2 conditions. Average
chemical composition of the sulfides shows certain characteristics compared with those
from the MORs. The average composition of sulfide samples of the North Fiji Basin, which
is in the most advanced stage of backarc spreading, closely resembles that of the EPR. On
the other hand, that of the Manus Basin (PACMANUS site) shows slight enrichment in gold
and arsenic. Such "epithermal elements" were likely derived from underlying dacite
magma, by an analogy of the terrestrial gold mineralization at Osorezan volcano, northern
Japan. The composition of sulfide samples from the Lau Basin, Mariana Trough, and
Minami-Ensei matches closely that of the average black ore (Zn-Pb-Ba massive ore) of the
Kuroko deposits in Japan, a strong indication of their common mechanism of formation.
Significant contribution of magmatic components to the hydrothermal fluids is also
manifested by volatile elements such as sulfur, carbon, and helium. Sulfur-isotopic values
of the known sulfide samples fall in the range 0%0 to 8%0, which corresponds to the
compositional range of magmatic sulfur.
Enough samples of high-temperature fluid (around 300°C) have been collected and
analyzed at six sites to estimate the chemical composition of end-member fluid. The fluids
have some characteristics in common, such as the depletion of Mg and SO4' the enrichment
in cations relative to seawater, low pH, and reduced conditions, regardless of the type of
hydrothermal systems or the tectonic setting. The spectrum of major element compositions
does not diverge more than an order of magnitude from those of the MOR sites. The Si02
concentration of some arc-related hydrothermal systems, such as Lau Basin, Suiyo sea-
mount, Minami-Ensei knoll, and Izena cauldron sites, lies along quartz solubility isobars
very close to the pressure of the seafloor, which suggests the existence of very shallow
hydrothennal circulation cells driven by a near-surface magmatic body.
The process of subseafloor phase separation appears to be more common in arc-related
hydrothennal systems. This process drastically changes the chemical composition of the
hydrothennal fluid and induces subseafloor sulfide precipitation. The White Lady site in the
North Fiji Basin was the first example of a phase-separated hydrothennal system described
in the western Pacific. Fluid chemistry is characterized by low concentrations in all of the
major elements and depletion in metal elements, which corresponds to sulfate chimney
fonnation. Several examples of arc-related hydrothennal systems are located at shallower
water depths than those at MaR system, which gives rise to P-T conditions close to the two-
phase boundary. Also, the added contribution of volatiles from arc magmas into the
hydrothennal fluid would promote the initiation of boiling. In several deposit samples
within the Okinawa Trough, there is evidence for explosions that are attributed to sub-
seafloor boiling.
Gas species are shown to be a good tracer for the source of the hydrothennal fluid
during subseafloor circulation. Compilation of the helium isotopic composition of various
fluids provides infonnation on the isotopic signature of their magmatic sources beneath the
hydrothennal system. The CO2 concentration (15-200 mM) in the hydrothennal fluids of
arc-related systems is significantly higher than that of the MaR hydrothennal fluids (ave.
10 mM) and corresponds well with that (50-350 mM) observed in fluid inclusions from the
stockwork part of Kuroko deposits. This is attributed to the volatile-rich nature of arc
magmas, since the gas species are dominantly derived from a magmatic source. The
hydrothennal systems located in the western Pacific region are variable, reflecting their
diverse geologic environment and tectonic setting. Nevertheless, a magmatic contribution
to the systems is more obvious than those in MaR, where several lines of evidence also
suggest direct interactions between the magma chamber and the hydrothennal fluids. This
is consistent with a hypothesis of magmatic hydrothennal origin of Kuroko-type deposits.
Acknowledgments
We are grateful to Brian Taylor, Steven Scott, Toshitaka Gamo, and Cornel deRonde
for their critical comments and improvement of the English in an early version of the
manuscript. Discussion with Hitoshi Sakai, Gary Massoth, Masahiro Aoki, Manabu Tana-
hashi, Hitoshi Chiba, and Urumu Tsunogai was quite helpful during the preparation of this
text. Ms. Hinako Shimizu helped in the word processing of the text. We particularly wish to
acknowledge the captains and crews of the various cruises and operational teams of DSRV s
whose cooperation enabled us to explore the wonder of seafloor hydrothennal activity.
REFERENCES
Agency of Natural Resources and Energy. 1993. Report of the Survey on the Occurrence of Rare Metal Resources;
Evaluation of Its Potential, Ministry of International Trade and Industry, Tokyo. (in Japanese).
Aoki, M. 1992a. Active gold mineralization in the Osorezan caldera, in 29th IGC Field Trip Guide Book, Vol. 6,
pp. 69-75, Society of Resource Geology, Tokyo.
Aoki, M. 1992b. Magmatic fluid discharging to the surface from the Osorezan geothermal system, Northern
Honshu, Japan, in Magmatic Contributions to Hydrothermal Systems (1. W. Hedenquist, ed.), pp. 16-21,
Geological Survey of Japan Report No. 279.
Aoki, M., Matsumoto, T., Kimura, M., and Fujioka, K. 1993. Hydrothermal activity and topographic features in the
western part of the Minami-Ensei Knoll, Northern Okinawa Trough, in Proc. JAMSTEC Symp. Deep Sea
Res., 9:309-320. (in Japanese with English abstract).
Aoki, M., and Nakamura, K. 1989. The occurrence of chimneys in Izena Hole No.2 ore body and texture and
mineral composition of the sulfide chimneys, in Proc. JAMSTEC Symp. Deep Sea Res., 5:197-210. (in
Japanese with English abstract).
Aoki, M., and Thompson, M. 1990. The Osorezan hydrothermal system, Japan: Gold bearing hot springs,
Geotherm. Resour. Council, Trans. 14(2):1365-1369.
Appel, P. W. U. 1979. Stratabound copper sulfides in a banded iron-formation and in basaltic tuffs in the early
Precambrian Isua supracrustal belt, west Greenland, Econ. Geol. 74:45-52.
Aramaki, S., Oshima, 0., Uto, K., and Kawanabe, Y. 1992. Izu-Oshima and Hakone volcanoes, in 29th IGC Field
Trip Guide Book, Vol. 4., pp. 111-148, Geological Survey of Japan.
Astakhova, N. V.1992. Hydrothermal barite in the Okhotsk Sea, in Proc. 29th IGC 1992 (S. Ishihara, T. Urabe, and
H. Ohmoto eds.), Resource Geol., Special Issue No. 17, pp. 169-172.
Auzende, J.-M., Bendel, v., Fujikura, K., Geistdoerfer, P., Gracia-Mont, E., Joshima, M., Kisimoto, K., Mitsu-
zawa, K., Murai, M., Nojiri, Y., Ondreas, H., Pratt, C., and Ruellan, E. 1992. Resultats preliminaires des
plongees du Shinkai 6500 sur la dorsale du Bassin Nord-Fidjien (SW Pacifique)-Programme STARMER,
C.R. Acad. Sci. Paris 314(II):491-498. (in French with English abstract).
Auzende, J.-M., Eissen, J. P., Lafoy, Y., Gente, P., and Charlou, J. L. 1988. Seafloor spreading in the North Fiji
Basin (South Pacific), Tectonophysics 146:317-351.
Auzende, J.-M., Honza, E., Boespflug, X., Deo, S., Eissen, J.-P., Hashimoto, J., Huchon, P., Ishibashi, J., Iwabuchi,
Y., Jarvis, P., Joshima, M., Kisimoto, K., Kuwahara, Y., Lafoy, Y., Matsumoto, T., Maze, J.-P., Mitsuzawa, K.,
Monma, H., Naganuma, T., Nojiri, Y., Ohta, S., Otsuka, K., Okuda, Y., Ondreas, H., Otsuki, A, Ruellan, E.,
Sibuet, M., Tanahashi, M., Tanaka, T., and Urabe, T. 1990. Active spreading and hydrothermalism in North
Fiji Basin (SW Pacific). Results of Japanese French cruise Kaiyo 87, Mar. Geophys. Res. 12:269-283.
Auzende, J.-M., and Urabe, T. 1994. The STARMER French-Japanese joint project, 1987-1992, Mar. Geol.
116:1-3.
Auzende, J.-M., Urabe, T., Bendel, v., Deplus, c., Eissen, J.-P., Grimaud, D., Huchon, P., Ishibashi, J., Joshima,
M., Lagabrielle, Y., Mevel, c., Naka, J., Ruellan, E., Tanaka, T., and Tanahashi, M. 1991. In situ geological
and geochemical study of an active hydrothermal site on the North Fiji Basin ridge, Mar. Geol. 98:259-269.
Auzende, J.-M., Urabe, T., Deplus, c., Eissen, J. P., Grimaud, D., Huchon, P., Ishibashi, J., Joshima, M.,
Lagabrielle, Y., Mevel, C., Naka, J., Ruellan, E., Tanaka, T., and Tanahashi, M. 1989. Geological setting of an
active hydrothermal site: preliminary results of the STARMER I cruise of the submersible Nautile in the
North Fiji Basin, C.R. Acad. Sci. Paris 309(II):1247-1254. (in French with English abstract).
Backer, H., and Lange, J. 1987. Recent hydrothermal metal accumulation: Products and conditions of formation,
NATO ASI Series C 194:317-337.
Baranov, E., and Levin, L. E. 1993. Geotectonic nature of localization of the Besshi and Kuroko type deposits in
the mobile belts of the continents and in island arcs, in Proc. 29th IGC 1992 (S. Ishihara, T. Urabe, and H.
Ohmoto, eds.), Resource Geol., Special Issue No. 17, pp. 331-335.
Baranov, B. v., Seliverstov, N. l., Murav' ev, A. v., and Muzurov, E. L. 1991. The Komandorsky Basin as a product
of spreading behind a transform plate boundary, Tectonophysics 199:237-269.
Bendel, V., Auzende, J.-M., Lagabrielle, Y., Grimaud, D., Fouquet, Y., and Urabe, T. 1993. The White Lady
hydrothermal site in the north Fiji Basin (STARMER), Econ. Geol. 88:2237-2249.
Berndt, M. E., Seyfried, W. E., Jr., and Beck, W. 1988. Hydrothermal alteration processes at midocean ridges:
Experimental and theoretical constraints from Ca and Sr exchange reactions and Sr isotope ratios, J.
Geophys. Res. 93:4573-4583.
Bertine, K. K., and Keene, J. B.1975. Submarine barite-opal rocks of hydrothermal origin, Science 188:150-152.
Binns, R. A, and Scott, S. D. 1993. Actively forming polymetallic sulfide deposits associated with felsic volcanic
rocks in the eastern Manus back-arc basin, Papua New Guinea, Econ. Geol. 88:2226-2236.
Binns, R. A, Scott, S. D., Bogdanov, Y. A., Lisizin, A. P., Gordeev, V. v., Gurvich, E. G., Finlayson, E. J., Boyd,
T., Dotter, L. E., Wheller, G. E., and Muravyev, K. G. 1993. Hydrothermal oxide and gold-rich sulfate
deposits of Franklin Seamount, Western Woodlark Basin, Papua New Guinea, Econ. Geol. 88:2122-2153.
Bischoff, J. L., and Rosenbauer, R. J. 1984. The critical point and two-phase boundary of seawater, 200-5OO°C,
Bischoff,1. L., and Rosenbauer, R J. 1985. An empirical equation of state for hydrothermal seawater, (3.2 percent
NaC\), Am. J. Sci. 285:725-763.
Both, R A., Crook, K., Taylor, B., Brogan, S., Chappell, B., Frankel, E., Liu, L., Sinton, J., and Tiffin, D. 1986.
Hydrothermal chimneys and associated fauna in the Manus back-arc basin, Papua New Guinea, EOS, Trans.
AGU 67:489-490.
Bowers, T. S. 1989. Stable isotope signatures of water-rock interaction in mid-ocean ridge hydrothermal systems:
Sulfur, oxygen, and hydrogen, J. Geophys. Res. 94:5775-5786.
Bowers, T. S., Campbell, A. c., Measures, C. I., Spivack, A. 1., Khadem, M. and Edmond, J. M. 1988. Chemical
controls on the composition of vent fluids at l3°-lioN and 210N, East Pacific Rise, J. Geophys. Res.
93:4522-4536.
Butterfield, D. A., Massoth, G. J., McDuff, R E., Lupton, 1. E., and Lilley, D. 1990. Geochemistry of hydrothermal
fluids from Axial Seamount Hydrothermal Emissions Study vent field, Juan de Fuca Ridge: Subseafloor
boiling and subsequent fluid-rock interaction, J. Geophys. Res. 95:12895-12921.
Campbell, A. C., Bowers, T. S., Measures, C. I., Falkner, K. K., Khadem, M., and Edmond, J. M. 1988. A time
series of vent fluid compositions from 210N, East Pacific Rise (1979, 1981, 1985), and the Guaymas Basin,
Gulf of California (1982, 1985), J. Geophys. Res. 93:4537-4549.
Campbell, A. C., and Edmond, 1. M. 1989. Halide systematics of submarine hydrothermal vents, Nature 342:
168-180.
Campbell, A. C., Edmond, 1. M., Colodner, D., Palmer, M. R, Falkner, K. K. 1987. Chemistry of hydrothermal
fluids from the Mariana Trough back arc basin in comparison to mid-ocean ridge fluids., EOS, Trans. AGU
68:1531 (abstract).
Cheminee, J. L., Stoffers, P., McMurtry, G. M., Richnow, H., Puteanus, D., and Sedwick, P. 1991. Gas-rich
submarine exhalations during the 1989 eruption of Macdonald Seamount, Earth Planet. Sci. Lett. 107:
318-327.
Chiba, H., Nakashima, K., Garno, T., Ishibashi, J., Tsunogai, U., and Sakai, H. 1993. Hydrothermal activity at the
Minami-Ensei Knoll, Okinawa Troguh: Chemical characteristics of hydrothermal solutions, in Proc.
JAMSTEC Symp. Deep Sea Res., 9:271-282. (in Japanese with English abstract)
Chiba, H., Sakai, H., Garno, T., Ishibashi, J., Nakashima, K., Minami, H., and Dobashi, F. 1992. Chemistry and
isotopic composition of bubbles emerging from the seafloor at the Minami-Ensei Knoll, Okinawa Trough, in
Proc. JAMSTEC Symp. Deep Sea Res., 8:81-87. (in Japanese with English abstract).
Christie, D. M., and Sinton, J. M. 1981. Evolution of abyssal lavas along propagating segments of the Galapagos
spreading center, Earth Planet. Sci. Lett. 56:321-355.
Compston, W., Kinny, P. D., Williams, I. S., and Foster, J. J. 1986. The age and Pb loss behavior of zircons from
Isua supracrustal belt as determined by ion microprobe, Earth Planet. Sci. Lett. 80:71-81.
Craig, H., Craig, V. K., and Kim, K. R. 1987a. PAPATUA Expedition 1: Hydrothermal vent surveys in back-arc
Basins: The Lau, N. Fiji, Woodlark, and Manus basins and Havre Trough, EOS, Trans. AGU 68:100
(abstract).
Craig, H., Horibe, Y., Farley, K. A., Welhan, J. A., Kim, K.-R. and Hey, R N. 1987b. Hydrothermal vents in the
Mariana Trough; Results of the first Alvin dives, EOS, Trans. AGU 68:1531 (abstract).
De Carlo, E. H., McMurtry, G. M., and Yeh, H.-W. 1983. Geochemistry of hydrothermal deposits from Loihi
submarine volcano, Hawaii, Earth Planet. Sci. Lett. 66:483-499.
Desbruyeres, D., Alayse-Danet, A.-M., Ohta, S., and Scientific Parties ofBIOLAU and STARMER Cruises. 1994.
Deep-sea hydrothermal communities in southwestern Pacific back-arc basins (the North Fiji and Lau basins):
Composition, microdistribution and food web, Mar. Geol. 116:227-242.
Drummond, S. E., and Ohmoto, H. 1985. Chemical evolution and mineral deposition in boiling hydrothermal
system, Econ. Geol. 80:126-147.
Eissen, 1.-P., Lefevere, P., Maillet, P., Morvant, G., and Nohara, M.1991. Petrology and geochemistry of the central
North Fiji Basin spreading center (SW Pacific) between 16°S and 22°S, Mar. Geol. 98:201-239.
Eissen, J.-P., Nohara, M., Cotten, J., and Hirose, K. 1994. North Fiji Basin basalts and their magma sources. Part 1:
Incompatible element constraints, Mar. Geol. 116:153-178.
Fouquet, Y., and the NAUTILAU Group. 1990. Hydrothermal activity in the Lau Basin, EOS, Trans. AGU 71:
678-679.
Fouquet, Y., von Stackelberg, U., Charlou, 1. L., Donval, J. P., Erzinger, J., Foucher, J. P., Herzig, P., Muhe, R.,
Soakai, S., Wiedicke, M., and Whitechurch, H. 1991a. Hydrothermal activity and metallogenesis in the Lau
back-arc basin, Nature 349:778-781.
Fouquet, Y., von Stackelberg, U., Charlou, J.-L., Donval, J.-P., Foucher, J.-P., Erzinger, J., Herzig, P., Muhe, R.,
Wiedicke, M., Soakai, S., and Whitechurch, H. 1991b. Hydrothermal activity in the Lau back-arc basin:
Sulfides and water chemistry, Geology 19:303-306.
Fouquet, Y., von Stackelberg, U., Charlou, J.-L., Erzinger, J., Herzig, P. M., Muhe, R., and Wiedicke, M. 1993.
Metallogenesis in back-arc environments: The Lau Basin example, Econ. Geol. 88:2154-2181.
Francheteau, J., Needham, H. D., Choukroune, P., Juteau, T., Segret, M., Ballard, R. D., Fox, P. J., Normark, W.,
Carranza, A., Cordova, D., Guerrero, J., Ranjin, c., Bougault, H., Gambon, P., and Hekinian, R. 1979.
Massive deep-sea sulfide ore deposits discovered on the East Pacific Rise, Nature 277:523-528.
Franklin, J. M., Lydon, J. W., and Sangster, D. E 1981. Volcanic-associated massive sulfide deposits, Econ. Geol.,
Special Issue, 75th Anniversary Volume: 485-627.
Fryer, P. 1993. The relationship between tectonic deformation, volcanism, and fluid venting in the southeastern
Mariana convergent plate margin, in Proc. JAMSTEC Symp., Deep Sea Res., 9:161-179.
Fryer, P., Gill, J., Jackson, M., Fiske, R., Hochstaedter, A., McMurtry, G., Sedwick, P., Malahaoff, A., Mouginins-
Mark, P., and Horikoshi, E. 1987a. Results of recent Alvin studies of the Kasuga volcanoes, northern Mariana
Are, EOS, Trans. AGU 68:1531 (abstract).
Fryer, P., Haggerty, J., Tilbrook, B., Sedwick, P., Johnson, L., Saboda, K., Newsom, S., Karig, D., Uyeda, S., and
Ishii, T. 1987b. Results of Alvin studies of Mariana forearc serpentinite diapirism, EOS, Trans. AGU 68: 1534
(abstract).
Fryer, P., Saboda, K. L., Johnson, L. E., Mackay, M. E., Moore, G. E, and Stoffers, P. 1990. Conical seamount:
SeaMARCII, Alvin submersible, and seismic-reflection studies, Proc. ODP, Init. Repts., 125, (P. Fryer, J. A.
Pearce, L. B. Stokking et aI., eds.), pp. 69-80, Ocean Drilling Program, College Station, TX.
Fujikura, K., Hashimoto, J., Segawa, S., and Fujiwara, Y. 1993. Thermal tolerance of White Blind Crab,
Bythograeidea, inhabited at hydrothermal vents, in Proc. JAMSTEC Symp. Deep Sea Res., 9:383-391. (in
Japanese with English abstract)
Fujioka, K. 1983. Where were the "Kuroko deposits" formed, looking for the present day analogy, Mining Geol.,
Special Issue, No. 11:55-68. (in Japanese with English abstract)
Garno, T., Chiba, H., Fujioka, K., Ishibashi, J., Kelly, K., Masuda, H., Ohta, S., Reysenbach, A.-L., Rona, P.,
Shibata, T., Tamaoka, J., Tanaka, H., and Yamaguchi, T. 1993b. Alice springs hydrothermal field, mid-
Mariana Trough 1992 revisit, EOS, Trans. AGU 43:361 (abstract).
Garno, T., Ishibashi, J., Sakai, H., Kodera, M., Igarashi, G., Ozima, M., Akagi, T., and Masuda, A. 1987.
Geochemistry of hydrothermal solutions in the Okinawa Trough: Report on Dive 235 of the SHINKAI2000,
in Proc. JAMSTEC Symp. Deep Sea Res., 3:213-224. (in Japanese with English abstract).
Garno, T., Sakai, H., Ishibashi, J., Oomori, T., Chiba, H., Shitashima, K., Nakashima, K., Tanaka, Y., and Masuda,
H. 1991b. Growth mechanism ofthe hydrothermal mounds at the CLAM site, Mid Okinawa Trough, inferred
from their morphological, mineralogical, and chemical characteristics, in Proc. JAMSTEC Symp. Deep Sea
Res. 7:163-184. (in Japanese with English abstract)
Garno, T., Sakai, H., Ishibashi, J., Nakayama, E., Isshiki, K., Matsuura, H., Shitatshima, K., Takeuchi, K., and
Ohta, S. 1993c. Hydrothermal plumes in the eastern Masnus Basin, Bismarck Sea: CH4 , Mn, AI and pH
anomalies, Deep-Sea Res. 40:2335-2349.
Garno, T., Sakai, H., Kim, E. S., Shitashima, K., and Ishibashi, J. 1991a. High alkalinity due to sulfate reduction in
the CLAM hydrothermal field, Okinawa Trough, Earth Planet. Sci. Lett. 107:328-338.
Garno, T., and Shipboard Scientific Party. 1993a. Revisits to the mid-Mariana Trough hydrothermal site and
discovery of new venting in the southern Mariana region by the Japanese Submersible Shinkai 6500,
InterRidge News 2:11-14.
Gracia, E., Ondreas, H., Bendel, v., and STARMER Group. 1994. Multi-scale morphologic variability of the North
Fiji Basin ridge (southwest Pacific), Mar. Geol. 116:133-151.
Grimaud, D., Ishibashi, J., Lagabrielle, Y., Auzende, 1. M., and Urabe, T. 1991. Chemistry of hydrothermal fluids
from the I7°S active site on the North Fiji Basin ridge (SW Pacific), Chem. Geol. 93:209-218.
Haggerty, J. A. 1987a. Cold-water, Deep-sea chimneys from the Mariana forearc serpentinite seamounts, EOS,
Trans. AGU 68:1534 (abstract).
Haggerty, J. A. 1987b. Petrology and geochemistry of Neogene sedimentary rocks from Mariana forearc
seamounts, in Seamounts, Islands and Atolls (B. H. Keating, P. Fryer, R. Batiza, and G. W. Boehlert, eds.),
Geophys. Monogr. Ser., Vol. 43, pp. 175-186, American Geophysical Union, Washington, DC.
Halbach, P., Nakamura, K.I., Wahsner, M., Lange, J., Sakai, H., Kaeselitz, L., Hansen, R. D., Yamano, M., Post, J.,
Prause, B., Seifert, R., Michaelis, W., Teichmann, E, Kinoshita, M., Maerten, A., Ishibashi, J., Czervinski, S.,
and Bulm, N. 1989. Probable modem analogue of Kuroko-type massive sulphide deposits in the Okinawa
Trough back-arc basin, Nature 338:496-499.
Halbach, P., Pracejus, B., and Andreas M. 1993. Geology and mineralogy of massive sulfide ores from the central
Okinawa Trough, Japan, Econ. Geol. 88:2210-2225.
Hannington, M. D. 1989. The geochemistry of gold in modern sea-floor hydrothermal systems and implications for
gold mineralization in ancient volcanogenic massive sulfides, Ph.D thesis, Univ. of Toronto, 554 p.
Hashimoto, J., Fujikura, K., and Hotta, H. 1990. Observation of deep sea biological communities at the Minami-
Ensei knoll, in Proc. JAMSTEC Symp. Deep Sea Res. 6:167-180. (in Japanese with English abstract)
Hashimoto, 1., Fujikura, K., Ohta, S., and Miura, T. 1993. Observations of hydrothermal vent communities at the
Minami-Ensei Knoll-2, in Proc. JAMSTEC Symp. Deep Sea Res. 9:327-336. (in Japanese with English
abstract)
Hawkins, J. 1986. "Black smoker" vent chimneys, EOS, Trans. AGU 67:430.
Hawkins, J., and Helu, S. 1986. Polymetallic sulphide deposits from "black smoker" chimney, Lau Basin, EOS,
Trans. AGU 67:378.
Hawkins, J. w., Lonsdale, P. E, Macdougall, J. D., and Vope, A. M. 1990. Petrology of the axial ridge of the
Mariana Trough back arc spreading center, Earth Planet. Sci. Lett. 100:226-250.
Hedenquist, J. w., White, N. C., and Siddeley, G. (eds.). 1990. Epithermal Gold Mineralization of the Circum-
Pacific, Association of Explorative Geochemistry, Special Publication No. 16ab.
Herzig, P. M., Hannington, M. D., Fouquet, Y., von Stackelberg, U., and Petersen, S. 1993. Gold-rich polymetallic
sulfides from the Lau back arc and implications for the geochemistry of gold in sea-floor hydrothermal
systems of the Southwest Pacific, Econ. Geol. 88:2182-2225.
Herzig, P., Hannington, M., McInnes, B., Stoffers, P., Villinger, H., Seifert, R., Binns, R., and Liebe, T. 1994.
Submarine volcanism and hydrothermal venting studied in Papua, New Guinea, EOS, Trans. AGU 75:
513-516.
Herzig, P., von Stackelberg, U., and Petersen, S. 1990. Hydrothermal mineralization from the Valu Fa Ridge, Lau
backarc basin (SW Pacific), Mar. Mining 9:271-301.
Hessler, R. R., France, S. C., and Boudrias, M. A. 1987. Hydrothermal vent communities of the Mariana back-arc
basin, EOS, Trans. AGU 68:1531 (abstract).
Hessler, R. R., and Lonsdale, P. E 1991. The biogeography of the Mariana Trough hydrothermal vents, in: Marine
Biology-Its Accomplishment and Future Prospect (J. Mauchline and T. Nemoto, eds.), pp. 165-182,
Hokusensha, Tokyo.
Hessler, R. R., Lonsdale, P., and Hawkins, J. 1988. Patterns on the ocean floor, New Sci. 117:47-51.
Hilton, D. R., Hammerschimdt, K., Loock, G., and Friederichsen, H. 1993. Helium and argon isotope systematics
of the central Lau Basin and Valu Fa Ridge: Evidence of crust/mantle interactions in a back-arc basin,
Hochstaedter, A. G., Gill, J. B., Kusakabe, M., Newman, S., Pringle, M., Taylor, B., and Fryer, P. 1990. Volcanism
in the Sumisu rift. I: Major element volatile, and stable isotope geochemistry, Earth Planet. Sci. Lett.
100: 179-194.
Honza, E. 1991. The Tertiary arc chain in the western Pacific, Tectonophysics 187:285-303.
Honza, E., Auzende, J. M., and KAIY088 Shipboard Party. 1989. Geology of the rift system in the North Fiji
Basin: Results of Japan-France cooperative research on board KAIY088, La Mer 27:53-61.
Honza, E., and Tamaki, K. 1985. Bonin arc, in The Ocean Basins and Margins, 7A, The Pacific Ocean (A. E. M.
Nairn, E G. Stehli, and S. Uyeda, eds.), pp. 459-502, Plenum Press, New York.
Horibe, Y., Kim, K. R., and Craig, H. 1986. Hydrothermal methane plumes in the Mariana back-arc spreading
centre, Nature 324:131-133.
Hussong, D. M., and Uyeda, S. 1981. Tectonic processes and the history of the Mariana arc: A systhesis of the
results of Deep Sea Drilling Project Leg 60, in In it. Repts. DSDP, 60 (D. M. Hussong and S. Uyeda, et aI.,
eds.), pp. 909-929, Deep Sea Drilling Project, Washington, DC.
Iizasa, K. 1993a. Petrographic investigations of seafloor sedeiments from the Kita-Bayonnaise submarine caldera,
Shichito-Iwo Jima Ridge, Izu-Ogasawara arc, northwestern Pacific, Mar. Geol. 112:271-290.
Iizasa, K. 1993b. Assessment of the hydrothermal contribution to seafloor sediements in the Myojinsho submarine
caldera, Shichito-Iwo Jima Ridge, Izu-Ogasawara arc, Japan, Mar. Geol. 114:119-132.
Iizasa, K., Terashima, S., Sasaki, M., and Marumo, K. 1993. Hydrothermal mineralization in the Kita-Bayonnaise
submarine caldera, Izu-Ogasawara arc, in Proc. JAMSTEC Symp. Deep Sea Res. 9:105-115. (in Japanese
with English abstract)
Iizasa, K., Yuasa, M., and Yokota, S. 1992. Mineralogy and geochemistry of volcanogenic sulfides from the
Myojinsho submarine caldera, the Shichito-Iwo Jima Ridge, Izu-Ogasawara arc, northwestern Pacific. Mar.
Geol. 108:39-58.
Ishibashi, J. 1991. Geochemical studies of hydrothermal fluids in the middle Okinawa Trough back arc basin, Ph.D
Thesis, Univ. Tokyo, l3lp.
Ishibashi, J., Grimaud, D., Nojiri, Y., Auzende, J. M., and Urabe, T. 1994a. Fluctuation of chemical compositions
of the phase-separated hydrothermal fluid from the North Fiji Basin Ridge, Mar. Geol. 116:215-226.
Ishibashi, J., Sano, Y., Wakita, H., Garno, T., Tsutsumi, M., and Sakai, H. 1990. Geochemical studies on the
hydrothermal activity in the mid-Okinawa Trough: Characterization of hydrothermal fluids from chemical
and isotopical composition of the gas components, in Proc. JAMSTEC Symp. Deep Sea Res. 6:63-68. (in
Japanese with English abstract)
Ishibashi, J., Tsunogai, U., Wakita, H., Watanabe, K., Kajimura, T., Shibata, A., Fujiwara, Y., and Hashimoto, J.
1994b. Chemical composition of hydrothermal fluids from the Suiyo and the Mokuyo Seamounts, lzu-Bonin
Arc, in JAMSTEC Journal of Deep Sea Res. 10, pp. 89-97. (in Japanese with English abstract)
Ishibashi, 1., Wakita, H., Nojiri, Y., Grimaud, D., Jean-Baptiste, P., Garno, T., Auzende, J. M., and Urabe, T. 1994c.
