Se 10 1181 2019

Solid Earth, 10, 1181–1217, 2019
https://doi.org/10.5194/se-10-1181-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.
A revised map of volcanic units in the Oman ophiolite: insights into

the architecture of an oceanic proto-arc volcanic sequence
Thomas M. Belgrano1 , Larryn W. Diamond1 , Yves Vogt1 , Andrea R. Biedermann1,2 , Samuel A. Gilgen1 , and
Khalid Al-Tobi3
1 Institute
of Geological Sciences, University of Bern, Baltzerstrasse 3, 3012 Bern, Switzerland
2 Institute
for Rock Magnetism, University of Minnesota, 100 Union St SE, 55455 Minneapolis, USA
3 National Earth Secrets Co., P.O. Box 1242, PC 130 Athaibah, Muscat, Sultanate of Oman
Correspondence: Thomas M. Belgrano ([email protected])
Received: 30 March 2019 – Discussion started: 18 April 2019

Revised: 13 June 2019 – Accepted: 3 July 2019 – Published: 29 July 2019
Abstract. Numerous studies have revealed genetic similari- 45 vol % and ubiquitously cover the underlying axial crust.
ties between Tethyan ophiolites and oceanic “proto-arc” se- Highly depleted boninitic members of the Lasail unit locally
quences formed above nascent subduction zones. The Semail occur within and directly atop the axial sequence, marking an
ophiolite (Oman–U.A.E.) in particular can be viewed as an earlier onset of boninitic magmatism than previously known
analogue for this proto-arc crust. Though proto-arc magma- for the ophiolite. The vast majority of the Semail boninites,
tism and the mechanisms of subduction initiation are of great however, belong to the Boninitic Alley unit and occur as dis-
interest, insight is difficult to gain from drilling and limited continuous accumulations up to 2 km thick at the top of the
surface outcrops in marine settings. In contrast, the 3–5 km ophiolite sequence and constitute ∼ 15 vol % of the upper
thick upper-crustal succession of the Semail ophiolite, which crust. The new map provides a basis for targeted exploration
is exposed in an oblique cross section, presents an opportu- of the gold-bearing VMS deposits hosted by these boninites.
nity to assess the architecture and volumes of different vol- The thickest boninite accumulations occur in the Fizh block,
canic rocks that form during the proto-arc stage. To deter- where magma ascent occurred along crustal-scale faults that
mine the distribution of the volcanic rocks and to aid explo- are connected to shear zones in the underlying mantle rocks,
ration for the volcanogenic massive sulfide (VMS) deposits which in turn are associated with economic chromitite de-
that they host, we have remapped the volcanic units of the posits. Locating major boninite feeder zones may thus be an
Semail ophiolite by integrating new field observations, geo- indirect means to explore for chromitites in the underlying
chemical analyses, and geophysical interpretations with pre- mantle.
existing geological maps. By linking the major-element com-
positions of the volcanic units to rock magnetic properties,
we were able to use aeromagnetic data to infer the extension
of each outcropping unit below sedimentary cover, resulting 1 Introduction
in a new map showing 2100 km2 of upper-crustal bedrock.
Whereas earlier maps distinguished two main volcanos- A growing body of geochemical and geochronological evi-
tratigraphic units, we have distinguished four, recording dence indicates that the crustal sequences of many Tethyan
the progression from early spreading-axis basalts (Geo- ophiolites – including the Oman–U.A.E. (Semail) example
times), through axial to off-axial depleted basalts (Lasail), to – formed in response to upper-plate extension during the
post-axial tholeiites (Tholeiitic Alley), and finally boninites initiation of lower-plate subduction and roll back (Belgrano
(Boninitic Alley). Geotimes (“Phase 1”) axial dykes and and Diamond, 2019; Dilek et al., 2007; Dilek and Furnes,
lavas make up ∼ 55 vol % of the Semail upper crust, whereas 2009; Guilmette et al., 2018; Ishikawa et al., 2002; MacLeod
post-axial (“Phase 2”) lavas constitute the remaining ∼ et al., 2013; Metcalf and Shervais, 2008; Rollinson, 2009;
Rollinson and Adetunji, 2015; Shafaii Moghadam et al.,
Published by Copernicus Publications on behalf of the European Geosciences Union.

1182 T. M. Belgrano et al.: A revised map of volcanic units in the Oman ophiolite
2013; Stern and Bloomer, 1992). These findings have led