Helium and carbon geochemistry of hydrothermal fluids from the North Fiji Basin spreading ridge, Southern
Pacific, Earth Planet. Sci. Lett. 128:183-197.
Ishizuka, H., Kawanobe, Y., and Sakai, H. 1990. Petrology and geochemistry of volcanic rocks dredged from the
Okinawa Trough, an active back-arc basin, Geochem. J. 24:75-92.
Izawa, E., and Aoki, M. 1991. Geothermal activity and epithermal gold mineralization in Japan, Episodes 14:
269-273.
Izawa, E., Motomura, Y., Tanaka, T., and Kimura, M. 1991. Hydrothermal carbonate chimneys in the Iheya Ridge
of the Okinawa Trough, in Proc. JAMSTEC Symp. Deep Sea Res. 7:185-192. (in Japanese with English
abstract)
Japanese DELP Research Group on Backarc Basins. 1986. Report on DELP 1984 cruises in the middle Okinawa
Trough. Part 1: General outline, Bull. Earthquake Res. lnst. Univ. Tokyo 61:159-165.
Japanese DELP Research Group on Backarc Basins. 1991. Report on DELP 1988 cruises in the southern Okinawa
Trough. Part I: General outline, Bull. Earthquake Res. Inst. Univ. Tokyo 66:1-16.
Jarvis, P., Hughes-Clarke, J., Tiffin, D., Tanahashi, M., and Kroenke, L. 1994. The western Fiji transform fault and
its role in the dismemberment of the Fiji Platform, Mar. Geol. 116:57-68.
Jean-Baptiste, P., Charlou, J. L., Stievenard, M., Donval, J. P., Bougault, H., and Mevel, C. 1991. Helium and
methane measurements in hydrothermal fluids from the mid-Atlantic Ridge: The Snake Pit site at 23°N,
Jenner, G. A., Cawood, P. A., Rautenschlein, M., and White, W. M. 1987. Composition of back-arc basin
volcanics, Valu Fa Ridge, Lau Basin: Evidence for slab-derived component in their mantle source, J.
Johnson, L. E., Fryer, P., Masuda, H., Ishii, T., Garno, T. 1993. Hydrothermal vent deposits and two magam sources
for volcanoes near l3°20'N in the Mariana backarc: A view from Shinkai 6500, EOS, Trans. AGU 74:681
(abstract) .
Kanehira, K., and Tatsumi, T. 1970. Bedded cupriferous iron sulphide deposits in Japan, a review, in Volcanism
and Ore Genesis, pp. 51-76, Tokyo Press, Tokyo.
Karig, D. E. 1985. Kinematics and mechanism of deformation across some accreting forearcs, in Formation of
Active Ocean Margins (N. Nasu et aI., eds.), pp. 155-177, Terra Scientific, Tokyo.
Karpov, G. A., and Naboko, S. I. 1990. Metal contents of recent thermal waters, mineral precipitates and
hydrothermal alteration in active geothermal fields, Kamchatka, J. Geochem. Explor. 36:57-71.
Kastner, M., Craig, H., and Sturz, A. 1987. Hydrothermal deposition in the Mariana Trough; Preliminary
mineralogical investigations, EOS, Trans. AGU 68:1531 (abstract).
Kasuga, S., and Kato, Y. 1992. Discovery of hydrothermal ore deposits in the crater of the Suiyo Sm!. on the Izu-
Ogasawara arc, in Proc. JAMSTEC Symp. Deep Sea Res. 8:249-255. (in Japanese with English abstract)
Kimura, M., Kaneoka, I., Kato, Y., Yamamoto, S., Kushiro, I., Tokuyama, H., Kinoshita, H., Isezaki, N., Masaki,
H., Ohsida, A., Uyeda, S., and Hilde, T. W. C. 1986. Report on DELP 1984 cruise in the middle Okinawa
Trough. 5: Topography and geology of the central grabens and their vicinity, Bull. Earthquake Res. Inst. Univ.
Tokyo 61:269-310.
Kimura, M., Tanaka, T., Kyo, M., Ando, M., Oomori, T., Izawa, E., and Yoshikawa, 1.1989. Study of topography,
hydrothermal deposits and animal colonies in the middle Okinawa Trough hydrothermal areas using the
submersible "SHINKAI2000" system, in Proc. JAMSTEC Symp. Deep Sea Res. 5:223-244. (in Japanese
Kimura, M., Uyeda, S., Kato, Y., Yarnano, M., Garno, T., Sakai, H., Kato, S., Izawa, E., and Oomori, T. 1988.
Active hydrothermal mounds in the Okinawa Trough backarc basin, Japan, Tectonophysics 145:319-324.
Kinoshita, M. 1990. Heat flow anomaly in some western Pacific trench-arc-backarc systems associated with
interstitial water circulation, Ph.D. thesis, Univ. Tokyo, 74p.
Kroenke, L. W. 1984. Cenozoic tectonic development of the Southwest Pacific, U.N. ESCAP, CCOP/SOPAC
Tech. Bull., vol. 6.
Kusakabe, M., Mayeda, S., and Nakamura, E. 1990. S, 0 and Sr isotope systematics of active vent materials from
the Mariana backarc basin spreading axis at 18°N, Earth Planet. Sci. Lett. 100:275-282.
Lafoy, Y., Auzende, J.-M., Ruellan, E., Huchon, P. and Honza, E. 1991. The 16°40'S triple junction in the North
Fiji Basin (SW Pacific), Mar. Geopys. Res. 12:285-296.
Leinen, M., McDuff, R., Delaney, J., Becker, K., and Schultheiss, P. 1987. Off-axis hydrothermal activity in the
Mariana mound field, EOS. Trans. AGU 68:1531 (abstract).
Lisitsyn, A. P.• Binns, R. A., Boganov, Yu. A .• Scott, S., Zonenshayn, L. P., Gordeyev. V. V.• Gurvich, Yeo G.,
Murav'yev, K. G., and Serova, V. V. 1991. Active hydrothermal activity at Franklin seamount, western
Woodlark Sea (Papua New Guinea), Int. Geol. Rev. 33:914-929.
Lisitsyn, A. P., Crook, K. A. w., Bogdanov, Yu. A., Zonenshayn, L. P., Murav'yev, K. G., Tufar, w., Gurvich, Yeo
G., Gordeyev, V. V.. and Ivanov, G. V. 1993. A hydrothermal field in the rift zone of the Manus Basin,
Bismark Sea, Inter. Geol. Rev. 35:05-126.
Lisitsyn, A. P., Malahoff. A. R., Bogdanov, Yu. A., Sione Soakai, Zonenshayn, L. P., Gurvich, Yeo G., Murav'yev,
K. G., and Ivanov, G. V. 1992. Hydrothermal formations in the northern part of the Lau Basin, Pacific Ocean,
Inter. Geol. Rev. 34:828-847.
Lupton, J. E., Klinkhammer, G. P.• Normark, W. R., Haymon, R., MacDonald, K. C .• Weiss, R. E, and Craig, H.
1980. Helium-3 and manganese at the 21°N East Pacific Rise hydrothermal site, Earth Planet. Sci. Lett.
50:115-127.
Massoth, G. 1., Butterfield, D. A., Lupton, 1. E., McDuff, R. E., Lilley, M. D., and Jonasson, I. R. 1989. Submarine
venting of phase-separated hydrothermal fluids at Axial volcano, Juan de Fuca Ridge, Nature 340:702-705.
Masuda, H., Garno, T., Freyer, P., Ishii, T., Jonson, L.E., Tanaka, H., Tsunogai. U .• Matsumoto, S., Masumoto, S.,
and Fujioka, K. 1993. The major element chemistry of submarine volcanic rocks from the Southern Mariana
Trough and its relation to the topography, in Proc. JAMSTEC Symp. Deep Sea Res. 9:181-189. (in Japanese
Masuda, H .• Ishibashi, J., Kato, Y., Garno. T., and Sakai, H. 1987. Oxygen isotope ratio and trace element
composition of hydrothermal sediments from Okinawa Trough, collected with SHINKAI2000, dive 231, in
Proc. JAMSTEC Symp. Deep Sea Res. 3:225-231. (in Japanese with English abstract)
McBirney, A. R. 1984. Igneous Petrology, Freeman & Cooper, San Francisco.
McMurtry, G. M., Sedwick, P., Fryer, P.• Yonder Haar, D. L., and Yeh, H. W. 1993. Unusual geochemistry of
hydrothermal vents in submarine arc volcanoes: Kasuga seamounts, northern Mariana are, Earth Planet. Sci.
Lett. 114:517-528.
Merlivat, L., Pineau, E, and Javoy, M. 1987. Hydrothermal vent waters at 13°N on the East Pacific Rise: Isotopic
composition and gas concentration, Earth Planet. Sci. Lett. 84:100-108.
Michard, A., Michard. G., Stiiben, D., Stoffers, P., Cheminee, J.- L., and Binard, N. 1993. Submarine thermal
springs associated with young volcanoes: The Teahitia vents, Society Islands, Pacific Ocean. Geochim.
Cosmochim. Acta 57:4977-4986.
Miller, A. R., Densmore, C. D .• Degens, E. T., Pocklington, R., and Jokela, A. 1966. Hot brines and recent iron
deposits in deeps of the Red Sea, Geochim. Cosmochim. Acta 30:341-359.
Mitchell, A. H. G., and Bell, J. D. 1973. Island-arc evolution and related mineral deposits, J. Geol. 81:381-405.
Mitsuzawa, K., Momma. H., Hotta, H., and Deep Sea Research Group. 1989. Measurements of water temperature
and current around the submarine caldera in the KC Kaikata seamount, in Proc. JAMSTEC. Symp. Deep Sea
Momma, H., Hashimoto, J., Tanaka, T., and Deep Sea Research Group. 1989. Preliminary report of deep tow
surveys in the Okinawa Trough, DK88-2-0KN-LEGl,2, JAMSTEC Technical Report, Vol. 21, pp. 203-221.
(in Japanese with English abstract)
Moorby, S. A., Cronan. D. S., and Glasby, G. P. 1984. Geochemistry of hydrothermal Mn-oxide deposits from the
SW Pacific island arc, Geochim. Cosmochim. Acta 48:433-441.
Moore, W. S., and Stakes, D. 1990. Ages of barite-sulfide chimneys from the Mariana Trough, Earth Planet. Sci.
Lett. 100:265-274.
Mottl, M. 1. 1992. Pore waters from serpentinite seamounts in the Mariana and Izu-Bonin forearcs, Leg 125:
Evidence for volatiles from the subducting slab, in Proc. OD?, Sci. Results, 125 (P. Fryer, J. A. Pearce, L. B.
Stokking et al., eds.), pp. 373-385, Ocean Drilling Program, College Station, TX.
Nagaoka, N., Kasuga, S., and Kato, Y. 1992. Geology of Mokuyo smt. Doyo smt., and Suiyo smt. in the Sichiyo
seamounts on the Ogasawara arc, in Proc. JAMSTEC Symp. Deep Sea Res. 8:237-248. (in Japanese with
English abstract)
Nagaoka, N., Okino, K., and Kato, S. 1991. Landfonns of submarine volcanoes in central part of the Izu-
Ogasawara arc, by multi-beam sounding system, Rep. Hydrogr. Res. 27:145-172. (in Japanese with English
abstract)
Naka, J., and Deep Sea Research Group. 1989a. Sea bottm observation around the KC peak of the Kaikata sea-
mount, Bonin Islands, in Proc. JAMSTEC Symp. Deep Sea Res. 5:57-65. (in Japanese with English abstract)
Naka, 1., and Deep Sea Research Group. 1989b. Volcanic products of the Iheya Ridge, central Okinawa Trough, in
Nakamura, K., Kato, Y., Kimura, M., Ando, M., and Kyo, M. 1989. Occurrence and distribution of the hydrother-
mal ore deposits at the Izena Hole in the Okinawa Trough-Summary of the knowledge in 1988, in Proc.
JAMSTEC Symp. Deep Sea Res., pp. 183-189. (in Japanese with English abstract)
Nakamura, K., Marumo, K., and Aoki, M. 1990. Discovery of a black smoker vent and a pockmark emitting CO2 -
rich fluid on the seafloor hydrothennal mineralization field at the Izena cauldron in the Okinawa Trough, in
Nakao, S., Yuasa, M., Nohara, M., and Vsui, A. 1990. Submarine hydrotheraml activity in the Izu-Ogasawara arc,
western Pacific, Rev. Aqua. Sci. 3:95-115.
Nakashima, K., Sakai, H., Yoshida, H., Chiba, H., Tanaka, Y., Gamo, T., Ishibashi, J., and Tsunogai, V. 1993.
Mineralogical and fluid inclusion studies on some hydrothennal deposits at the Iheya Ridge and Minami-
Ensei Knoll, Okinawa Trough: Report of research dive 621 and hydrothennal precipitates collected by dives
487 and 622 of "Shinkai2000," in Proc. JAMSTEC Symp. Deep Sea Res. 9:253-269. (in Japanese with
English abstract)
Nedachi, M., Veno, H., Oki, K., Shiga, Y., Hayasaka, S., Ossaka, 1., Nogami, K., Ito, N., and Hashimoto, J. 1991.
Sulfide veinlets and the surrounding marine sediments in the fumarole area in the Wakamiko caldera,
northern Kagoshima Bay, in Proc. JAMSTEC Symp. Deep Sea Res. 7:235-243. (in Japanese with English
abstract)
Nedachi, M., Veno, H., Ossaka, J., Nogami, K., Hashimoto, J., Fujikura, K., and Miura, T. 1992. Hydrothennal ore
deposits on the Minami-Ensei Knoll of the Okinawa Trough-Mineral Assemblages, in Proc. JAMSTEC
Symp. Deep Sea Res. 8:95-106. (in Japanese with English abstract)
Nishimura, A., Mita, N., and Nohara, M. 1992b. Pelagic and hemipelagic sediments of the Izu-Bonin region, Leg
126: Geochemical and compositional features, in Proc. ODp, Sci. Results, 126 (B. Taylor, K. Fujioka et aI.,
Nishimura, A., and Murakami, F. 1988. Sedimentation of the Sumisu rift, Izu-Ogasawara arc, Bull. Geol. Surv.
Jpn. 39:39-61.
Nishimura, A., Rodolfo, K. S., Koizumi, A., Gill, J., and Fujioka, K. 1992a. Episodic deposition of Pliocene-
Pleistocene pumice from the Izu-Bonin arc, Leg 126, in Proc. ODp, Sci. Results, 126 (B. Taylor, K. Fujioka
et aI., eds.), pp. 3-21, Ocean Drilling Program, College Station, TX.
Nohara, M., Hirose, K., Eissen, J.-P., Vrabe, T., and Joshima, M. 1994. The North Fiji Basin basalts and their
magma sources. Part 2: Sr-Nd isotopic and trace element constraints, Mar. Geol. 116:179-195.
Nojiri, Y., Ishibashi, J., Kawai, T., Otsuki, A., and Sakai, H. 1989. Hydrothennal plumes along the North Fiji Basin
spreading axis, Nature 342:667-670.
Ohmoto, H., Tanimura, S., Date, J., and Takahashi, T. 1983. Geologic setting of the Kuroko deposits, Japan, Econ.
Geol. Monogr. 5:9-54.
Ohta, S. 1990. Deep-sea submersible survey of the hydrothennal vent community on the northeastern slope of the
Iheya Ridge, the Okinawa Trough, in Proc. JAMSTEC Symp. Deep Sea Res. 6:145-156. (in Japanese with
English abstract)
Okutani, T., and Ohta, S. 1988. A new gastropod mollusk assciated with hydrothennal vents in the Mariana back-
arc basin, western Pacific, Venus (Jpn. J. Malacology) 47:1-9.
Palmer, M. R. 1991. Boron isotope systematics of hydrothenn fluids and tounnalines: a synthesis, Chem. Geol.
(Isotope Geosci. Sect.) 94:111-121.
Palmer, M. R., and Edmond, J. M. 1989. The strontium isotope budget of the modem ocean, Earth Planet. Sci. Lett.
92:11-26.
Pisutha-Arnond, v., and Ohmoto, H. 1983. Thennal history, and chemical and isotopic compositions of the ore-
forming fluids responsible for the Kuroko massive sulfide deposits in the Hokuroku district of Japan, Econ.
Geol. Monogr. 5:523-558.
Puteanus, D., Glasby, G. P., Stoffers, P., and Kuznendorf, H. 1991. Hydrothermal iron-rich deposits from the
Teahitia-Mehetia and Macdonald hot spot areas, Southwest Pacific, Mar. Geol., 98:389-409.
Rona, P. A. 1983. Exploration for hydrothermal mineral deposits at seafloor spreading centers, Mar. Mining 4:7-38.
Rona, P. A. 1984. Hydrothermal mineralization at seafloor spreading centers, Earth Sci. Review 20:1-104.
Rona, P. A. 1988. Hydrothermal mineralization at oceanic ridges, Can. Mineral. 26:431-465.
Rona, P A., and Scott, S. D. 1993. Preface for a special issue on sea-floor hydrothermal mineralization: New
perspectives, Econ. Geol. 88:1935-1976.
Sakai, H., Garno, T., Kim, E.-S., Tsutsumi, M., Tanaka, T., Ishibashi, J., Wakita, H., Yamano, M., and Oomori, T.
1990b. Venting of carbon dioxide-rich fluid and hydrate formation in mid-Okinawa Trough backarc basin,
Science 248:1093-1096.
Sakai, H., Garno, T., Ishibashi, J., Shitashima, K., Kim., E. S., Yanagisawa, F., Tsutsumi, M., Sano, Y., Wakita, H.,
Tanaka, T., Matsumoto, T., Naganuma, T., and Mitsuzawa, K. 1990a. Unique chemistry of the hydrothermal
solutions in the mid-Okinawa Trough backarc basin, Geophys. Res. Lett. 17:2133-2136.
Sawkins, F. J. 1972. Sulfide ore deposits in relation to plate tectonics, J. Geol. 80:377-397.
Sawkins, F. J. 1976. Massive sulphide deposits in relation to geotectonics, Geol. Assoc. Can. Special Paper
14:221-240.
Scott, S. D. 1985. Seafloor polymetallic sulfide deposits: Modern and ancient, Mar. Mining 5:191-212.
Scott, S. D. 1987. Seafloor polymetallic sulfides: Scientific curiosities or mines of the future?, in Marine Minerals
(P. G. Teleki et aI., eds.), NATO ASI Series C, Vol. 194, pp. 277-300.
Scott, S.D., and Binns, R. A. 1992. An actively-forming, flesic volcanic-hosted polymetallic sulfide deposit in the
southeast Manus back-arc basin of Papua New Guinea, EOS. Trans. AGU 73:626 (Abstract).
Sedwick, P. N., Garno, T., and McMurtry, G. M. 1990. Manganese and methane anomalies in the North Fiji Basin,
Deep-Sea Res. 37:891-896.
Sedwick, P. N., McMurtry, G. M., Hilton, D. R., and Goff, F. 1994. Carbon dioxide and helium in hydrotbrmal
fluids from Loihi seamount, Hawaii, USA: Temporal variability and implications for the release of mantle
volatiles, Geochim. Cosmochim. Acta 58:1219-1227.
Sedwick, P. N., McMurtry, G. M., and Macdougall, 1. D. 1992. Chemistry of hydrothermal solutions from Pele's
vent, Loihi seamount, Hawaii, Geochim. Cosmochim. Acta 56:3643-3667.
Seyfried, W.E. Jr. 1987. Experimental and theoretical constraints on hydrothermal alteration processes at mid-
ocean ridges, Ann. Rev. Earth Planet. Sci. 15:317-335.
Seyfried, W. E. Jr., Janeckey, D. R., and Mottl, M. 1. 1984. Alteration of the oceanic crust: Implications for
geochemical cycles of lithium and boron, Geochim. Cosmochim. Acta 48:557-569.
Shadlum, T. N., Bortnikov, N. S., Bogdanov, Yu. A., Tufar, W., Murav'yev, K. G., Gurvich, Yeo G., Muravitskaya,
G. N., Korina, Yeo A., and Topa, T. 1993. Mineralogy, textures, and formation conditions of modern sulfide
ores, Manus Basin rift zone, Int. Geol. Rev. 35(2):127-145.
Shitashima, K. 1994. Distribution and behavior of trace metals on the Okinawa Trough, in JAMSTEC Journal of
Deep Sea Res. 10, pp. 291-298. (in Japanese with English abstract)
Sibuet, J. c., Letouzey, 1., Barbier, F., Charvet, J., Foucher, J. P., Hilde, T. W. C., Kimura, M., Chiao, L. Y.,
Marsset, B., Muller, C., and Stephan, J. F. 1987. Back arc extension in the Okinawa Trough, J. Geophys. Res.
92:14,041-14,063.
Sillitoe, R. H. 1972. Formation of certain massive sulphide deposits at sites of sea-floor spreading, Inst. Mining
Metall. Trans. 81:B141-148.
Spiess, F. N., and RISE Group. 1980. East Pacific Rise; hot springs and geophysical experiments, Science
297:1421-1433.
Stolper, E. M., and Newman, S. 1992. Fluids in the source regions of subduction zone magmas: Clue from the
study of volatiles in Mariana Trough magmas, in Magmatic Contributions to Hydrothermal Systems (1. W.
Hedenquist, ed.), pp. 161-169, Geol. Surv. Jpn Rep. No. 279.
Stiiben, D., Bloomer, S. H., Taibi, N. E., Neumann, Th., Bendel, v., Piischel, U., Barone, A., Lange, A., Shiying,
W., Cuizhong, L., and Deyu, Z. 1992. First results of study of sulphur-rich hydrothermal activity from an
island-arc environment: Esmeralda Bank in the Mariana arc, Mar. Geol. 103:521-528.
Stiiben, D., Taibi, N. S., McMurtry, G. M., Scholten, J., Stoffers, P., and Zhang, D. 1994. Growth history of a
hydrothermal silica chimney from the Mariana backarc spreading center (southeast Pacific, 18°13 'N), Mar.
Geol. 113:273-296.
Taira, A., and Pickering, K. T. 1991. Sediment deformation and fluid activity in the Nankai, Izu-Bonin and Japan
forearc slopes and trenches, in The Behaviour and Influence of Fluids in Subduction Zones (J. Tamey, K. T.
Pickering, R. J. Knipe, and J. F. Gewey, eds.), pp. 63-88, The Royal Society, London.
Tamaki, K., and Honza, E. 1992. Global tectonics and formation of marginal basins: Role of the western Pacific,
Episodes 14:224-230.
Tanahashi, M., Kisimoto, K., Joshima, Jarvis. P., Iwabuchi, Y., Ruellann, E., and Auzende, J. M. 1994. 800 km
long N-S spreading system of the North Fiji Basin, Mar. Geol. ll6:5-24.
Tanahashi, M., Kisimoto, K., Joshima, M., Lafoy, Y., Honza, E., and Auzende, J. M. 1991. Geological structure of
the central spreading system, North Fiji Basin, Mar. Geol. 98:187-200.
Tanaka, T., Hotta, H., Sakai, H., Ishibashi, J., Oomori, T., Izawa, E., and Oda, N. 1990. Occurrence and distribution
of hydrothermal deposits in the Izena Hole, central Okinawa Trough, in Proc. JAMSTEC Symp. Deep Sea
Tanaka, T., Mitsuzawa, K., and Hotta, H. 1989. "SHINKAl2000" diving surveys in the east of Iheya Small ridge
in the central Okinawa Trough, in Proc. JAMSTEC Symp. Deep Sea Res. 5:267-281. (in Japanese with
English abstract)
Taylor, B. 1979. Bismark Sea; evolution of a back arc basin, Geology 7:171.
Taylor, B., Brown, G., Fryer, P., Gill, J., Hochstaedter, A., Hotta, H., Langmuir, c., Leinen, M., Nishimura, A., and
Urabe, T. 1990. ALVIN-SeaBeam studies of the Sumisu rift, Izu-Bonin arc, Earth Planet. Sci. Lett. 100:
127-147.
Taylor, B., and Kamer, G. D.1983. On the evolution of marginal basins, Rev. Geophys. Space Phys. 21:1727-1741.
Tsunogai, U., Ishibashi, J., Wakita, H., Garno, T., Watanabe, K., Kajimura, T., Kanayama, S., and Sakai, H. 1994.
Peculiar features of Suiyo seamount hydrothermal fluids, Izu-Bonin arc: Differences from subaerial volca-
nism, Earth Planet. Sci. Lett. 126:289-301.
Tsunogai, U., Ishibashi, J., Wakita, H., Watanabe, K., Kazimura, T., and Hashimoto, J. 1993. Dissolved gas
geochemistry of submarine hydrothermal fluids from the Suiyo and the Mokuyo seamount, Izu-Bonin,
Ogasawara arc, in Proc. JAMSTEC Symp. Deep Sea Res. 9:92-103. (in Japanese with English abstract)
Tufar, W. 1990. Modern hydrothermal activity, formation of complex massive sulfide deposits and associated vent
communities in the Manus back-arc basin, Bismark Sea (Papua New Guinea), Mitt. asterr. geol. Ges.
82:183-210.
Urabe, T. 1987. Kuroko deposit modeling based on magmatic hydrothermal theory, Mining Geol. 37:159-176.
Urabe, T. 1989. Mineralogical characteristics of the hydrothermal ore body at Izena Hole No.1 in comparison with
Kuroko deposits, in Proc. JAMSTEC Symp. Deep Sea Res. 5:191-196. (in Japanese witlt"English abstract)
Urabe, T., and Kusakabe, M. 1990. Barite silica chimneys from the Sumisu rift, lzu-Bonin arc: Possible analog
tohematitic chert associated with Kuroko deposits, Earth Planet. Sci. Lett. 100:283-290.
Urabe, T., and Marumo, K. 1991. A new model for Kuroko-type deposits of Japan, Episodes 14:246-251.
Urabe, T., Yuasa, M., Nakao, S., and On-board Scientists. 1987. Hydrothermal sulfides from a submarine caldera
in the Shichito-Iwo Jima Ridge, northwestern Pacific, Mar. Geol. 74:295-299.
Usui, A., Mellin, T.A., Nohara, M., and Yuasa, M. 1989. Structural stability of marine lOA manganates from the
Ogasawara, Bonin, arc: Implication for low-temperature hydrothermal activity, Mar. Geol. 86:41-56.
Usui, A., and Nishimura, A. 1992. Submersible observations of hydrothermal manganese deposits on the Kaikata
seamount, lzu-Ogasawara, Bonin arc, Mar. Geol. 106:203-216.
Usui, A., Yuasa, M., Yokota, S., Nohara, M., Nishimura, A., and Murakami, F. 1986. Submarine hydrothermal
manganese deposits from the Ogasawara (Bonin) arc, off the Japan islands, Mar. Geol. 73:311-322.
Von Damm, K. L. 1990. Seafloor hydrothermal activity: Black smoker chemistry and chimneys, Ann. Rev. Earth
Planet. Sci. 18:173-204.
Von Damm, K. L., and Bischoff, J. L. 1987. Chemistry of hydrothermal solutions from the southern Juan de Fuca
Ridge, J. Geophys. Res. 92:11,334-11,346.
Von Damm, K. L., Bischoff, J. L., and Rosenbauer, R. J. 1991. Quartz solubility in hydrothermal seawater: an
experimental study and equation describing quartz solubility for up to 0.5 M NaCI solutions, Am. J. Sci.
291:977-1007.
Von Damm, K. L., Edmond, J. M., Grant, B., Measures, C. I., Walden, B., and Weiss, R. F. 1985a. Chemistry of
submarine hydrothermal solutions at 21°N, East Pacific Rise, Geochim. Cosmochim. Acta 49:2197-2220.
Von Damm, K. L., Edmond, J. M., Measures, C. I., and Grant, B. 1985b. Chemistry of submarine hydrothermal
solutions at Guaymas Basin, Gulf of California, Geochim. Cosmochim. Acta 49:2221-2237.
von Stackelberg, U., Marchig, v., Muller, P., and Weiser, T. 1990. Hydrothermal mineralization in the North Fiji
basins, Geologisch. Jahrbuch D92:547-613.
von Stackelberg. U .• and Shipboard Scientific Party. 1985. Hydrothermal sulfide deposits in back-arc spreading
centers in the Southwest Pacific. BGR eirc .• Vol. 2. pp. 3-14. Bundesanst. Geowiss. Rohstoff.• Han-
nover. FRG.
von Stackelberg. U .• and Shipboard Scientific Party. 1988. Active hydrothermalism in the Lau back-arc basin (SW
Pacific): First results of the SONNE48 Cruise. 1987. Mar. Mining 7:431-442.
Watanabe. K.. and Kajimura. T. 1993. Topography. geology and hydrothermal deposits at Suiyo seamount. in
Watanabe. K.. and Kajimura. T. 1994. The hydrothermal mineralization at Suiyo seamount. in the Izu-Ogasawara
arc. Resource Geol. 44:133-140. (in Japanese with English abstract)
Watanabe. K .• Shibata. A.. Kajimura. T.. Ishibashi. J .• Tsunogai. U .• Aoki. M .• and Nakamura. K. 1994. Survey
method about the submarine volcano and its sea-floor hydrothermal ore deposit-A example of Suiyo
seamount in the Izu-Ogasawara arc with the submersible "Shinkai2000." J. Jpn. Soc. Mar. Surv. Tech. 6:
29-44. (in Japanese with English abstract)
Welhan. J. A .• and Craig. H. 1983. Methane. hydrogen and helium in hydrothermal fluids at 21°N on the East
Pacific Rise. in Hydrothermal Processes at Seafloor Spreading Venters (P. A. Rona. K. Bostrom. L. Laubier.
and K. L. Smith. eds.). pp. 391-409. Plenum Press. New York.
Welhan. J. A.. and Lupton. J. E. 1987. Light hydrocarbon gases in Guaymas Basin hydrothermal fluids: Thermo-
genic versus abiogenic origin. Am. Assoc. Petrol. Geol. Bull. 71:215-223.
Wheat. C. G .• and McDuff. R. E. 1987. Advection of pore waters in the Marianas mounds hydrothermal region as
determined from nutrient profiles. EOS. Trans. AGU 68:1531 (abstract).
Wheat. C. G .• and McDuff. R. E. 1994. Hydrothermal flow through the Mariana Mounds: Dissolution of
amorphous silica and degradation of organic matter on a mid-ocean ridge flank. Geochim. Cosmochim. Acta
58:2461-2475.
Wheller. G .• Binns. R .• and Scott. S. 1992. The PACMANUSIPACLARK program: Search for modem "Kuroko-
type" analogues in the SW Pacific. InterRidge News 1:7-9.
Yamano. M .• Uyeda. S .• Foucher. J. P.• and Sibuet. 1. C. 1989. Heat flow anomaly in the middle Okinawa Trough.
Yamano. M .• Uyeda. S .• Kinoshita. H .• and Hilde. T. W. C. 1986. Report on DELP1984 cruises in the middle
Okinawa Trough. 4. Heat flow measurements. Bull. Earthquake Res. lnst. Univ. Tokyo 61:251-267.
Yamazaki. T. 1988. Heat flow in the Sumisu rift. IZU-Ogasawara (Bonin) arc. Bull. Geol. Surv. Jpn. 39:63-70.
You. c.-F.. Butterfield. D. A.. Spivack. A. J .• Gieskes. J. M .• Gamo. T.. and Campbell. A. 1. 1994. Boron and halide
systematics in submarine hydrothermal systems: Effects of phase separation and sedimentary contributions.
Earth Planet. Sci. Lett. U3:227-238.
Yuasa. M. 1985. Sofugan tectonic line. a new tectonic bounary separating northern and southern parts of the
Ogasawara. Bonin arc. northwest Pacific. in Formation of Active Ocean Margins (N. Nasu et al.. eds.). pp.
483-496. Terra Scientific. Tokyo.
Yuasa. M .• Murakami. F.. Saito. E .• and Watanabe. K. 1991. Submarine topography of seamounts on the volcanic
front of the Izu-Ogasawara. Bonin arc. Bull. Geol. Surv. Jpn 42:703-743.
Yuasa. M .• Urabe. T.. and Murakami. F. 1988. A submersible study of hydrothermal fields at the Kaikata seamount.
Izu-Ogasawara arc. in Proc. JAMSTEC Symp. Deep Sea Res. 4:129-139. (in Japanese with English abstract)
14
Tectonic and Magmatic Controls