to reinterpretation of many ophiolites as counterparts to
difficult-to-access magmatic records of subduction initiation
preserved in submarine forearcs and arc basement (“proto-
arcs”; Stern et al., 2012; Stern and Bloomer, 1992; What-
tam and Stern, 2011). The significance of the Semail ophio-
lite as a proto-arc analogue has recently been highlighted by
tectonic modelling and metamorphic geochronological stud-
ies, indicating that the Semail outcrops may constitute a rare
example of proto-arc crust produced by the tectonically in-
duced mode of subduction initiation rather than by sponta-
neous gravitational collapse (Duretz et al., 2016; Guilmette
et al., 2018; Stern and Gerya, 2017).
Though apparently dissimilar in their modes of onset, both
the mid-Cretaceous Tethyan and Eocene west Pacific sub-
duction initiation events appear to have formed vast tracts
of proto-arc crust thousands of kilometres along strike and
hundreds of kilometres wide, all within a few million years
(Fig. 1; Arculus et al., 2015; Hickey-Vargas et al., 2018;
Ishizuka et al., 2011; Meffre et al., 2012; Moghadam et al.,
2010; Reagan et al., 2019; Shafaii Moghadam et al., 2013).
As a potential remnant of this crust, the Semail ophiolite, ex-
posed over ∼ 20 000 km2 , presents an excellent opportunity
to assess the bulk composition and lateral continuity of these Figure 1. Map of the Semail ophiolite structural blocks simpli-
swathes and their underlying mantle. fied after Nicolas et al. (2000) with the additional differentiation of
the structurally separate western (W.) and eastern (E.) Fizh blocks,
Differential uplift and erosion of the Semail ophiolite has
showing the area mapped in this study. Inset: map of Tethyan ophi-
exposed a 300 km long strip of volcanic rocks that dip 20– olites adapted after Dilek et al. (2007) with U–Pb zircon ages of ini-
50◦ E along the north-eastern margin of the Oman mountains tial accretion indicated for the Albian–Turonian ophiolite chain of
(Fig. 1). This simple structure, the excellent exposure, and 1 Troodos, 2 Kizildag, 3 Nain, 4 Deshir, and 5 Semail after Shafaii
the wealth of previous petrogenetic studies means that the Moghadam et al. (2013) and references therein.
distribution and approximate volume of rocks produced by
each volcanic episode should be evident in map view.
Four regionally extensive mappable volcanic units have overprint, which is debated (de Graaff et al., 2019). Further-
been recognised in the Semail ophiolite (Geotimes, Lasail, more, a map differentiating the ophiolite’s boninites facili-
Alley, and Boninitic Alley), with an increasingly pronounced tates exploration for the gold-bearing volcanogenic massive
subduction influence recorded by each successively erupted sulfide (VMS) ore deposits hosted by these lavas (Gilgen et
unit (Alabaster et al., 1982; Belgrano and Diamond, 2019; al., 2014). Finally, a map perspective of the Semail upper
Gilgen et al., 2014; Ishikawa et al., 2002; Kusano et al., crust provides an areally extensive and well-studied compar-
2012, 2014, 2017). However, as previous regional mapping ison to drilled sections from the Izu–Bonin–Mariana (IBM)
was carried out prior to the recognition of boninites, and the proto-arc basement (Belgrano and Diamond, 2019; Ishikawa
Lasail unit was either lumped in with Geotimes (e.g. BME, et al., 2002; MacLeod et al., 2013; Pearce et al., 1992; Rea-
1987a; Umino et al., 1990) or Alley (e.g. BRGM, 1993a), gan et al., 2013, 2017; Shervais et al., 2019; Stern et al.,
the existing regional maps distinguish only two regionally 2012).
distributed units: the early (V1) and late (V2) extrusives. Up- In the following we review the previous work on the
dating the geological maps of the northern ophiolite to the Semail volcanic units and document the combined field,
four-unit volcanostratigraphy is the aim of the present study. geochemical, and aeromagnetic approach used to distin-
This update is necessary to incorporate the findings of nu- guish them during mapping. We then present a new map of
merous detailed investigations of isolated volcanic sections the northern Semail volcanics, displaying both outcrop and
(e.g. Einaudi et al., 2003; Godard et al., 2006; Kusano et sediment-covered bedrock occurrences, split for the first time
al., 2012, 2014). Variations in volcanostratigraphy can then into the major basal, arc-like, and boninitic volcanic groups.
be used to test previously proposed along-strike variations This will serve as a renewed basis for volcanological studies
in subduction-zone influence (e.g. Goodenough et al., 2010; and for ore exploration in the Semail ophiolite. The volcanos-
Python et al., 2008) and hydrothermal processes (Alabaster et tratigraphic implications of the new map are then discussed
al., 1982; Gilgen et al., 2016; MacLeod and Rothery, 1992), and briefly compared to the IBM proto-arc record.
as well as the overall significance of the “Phase 2” magmatic
Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

T. M. Belgrano et al.: A revised map of volcanic units in the Oman ophiolite 1183
1.1 Approach 1.3 Geology and petrogenesis of the Semail