on Backarc Basin Sedimentation
The Mariana Region Reexamined
Kathleen M. Marsaglia and Kathleen A. Devaney
ABSTRACT
Previous models for intraoceanic backarc basin sedimentation are static in that they focus
on a long-lived, stationary, magmatic source along the arc axis. These models are adequate
in the case of backarc rifting, but a more dynamic model is needed in the case of forearc
rifting. We present a revised model for backarc basin sedimentation that reflects progressive
shifts in volcanism and associated depocenters through the history of a backarc basin
developed by forearc rifting. This model is based on recent Ocean Drilling Program results
from the Lau Basin and a reexamination of the sedimentary sequences recovered on Deep
Sea Drilling Project legs across the Mariana arc and backarc system. We report new
petrographic data for volcaniclastic sand from the Mariana forearc and backarc regions and
compare them to data collected from other backarc basins. The proportions of colorless,
brown and black glassy volcanic fragments within coarse, sandy intervals associated with
these arc systems suggest that the early rift phase of backarc basin formation is charac-
terized by felsic (bimodal) volcanism and that volcanic centers become progressively more
intermediate with time. This trend is consistent with the geochemical evolution of arc lavas.
In the case of forearc rifting, this compositional trend may coincide with a shift in
volcanism from the protoremnant arc, across the nascent backarc basin, toward the fron-
tal arc.
1. INTRODUCTION
The depositional history of backarc basins is closely tied to regional tectonics and
volcanism; hence it is appropriate that they be discussed in this volume. Proposed models
for the development of, and sediment distribution within, backarc basins have primarily
Kathleen M. Marsaglia and Kathleen A. Devaney • Department of Geological Sciences, University of