upper-crustal units
Improvement of the excellent pre-existing regional maps re-
quired a multidisciplinary approach, which is summarised The upper crust of the Semail ophiolite consists of a con-
below: tinuous sheeted dyke complex (SDC) 1–1.5 km thick which
is conformably overlain by a 1–3 km thick sequence of ex-
1. Previously published geological maps (listed in Sect. 6) trusives (Lippard et al., 1986). Though the Semail magmatic
and locations of geochemically analysed lava samples history can be broadly split into two magmatic phases (Phase
were georeferenced, digitised, and a correspondence 1 and Phase 2; Goodenough et al., 2014), geochemical indi-
was established between their volcanostratigraphies and cators of subduction-zone input gradually increase upwards
that employed here (Table 1). through the volcanic stratigraphy (A’Shaikh et al., 2005; Al-
2. Field mapping was conducted between 2014 and 2019 abaster et al., 1982; Belgrano and Diamond, 2019; Ernewein
by teams of two to four geologists during six field cam- et al., 1988; Godard et al., 2003; Kusano et al., 2012, 2014,
paigns each of 1-month duration. We collected 190 new 2017; Lippard et al., 1986; Umino et al., 1990). This progres-
lava samples at strategic locations and analysed their el- sion is manifested by the more-or-less sequential eruption
emental compositions. This allowed us to assign each of four regionally distributed and mappable volcanic units
sample to a volcanostratigraphic unit by comparison (Fig. 2):
with compositional fields defined by previously pub- 1. Geotimes. This unit comprises basal basalts and basaltic
lished stratigraphically confirmed analyses. These as- andesites produced during the axial-spreading stage of
signments in turn allowed the field discrimination cri- ophiolite formation (Phase 1). The Geotimes lavas were
teria for the lava units to be iteratively refined during fed and contemporaneously emplaced with the comag-
each campaign. matic SDC predominantly between 96.5 and 95.5 Ma
(Alabaster et al., 1982; Rioux et al., 2013; Umino et
3. The bulk magnetic and magnetic–mineralogical proper-
al., 1990). The petrogenetic affinity of Geotimes is dis-
ties of a subset of samples from each unit were analysed
puted between mid-ocean ridge basalts (MORB), back-
to aid interpretation of the regional aeromagnetic sur-
arc basin basalts, and forearc basalts (Alabaster et al.,
vey of the volcanic sequence (Isles and Witham, 1993).
1982; Ernewein et al., 1988; Godard et al., 2006; Ku-
This interpretation allowed us to infer which volcanos-
sano et al., 2012; MacLeod et al., 2013). In comparison
tratigraphic units are buried beneath the supra-ophiolitic
to MORB, the geochemical composition of Geotimes
sedimentary cover.
is marked by negative Nb–Ta anomalies and by major-
4. These datasets were integrated into a geological map element fractionation trends indicative of elevated water
covering ∼ 950 km2 of exposed upper-ophiolitic crust contents (Alabaster et al., 1982; MacLeod et al., 2013).
and showing a further ∼ 1200 km2 of buried volcanic
2. Lasail. This is dominantly a low-Ti primitive basaltic
bedrock.
unit intercalated with, but mostly overlying, the Geo-
times unit. It occurs as discontinuous off-axis accumu-
1.2 Nomenclature of volcanostratigraphic units lations (Alabaster et al., 1982; Belgrano and Diamond,
2019; Kusano et al., 2012). The basal intercalations with
Different nomenclatures for the volcanostratigraphy have Geotimes demonstrate a clear stratigraphic association
been employed by different research groups (Table 1). We between Lasail and Geotimes/V1 (Belgrano and Dia-
use the names assigned by the seminal studies of the volcanic mond, 2019; Kusano et al., 2012; Umino et al., 1990).
sequence (Alabaster et al., 1980, 1982), i.e. Geotimes, Lasail However, Lasail is closer to Alley/V2 in terms of incom-
and Alley, with the addition of Boninitic Alley from Gilgen et patible trace-element depletion (Alabaster et al., 1982;
al. (2014). As well as being easy to distinguish, these names Belgrano and Diamond, 2019; Ernewein et al., 1988;
carry no connotations of chronological order, and hence they Gilgen et al., 2014). This mismatch has led to varying
remain practical when mapping intercalated sequences. For interpretations as to whether the Lasail unit belongs to
clarity in the present article, when referring to previous work Phase 1 or Phase 2. Lasail’s rare-earth element deple-
we append to our names the term used in the cited study tion can be modelled by elevated (∼ 30 wt %) partial
(e.g. Geotimes/LV1). Notwithstanding these differences in melting of the same depleted-MORB mantle source that
nomenclature and some disagreement regarding petrogenetic produced Geotimes (Godard et al., 2006). Such high-
groupings, these unit divisions are clearly defined by field degree melting fits with the distinctly primitive compo-
observations and are widely accepted (Alabaster et al., 1982; sition of the Lasail lavas. Lasail’s elevated Th/Nb ra-
Gilgen et al., 2014; Godard et al., 2006; Kusano et al., 2017; tios indicate that this melting was assisted by additions
MacLeod et al., 2013; Umino et al., 1990). of a high-temperature subduction component (Belgrano
and Diamond, 2019). Thus, high-degree melting of a
www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Table 1. Comparison of different naming schemes for the Semail volcanostratigraphy and the closest-equivalent rock types identified in the
Izu–Bonin–Mariana (IBM) proto-arc record.
Alabaster BME (1987a) JICA (2000) Godard Kusano Gilgen et al. (2014); Goodenough Rock Similar IBM
et al. (1980) & BGRM (1986a) et al. (2003) et al. (2012, 2014) this study et al. (2014) types proto-arc units*
Salahi SE3 V3 V3 V3 Salahi – Alkali –
basalt
– SE2 V2 V2 Type II UV2 Boninitic Phase 2 Boninite, Low-Si
Alley high-Mg andesite boninite series
Cpx-phyric SE2P – – – – Phase 2 Clinopyroxene- –
phyric basalt
Alley SE2 V2 V2 Type I Tholeiitic Alley Phase 2 Basalt, high-Mg Forearc
andesite, basalt
dacite, rhyolite series
Lasail SE2, SE1 V1 type 2 – UV1 Lasail Phase 2 Primitive, low-Ti Primitive forearc
basalt basalt
Geotimes SE1 V1 type 1 V1 MV1 Geotimes Phase 1 Basalt, basaltic No exact equivalent:
LV1 andesite, andesite high-Ti/V axial basalts
– means not included in that study; ∗ Reagan et al. (2017) and Shervais et al. (2019), see Sect. 8 for comparison.
supra-subduction MORB-mantle source, triggered by

localised infiltration of a slab-derived fluid, succinctly
explains Lasail’s geochemical features as well as its dif-
ferences with respect to the overlying Alley lavas (Bel-
grano and Diamond, 2019).
3. Tholeiitic Alley. This is part of the original “Alley”