Texas at El Paso, El Paso, Texas 79968.
497
498 KATHLEEN M. MARSAGLIA AND KATHLEEN A. DEVANEY
,-------~~~--~~~~--~o
1
2
3
4
New Hebrides
,------------------,~~~__,o
1
2
3
4
5
Mariana Basin E
6
.s::.
C.
(\)
,---------------fo~=ro Cl
1
2
3
>. 150 4
E 100 5
E 50 6
Northern Lau Basin
.-----------------------~o
1
2
3
4
,;. 8 5
E 6 6
'- E_~ ;~- =:-=:- .".-=:- =- ~- =-~:. .,. .,-= :-:-=-= :"-~"'-:- =- =-~- =: .-=~.:-:;,: -=:-~: ~-~:;-:-=:~ ~-~"-'~" '.:._ - ' ~
a Parece Vela Basin
FIGURE 14.1. Models of backarc basin evolution and sedimentation from (a) Karig and Moore (1975) and
(b) Carey and Sigurdsson (1984). (a) The distribution, thickness and type of surficial sediments are indicated on a
series of four idealized cross sections based on various backarc basins. Sedimentation rates (m1m.y.) and
distribution of surficial sediment types are illustrated below each cross section. Symbols are as follows: dashes =
clay, barbed dashes = nannofossil ooze, dots = clastics, and triangles = coarse-grained volcaniclastic material.
(b) Facies distribution during various stages of backarc basin formation: early rifting (Stage 1), backarc spreading
(Stage 2), basin maturity (Stage 3), and cessation of backarc spreading and initiation of a new cycle of backarc
rifting (Stage 4). Bold curved arrows indicate extension due to rifting or seafloor spreading.
stemmed from Deep Sea Drilling Project (DSDP) drilling results (Karig and Moore, 1975;
Klein, 1975a,b, 1985a,b; Carey and Sigurdsson, 1984). The most detailed volcano-tectonic
model, by Carey and Sigurdsson (1984; Fig. 14.1) is largely based on drilling results from
DSDP Legs 59 and 60; during these legs, a series of sites was drilled on a transect across the
Parece Vela Basin and Mariana Trough. In this chapter we present new petrographic data on
the proportions of varieties of volcaniclastic lithic fragments in samples from the Mariana
backarc and forearc and use these data to help unravel the history of backarc basin
evolution. We then reevaluate the history of sedimentation across the Mariana "model"
region in light of recent Ocean Drilling Program (ODP) drilling results in other backarc
basins and propose a revised model for sedimentation in backarc basins formed by forearc
rifting.
MARIANA REGION REEXAMINED 499
VOLCANIC ARC
STAGE I
SL
STAGE 2
STAGE 3
SL
I
HYDROTHERMAL VOLCANICLASTIC
DEPOSIT APRON
~ PELAGIC FACIES
,\t
STAGE 4
2. CHARACTERISTICS OF AND CONTROLS ON INTRAOCEANIC BACKARC