volcanic group formalised by Smewing et al. (1977)
and Alabaster et al. (1982). The group includes
tholeiite-series lavas that we differentiate here from
their boninitic counterparts. The Tholeiitic Alley unit
spans a fractionation series from basalt through high-
magnesium andesite to rhyolite, and it is generally ac-
cepted as belonging to Phase 2 of the Semail magmatic
history (Alabaster et al., 1982; Gilgen et al., 2014; Ku-
sano et al., 2014, 2017; Umino et al., 1990). The mod-
erate incompatible-element depletion and fluid-mobile
element enrichment of Tholeiitic Alley/LV2 lavas and
glasses are interpreted as the result of flux melting of
the axial mantle source residue by a slab-derived hy-
drous fluid (Alabaster et al., 1982; Kusano et al., 2014,
2017).
Figure 2. Semail ophiolite volcanostratigraphy with capping
pelagic sediments (Suhaylah Formation) and sheeted dyke complex
4. Boninitic Alley. High-Ca boninitic lavas were first dis-
(SDC), with approximate stratigraphic thicknesses revised accord-
covered at the top of the stratigraphy in the Wadi Jizi
ing to field observations in this study after Alabaster et al. (1982),
area (Ishikawa et al., 2002). Kusano et al. (2014) then Gilgen et al. (2014), and Belgrano and Diamond (2019).
showed that boninites occur as a ∼ 140 m thick package
(termed UV2) overlying and intercalated with Tholei-
itic Alley/LV2 lavas along Wadi Bidi (Hilti block). Re- tions, their source composition and geochemical frac-
gional sampling by Gilgen et al. (2014) and Kusano et tionation trends are clearly distinct from one another
al. (2017) further showed that boninites occur at numer- (Gilgen et al., 2014; Kusano et al., 2014), thereby jus-
ous locations throughout the northern ophiolite. How- tifying their definition as distinct volcanostratigraphic
ever, the volume and continuity of the boninites was un- units.
known prior to this study. While Boninitic Alley and
Tholeiitic Alley have previously been grouped on the Other locally distributed volcanic units were also de-
basis of their appearance and stratigraphic intercala- scribed by Alabaster et al. (1982). The Clinopyroxene-phyric

unit occurs locally around Wadi Jizi and is compositionally 1.4.2 Aeromagnetic map
intermediate between Tholeiitic Alley and Boninitic Alley
(Gilgen et al., 2014). Owing to its small extent and interme- In addition to the above geological maps, we used the Bati-
diate nature, we have not included it in our map. The mildly nah aeromagnetic survey performed in 1992 by the World
alkaline, probably obduction-related Salahi volcanics (V3; Geoscience Cooperation (Isles and Witham, 1993) and made
Alabaster et al., 1982) have so far been found only in the available by the Oman Public Authority for Mining, Mus-
Wadi Salahi area (Hilti block), where their extent is well- cat. The survey was flown with a line spacing of 200 m at a
defined and where they are separated from the underlying mean terrain clearance of 80 m. This corresponds to an ap-
Alley group volcanics by either an eroded base or by sev- proximate resolution of 80 m, where magnetic bodies that
eral metres of pelagic sediments (BME, 1987b; Kusano et are less than 80 m wide or that are separated by less than
al., 2014; Umino, 2012). Our map displays the previously 80 m will be detected, but not necessarily resolved (Flint et
known extent of the Salahi lavas without modification. al., 1999). The acquired data were processed by reduction-to-
pole (RTP) transformation prior to distribution but details of
1.4 Previous works and scope of new map this transformation are unfortunately unavailable. Such RTP
transformations are performed to compute the anomalies that
1.4.1 Previous geological maps would be generated by the same source bodies in a verti-
cally orientated magnetic field, thus better situating anoma-
The first regional-scale geological map of the Semail ophio-
lies over their sources and facilitating geological interpreta-
lite with a differentiated lower and upper extrusive sequence
tion (Blakely, 1995; Clark, 2014).
was compiled by Lippard et al. (1986) at a scale of 1 :
250 000. This work was expanded upon by the 1 : 50 000 and
1.4.3 Scope of new map
1 : 100 000 maps produced by Japanese and French teams to-
gether with the Ministry of Petroleum and Minerals (BME,
The present study focuses exclusively on the northern ophio-
1987a, b, c; BRGM, 1986a, b, 1993a, b). This collection cov-
lite, where the vast majority of the ophiolite’s volcanic rocks
ers our entire mapping area and is referred to hereafter as the
occur (Fig. 1). Volcanic rocks also occur outside of this
“regional map set”. The regional map set displays the distri-
mapped area: as a continuation of the Aswad block into the
bution of an essentially two-unit extrusive stratigraphy (Se-
U.A.E., as discontinuous outcrops along the western flank of
mail Extrusive 1 and Extrusive 2 for lower and upper, respec-
the northern ophiolite, and in the south-eastern Semail and
tively), with the addition of the Salahi unit (Semail Extrusive
Tayin blocks. These occurrences fell beyond the scope of
3) as well as dyke-cut and felsic subunits. The different re-
this study but they would be amenable to mapping using the
gional maps generally agree well along their margins, with
methodology outlined here.
the exception of a mismatch at the gabbro–SDC contact be-
As our mapping was conducted on GPS-guided digital
tween the Fizh (BRGM, 1993b) and Wadi Bani Umar (BME,
tablets and the coverage of pre-existing map information was
1987c) map sheets. Here, we used the more recent field map
variable, the detail achieved in our map varies somewhat
of Adachi and Miyashita (2003) to correct the contact trace.
over the mapping area and cannot be defined by a traditional
The map presented in this study primarily builds on this re-
scale. In the northern blocks, the detail of mapping is ap-
gional map set. In addition, numerous local maps have since
proximately equivalent to the 1 : 50 000 scale regional base
been published together with articles and reports on the ex-
maps (e.g. BME, 1987). However, in the Sarami and Hay-
trusives. In particular, the detailed field maps around VMS
layn blocks the coarser scale of the base maps (1 : 100 000;
prospects published by JICA (2000, 2002) follow a similar
BRGM, 1986a, b) is inherited to some extent, but improved
stratigraphic scheme to the one used in the present study, and
to approximately 1 : 50 000 equivalence in areas where we
those maps provided useful constraints for our map in the
focussed our field mapping, and it was generally improved
Sarami, Haylayn, and Yanqul areas.
by redigitisation with modern satellite imagery. Use of the
The outcrop outlines on the existing regional map set are
digital map file provided in the Supplement to this article al-
commonly offset from their true geodetic positions by 50–
lows a wide range of scales to be displayed.
100 m. To remedy this, we retraced the outcrop outlines
To provide geological context below the volcanic se-
of the sheeted dyke complex and extrusives using 1 m res-
quence, the sheeted dyke complex (SDC) and upper-crustal
olution OrbView-3 satellite images made available by the
intrusions (above the base of the SDC) were traced from the
United States Geological Survey through the Earth Explorer
regional map set and where necessary their outlines were
platform. We verified that the registration accuracy of these
adapted to match new observations and satellite imagery. Be-
images is within 20 m by comparison with GPS positions
low the SDC, the lower crustal units were grouped and sim-
recorded at outcrops and road intersections.
plified based on the regional map set, as was the mantle sec-
tion. Selected dyke swarms and umbers at the unit contacts
are newly drawn or reproduced on our map but we have omit-
ted many other small-scale umbers, dykes, veins, and geo-

graphical features that are shown in the regional map set. drous trace-element composition to be calculated by closure
Accordingly, we recommend using the new map in tandem to 100 wt % oxides and trace elements. Accuracy was moni-
with the existing regional map set. tored through measurements of basalt standard BRP-1 (Cotta
The top of the volcanic sequence marked in our new map and Enzweiler, 2008) and highly depleted komatiite stan-
corresponds to these cases: the true stratigraphic top, which dard OKUM (Kane et al., 2007) prepared using an identical
is conformably overlain by pelagic sediments of the Suhay- method to the samples.
lah Formation (Robertson and Woodcock, 1983b); or an Bulk magnetic susceptibility was determined by two meth-
eroded surface that is overlain by post-Suhaylah sediments; ods: a handheld Exploranium® KT-5 kappameter and a desk-
or faulted contact with overthrust sheets of the Batinah com- top Magnon kappameter at the Institute for Rock Magnetism
plex (Woodcock and Robertson, 1982). To mark the top of (IRM), University of Minnesota. The KT-5 measurements
the volcanic sequence where it crops out, we adapted the out- were made on flat-sawn sample surfaces. The Magnon mea-
lines of the Suhaylah Formation and the volcanic conglom- surements were made on 25 mm diameter rock cores 20–
erates of the Zabyat Formation (Robertson and Woodcock, 30 mm in length. The comparability of the two methods is in-
1983b) from the regional map set. All other supra-ophiolitic dicated by a good correlation (1 : 0.97, R 2 = 0.97) between
sediments (e.g. olistostrome, gravels) that appear on previous measurements of the same samples (Fig. S1). Natural rema-
maps are undifferentiated in our map. nent magnetisation (NRM) was determined on the same rock
Where permitted by the aeromagnetic survey, we inferred cores on a 2G Enterprises 760 RF SQUID superconducting
the occurrence of the upper-crustal ophiolitic units, together rock magnetometer at the IRM.
with the volcaniclastic Zabyat Formation, under all other High-temperature magnetic susceptibility was measured
post-volcanic sedimentary cover (Suhaylah sediments, olis- on rock powders (0.2–0.5 g) in air with an AGICO Kap-
tostrome, gravels). The areas of the map showing such infer- pabridge at the IRM. A magnetic KLY-2 susceptometer was
ences are to be viewed as a best estimate of the identity of used, operating at a frequency of 920 Hz from room tem-
the bedrock units at the upper surface of the Semail igneous perature to 700 ◦ C and back to room temperature in steps
sequence. No attempt was made to infer the presence of vol- of ∼ 3 ◦ C. Initial attempts in an argon atmosphere resulted
canic units beneath surficial fault traces. in significant artificial magnetite production during the ex-
periment. Low-temperature measurements were conducted
at the IRM on 0.2–0.5 g rock chips using a Quantum Designs
2 Analytical methods magnetic properties measurement system (MPMS) of satura-
tion magnetisation as a function of temperature (field 2.5 T),
Major-element compositions of whole-rock samples were field-cooled (FC) and zero-field-cooled (ZFC) remanences,
determined by X-ray fluorescence (XRF) at ETH Zurich us- and temperature cycling of a room temperature isothermal
ing the same method as Gilgen et al. (2014, 2016). Electron remanent magnetisation (RT-IRM) down to 20 K and back to
microprobe (EMP) analyses were conducted on a Jeol JXA- room temperature.
8200 EMP at the University of Bern using the same stan-
dardisation described in Belgrano and Diamond (2019). The
major-element compositions of volcanic glass samples were 3 Field descriptions
determined by EMP with a beam voltage of 15 kV, a beam
current of 10 nA, and a beam diameter of 10 µm. This low The field characteristics of the Semail volcanic units were re-
current density (0.13 nA µm−2 ) was chosen to minimise Na cently reviewed by Gilgen et al. (2014). Here we summarise
loss during measurement (Morgan and London, 2005) and and expand on the key field features which aid in unit dis-
current across the specimen was monitored as stable during crimination, adding our own observations of potentially mis-
measurement, indicating that minimal diffusion of Na oc- leading exceptions and complicating features.
curred. The compositions of igneous clinopyroxenes were
determined by EMP with a beam voltage of 20 kV, a beam 3.1 Geotimes
current of 15 nA, and a beam diameter of 3 µm.
Trace-element analyses were conducted on a subset The Geotimes unit is typically made up of monotonous se-
of samples from each unit using the pressed-powder- quences of basaltic to basaltic-andesitic pillow lavas (Fig. 3a)
pellet laser-ablation inductively coupled plasma spectrome- with occasional columnar-jointed massive flows up to 20 m
try (PPP-LA-ICP-MS) method described by Peters and Pet- thick and pillow breccias. Pillowed outcrops weather to
tke (2016). This was performed using a GeoLas-Pro 193 nm colours of hematitic red to dark brownish-grey and tend to
ArF excimer laser system in combination with an ELAN be more topographically prominent than the other volcanic
DRC-e quadrupole mass spectrometer at the University of units. Magnetism, as tested with a small field magnet, is
Bern, with USGS basalt glass GSD-1G as the primary cali- generally strong for Geotimes lavas. Geotimes pillows are
bration standard (reference values from Jochum et al., 2005). usually weakly vesicular and aphyric. Where present, vesi-
All major elements were also analysed, allowing the anhy- cles are mostly filled with the greenschist-facies minerals