BASIN SEDIMENTATION
The following summary of intraoceanic backarc basin sedimentation is largely based

on previous syntheses (Karig and Moore, 1975; Carey and Sigurdsson, 1984; Klein 1985a,b;
Marsaglia, in press). Additional information comes from recent drilling in the western
Pacific during ODP Legs 124 (Sulu Sea; Rangin et ai., 1990; Silver and Rangin, 1991), 126
(lzu-Bonin arc; see Klaus et ai., 1992; Taylor, 1992), and 135 (Lau Basin; see Hawkins,
Chapter 3 this volume; Parson et ai., 1992).
Sedimentary lithofacies in intraoceanic backarc basins are quite variable and include
pyroclastic and volcaniclastic deposits, pelagic and resedimented carbonates, biogenic
siliceous sediments, sandy to silty turbidites, and debris flows (Klein, 1985a,b). The
distribution of these lithofacies is partly a function of latitude (carbonate vs. silica produc-
tivity), ocean current circulation, and proximity to subaerial land sources of hemipelagic
clay and turbidites (Klein, 1985a). Eustacy may also affect sediment supply into backarc
basins (e.g., Shipboard Scientific Party, 1990; Betzler et ai., 1991).
Tectonism and arc magmatism play the most important roles in intraoceanic backarc
basin sedimentation. Tectonism dictates basin depth (subsidence) and hence the localiza-
tion of depocenters. In incipient backarc basins such as the Sumisu rift, subsidence is
localized between arc volcanoes, and basins have half- to full-graben configurations
(Taylor et aI., 1991; Klaus et aI., 1992). Rapid uplift of the Sumisu rift-flank margins
coincides with differential subsidence in the basin (Taylor et aI., 1990a; Taylor, 1992).
Subsidence also controls water depth, which directly affects the preservation of pelagic
carbonate facies (Klein, 1985a). The rate of tectonic uplift related to volcano building and
changes in subduction and plate boundary tectonics may also affect sediment input into
backarc basins. Klein (l985b) found that uplift rates of 400 mlm.y. or more along the arc
were required to support canyon-fed submarine fan development in the adjacent backarc
basin. In contrast, rapidly building volcanic edifices along a dominantly submerged arc are
characterized by the development of submarine volcanic aprons. Sedimentation on these
aprons is a function of multiple-sourced, coalesced depositional systems and so is charac-
terized by rare systematic facies transitions (Carey and Sigurdsson, 1984). The loose
pyroclastic materials that make up these aprons accumulate rapidly during periodic vol-
canic eruptions and are subject to downslope redistribution into backarc and interarc areas
as mass-flow deposits (Nishimura et aI., 1991).
Possible sources of volcaniclastic material in nascent backarc basins include the
protoremnant arc, intrarift volcanic centers and the active arc. During the early rift phase of
backarc basin formation intrarift volcanism may be concentrated along fault-controlled
transfer zones such as those documented in the Izu-Bonin arc system (Taylor et aI., 1990b).
The central ridge in the Sumisu rift is one such transfer zone; it is composed primarily of
basaltic lavas with lesser rhyolite and dacite (Fryer et aI., 1990; Hochstaedter et al., 1990a).
Volcanic vents are also present on the flanks of the volcanoes that compose the Sumisu rift
protoremnant arc, but their relative sedimentary contribution is unknown. Overall, the main
source of volcanic material in the Sumisu rift appears to be the active magmatic arc
(Marsaglia, 1992).
As a backarc basin matures, dike-fed intrabasinal volcanic ridges may develop into
seafloor spreading centers that supply minor amounts of pyroclastic material. In mature
backarc basins, the arc axis, rather than the inactive remnant arc, is the dominant source of
pyroclastic debris. Other minor sources of volcanic material include chains of small
volcanoes that develop on fractures extending from the arc into the backarc basin, such as
those found in the Mariana backarc basin (Hussong and Fryer, 1983; Jackson and Fryer,
1986; Bloomer et aI., 1989).
The nature and volume of volcaniclastic debris supplied to a backarc basin, therefore,
are controlled by the distribution, composition, eruption rate, and mode of eruption of
volcanoes along the arc axis (Fisher, 1984). This volcaniclastic material ranges in size from
fine ash and silt to coarse lapilli and gravel, and from basaltic to rhyolitic in composition.
Mafic eruptions likely produce minor fragmental material, whereas intermediate volcanoes
and felsic calderas may produce large volumes of pyroclastic debris. The pyroclastic debris
are variably vesicular and crystalline, depending on the eruption mode. For example,
submarine eruptions may produce thick pumice deposits, especially during the early phases
of rift development (Cashman and Fiske, 1991; Nishimura et aI., 1991). The localization of
coarse sandy material is also a function of water and atmospheric currents; Sigurdsson et al.
(1980) have shown that along the modem Lesser Antilles arc the fine ash is carried to the
east, whereas coarse clastic materials are preferentially carried to the west and deposited in
the backarc basin. The amount of epiclastic debris is a function of the degree of volcano
emergence, which is related to rates of tectonic uplift and volcanism (volcano size).
Backarc spreading has been tied to both minima and maxima in volcanism (Karig,
1975, 1983; Kroenke et aI., 1980; Scott and Kroenke, 1980; Taylor, 1992). According to
Taylor (1992), many arc segments (excluding the Mariana arc) show cyclic volcanism with
maxima before rifting, during early rifting, and middle to late backarc spreading, and show
minima during latest rifting and early backarc spreading. In addition, along-strike transport
of pyroclastic debris can camouflage local lulls in volcanism (Taylor, 1992). Temporal
variations are affected by the locus of rifting, forearc versus backarc; in the case of forearc
rifting, a temporal gap in volcanic activity may result from the shift in volcanism from the
arc toward the rifted frontal arc. The effects of forearc versus backarc rifting on the
localization of volcaniclastic debris are discussed below.
Karig and Moore (1975) constructed the first tectonic and volcanic model for backarc
basin sedimentation (Fig. 14.1). It is a simple model that shows the distribution of sediment
types and range of sedimentation rates in various backarc basins, from young (e.g., New
Hebrides), to mature (Mariana and Lau), to inactive (Parece Vela Basin) stages. In this
model, volcaniclastic sediments and higher sediment accumulation rates are localized near
the frontal arc. Carey and Sigurdsson (1984) modified the Karig and Moore (1975) tectonic
model to better reflect subsequent DSDP drilling results from the western Pacific (Legs 59
and 60) and their research on volcaniclastic sedimentation in the Grenada backarc basin.
This model shows more detailed facies distributions within a hypothetical, evolving
backarc basin (Fig. 14.1). These models indicate that proximal coarse-grain sedimentary
packages, as opposed to more distal facies, provide the most detailed information on the
tectonic and volcanic evolution of the arc during backarc basin formation.
3. MARIANA DRILLING RESULTS
During Legs 59 and 60, a series of sites were drilled along a transect across the West
Mariana Ridge, the Mariana Trough, and the Mariana arc. Significant coarse volcaniclastic
sections were recovered at a number of these sites, and the stratigraphic relationship and
depositional history of these sandy units are summarized below. Much of this information
has been taken from Kroenke et al. (1980) and Hussong et al. (1981).
Site 450 is located on the eastern side of the Parece Vela Basin, west of the West
Mariana Ridge (Fig. 14.2). Basaltic basement at this site is overlain by Middle to Upper
Miocene vitric tuffs and minor volcaniclastic conglomerates that pass upward in an
interbedded fashion into an Upper Miocene to Pleistocene pelagic clay sequence (Kroenke
et aI., 1980). The volcaniclastic units at Site 450 are thought to have been derived from the
West Mariana Ridge, prior to an eastward shift of volcanism during extension and forma-
tion of the Mariana Trough (Scott et aI., 1980). An upsection decrease in carbonate content
at this site is attributed to thermal subsidence through the CCD (Klein, 1985a).
Site 451 lies on the eastern edge of the Western Mariana Ridge (Fig. 14.2). The
stratigraphic sequence at this site consists of approximately 900 m of Upper Miocene vitric
tuff and andesitic to basaltic breccia and conglomerate, gradationally overlain by a thin (36
m) Plio-Pleistocene cover of calcareous ooze (Kroenke et aI., 1980). Geochemical analysis
of basalt clasts from this site indicates that they are evolved island-arc tholeiites with strong
calc-alkalic affinities (Mattey et aI., 1980; Wood et aI., 1980). The degree of clast rounding
and the presence of shallow-water bioclasts and lignite fragments suggest that the source of
the Upper Miocene section was an at-least-partly emergent volcanic arc along the West
Mariana Ridge (Kroenke et al., 1980). Klein (1985a) interpreted this sequence as volcani-
clastic apron deposits.
Site 453 was drilled just east of Site 451 in a sediment pond, in what is considered the
FIGURE 14.2. Location map of DSDP and ODP sites for which data are presented in this study (from Marsaglia,
1992).
oldest part of the Mariana Trough (Fig. 14.2). The stratigraphic sequence begins with
approximately 150 m of gabbro-metabasalt polymict breccia of unknown age derived from
the rifted roots of a calc-alkaline magmatic arc (Natland and Tarney, 1981). The breccia is
in tum overlain by Pliocene to Pleistocene volcaniclastic turbidites and pelagic to hemi-
pelagic muds. Klein (1985a) interpreted this volcaniclastic sequence as submarine fan
deposits. The age of the sediments and basement rocks recovered at this site constrain the
time of initial rifting in the Mariana Trough to approximately 5 Ma. The frequency and
grain size of sandy units decrease from the Lower to the Upper Pliocene sections; associ-
ated changes in sedimentary structures within these turbidites indicate a shift from a more
proximal to a more distal setting. Given the paucity of pyroclastic debris in equivalent
sections to the west at Site 451, the source of volcaniclastic material at Site 453 was either
the Mariana arc (Hussong et aI., 1981) or, perhaps, intrabasinal volcanoes. According to
Hussong et al. (1981), the progressive decrease in arc input upsection could be a function of
sequential westward displacement of the site with respect to the volcanic source due to
backarc spreading.
The presence of unconsolidated coarse volcaniclastic sediment at arc-proximal sites
(455 and 457; Fig. 14.2) resulted in poor recovery (30-37%), unstable hole conditions, and
early abandonment of the sites. Site 455 is located near the eastern margin of the Mariana
Trough on the sediment apron of the Mariana arc. The 104 m of sedimentary section
penetrated at this site consists of Pleistocene vitric mud and mudstone, vitric nannofossil
ooze and chalk, vitric ash and tuff, and unconsolidated volcanic gravel and sand. Site 457 is
located on the arc axis, near Alamagan Island. A 61-m sequence of coarse volcanic
sediment and volcanic rock (drilling breccia) was penetrated at this site; the volcaniclastic
units contain red scoria fragments suggesting a subaerial source (Alamagan Island).
Site 458 is located in the Mariana forearc region approximately 130 km east of the
active arc. At this site a 256-m, Oligocene to Pleistocene sedimentary section rests on
pillowed to massive flows of high-MgO bronzite andesite and basalt. The sedimentary
sequence is composed of Oligocene and Miocene chalks, tuffs, and vitric siltstones and
sandstones (turbidites), overlain by Neogene to Pleistocene siliceous and calcareous oozes
with variable ash content. Ash is present throughout the sequence, with maxima in the
Lower Oligocene and Upper Pliocene to Pleistocene sections. The Lower Oligocene to
Miocene volcanic input was likely derived from both the Palau-Kyushu and West Mariana
ridges prior to rifting, whereas the upper Pliocene to Pleistocene sections were likely
derived from the modem Mariana arc. A number of hiatuses occur in the section, the
longest from 7 to 3 Ma.
At Site 459, 31 km to the east of Site 458, a similar section was drilled. The
stratigraphy at this site begins with basalt, overlain by a thin sequence of Upper Eocene
claystone and chert, followed by a thick sequence of Oligocene to Miocene sandy to silty
tuffaceous turbidites, chalk and ooze and mudstone, and then a thin (50 m) Pliocene to
Pleistocene section of vitric and siliceous mud, ash, and calcareous ooze. As at Site 458,
this sequence is characterized by a series of hiatuses, the longest in duration from 10 to
3 Ma. The major Mio-Pliocene hiatuses at Sites 458 and 459 likely correspond to the advent
of rifting that formed the Mariana Trough.
From west to east, the volcaniclastic units cored during DSDP Legs 59 and 60 on the
Mariana transect record volcanic activity associated with the prerift to seafloor spreading
stages of Mariana Trough formation. Coarse volcaniclastic sediment packages at these sites
young from west to east, recording an eastward shift of volcanism associated with rifting
processes in the Mariana arc: Site 450 (Middle to Upper Miocene), Site 451 (Upper
Miocene), Site 453 (Pliocene), Site 455 (Pleistocene), and Site 457 (Pleistocene). The
forearc sections at Sites 458 and 459 provide discontinuous records, with significant coarse
volcaniclastic sediments restricted to the Miocene and Oligocene sections. In combination
with lava geochemistry, the volcaniclastic record at these sites should provide a means of
determining evolutionary trends in the composition and distribution of magmatism during
backarc basin formation.
4. PREVIOUS PETROGRAPHIC STUDIES OF BACKARC BASIN SAND
Packer and Ingersoll (1986) determined detrital modes for 60 volcaniclastic sand and
sandstone samples from DSDP Sites 450-459 (Fig. 14.2). They found sands from these
sites were relatively homogeneous in comparison to sands derived from the Japan arc and
did not document any trends along this transect. Marsaglia and Ingersoll (1992) combined
the Packer and Ingersoll (1986) data set with data from other circum-Pacific intraoceanic-
arc systems and found that active intraoceanic and remnant arcs have very similar composi-
tional ranges. Petrographic data for these systems tend to fall along the base of a QFL
ternary diagram and cluster at the plagioclase and volcanic apices of QmKP and LmLvLs
ternary diagrams, respectively (Packer and Ingersoll, 1986; Marsaglia, 1992; Marsaglia and
Ingersoll, 1992).
504 KATHLEEN M. MARSAGUA AND KATHLEEN A. DEVANEY
Studies of the glass fractions of these sediments provide additional information on the
provenance of backarc sediment. Rodolfo and Warner (1981) examined the coarse fraction
of 25 ash-bearing samples from Site 450 (predominantly very fine sand), and Schmincke
(1981) examined the very fine sand-sized (65-125 J..Lm) glass popUlations of 32 samples
from Sites 453, 454, 455, 458, and 459. Both studies found a wide range of glass colors and
refractive indices in these samples, and both demonstrate a link between glass color,
refractive index, and silica content. Glass fragments with refractive indices of <1.53 are
colorless and have high silica contents, whereas those with refractive indices of > 1.56 are
dark brown and have low silica contents. Light brown to tan to green glass fragments
exhibit intermediate refractive indices and silica contents. According to Schmincke (1981),
grain shape and vesicularity are not diagnostic of composition, but of eruption mode or
mechanism.
Because standard detrital modes of intraoceanic-arc volcaniclastic sands are very
similar and offer little insight into variations in sand provenance, provenance information
might best be determined by subdividing the volcanic lithic populations according to
texture (crystallinity) and color (composition). Glassy volcanic lithic fragments can be
categorized as vitric (holohyaline), microlitic (silt-sized crystals), or lathwork (sand-sized
crystals), depending on the presence or absence and size of crystal inclusions. The glassy
groundmass may range in color from colorless to brown to black. Marsaglia (1992)
developed this classification system in order to maximize petrographic information for Leg
126 sand and sandstone samples from the Izu-Bonin arc. Based on the studies by Rodolfo
and Warner (1981) and Schmincke (1981), glassy fragments are categorized as either
colorless or brown (includes tan, brown, and green varieties), providing information on the
proportion of felsic and intermediate to mafic components respectively. Black tachylitic
glassy fragments ("lithics") are also differentiated. Tachylitic glass is black (opaque) in
plane-polarized light due to the presence of fine Fe-Ti opaque minerals. This texture is
thought to be produced by lower cooling rates of mafic lava in both subaerial and
subaqueous conditions (Macdonald, 1972; Fisher and Schmincke, 1984; Cas and Wright,
1987). In the Leg 126 example (Marsaglia, 1992), this petrographic method provided
information on source lava composition and the effects of mixing and segregation during
sediment transport.
Standard modal analysis of sand from Leg 126 and Mariana transect (Legs 59 and 60)
sites shows them to be very similar and relatively uniform in composition (Fig. 14.3). In this
chapter, we present a petrographic data set acquired using the scheme outlined in Marsaglia
(1992), which allows us to compare and contrast the Izu-Bonin and Mariana arc systems.
The Leg 126 samples represent an early rift phase of backarc basin formation, whereas the
Legs 59 and 60 samples represent prerift to mature stages of backarc basin formation.
When combined, these data provide a detailed picture of changes in volcaniclastic prove-
nance during backarc basin evolution.
5. METHODS
Thin sections examined in this study were originally prepared and analyzed by Packer
and Ingersoll (1986). Their preparation techniques included sieving unconsolidated sam-
ples for the sand fraction (0.0625-2 mm), epoxying the sand concentrates to glass slides,
and then grinding them to 30-J..Lm thickness. Thin sections of sandstone and unconsolidated
r-----------~50%Q
FIGURE 14.3. Right-lower portion of QFL ternary plot of

Mariana (Legs 59 and 60) site means (filled circles) and Izu-
Bonin (Leg 126) site means (stars) by age from Marsaglia
(1992). Apices of ternary plot defined as Q =total quartz, F =
total feldspar, and L = total lithic fragments except poly-
crystalline quartz. Compositional fields taken from Dickin-
son et at. (1983): RO =recycled orogenic, DA =dissected arc,
TA =transitional arc, and VA =undissected arc. Age desig-
nations follow site numbers: Q = Quaternary, P = Pliocene,
and M = Miocene.
sand were subsequently stained for both calcium- and potassium-rich feldspars according
to the method outlined in Marsaglia and Tazaki (1992). Petrographic data for these samples
can be found in Packer and Ingersoll (1986). They used the Gazzi-Dickinson method of
point counting (Dickinson, 1970; Ingersoll et ai., 1984) and divided volcanic lithic clasts
into various textural types, including vitric, micro Ii tic, lathwork, and felsitic categories
(Dickinson, 1970). The volcanic fraction of these thin sections was recounted in this study,
further separating volcanic lithic grains according to glass color under plane-polarized light
(i.e., colorless, tan or brown, and black). All nonopaque "colored" glassy fragments were
grouped into the brown glass category, in part, because color is a function of grain
thickness. The Gazzi-Dickinson method was used and 300 volcanic lithic grains were
counted per thin section. The categories and recalculated parameters used in this study are
defined in Table I and illustrated in Fig. 14.4. Note that in accordance with the Gazzi-
Dickinson method, sand-sized (0.0625-2 mm) phenocrysts in volcanic lithic fragments
were skipped and not included in the lithic categories. Recalculated data are presented
in Table II.
6. PETROGRAPHIC RESULTS
Petrographic data were obtained from sand-rich sections recovered at Sites 450, 451,
453,455,457,458, and 459. The age of the main volcaniclastic sand and sandstone sections
encountered at each site varied from Oligocene (459) to Miocene (450, 451, and 458), to
Pliocene (453) to Pleistocene (455, 457). The proportions of colorless, brown, and black
glassy fragments for samples from these sites are shown in Fig. 14.5. Overall, these
volcaniclastic sand samples are dominated by brown and colorless glassy fragments. Black
glassy fragments are most prevalent in the Upper Miocene section at Site 451 (Fig. 14.5).
The distributions of vitric, microlitic, and lath work textures for colorless, brown, and black
glassy populations by site are shown in Figs. 14.6 and 14.7. Sites 450 and 453 show similar
vitric- and microlitic-dominated colorless, brown, and black glass populations, with a
lesser lath work component. At each of these sites, mean values for brown glassy fractions
are slightly enriched in microlitic component with respect to the colorless and black glassy
fractions, and the black glassy fraction shows a higher range in percentage of lathwork
TABLE I
Counted and Recalculated Parameters
Counted Parameters
Lvv: Vitric volcanic lithic fragments
brgl = brown glass
clgl = colorless glass
blgl = black glass
Lvml:Volcanic lithic fragments with microlitic texture
brgl = brown glass
blgl = black glass
Lvi: Volcanic lithic fragments with lathwork texture
brgl = brown glass
blgl = black glass
(altered glass counted but not included in categories)
Recalculated Parameters
Total colorless glass = ColBlkBrwn%col = [Lvv(clgl) + Lvml(clgl) + Lvl(clgl))/[Lvv(clgl) + Lvml(clgl) +
Lvl(clgl) + Lvv(brgl) + Lvml(brgl) + Lvl(brgl) + Lvv(blgl) + Lvml(blgl) + Lvl(blgl))
Total brown glass = ColBlkBrwn%brwn = [Lvv(brgl) + Lvml(brgl) + Lvl(brgl))/[Lvv(clgl) + Lvml(clgl) +
Total black glass = ColBlkBrwn%blk = [Lvv(blgl) + Lvml(blgl) + Lvl(blgl))/[Lvv(clgl) + Lvml(clgl) +
LvvLvmLvl%Lvv = lOO*Lvv/(Lvv + Lvml + Lvi)
LvvLvmlLvl%Lvml = lOO*LvmV(Lvv + Lvml + LvI)
LvvLvmlLvl%Lvl = lOO*LvV(Lvv + Lvml + LvI)
(LvvLvmlLvl percentages were calculated separately for ColoriesslPumice, Brown and Black glassy grains)
grains. A second group of samples from Sites 451, 455, and 457 are enriched in grains
exhibiting microlitic and lathwork textures. Sites 458 and 459B distributions are interme-
diate between these two groups.
7. DISCUSSION
7.1. Spatial and Temporal Trends in Sand Composition across the Mariana
Trough
There appear to be both spatial and temporal shifts in the locus of arc volcanism across
the Mariana arc and backarc region, and, as discussed below, these shifts are reflected in the
nature of associated volcaniclastic sediment. Sites that penetrated significant coarse clastic
sedimentary sequences, or arc proximal deposits, are limited to (1) the forearc (Palau-
Kyushu and West Mariana ridges) in the Oligocene and early Miocene, (2) the proto-
remnant arc (West Mariana Ridge) in the Middle to Upper Miocene, (3) within the Mariana
Trough in the Pliocene, and (4) along the modern arc axis (Mariana Ridge) in the
Pleistocene. This shift in volcanism could be explained as an artifact of drill site location
and degree of penetration, or it could be due to a physical shift in the locus of volcanism
from the now "remnant" arc, across the Mariana Trough to the present arc axis. Models
based on backarc spreading (i.e., Fig. 14.1) imply that the locus of arc volcanism remains
~
TABLE /I lJ
);
Recalculated Petrographic Data
~
Colorless Brown Black % Total
LvvLvrnlLvl% LvvLvrnlLvl% LvvLvrnlLvl% ColBlkBrwn% ffi
Site" Core Section Interval (crn) Age Lvv Lvrnl Lvi Lvv Lvrnl Lvi Lvv Lvrnl Lvi col blk brwn ~
lJ
450 59.11 12 25-30 Mio 68 29 2 5 91 5 11 79 10 30 32 38 m
450 59.12 13 1 136-140 Mio 94 6 o 72 27 72 28 o 38 11 51 ~
450 59.13 14 6 12-14 Mio 89 10 1 52 41 7 43 48 9 27 31 42 ~
450 59.14 16 1 129-131 Mio 84 16 o 32 66 2 42 47 11 40 25 35 <:
450 59.15 18 3 141-143 Mio 83 17 o 58 37 5 63 33 4 36 27 37 !:!:l
450 59.15a 18 3 143-145 Mio 94 6 o 88 12 o 64 25 11 42 9 48
450 59.15b 20 54-56 Mio 94 6 o 89 11 o 86 14 o 53 7 40
450 59.17 28 1 29-31 Mio 94 6 o 85 14 89 5 5 36 19 45
450 59.18 30 2 130-134 Mio 94 6 o 35 60 4 71 20 9 42 27 31
450 59.19 35 3 11-14 Mio 89 10 32 66 2 27 52 22 47 22 32
451 59.20 32 24-27 Mio 4 93 4 88 11 11 59 30 23 32 45
451 59.21 34 90-94 Mio 38 58 4 17 74 9 32 57 11 19 24 57
451 59.22 35 17-21 Mio 36 64 o 3 86 10 20 48 32 21 35 44
451 59.23 38 18-22 Mio 43 50 7 15 69 16 28 52 20 6 38 57
451 59.24 40 2 20-25 Mio 40 60 o 4 94 2 33 57 10 20 42 37
451 59.25 41 63-67 Mio 46 54 o 12 81 6 32 46 21 23 29 48
451 59.27 46 1 88-92 Mio 35 65 o 15 75 10 28 61 11 11 34 55
451 59.28 48 CC Mio 6 94 o 7 84 9 33 45 22 12 26 61
451 59.35 66 79-83 Mio 83 17 o 2 84 13 32 54 13 19 39 42
451 59.36 68 10-14 Mio 18 79 4 6 63 32 25 49 26 11 36 53
451 59.37 69 1 7-11 Mio 80 18 2 7 83 10 30 45 25 40 29 31
451 59.37a 70 2 140-143 Mio 91 8 1 35 55 9 47 31 23 35 26 40
451 59.38 72 1 134-140 Mio 100 o o 39 56 6 37 28 35 9 43 48
451 59.41 79 2 24-27 Mio 92 6 2 11 74 15 22 45 33 32 36 32
451 59.42 80 2 57-60 Mio 94 6 o 14 64 22 26 33 41 18 32 50
451 59.43 83 62-65 Mio 84 16 o 14 78 9 62 22 16 34 38 28
451 59.44 84 21-26 Mio 93 6 24 69 7 53 25 22 25 48 27
451 59.45 86 2 57-61 Mio 57 43 o 13 72 15 15 54 31 17 43 40 ~
"I
451 59.45a 86 3 75-78 Mio 86 14 o 21 66 14 52 23 25 5 44 51
451 59.45b 87 2 139-142 Mio 73 27 o 20 55 25 39 27 35 4 47 50 ~
451 59.46 89 144-148 Mio 74 26 o 22 73 5 52 37 11 28 29 43
451 59.48 93 14-20 Mio 100 o o 14 80 6 35 50 15 2 53 46
451 59.49 94 144-150 Mio 71 29 o 11 86 3 25 66 9 2 56 42
453 60.01 29 32-33 Plio 38 58 4 27 70 4 62 33 5 18 33 48
453 60.02 32 2 23-25 Plio 22 76 2 12 74 14 27 54 19 51 34 15
453 60.03 33 2 72-74 Plio 69 31 o 53 46 71 21 7 54 5 41
453 60.04 34 2 40-43 Plio 79 21 o 85 15 78 22 o 49 3 48
453 6O.06a 39 3 108-112 Plio 94 6 o 80 18 2 68 21 11 75 6 19
453 60.07 41 3 31-34 Plio 75 21 4 52 39 10 78 15 7 80 9 11
453 60.08 42 2 20-22 Plio 43 54 3 45 51 4 44 24 32 37 24 39
455 60.13 1 6 35-37 Pleist 25 71 4 39 59 2 47 39 14 39 20 40
455 60.14 3 3 55-58 Pleist 38 54 8 34 64 2 66 32 2 59 20 21
455 60.15 4 CC Pleist 26 43 31 16 68 16 16 51 33 20 28 51
455 60.16 5 95-98 Pleist 18 55 27 12 66 22 26 57 17 47 26 27 ~
455 60.17 7 1 7-10 Pleist 19 79 2 10 63 27 22 44 33 41 7 53 ~
r-
455 60.18 8 CC Pleist 14 80 6 14 57 29 19 52 29 43 13 44 m
455 60.19 9 CC Pleist 4 71 25 6 52 42 o 64 36 18 10 73 <:
455 60.20 10 2 66-68 Pleist 10 84 5 6 58 36 4 68 29 21 10 70 ~
457 60.21 2 3 18-20 Pleist 10 68 23 43 27 30 20 43 37 11 32 57
457 60.22 2 64-67 Pleist 11 47 42 16 58 26 20 24 7
~
6 56 32 61 II
458 60.23 5 3 88-90 Mio 63 37 1 28 60 12 35 59 6 55 6 39 ~
459b 60.25 12 CC Mio 66 34 o 26 69 5 22 56 22 58 6 36 ~
459b 60.26 15 59-61 Mio 56 42 2 39 58 3 40 55 5 31 29 40 S;;
).
459b 60.27 19 56-58 Mio 100 o o 93 6 1 43 48 9 2 8 90
459b 60.28 21 1 74-76 Mio 74 26 o 37 57 6 24 76 o 27 23 49 ~
459b 60.30 25 2 98-100 Mio 50 47 3 29 62 10 12 65 24 76 6 18 ~
459b 60.33 32 59-61 Mio 79 19 2 32 62 6 29 65 6 39 6 55 ~
4
,....
459b 60.34 36 71-73 Mio 95 40 57 4 25 71 4 55 8 37
459b 60.35 37 1 126-128 Mio 88 11 1 28 65 7 12 58 31 63 9 28 m
<:
459b 60.36 39 2 100-102 Mio 93 7 o 18 76 6 o 100 o 82 17 ;to
459b 60.37 49 39-41 Oligo 88 12 o 26 71 3 32 58 11 74 6 20
459b 60.38 50 2 10-12 Oligo 50 45 6 44 40 17 45 48 7 61 15 24 ~
§
aSample numbers from Packer and Ingersoll (1986).
~
-.::
FIGURE 14.4. Photomicrographs of