chlorite, epidote, or quartz. Exceptions to this typical ap- and contrasts with the rare mainly pore-filling celadonite
pearance include occasional pore-filling sub-greenschist al- found in the Geotimes unit. Irregularly shaped fingers, pods,
teration mineralogy (e.g. celadonite and zeolite) in the area a and tabular zones of greenish-yellow pumpellyite alteration
few kilometres south of Wadi Fizh (sample TB3-11F) and in (Pflumio, 1991) are common in both Tholeiitic Alley and
the up-faulted block north-east of the Lasail mine (e.g. TB3- Boninitic Alley, but rare in Geotimes (where epidotisation
07A2). Geotimes lavas in the Aswad block also commonly dominates; Gilgen et al., 2016). Pillow flows are typically
feature millimetric sub-equant phenocrysts of relatively fresh pervasively vesicular, ranging from millimetric vesicles in
plagioclase. Evolved andesitic Geotimes lavas tend to form pillows (Fig. 3c, d) to fist-sized cavities in evolved flows.
large, flat, rusty-brown pillows and lobate flows, and are par- These vesicles are commonly filled by quartz, chalcedony,
ticularly common around Wadi Jizi and the Yanqul areas. celadonite, zeolite, and late calcite, giving outcrops a char-
acteristic white-spotted appearance. Primitive Tholeiitic Al-
3.2 Lasail ley lavas take on a pale greyish-green appearance (Fig. 3d),
which may have been interpreted as Lasail lavas around the
Lasail lavas typically consist of small bun-shaped pillows Lasail mine in the past (as shown by Gilgen et al., 2014).
and pahoehoe flows, often with more irregular shapes than In this case, Tholeiitic Alley can be distinguished by its
is typical for Geotimes or Tholeiitic Alley (Fig. 3b). These spotted, dark grey spherules (and by its clinopyroxene com-
pillows may be interspersed with occasional massive flows positions; Gilgen et al., 2014). Thick accumulations of an-
up to 10 m thick. Pillowed outcrops are typically pale pastel desitic to dacitic Alley lavas are common in the Haylayn
green to light grey in colour. Lasail lavas are either weakly or block as well as in the Wadi Jizi area west of the Lasail
non-magnetic when tested with a small field magnet. White mine. Where continuous enough to form mappable units,
variolites a few millimetres across tend to concentrate in the we differentiate these as Felsic Alley: a subunit of Tholei-
pillow rims. Lasail pillows are usually only weakly vesicu- itic Alley. These lavas are generally composed of strongly
lar, with chlorite, epidote, and quartz fillings. Slightly more celadonitic or hematitic columnar-jointed andesite to dacite
evolved Lasail lavas and those transitional with Geotimes massive flows and local rhyolite flows with relict pods of un-
may take on a darker appearance similar to Geotimes (e.g. altered obsidian (Alabaster et al., 1982).
wadis Fizh and Ashar; Belgrano and Diamond, 2019; Kusano
et al., 2012). In this case, field discrimination is challenging. 3.4 Boninitic Alley
Andesites and dacites purportedly belonging to the Lasail
unit appear to be limited to the Wadi Jizi area (Alabaster et Boninite lavas in the Semail ophiolite are composed of mixed
al., 1982). However, our observations in the area indicate that sequences of pillow, pahoehoe, and massive flows with lo-
these evolved lavas always overlie the Lasail basalts, and the cal accumulations of hyaloclastite breccias up to tens of
associated dyke sheets and plutonic complex cut the overly- metres thick. The pillows are of various sizes and shapes
ing Tholeiitic Alley lavas. The evolved “Lasail”-affinity lavas even in the same outcrop, and they fit snugly together, in-
and sheets around Wadi Jizi are therefore probably related to dicating that they formed from low-viscosity lavas (Fig. 3e).
Alley-stage volcanism, consistent with other large complexes Thick (∼ 150 m) sequences of blocky breccias interspersed
of evolved lavas elsewhere in the ophiolite. with boninite lava flows and calcareous sediments locally
top the sequence in the Wadi Jizi area sampled by Ishikawa
3.3 Tholeiitic Alley et al. (2002) and along the section between Wadi Hayl and
Wadi Bani Umar (Fizh block). The Semail boninites com-
The Tholeiitic Alley unit is composed of mixed sequences of monly have abundant macroscopic olivine phenocrysts or
basaltic to high-magnesium andesitic pillow lavas (Fig. 3c, their pseudomorphs preserved in their flow rims. They also
d) interspersed with occasional high-magnesium andesite– tend to be highly vesicular (Fig. 3f), even with micro- or
dacite columnar-jointed massive flows. These sequences are macroscopically foamy textures, though weakly vesicular ex-
locally intercalated with lenses of hyaloclastite breccias up to amples also exist (Fig. 3e). Boninitic Alley lavas are typ-
tens of metres thick, which occasionally preserve fragments ically only weakly attracted to small field magnets where
of volcanic glass in a palagonitised matrix (Kusano et al., they are darker and fresher. Depending on their thickness and
2017). Basaltic to andesitic Tholeiitic Alley lavas are typi- thus burial depth, which ranges from 0–2 km, the metamor-
cally strongly attracted to small field magnets. Tholeiitic Al- phic grade and secondary alteration mineralogy of Boninitic
ley pillows often have ∼ 1–10 mm diameter dark-grey spots Alley varies between greenschist, prehnite–pumpellyite, and
or spherules spaced evenly throughout their cross sections brownstone facies (i.e. clay rich; Alabaster and Pearce,
(Fig. 3c, d). The alteration mineralogy is dominantly of zeo- 1985). The outcrop appearance of the boninites strongly
lite and pumpellyite-prehnite facies; however, minor chlorite depends on this alteration grade, with brownstone-altered
can commonly be observed in thin sections (Alabaster and boninites generally being indistinguishable from Tholeiitic
Pearce, 1985; Pflumio, 1991). Celadonite alteration pervad- Alley equivalents, having weathered to a mixture of dull
ing the rock matrix is common, especially in massive flows, greys, browns, and greens reflecting various clays (e.g.

Figure 3. Field photographs of typical outcrops of the mapped volcanic units, with UTM 40N location coordinates (orientations: east, north;
unit: metres). Hammer for scale is 50 cm long. (a) Geotimes greenschist-facies spilitic pillow flows: Wadi Jizi type locality (438 288 mE,
2 685 876 mN). (b) Lasail greenschist-facies spilitic pillow and pahoehoe flows showing abundant chlorite (Chl): Wadi Lasail (440 680 mE,
2 683 521 mN). (c) Tholeiitic Alley zeolite/pumpellyite-facies spilitic pillow flows and hyaloclastite breccia (light streak in centre pillow is
a road-cutting artefact). Note sea-green celadonite (Cel) alteration in inter-pillow and breccia, yellowish-beige pumpellyite alteration around
pillow rim, abundant zeolite-filled vesicles, and dark grey spots (spherules) throughout pillow (sample TB4-17J): Suhaylah section, blasted
roadcut around electricity pylon (434 460 mE, 2 683 370 mN). (d) Primitive Tholeiitic Alley pillows with highly vesicular pillow rims and
dark spots throughout the pillow cross section: roadcut near Maqail South mine (453 534 mE, 2 661 208 mN). (e) Boninitic Alley pillow
lavas with weak pumpellyite–clay altered groundmass and zeolite altered (Zeo) inter-pillows (sample TB2-45B; 448 940 mE, 2 712 935 mN).
(f) Boninitic Alley greenschist-facies pillow lavas with three pillow triple-junctions marked by dotted lines. Note pale chlorite–albite alter-
ation with chlorite–epidote replacing olivine in pillow rims and large epidote-filled vesicles (Ep) through pillow core: Wadi Zab’in (approx.
436 040 mE, 2 714 170 mN). (g) Boninitic Alley pillow lava showing globular/spherulitic texture typical of Alley (particularly Boninitic
Alley) lavas: Wadi Bidi (455 370 mE, 2 667 380 mN).
Wadi Jizi, Wadi Bidi). At prehnite-pumpellyite facies, the sail (Fig. 3f). A key difference to Lasail is that these pale
boninites take on a light-brown colour (Fig. 3e). In areas boninites generally have highly spheroidal volcanic textures
of thicker boninite accumulations (e.g. Rajmi), greenschist- (Fig. 3g) with abundant white or grey spherules which in-
facies alteration generally transforms the boninites to pale crease in size and coalesce towards the pillow cores, presum-
green, reflecting abundant chlorite and albite, similarly to La- ably related to fluid saturation during solidification (Ballhaus