various volcanic lithic fragments, all
plane-polarized light in unstained portion
of slide. (a) Sample 60-20, (b) Sample
59-19, (c) Sample 60-20. See Table II for
sample information. Lithic fragment
types are keyed as follows: A =colorless
vitric (Lvv), B = colorless microlitic
(Lvml), C = brown vitric (Lvv), D =
brown microlitic (Lvml), E =brown lath-
work (Lvi), F = black vitric (Lvv), G =
black microlitic (Lvml), H = black lath-
work (LvI). Colorless and brown vitric
fragments are commonly vesicular. Sand-
sized phenocrysts in lath work fragments
include plagioclase (p) and pyroxene (x);
the groundmass of lathwork fragments
shown here are microlitic, but may also
be glassy (vitric) in other instances.
Edges of colorless grains may appear
dark due to adhesion of matrix material.
Note that in stained portion of slide, the
stain often masks nonopaque glass color.
COLORLBSS
GLASS
BLACK ai441e lIi_eDe upper Miocene upper upper upper lower- lower BROIfII
GLASS pli_eDe Plei8t. Plei8t. .1441e Oligo. GLASS
lIi_eDe
FIGURE 14.5. Ternary plots of colorless, brown, and black glassy lithic proportions by site and age. Polygons
represent fields of variation defined by one standard deviation on either side of mean ("circle"). See Table I for
definition of recalculated parameters and Table II for recalculated data.
fixed along the frontal arc. Therefore a temporal or spatial shift might be consistent with
forearc rifting rather than backarc rifting along this arc segment. Sedimentary sequences
cored in the Mariana forearc also present an incomplete history of arc volcanism. Nascent
Mariana arc volcanoes appear to have developed at the boundary between backarc and old
frontal arc crust, or on downfaulted frontal arc crust, to the west of and structurally lower
than the frontal arc scarp (Hussong and Uyeda, 1981; Bloomer et aI., 1989). The active arc is
largely submerged and located to the west of the frontal arc scarp, which acts as a
topographic barrier; hence coarse volcaniclastic materials associated with these volcanoes
are probably preferentially distributed in the backarc basin. A successful deep penetration
of Mariana arc apron deposits in the backarc region would help document the longevity of
the modem Mariana arc and test the theory of spatial shifts in volcanism through time
across the Mariana Trough.
Temporal trends in volcaniclastic composition across the Mariana Trough are por-
trayed in Fig. 14.7, in which site means are plotted by age. These data indicate a gradual
shift from more felsic populations in the Oligocene (Site 459) and early Miocene (Sites 458
and 459) to more intermediate to mafic populations (brown glass) in the middle (Site 450)
and late Miocene (Site 451). A subsequent shift to more felsic components in the Pliocene
(Site 453) is coincident with rifting and the inception of seafloor spreading in the Mariana
Trough. Samples from Site 453 are also slightly enriched in quartz with respect to the other
sites, suggesting more felsic input (Fig. 14.3). Schmincke (1981) noted the dominance of
intermediate to silica-rich volcaniclastic material in the Pliocene sections at Site 453 and
found Pleistocene vitric muds at Sites 454, 456, 458, and 459B to be rich in glass shards of
basaltic andesitic composition. A corresponding trend from more felsic Pliocene composi-
tions toward more intermediate to mafic compositions (brown and black glass) in the
Pleistocene is also seen in the glassy proportions determined in this study (Fig. 14.7).
By focussing on coarser-grain samples from more volcaniclastic-rich sections, we
hope we have circumvented provenance problems related to widespread distribution of
airborne ash. Note that paleoceanographic currents may have favored accumulation of
coarse clastic debris in the backarc region, similar to the modem Lesser Antilles arc as
described by Carey and Sigurdsson (1984). Schmincke (1981) discusses the likely input at
Leg 60 sites from distant versus proximal volcanoes and suggests that finer-grain felsic
components produced by highly explosive volcanic eruptions are more apt to be far
traveled. Klein (1985a) indicates that wind patterns in the Mariana region would likely
result in along-strike transport of airborne ash, from the north or south, depending on the
season. Some of the apparent trends in provenance that Schmincke (1981) outlines in the
younger sections at Site 453 may reflect the interplay of distal versus proximal sources, as
opposed to actual changes in volcanic provenance from calc-alkaline to bimodal rhyolite-
basalt. Whereas Schmincke (1981) sees a change from more mafic to felsic compositions
from Cores 36 to 19 at Site 453, data presented in Table II indicate a shift from more
colorless (felsic) to more brown (intermediate to mafic) glass from Cores 48 to 29.
Discrepancies between our data sets are understandable given the differences in depth
range, grain size, and number of samples.
The volcaniclastic history of Site 453 is key to understanding the evolution of the
Mariana Trough. The thick Pliocene volcaniclastic section present at Site 453 is represented
by only traces of coarse pyroclastic debris in the Pliocene sections at Site 451. The fact that
Site 451 is more proximal to West Mariana Ridge than Site 453 suggests that the likely
source of the Pliocene volcaniclastic units was to the east, derived from the Mariana arc
during the early phases of seafloor spreading (Hussong et aI., 1981). Thin layers of
tachylitic glass present in this section (Hussong et aI., 1981) may have been derived from
the nascent spreading center or intrabasinal volcanoes, but apparently these layers were not
sampled by Packer and Ingersoll (1986), and so were not counted in this study. Hussong
et al. (1981) believed the primary source of the volcanic material at Site 453 to be the
Mariana arc, but thought that this material had probably been reworked from the highs
flanking the small basin in which Site 453 was drilled. The possibility of intrabasinal
volcanic sources for the main portion of the volcaniclastic section was not discussed by
Hussong et al. (1981). However, in light of recent drilling in the Lau Basin (Parson et aI.,
1992; Hawkins, Chapter 3 this volume), the possibility of an intrabasinal volcanic source
for volcaniclastic materials at Site 453 should not be ruled out (see discussion below).
7.2. Relationship between Lava Geochemistry and Detrital Modes

The petrologic evolution of Mariana arc and backarc igneous rocks is reviewed by
Natland and Tamey (1981). They outline the dominant lava compositions as follows:
Oligocene-arc tholeiite; Miocene-calc-alkalic; Pliocene-arc tholeiite; and Quater-
nary-arc tholeiite to calc-alkalic. Tholeiitic volcanic rocks characterize the early rift and
spreading phases of backarc basin formation, whereas calc-alkalic volcanism associated
with fractional crystallization characterizes the more mature arc and backarc systems
(Hussong and Uyeda, 1981; Natland and Tamey, 1981). Thus these compositional shifts are
related to backarc basin evolution (Natland and Tamey, 1981).
There is a general correlation between the volcanic lithic proportions outlined in Fig.
14.5 and the temporal trends outlined in Natland and Tamey (1981): periods of calc-alkaline
volcanism (Miocene and Quaternary-Recent) are characterized by glass populations domi-
nated by brown glassy fragments, whereas periods of arc tholeiite volcanism (Oligocene-
early Miocene and Pliocene-Pleistocene) are characterized by glass populations dominated
by colorless glassy fragments. The high felsic and minor mafic components associated with
periods of tholeiitic volcanism can be best explained by bimodal volcanism during the early
rift phases of backarc basin formation. After the onset of seafloor spreading and concurrent
with arc evolution, volcanism becomes more intermediate and brown glass dominates.
Lvl
a Lftll Lv!
FIGURE 14.6. Ternary plots ofvitric (Lvv), rnicrolitic (Lvrnl), and lathwork (LvI) proportions in glassy fractions
(colorless, brown, and black) by site. (a) Sites 450, 451, and 453. (b) Sites 455, 457, 458, and 459b. Means and
polygons as in Fig. 14.5.
Electron microprobe analyses of glass and mineral fragments from Site 453 by Packham
and Williams (1981) demonstrate the Pliocene to Recent trend from arc-tholeiite to calc-
alkaline volcanism. Glass compositions in the Lower Pliocene volcaniclastic units indicate
a broad spectrum of silica contents (49-73%), with a bimodal distribution in some intervals
(Packham and Williams, 1981). The dominance of felsic components in these sediments has
been ascribed to preferential eruption of differentiated felsic magmas or explosive volca-
nism (Packham and Williams, 1981; Schmincke, 1981). The glass proportions presented in
Fig. 14.5 suggest that bimodal volcanism, such as that ofthe incipiently rifting Izu-Bonin
are, may have characterized the early history of both the Mariana Trough and the Parece
Vela Basin.
7.3. Comparison of Mariana Sand Composition to That of Other Backarc Basins

The Sumisu rift is a well-studied example of a nascent backarc basin within the Izu-
Bonin arc system. It is an asymmetrical structural graben located between two submarine
L~L---~--~------~----------------~Lvl
L~ L -______~ ______'---________________....:. LvI
Lvv L.. L..
• IUII
U8. U'b
b L~ -L--=-==--__________
L -_ _ _ _----''--_ _ _ _ _ _ ~ Ld
arc calderas (Taylor et aI., 1990b, 1991; Klaus et aI., 1992). Arc volcanism prior to and
during the present rifting of the Izu-Bonin arc has been dominantly silicic (subalkaline
andesite-rhyolite) with tholeiitic basalt limited to the backarc region (Fryer et al., 1985;
Fujioka et aI., 1992; Gill et aI., 1992; Rodolfo et aI., 1992; Taylor, 1992). Pliocene and
Pleistocene sand detrital modes across the Izu-Bonin arc are consistent with this picture of
volcanism. Pliocene volcaniclastic deposits from the rift-flank uplift (Site 788) are domi-
nantly composed of colorless glassy fragments, whereas samples from intrarift sites (790
and 791) are extremely heterogeneous, showing wide ranges of basaltic and intermediate
brown and black glassy components (Fig. 14.8; Marsaglia, 1992). Their heterogeneous
nature can be explained by either syneruptive mixing or mixing during mass transport
within the basin. The proximity of Sites 790 and 791 to intrarift mafic volcanic centers
suggests that sand compositions could have been produced by mixing of intrabasinal
basaltic materials during mass flow of siliceous debris from nearby calderas or by rework-
ing off rift flank uplifts (Marsaglia, 1992). A similar compositional range is also observed in
Quaternary deposits cored at Izu-Bonin forearc sites, but mean forearc-site compositions
are much more enriched in intermediate to mafic materials (Fig. 14.9; Marsaglia, 1992).
These sites are structurally isolated from backarc basaltic volcanoes, and therefore only
FIGURE 14.7. Ternary plot of mean proportions of colorless, black,

and brown glassy fragments for Mariana sites. Temporal trends indi-
cated by arrows connecting Oligocene to Miocene means (filled cir-
cles), and Pliocene to Pleistocene means (open circles). Lower Oligo-
cene = Site 459; Lower to Middle Miocene = Sites 458/459; Middle
= =
Miocene Site 450; Upper Miocene Site 451; Upper Pliocene Site =
453; Pleistocene = Sites 455 and 457.
syneruptive mixing of felsic and mafic components could be responsible for these mixtures.
In a syneruptive mixing scenario, the tachylitic fraction might represent arc lavas frag-
mented during explosive caldera eruptions. According to Marsaglia (1992), the tachylitic
sand fragments at Site 787 (Fig. 14.8) may be epiclastic detritus eroded from an emergent
arc volcano and concentrated during transport down the submarine canyon in which Site
787 is located. Detailed geochemical analyses oftachylite grains that might constrain their
provenance are not available.
The Lower Pliocene volcaniclastic units at Site 453 within the Mariana Trough could
represent ancient analogs to Quaternary sediments deposited at Sites 790 and 791 in the
Sumisu rift. Sediments from all three sites are enriched in felsic components such as quartz
(Fig. 14.3) and colorless glassy fragments (Figs. 14.5 and 14.8). Their mean glassy propor-
tions are almost identical (Figs. 14.5 and 14.8).
Sequences drilled at a number of sites in the Lau Basin on Leg 135 are remarkably
similar to sequences recovered at Site 453 in the Mariana Trough and, in some respects, to
those in the Sumisu rift. Sites 834 through 839 drilled a series of fault-bounded subbasins,
initially filled by Pliocene to Pleistocene coarse volcaniclastic turbidites and debris flows,
which are overlain by fine clayey pelagic units (Shipboard Scientific Party, 1992). Se-
Colorlelu.
GIIIIUI
Br_ Black
(;I1alllill Ole",s
FIGURE 14.8. Ternary plots of colorless, brown and black glassy lithic proportions by site and age, for samples
from ODP Leg 126, Izu-Bonin arc Sites 787,789,790-793. Polygons represent fields of variation defined by one
standard deviation on either side of mean (circle). See Table I for definition of recalculated parameters and
Marsaglia (1992) for recalculated data.
STAGE 1
lr
SPREADING
CENTER
STAGE 2
EXTENSION
STAGE 3
OR
"
NASCENT ARC
BACKARC BASIN
STAGE 4
1~2
;:,r REMNANT'Y
ARC
SPREADING
CENTER
Temporal/Spatial Shift in Volcanism

and Accumulation of Coarse Volcaniclastic Deposits
FIGURE 14.9. Model of backarc basin evolution (Stage I [oldest) to Stage 4 [youngest)) due to forearc rifting.
Main volcanic centers depicted by erupting subaerial (Stages 1, 2, and 4) and submarine (Stage 3) volcanoes.
Associated coarse volcaniclastic apron or fan deposits are numbered according to the stage in which they are
produced. Temporal and spatial shifts are summarized at base of diagram. Model concept modified from Karig and
Moore (1975) and Carey and Sigurdsson (1984). Filled triangle denotes sea level.
quences within these basins are similar to those at Sites 453, 790, and 791 not only in terms
of their stratigraphy but also in terms of their apparent volcaniclastic populations. No
detrital modes for these sediments have been reported, but general descriptions from the
initial scientific reports (Shipboard Scientific Party, 1992) indicate that the sandy turbidites
are predominantly composed of colorless shards and pumice fragments, dominantly rhyo-
dacitic (70-71% Si02 ) in composition. Less common layers of scoriaceous basaltic sand
and hyaloclastite layers of brown to brownish green glass ranging from basaltic andesite to
andesite (52-57% Si02) in composition are also present. The majority of these materials
appear to have been derived from (an) unidentified intrabasinal or arc-related silicic
volcanic center(s) (Shipboard Scientific Party, 1992). Deposition occurred during a geo-
graphically broad extensional event, coeval with magmatism and eastward migration of
volcanic centers. This migration is indicated by eastward younging of both volcanic
basement and coarse clastic packages at Leg 135 sites (Shipboard Scientific Party, 1992).
7.4. Implications of Forearc versus Backarc Rifting on Sedimentary Models

The Lau and Mariana backarc basins have been proposed as possible examples of
"backarc" basins initiated by forearc spreading (Hawkins et ai., 1984). The forearc
spreading model of Hawkins et al. (1984) predicts a gap in volcanism during the time of
backarc basin initiation. Volcanic activity along the main arc axis is thought to wane during
forearc rifting, and then the center of volcanism shifts to the forearc "remnant" ridge. In
contrast, Leg 135 documented a record of continuous volcanism across the Lau Basin, but
with an eastward migration of the locus of volcanism (Shipboard Scientific Party, 1992).
We suggest that such a shift in volcanism may also be recorded in the temporal and spatial
distribution of coarse proximal volcaniclastic facies across the Mariana Trough.
In both the Karig and Moore (1975) and Carey and Sigurdsson (1984) models, the
effects of forearc rifting are not considered. According to these models (Fig. 14.1), the
history of arc magmatism is best recorded in the proximal deposits of the arc apron. The fine
ash component is not as useful because of probable transport from along strike or from
nearby arc volcanoes. To date, arc aprons, particularly those along the Mariana arc system,
have proved unsuitable for drilling, and only minimal penetration of this facies has been
achieved at any given site. The complete history of Mariana volcanism is not preserved at
anyone site drilled during Legs 59 and 60; in part because of unsuccessful drilling,
unconformities in the forearc sections, and an apparent shift in the locus of magmatism, this
history is only preserved piecemeal at a combination of sites.
The backarc basin sedimentation models presented in Fig. 14.1 can be modified to
represent the possible effects of forearc rifting on the localization of coarse volcaniclastic
sedimentation. If rifting begins in the backarc region, the arc is static and semicontinuously
supplies volcaniclastic material into the forearc and backarc basins. If rifting initiates just in
front of the arc axis, in the forearc region, then the locus of volcanism and volcaniclastic
sedimentation shifts temporally and spatially from the protoremnant arc, across the basin
concurrent with extension, rifting, and seafloor spreading. A simplified model depicting
such a scenario is shown in Fig. 14.9. It is loosely based on the Mariana and Lau examples
discussed in this chapter and shows the development of coarse, volcaniclastic-apron facies
across the region with time. Stage 1 represents the prerift phase, in which volcaniclastic
aprons flank a mature intraoceanic arc. Stage 2 indicates a shift in the locus of volcanism
and volcaniclastic sedimentation away from the arc axis associated with early extension of
the forearc region. Two scenarios are presented in Stage 3, one with volcanism shifting
directly from the protoremnant arc to the protoarc, and another with volcanism moving
across the nascent backarc basin. In Stage 4, the volcanic arc is reestablished and seafloor
spreading has commenced. The likely spatial distribution of coarse volcaniclastic sedi-
ments produced during forearc rifting is depicted for each stage (sequentially numbered 1
through 4).
In the Mariana example, Stage 1 would represent the middle Miocene, Stage 2 would
correspond to the late Miocene, Stage 3 would represent the Pliocene, and Stage 4 would
depict the Pleistocene to Recent setting. According to the petrographic data presented in
Figs. 14.5 and 14.7, volcaniclastic apron deposits in Stages 1, 2, and 4 are characterized by
intermediate volcaniclastic debris, whereas Stage 3 deposits are characterized by felsic to
bimodal volcaniclastic debris. Felsic to bimodal volcaniclastic sequences cored on Leg
135 in the western Lau Basin also represent Stage 3 sedimentation and, as discussed
previously, appear to be compositionally similar to Site 453 (Stage 3) volcaniclastic debris.
Sumisu rift sediments cored on Leg 126 correspond to intra-arc sedimentation during
Stages 1 and 4 of the Carey and Sigurdsson (1984) backarc-rifting model shown in Fig. 14.1.
8. CONCLUSIONS
This chapter presents a nonclassical view of intraoceanic volcaniclastic sedimentation

focusing on the populations of sand-sized volcanic lithic fragments rather than on bulk and
individual grain geochemistries. The proportion of volcaniclastic components yields infor-
mation on magma composition and is also influenced by posteruption depositional pro-
cesses. By focusing on the sand-size components of intraoceanic backarc sequences, we
have decreased the effects of long-distance transport of fine ash in the atmosphere and have
emphasized proximal volcanism. Compositional variations between and across these re-
gions may be linked to the stage (incipient vs. mature) of backarc basin formation. The
proportions of colorless, brown, and black glassy sand fragments within coarse volcaniclas-
tic intervals suggest that the early rift phase of backarc basin formation is characterized by
felsic (bimodal) volcanism and that volcanic centers become progressively more intermedi-
ate with time. This trend is consistent with the geochemical evolution of arc lavas. In the
case of forearc spreading, these igneous and volcaniclastic compositional trends coincide
with a shift in volcanism from the backarc, across the nascent backarc basin, toward the
frontal arc.
Acknowledgments
We thank Bonnie Packer and Raymond Ingersoll for use of their thin sections and
Loren Kroenke and Neil Lundberg for their thoughtful reviews.
REFERENCES
Betzler, C., Nederbragt, A. J., and Nichols, G. J. 1991. Significance of turbidites at Site 767 (Celebes Sea) and Site
768 (Sulu Sea), in Proc. DDP, Sci. Results, 124 (A. E. Silver, C. Rangin, T. von Breymann et aI., eds.), pp.
431-446, Ocean Drilling Program, College Station, TX.
Bloomer, S. H., Stem, R. J., and Smoot, N. C. 1989. Physical volcanology of the submarine Mariana and Volcano
arcs, Bull. Volcanol. 51:210-224.
Carey, S., and Sigurdsson, H. 1984. A model of volcanogenic sedimentation in marginal basins, in Marginal Basin
Geology (B. P. Kokelaar and M. E Howells, eds.), pp. 37-58, Geological Society Special Publ., Vol. 16,
Oxford.
Cas, R. A., and Wright, J. V. 1987. Volcanic Successions, Allen and Unwin, London.
Cashman, K. V., and Fiske, R. S. 1991. Fallout of pyroclastic debris from submarine volcanic eruptions, Science
253:275-280.
Dickinson, W. R. 1970. Interpreting detrital modes of graywacke and arkose, J. Sediment. Petrol. 40:695-707.
Dickinson, W. R., Beard, L. S., Brakeoridge, G. R., Erjavec, J. L., Ferguson, R. c., Inman, K. E, Knepp, R. A.,
Lindberg, E A., and Ryberg, P. T. 1983. Provenance of North American Phanerozoic sandstones in relation to
tectonic setting, Geol. Soc. Am. Bull. 94:222-235.
Fisher, R. V. 1984. Submarine volcaniclastic rocks, in Marginal Basin Geology (B. P. Kokelaar and M. E Howells,
eds.), pp. 5-27, Geological Society Special Publ., Vol. 16, Oxford.
Fisher, R. Y., and Schmincke, H.-U. 1984. Pyroclastic Rocks, Springer-Verlag, New York.
Fryer, P., Langmuir, C., Taylor, B., Zhang, Y., and Hussong, D. 1985. Rifting of the Izu arc. III: Relationship of
chemistry to tectonics, Eos 66:421.
Fryer, P., Taylor, B., Langmuir, C. H., and Hochstaedter, A. G. 1990. Petrology and geochemistry oflavas from the
Fujioka, K., Nishimura, A., Matsuo, Y., and Rodolfo, K. S. 1992. Correlation of Quaternary tephras throughout the
Izu-Bonin areas, in Proc. ODp, Sci. Results, 126 (B. Taylor, K. Fujioka et al., eds.), pp. 23-46, Ocean
Gill, J. B, Seales, C., Thompson, P., Hochstaedter, A. G., and Dunlap, C. 1992. Petrology and geochemistry of
Pliocene-Pleistocene volcanic rocks from the Izu arc, Leg 126, in Proc. ODp, Sci. Results, 126 (B. Taylor, K.
Fujioka et al., eds.), pp. 383-404, Ocean Drilling Program, College Station, TX.
Hawkins, J. W., Bloomer, S. H., Evans, C. A., and Melchior, J. T. 1984. Evolution of intra-oceanic arc-trench
systems, Tectonophysics 102:175-205.
Hochstaedter, A. E, Gill., J. B., Kusakabe, M., Newman, S., Pringle, M., Taylor, B., and Fryer, P. 1990a. Volcanism
in the Sumisu Rift. I: Major element, volatiles and stable isotope geochemistry, Earth Planet. Sci. Lett.
100:179-194.
Hussong, D. M., and Fryer, P. 1983. Backarc seamounts and the SeaMARC II seafloor mapping system, Eos
64:627 -632.
Hussong, D. M., and Uyeda, S. 1981. Tectonic processes and the history of the Mariana arc: A synthesis of the
results of Deep Sea Drilling Project Leg 60, in Init. Repts, DSDP, 60 (D. M. Hussong, S. Uyeda et aI., eds.),
pp. 909-929, Deep Sea Drilling Project, U.S. Govt. Printing Office, Washington, DC.
Hussong, D. M., Uyeda, S. et al. 1981. Init. Repts, DSDP, 60, Deep Sea Drilling Project, U.S. Govt. Printing
Office, Washington, DC.
Ingersoll, R. V., Bullard, T. E, Ford, R. L., Grimm, J. P., Pickle, J. D., and Sares, S. W. 1984. The effect of grain
size on detrital modes: A test of the Gazzi-Dickinson point-counting method, J. Sediment. Petrol. 54:
103-116.
Jackson, M. c., and Fryer, P. 1986. The Kasuga volcanic cross-chain: A calc-alkaline basalt-dacite suite in the
northern Mariana island arc, Eos 67:276.
Karig, D. E. 1975. Basin genesis in the Philippine Sea, in Init. Repts, DSDP, 31 (D. E. Karig, 1. C. Ingle, Jr. et al.,
eds.), pp. 857-879, Deep Sea Drilling Project, U.S. Govt. Printing Office, Washington, DC.
Karig, D. E. 1983. Temporal relationships between back arc basin formation and arc volcanism with special
reference to the Philippine Sea, in The Tectonic and Geologic Evolution of Southeast Asian Seas and Islands.
Part II (D. E. Hayes, ed.), pp. 318-325, Geophys. Monogr. Ser, Vol. 27, American Geophysical Union,
Washington, DC.
Karig, D. E., and Moore, G. E 1975. Tectonically controlled sedimentation in marginal basins, Earth Planet. Sci.
Lett. 26:233-238.
Klaus, A., Taylor, B., Moore, G. E, MacKay, M. E., Brown, G. R., Okamura, Y., and Murakami, E1992. Structural
and stratigraphic evolution of Sumisu rift, Izu-Bonin arc, in Proc. ODp, Sci. Results, 126 (B. Taylor, K.
Fujioka et aI., eds.), pp. 555-574, Ocean Drilling Program, College Station, TX.
Klein, G. deY. 1975a. Sedimentary tectonics in southwest Pacific marginal basins based on Leg 30 Deep Sea
Drilling Project cores from the South Fiji, Hebrides, and Coral Sea basins, Geol. Soc. Am. Bull. 86:
1012-1018.
Klein, G. deV. 1975b. Depositional facies of leg 30 DSDP sediment cores, in Init. Repts, DSDP, 30 (J. E. Andrews,
G. Packham, G. et aI., eds.), pp. 423-442, Deep Sea Drilling Project, U.S. Gov!. Printing Office, Washing-
ton, DC.
Klein, G. deY. 1985a. The control of depositional depth, tectonic uplift, and volcanism on sedimentation processes
in the backarc basins of the western Pacific Ocean, J. Geol. 93:1-25.
Klein, G. deV. 1985b. The frequency and periodicity of preserved turbidites in submarine fans as a quantitative
record of tectonic uplift in collision zones, Tectonophysics 119:181-193.
Kroenke, L., Scott, R. et al. 1980. Init. Repts., DSDP, 59, U.S. Govt. Printing Office, Washington, DC.
Macdonald, G. A. 1972. Volcanoes, Prentice-Hall, Englewood Cliffs, N1.
Marsaglia, K. M. 1992. Petrography and provenance of volcaniclastic sands recovered from the Izu-Bonin arc,
Leg 126, in Proc. ODp, Sci. Results, 126 (B. Taylor, K. Fujioka et aI., eds.), pp. 139-154, Ocean Drilling
Program, College Station, TX.
Marsaglia, K. M., in press, Interarc and backarc basins, in Tectonics of Sedimentary Basins (R. V. Ingersoll, and
C. 1. Busby, eds.), Blackwell Scientific.
Marsaglia, K. M., and Ingersoll, R. Y. 1992. Compositional trends in arc-related, deep-marine sand and sandstone:
A reassessment of magmatic-arc provenance, Geol. Soc. Am. Bull. 104:1637-1649.
Marsaglia, K. M., and Tazaki, K. 1992. Diagenetic trends in ODP Leg 126 sandstones, in Proc. ODP, Sci. Results,
126 (B. Taylor, K. Fujioka et al., eds.) pp. 125-138, Ocean Drilling Program, College Station, TX.
Mattey, D. P., Marsh, N. G., and Tarney, J.1980. The geochemistry, mineralogy, and petrology of basalts from the
West Philippine and Parece Vela basins and from the Palau-Kyushu and West Mariana ridges, Deep Sea
Drilling Project Leg 59, in Init. Repts, DSDP, 59 (L. Kroenke, R. Scott et aI., eds.), pp. 753-800, Deep Sea
Drilling Project, U.S. Gov!. Printing Office, Washington, DC.
Natland,1. H., and Tarney, 1. 1981. Petrologic evolution of the Mariana arc and backarc system-a synthesis of
drilling results in the south Philippine Sea, in Init. Repts, DSDP, 60 (D. M. Hussong, S. Uyeda et al., eds.), pp.
877-908, Deep Sea Drilling Project, U.S. Govt. Printing Office, Washington, DC.
Nishimura, A., Marsaglia, K. M., Rodolfo, K. S., Colella, A., Hiscott, R. N., Tazaki, K., Janecek, T., Firth, J.,
Isiminger-Kelso, M., Herman, Y., Gill, 1. B., Taylor, R. N., Taylor, B., Fujioka, K., and Leg 126 Shipboard
Scientific Party. 1991. Pliocene-Quaternary submarine pumice deposits in the Sumisu rift area, Izu-Bonin
Arc, in Sedimentation in Volcanic Settings (R. Y. Fisher and G. A. Smith, eds.), pp. 201-208, SEPM Special
Publication 45.
Packham, G. H., and Williams, K. L. 1981. Volcanic glasses from sediments from Sites 453 and 454 in the Mariana
Trough, in Init. Repts., DSDP, 60 (D. M. Hussong, S. Uyeda et aI., eds.), pp. 483-495, Deep Sea Drilling
Project, U.S. Gov!. Printing Office, Washington, DC.
Packer, B. M., and Ingersoil, R. V. 1986. Provenance and petrology of Deep Sea Drilling Project sands and
sandstones from the Japan and Mariana forearc and backarc regions, Sediment. Geol. 51:5-28.
Parson, L., Hawkins, J., Allan, J., et al. 1992. Proc. ODP, Init. Repts., 135, Ocean Drilling Program, College
Station, TX.
Rangin, C., Silver, E. A., von Breyman, M. T. et al. 1990. Proc. ODp, Init. Repts., 124, Ocean Drilling Program,
Rodolfo, K. S., Solidum, R. U., Nishimura, A., Matsuo, Y., and Fujioka, K.1992. Major-oxide stratigraphy of glass
shards in volcanic ash layers of the Izu-Bonin arc-backarc sites (Sites 7881788 and 7901791), Proc. ODp, Sci.
Results, 126 (B. Taylor, K. Fujioka et al., eds.), pp. 505-517, Ocean Drilling Program, College Station, TX.
Rodolfo, K. S., and Warner, R. 1. 1981. Tectonic, volcanic, and sedimentologic significance of volcanic glasses
from Site 450 in the eastern Parece Vela basin, in Init. Repts., DSDP, 59 (L. Kroenke, R. Scott et aI., eds.), pp.
603-607, Deep Sea Drilling Project, U.S. Gov!. Printing Office, Washington, DC.
Schmincke, H.-U. 1981. Ash from vitric muds in deep sea cores from the Mariana Trough and fore-arc region
(South Philippine Sea)(Sites 453, 454, 455, 458, 459), in Init. Repts., DSDP, 60 (D. M. Hussong, S. Uyeda
et aI., eds.), pp. 473-481, Deep Sea Drilling Project, U.S. Govt. Printing Office, Washington, DC.
Scott, R., and Kroenke, L. 1980. Evolution of back arc spreading and arc volcanism in the Philippine Sea:
Interpretation of Leg 59 DSDP results, in The Tectonic and Geologic Evolution of Southeast Asian Seas and
Islands (D. E. Hayes, ed.), pp. 283-291, Geophys. Monogr. Ser., Vol. 23, American Geophysical Union,
Washington, DC.
Scott, R. B., Kroenke, L., and Zakariadze, G. 1980. Evolution of the South Philippine Sea: Deep Sea Drilling
Project Leg 59 drilling results, in Init. Repts., DSDP, 59 (L. Kroenke, R. Scott et al., eds.), pp. 803-815, Deep
Sea Drilling Project, U.S. Gov!. Printing Office, Washington, DC.
Shipboard Scientific Party. 1990. Introduction, background, and principal results of Leg 128 of the Ocean Drilling
Program, Japan Sea, in Proc. ODP, [nit. Repts., 128 (J. C. Ingle, Jr., K. Suyehiro, T. von Breymann et al.,
Shipboard Scientific Party. 1992. Introduction, background, and principal results of Leg 135, Lau Basin, in Proc.
ODP, [nit. Repts., 135 (L. Parson, J. Hawkins, J. Allan et al., eds.), pp. 5-47, Ocean Drilling Program,
Sigurdsson, H., Sparks, R. S., Carey, S. N., and Huang, T. C. 1980. Volcanogenic sedimentation in the Lesser
Antilles arc, J. Geol. 88:523-540.
Silver, E. A., and Rangin, C. 1991. Leg 124 tectonic synthesis, in Proc. ODp, Sci. Results, 124 (A. E. Silver, C.
Rangin, T. von Breymann et al., eds.), pp. 3-9, Ocean Drilling Program, College Station, TX.
Taylor, B. 1992. Rifting and the volcanic-tectonic evolution of the Izu-Bonin-Mariana arc, in Proc. ODP,
Sci. Results, 126 (B. Taylor, K. Fujioka et aI., eds.), pp. 627-651, Ocean Drilling Program, College Sta-
tion, TX.
Taylor, B., Brown, G., Fryer, P., Gill, J. B., Hochstaedter, A. G., Hotta, H., Langmuir, C. H., Leinen, M.,
Nishimura, A., and Urabe, T. 1990b. ALVIN-Sea Beam studies of the Sumisu rift, Izu-Bonin arc, Earth
Taylor, B., Klaus, A., Brown, G. R., and Moore, G. F. 1991. Structural development of Sumisu rift, Izu-Bonin arc,
J. Geophys. Res. 96:16,1l3-16,129.
Taylor, B., Fujioka, K., et al. 1990a. Proc. ODp, [nit. Repts., 126, Ocean Drilling Program, College Station, TX.
Wood, D. A., Mattey, D. P., Joron, J. L., Marsh, N. G., Tamey, J., and Treuil, M. 1980. A geochemical study of 17
selected samples from basement cores recovered at Sites 447,448,449, and 451, in [nit. Repts., DSDP, 59 (L.
Kroenke, R. Scott et al., eds.), pp. 743-752, Deep Sea Drilling Project, U.S. Govt. Printing Office,
Washington, DC.
Index
Academy of Sciences Rise, 422-424, 426 East Scotia Sea (cont.)

aesthenospheric convection, 66 geochemistry, 300, 30 I
Antarctic Peninsula, 316-321, 323, 328, 329, 337 gravity, 290-292
Aoba Basin(s), 192, 193 magnetic anomalies, 284, 285,289, 292, 293, 299-
305
Balmoral Reef and Ridge, 154, 155 seafloor spreading, 289, 293, 299-305, 310-
Braemar Ridge, 153, 154 312
Bransfield Strait, 315-342 Elephant Island, 320, 323
bathymetry, 320, 321, 326 'Eua,68
earthquake seismicity, 322-324
gravity, 322-325 fault-controlled intrusions, 246, 263, 264
heat flow, 326-328 Fiji, 140, 141, 168, 169, 179
hydrothermal activity, 328 forearc rifting, 73, 123, 126-128, 262, 515-517
petrology and geochemistry, 328-337 Fukujin, 243, 248, 264
seismic reflection and refraction, 324--326 Futuna Island, 142, 189-191
tectonic setting, 316-322
Bridgeman Island, 319-324, 329-333 GulfofCalifomia, 328, 337, 481
calcium carbonate compensation depth (CCD), 392 Havre Trough, 29-62

Caroline Ridge, 240 continent-ocean transition, 33, 37
central Scotia Sea, 283, 284, 287, 289, 298, 302, 305 magnetic anomalies, 43, 44
central volcanic region: see New Zealand central vol- Ngatoro rift, 34--38
canic region petrology and geochemistry, 44, 45, 48--58
Clarence Island, 319-321, 323 rifting model, 42-44
collision, 345, 370-372 seismicity, 39, 40
Colville Ridge, 30-33, 35, 37, 39, 42 volcanism, 38, 39, 41
Conical seamount, 457, 465, 466 Hazel Holme fracture zone and Ridge, 150-152,
Constantine Bank, 142 156,158,159,167-170,179-184,186,189,
Coriolis Trough(s): see New Hebrides backarc 227
troughs heat flow, 8, II, 12,22, 159,202,220,266,326-
328,364,400,410,411,428-430
Deception Island, 320-324, 329-332 Herdman Bank, 285, 289, 290, 292, 294, 298, 299,
d'EntrecasteauxRidgeandzone, 141, 179-181, 192, 302,303-306,309,311
193,227,229 Hero Fracture Zone, 316-318, 321
Discovery Bank, 285, 288, 290, 294, 295, 298, 303, HikurangiTrough,2,3,30,31
305, 306, 309, 311 hydrothermal activity, 451-495
Donna Seamount, 105, 106 Bransfield Strait, 328
Drake Passage, 283, 284, 287, 319 Conical seamount, 457, 465, 466
Duff Ridge, 151, 152, 184, 188 East Pacific Rise, 452, 475, 477, 479-482, 484
Gulf of California (Escanaba Trough and Guay-
East China Sea, 344, 345, 349, 367 mas Basin), 481
East Pacific Rise, 452, 475, 477, 479-482,.484 hot spots, 472
East Scotia Sea, 92, 281-314 Isua, Greenland, 452
backarc basin, 282, 285, 286, 289, 293, 298--305 Tzu-Bonin (Ogasawara) arc and rifts, 456, 457,
evolution, 284--286, 289, 308--312 466-468,472-484
exploration, 282-284 Kasuga seamounts, 464, 465, 475
521
522 INDEX
hydrothennal activity (cont.) Kuril (South Okhotsk) Basin (cont.)

Lau Basin, 95, 96,118-120,455,462,463,473- stratigraphy and sediment thickness, 433-435,
476,478-484 439-445
Manus Basin, 454, 460-462, 473-476, 478, 484 Kuril island arc and trench, 421-423, 430-432, 443,
Mid-Atlantic Ridge, 480, 482 444
Mariana Trough, 264-267, 273, 455, 463-465, Kyushu--Palau Ridge: see Palau--Kyushu Ridge
473-482,484
New Hebrides backarc troughs, 202, 220, 222 Lau Basin, 63-138, 337, 407
North Fiji Basin, 162-168,454,458-460,473- hydrothermal activity, 95, 96, 118--120,455,462,
477,479-482,485 463,473-476,478-484
off-axis, 266, 463, location, 30, 32, 65, 79, 179
Okinawa Trough, 364, 456, 468-471, 477-485 Mangatolu (Kings) triple junction, 75, 88, 90, 95-
Red Sea, 452 98,110,115,260
Woodlark Basin, 453, 455, 475 ODP Sites, 65, 79, 86, 87, 90, 91, 97-103
petrology and geochemistry, 51, 52, 55, 58, 77-
Han Plain, 345, 348, 364, 370 118,337
Izu--Bonin (Ogasawara) arc and trench, 117,381- propagation, 65, 73, 79,91,93,99,102,416,417
388,392-397,401,456,457,466-468,472- rifting, 42, 43, 73, 76
484,499,500,502,504,505,512-517 seafloor spreading, 71-73
Izu-Bonin rifts, 42, 55, 73, 117,229,239,241,248, seamounts, 74, 77, 90
357,358,360,388,456 spreading centers (central, eastern, northwestern)
morphology, 73-76
James Ross Island, 328-331, 333-335 petrography, 84, 85
Japan Basin, 408-415 petrology and geochemistry, 78--83, 85, 88-92,
Japan island arc and trench, 408, 411-413 97-99,110--118
Japan Sea, 382,396,407-420,445 Lau Ridge, 65, 68-73, 96, 179
bathymetry, 408, 409 volcanism, 73
crustal structure, 410-412 Low Island, 320, 321
heat flow, 410, 411 Luzon volcanic arc, 344, 345, 371
magnetic anomalies, 410, 414
propagation, 414-416 magma chamber, 93
rifting, 411-416 Mangatolu (Kings) triple junction, 65, 75, 88, 90, 95-
seafloor spreading, 416 98, 110, 115,260
strike-slip fault, 415, 416 Manila Trench, 344
Jean-Charcot (Vot Tande) troughs: see New Hebrides mantle diapirism, 66, 67
backarc troughs Manus Basin, 453, 454, 460-462, 473-476, 478, 484
Mariana island arc, 113,238--243,255-258,260--
Kamchatka-Komandorsky Basin, 445 263,457,500--517
Kasuga seamounts, 243, 248, 261, 262, 264, 265, Mariana Trough, 117,237-279,337,407
268-270,464,465,475 age of opening, 239-241, 249
Kermadec arc, Ridge and Trench, 2, 3, 30--33, 35, basalt, 77, 78, 84, 92,101,267-271
37,39-41,71,113 bathymetry, 70, 240, 243, 250, 253-256
petrology and geochemistry, 44-52 hydrothermal activity, 264-267, 273, 455, 463-
volcanic front and submarine volcanoes, 40-42 465,473-482,484
Kinan escarpment, 385, 387 petrology, 117,267-272,337
Kinan seamount chain, 384, 386, 388, 393, 398, rifting, 117,238-249,255,268--270,272
399 seafloor spreading, 241, 246--260, 272, 273
King George Island, 320, 329-331 sedimentation, 497-517
Korea Plateau, 408, 409, 412 volcanic cross-chains, 260--262, 500
Kuril (South Okhotsk) Basin, 421-449
bathymetry, 422-426 Nankai Trough,381-385, 389-391,395, 396, 399-
crustal structure, 431-433, 438, 439 400
dredged rocks, 435-438 New Hebrides backarc (Coriolis and Jean-Charcot)
earthquake seismicity, 429-431 troughs, 142, 150--152, 177-235,498,501
gravity, 428 ferromanganese crusts, 222-226
heat flow, 428-430 geochronology, 196,219
magnetic anomalies, 427 geochemistry, 196--222, 224, 225
INDEX 523
New Hebrides backarc (Coriolis and Jean-Charcot) Okinawa Trough (cant.)

troughs (cant.) crustal structure, 365-367
history of exploration, 185-187 hydrothermal activity, 364, 456, 468-471, 473,
hydrothermal activity, 202, 220, 222 475,477-485
location, 178-180, 183-186 magnetic anomalies, 348, 349, 353, 373-377
location of volcanic samples, 197-20 I normal faults and tilted fault blocks, 346-348,
morphology and structure, 187-192 350--364,376
petrology, 194-196 opening amounts and poles, 362-372, 375, 376
New Hebrides (Vanuatu) island are, 140, 155, 168- rifting phases, 346-355, 364, 368, 370--372, 376
170,178-181,193 sediments, 351-358
New Hebrides (Vanuatu) Trench, 178-180, 193 tectonic setting, 344-350
New Zealand, 1-28, 30--33 volcanism, 348-350, 375-377
Bay of Plenty faulting, 5, 6 Okushiri Ridge, 414
central volcanic region (CVR), 1-28; see a/so overlapping spreading center (OS C), 93, 94, 143,
Taupo Volcanic Zone 144,166
crustal structure, 8-10
deformation, 12-16 Pacific-Australian plate boundary, 2, 3, 30--32, 65,
gravity, 8, 10, II 69,140,141,166,178-180
heat flow, 8, II, 12 Pagan fracture zone, 249, 251, 254
kinematics, 17 Palau-Kyushu Ridge, 238, 239, 381-390, 392-402,
location, 3, 4 506
stratigraphy, 7 Parece Vela Basin, 238, 239, 383, 386, 397,498,
north island shear belt, 3, 4, 16 501,502,506,512
volcanic arcs, 3-5 Peggy Ridge, 65, 72-77, 97, 103-105, 118,455, 463
volcanic front, 4, 12, 17, 18 Penguin Island, 320, 321, 330, 332, 336, 337
Ngatoro rift and ridge: see Havre Trough Philippine Sea plate, 238, 239, 344-346, 382
Nikko seamount, 241, 243 Phoenix plate, 316-318, 323, 337
Nishi-8hichito Ridge, 381, 383-385, 388, 389, 393- propagation and propagating rift, 414-418
396, 399, 40 I Japan Basin and Sea, 414-416
Niuafo'ou Island, 65, 77, 90, 106, 107 Lau Basin, 65, 73, 79, 91,93,99, 102,416,417
North Fiji Basin (Plateau), 139-175, 179, 182, 183, Mariana Trough, 248, 258, 262, 418
196,416 North Fiji Basin, 143, 144, 148, 149, 166,416
bathymetry, 141-143 Shikoku Basin, 396-398, 416
central spreading ridge (CSR), 143-147, 158, 165, South China Sea, 416
166, 168-170
eastern basin and West Fiji Ridge, 142, 147-149, remnant are, 66, 238, 289, 298, 302, 400,515
158, 159, 166-170 rift-flank flexural uplift, 192
evolution, 165-170, 182, 183 ridge subduction and ridge-trench collision, 288,
geochemistry, 160--162, 166, 167 298,302,303-306,310,317,318
heat flow, 159 Rochambeau Bank, 105, 106
history of exploration, 139-141 rollback: see trench rollback
hydrothermal activity, 162-168,454,458-460, Ryukyu Islands and Trench, 345, 348-350, 364, 369,
473-477,479-482,485 371,374-376,399
magnetic anomaly lineations, 155-157
propagation, 143, 144, 148, 149, 166,416 Sakhalin, 422, 423, 427
northeastern basin, 153-155, 167-170 Sandwich plate, 283-286, 305, 316
northwestern basin, 142, 150--153, 167-170 Scotia plate, 283-286, 316, 323
seismicity, 157-158 Scotia Ridge, north and south, 283, 284, 289, 292,
western basin, 141-143, 167-170 316,323
North Fiji fracture zone, 153-155, 158, 168, 169, Scotia Sea, 283, 285, 288, 316, 319; see a/so East
179,454 Scotia Sea and central Scotia Sea
sedimentation, 120--122,497-520
Ogasawara Plateau, 239, 240 serpentinized peridotite, 307, 308, 310, 457
Okhotsk Sea, 421-446 Shackleton Fracture Zone, 316-320, 323
Okinawa platelet, 350, 361, 368-371, 374, 376 Shichito-lwojima Ridge: see Izu-Bonin arc
Okinawa Trough, 229, 239, 343-380 Shikoku Basin, 381-405, 416
bathymetry, 344, 345, 356-361, 372, 373 bathymetry, 384-389
524 INDEX
Shikoku Basin (cont.) Tofua are, 50, 52, 65, 67-69,

exploration history, 381-384 age, 67,110
gravity, 389, 390 geochemistry, 69, 91, 94, 95,113--117
magnetic anomalies, 394-398 Tonga arc, platform, Ridge and Trench, 65, 67-71,
petrology, 393 96,113, 178, 179
propagation, 396-398, 416 trench rollback, 66, 67, 240, 286, 287, 289, 306, 318,
stratigraphy, 389--392 319,337
slab window, 318 Tshima (Ullung) Basin, 408-410
South American-Antarctic plate boundary, 283--288,
296,302,303,316 Valu Fa Ridge, 65
South China Sea, 344, 407, 416-418 hydrothermal activity, 455, 462, 463, 473-476,
South Georgia, 283, 285, 290, 292, 294, 309 478-484
South Pandora Ridge, 150--152, 156, 158, 179, 182 morphology, 73-75, 93
South Sandwich Island Arc, 283, 290, 292-296, 306 petrology and geochemistry, 90, 91, 93--95, 115
South Sandwich Trench and forearc, 283--288, 290, spreading rate, 93
292,294,296-298,302,304-308 VAMP area, 349, 364
South Shetland Islands, Plate, and Trench, 283--285, Vanikoro-Torres Basin, 192-194
316-325,328-330,333,337 Vanuatu: see New Hebrides
Shikoku Basin, 92, 238, 239 Vening Meinesz fracture zone, 34
strike-slip fault, 415-418 Vitiaz arc and Trench (Trough), 140, 141, 153, 168,
subduction, 66, 71,267,286,287,310,317-319, 169,179,184
337,344-346,350,399-401,421,429-431, volcanic cross-chains
444 1m-Bonin arc (Nishi--Shichito Ridge), 385, 388
submarine canyons, slumps and slides, 424 Kuril are, 423, 424
Sumisu Rift, 73, 92,100, 117,229,357,358,456, Mariana arc, 260--262, 500
468,473,499,500,512-517 South Sandwich are, 295, 296
volcanic massive sulfide deposits (VMSDs),
Taiwan, 344-346, 370--372 Kuroko-, Cyprus-, and Besshi-types, 451,
Taiwan--Sinzi folded belt, 345, 346 453,469,473-475,477,478,482-485
Taupo Volcanic Zone (TVZ), 1-28, 30--35; see also Volcano Islands, 238, 240, 241
New Zealand central volcanic region
calderas and ignimbrites, 4,6,7,21-23 West Mariana Ridge, 238-243, 255, 383, 501-503,
chronology of volcanism, 6, 7 506,511
faulting (Taupo fault belt), 14-16, 23 West Philippine Basin, 345, 383, 389, 399, 502
heat source, 22 Woodlark Basin, 453, 455, 475
location, 3, 4
petrology and geochemistry, 6, 7, 17-22,45,48- Yamato Basin, 408-414
58 Yamato Rise, 408, 409, 412
tectonic erosion, 298
Tikopia and West Tikopia Ridges, 150--152, 156, Zenisu Ridge, 383, 388
184, 186,220 Zephyr Shoal, 105, 107

Backarc Basins - Tectonics and Magmatism

Uploaded by

Copyright:

Available Formats

Backarc Basins - Tectonics and Magmatism

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Backarc Basins - Tectonics and Magmatism

Uploaded by

Copyright:

Available Formats

Tectonics and Magmatism

Springer Science+Business Media, LLC

Backarc basins tectonics and magmatism / edited by Brian Taylor.

© 1995 Springer Science+Business Media New York

All rights reserved

Andrey A. Andreyev • Institute of Marine Geology and Geophysics, Yuzhno-Sakha-

Jean-Marie Auzende • IFREMER/CB, 29280 Plouzane, France. Present address:

Peter F. Barker • British Antarctic Survey, National Environment Research Council,

J. W. Cole • Department of Geology, University of Canterbury, Christchurch, New

Kathleen A. Devaney • Department of Geological Sciences, University of Texas at El

Jean-Philippe Eissen • Antenne ORSTOM-IFREMERlCB, 29280 Plouzane, France

Patricia Fryer • Hawaii ~nstitute of Geophysics and Planetology, School of Ocean

J. A. Gamble • Department of Geology, Research School of Earth Sciences, Victoria

Martine Gerard • ORSTOM, 93143 Bondy, France

Helios S. Gnibidenko • Shirshov Institute of Oceanology, Russian Academy of Sci-

Elena V. Gretskaya • Institute of Marine Geology and Geophysics, Yuzhno-Sakha-

James W. Hawkins, Jr. • Geological Research Division, Scripps Institution of Ocean-

Thomas W. C. Hilde • Geodynamics Research Institute, Texas A&M University,

Shu-Kun Hsu • IFREMER, Centre de Brest, 29280 Plouzane, France

Jun-ichiro Ishibashi • Laboratory for Earthquake Chemistry, Faculty of Science, Uni-

Jean-Louis Joron • Groupe des Sciences de la Terre, Laboratoire Pierre-SUe, CEN

Shigeru Kasuga • Hydrographic Department, Maritime Safety Agency, Chuo-ku,

Randall A. Keller • College of Oceanic and Atmospheric Sciences, Oregon State

Lawrence A. Lawver • Institute for Geophysics, University of Texas at Austin, Austin,

Char-Shine Liu • Institute of Oceanography, National Taiwan University, Taipei,

Patrick Maillet • ORSTOM, Centre de Brest-GDR "GEDO" 910, 29280 Plouzane,

Kathleen M. Marsaglia • Department of Geological Sciences, University of Texas

Setsuya Nakada • Kyushu University, Fukuoka 812, Japan

Kyoko Okino • Hydrographic Department, Maritime Safety Agency, Chuo-ku, Tokyo

Bernard Pelletier • ORSTOM, Noumea, New Caledonia

Alain Person • Laboratoire de Geologie des Bassins Sedimentaires, Universite Pierre

Richard C. Price • La Trobe University, Bundoora, Victoria 3083, Australia

Etienne Ruellan • CNRS Sophia-Antipolis, 06560 Valbonne, France

Chuen-Tien Shyu • Institute of Oceanography, National Taiwan University, Taipei,

Jean-Claude Sibuet • IFREMER, Centre de Brest, 29280 Plouzane, France

T. A. Stern • Research School of Earth Sciences, Victoria University of Wellington,

Jorge A. Strelin • Departamento de Ciencias de la Tierra, Instituto Antartico Argen-

Kensaku Tamaki • Ocean Research Institute, University of Tokyo, Nakano-ku, Tokyo

Tetsuro Urabe • Geological Survey of Japan, Tsukuba 305, Japan

I. C. Wright • New Zealand Oceanographic Institute, National Institute of Water and

CAUSE AND LOCATION

2. Geologic Setting of the Lau Basin ................................... 67

8. Discussion of the Isotope Data ..................................... 113

ABSTRACT ....................................................... 139

ABSTRACT ....................................................... 177

ABSTRACT ....................................................... 237

5.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 267

ABSTRACT ....................................................... 281

ABSTRACT ....................................................... 315

ABSTRACT ....................................................... 343

4. Kinematics of the Okinawa Trough Opening ........................... 362

ABSTRACT ....................................................... 381

ABSTRACT ....................................................... 407

ABSTRACT ....................................................... 421

ABSTRACT ....................................................... 451

ABSTRACT ....................................................... 497

7.4. Implications of Forearc versus Backarc Rifting on Sedimentary Models 516

Taupo Volcanic Zone and Central

J. W. Cole • Department of Geology, University of Canterbury, Christchurch, New Zealand. D. J.