et al., 2015). Excellent exposures of all of these alteration mineralogy as a function of their emplacement depth. Given
grades can be found in newly opened roadcuts along the Bati- the uniformity of these veins with depth, we conclude that
nah Expressway between Liwa and Hatta. they were emplaced after significant cooling of the ophio-
lite and probably following the warping of the ophiolite into
3.5 Late dyke swarms and sills its current anticlinal structure, i.e. during or after obduction
and hence after the Phase 2 magmatic stage. Other, possi-
Swarms of “V2” dykes cutting the axial volcanic strata are bly related, quartz–carbonate veins of similar dimensions but
described in several places throughout the ophiolite in the with indurated carbonation haloes are common as topograph-
regional map set. These swarms grade from spaced sets of ically prominent lineaments in the easternmost outcrops of
dykes to late fully sheeted complexes. In the Wadi Rajmi– the Aswad block. Similar carbonate veins form listvenites
Safwa area (Fizh block), boninite dyke swarms (samples where they intersect ultramafic intrusions near Wadi Shaffan
TB5-22A and TB5-22B; also dyke samples from Adachi and (Haylayn block; BRGM, 1986b).
Miyashita, 2003) feed thick accumulations of boninite lavas
(Ishikawa et al., 2002). In the Yanqul area, late sheeted dyke
swarms confirmed as Tholeiitic Alley by geochemistry (sam- 4 Geochemical assignment of samples
ples TB3-25H, TB4-20L) are several hundred metres wide
4.1 Samples
and run parallel to thick accumulations of blocky Tholeiitic
Alley breccias and their associated normal faults. Samples of 179 spilitic lavas, 9 sills and dykes, and 11 fresh
Late sills locally make up significant proportions of the volcanic glasses were collected throughout the mapping area
upper crust. In the eastern Fizh block, numerous columnar- for geochemical analysis to provide reference points for map-
jointed Tholeiitic Alley sills (confirmed by samples TB2- ping. The spilite samples were mostly collected from pil-
43C, TB2-44D) up to 20 m thick intrude the Lasail lavas. lowed lavas which were visually representative of the area,
Their character as sills rather than massive flows is indi- as well as occasional massive flows. Possible sills and dykes
cated by their planar basal contacts, their symmetric upper were avoided, except where deliberately sampled to deter-
and lower chilled margins, and their apophyses locally in- mine the unit affiliation of such features (these samples are
jecting the surrounding Lasail pillows. differentiated in our figures).
Previously published lava analyses from useful locations
3.6 Non-ophiolite rocks (Gilgen et al., 2014, 2016; Kusano et al., 2017) were also
considered and are differentiated as squares in our figures.
Triassic volcanics of the Haybi complex crop out in the same Though our interpretations mostly agree with these studies,
area as ophiolite lavas in the Wadi Hawasina and Harimah ar- we suggest a revised assignment for a minority of samples,
eas (Sarami, Haylayn blocks), where they are exposed in tec- which are asterisked in our figures and listed in Table S2
tonic windows, exist on top of the ophiolite as olistostrome in the Supplement. MacLeod et al. (2013) also documented
mega-clasts, or are part of the overthrust Batinah complex the major-element compositions of numerous V1 samples.
(BRGM, 1986a, b; Searle et al., 1980). Many of the Haybi With only major elements available for unit confirmation,
tholeiites are pillowed and have comparable geochemistry we restricted this dataset to samples with > 1 wt % TiO2 and
to Geotimes lavas, whereas celadonite-altered alkali Haybi > 0.1 wt % P2 O5 (i.e. the approximate lower limits of other
volcanics can resemble altered Alley lavas. Nevertheless, oc- reliably assigned Geotimes basalts).
currences of the Haybi volcanics can generally be recognised
because they are seldom continuous for more than a few hun- 4.2 Approach to identify units
dred metres outside of the Hawasina window; they are always
tectonically juxtaposed against the Semail volcanics and they Classical trace-element diagrams for the determination of
are often intercalated with limestones and cherts (Searle et tectonic setting serve as useful unit discriminants in the Se-
al., 1980). mail ophiolite (Alabaster et al., 1982; Ernewein et al., 1988;
Quartz–hematite veins (labelled Q0 in the regional map Gilgen et al., 2014; Kusano et al., 2014, 2017). However,
set), often with magnetised hematite and distinctively chlo- these diagrams are intended for characterising entire suites
ritised alteration haloes up to tens of metres wide, fre- of lavas (Pearce, 2014), for which interpretation is relatively
quently cut the extrusives throughout the mapping area. resistant to the presence of outliers. Multiple diagrams based
BRGM (1993a) suggested these veins are associated with the on independent principles are therefore required to best char-
Phase 2 intrusives. In fact, these veins cross cut the entire acterise individual samples.
crustal sequence from the layered gabbros to the Boninitic Alabaster et al. (1982), Umino et al. (1990), and later
Alley lavas (e.g. Aswad, Fizh blocks) with the same vein Gilgen et al. (2014, 2016) developed a geochemical work-
texture, mineralogy, and chloritic alteration haloes. Any syn- flow for identifying spilitised Semail lavas and dykes of un-
magmatic veins emplaced throughout the vertical extent of known affiliation by comparing their immobile element and
the in situ oceanic crust should show changes in texture and relict clinopyroxene compositions to those of lavas reliably

assigned to a unit on the basis of field relationships. We

firstly updated these geochemical discrimination fields to in-
clude new stratigraphically assigned datasets from sections
along Wadi Shaffan (Geotimes/V1; Einaudi et al., 2003),
Wadi Fizh (Geotimes/less-depleted LV1 and MV1 and La-
sail/UV1; Kusano et al., 2012), and Wadi Bidi (Tholeiitic
Alley/LV2 and Boninitic Alley/UV2; Kusano et al., 2014).
We then compared our samples to these fields and the dif-
ferentiation trends that define them. If a sample fell outside
of the predefined fields or within an overlapping area in one
plot, its position along trends in other plots was taken into
consideration along with any field-based constraints to best
assign it to a unit. The predefined fields were then expanded
to include the independently assigned sample, producing a
new set of slightly larger discriminatory fields. The possible
assignments of samples indicated by each diagram are sum-
marised in Table S1.
4.3 Geochemical results
The Semail volcanostratigraphy records a progressive deple-

tion in incompatible elements during the ophiolite’s forma-
tion (Fig. 4a, b; Alabaster et al., 1982; Ernewein et al., 1988;
Kusano et al., 2014, 2017). Incompatible-element concentra-
tions are also sensitive to magmatic fractionation and con-
sequently they tend to overlap with their under- and overly-
ing units (Fig. 4). Interestingly, Fig. 4 suggests that Hf/Zr
ratios tend to increase upwards through the volcanostratig-
raphy, which may prove instructive in narrowing down the
nature of the slab component contaminating these lavas.
Comparisons of compatible elements (Cr), or somewhat-
compatible elements (e.g. Ti, V), with incompatible elements
(Zr, Y) allow individual samples to be classified as poten-
tially evolved or primitive members of each unit (Fig. 5).
Firstly, the Zr–Ti diagram (Fig. 5a; Alabaster et al., 1982;
Pearce et al., 1981) allows for the assessment of magmatic Figure 4. Incompatible-element patterns normalised to N-MORB
Ti-bearing magnetite fractionation, a process which must be (Gale et al., 2013) for a subset of our samples. (a) Geotimes, tran-
ruled out before applying the Ti–V discriminant diagram sitional Geotimes–Lasail, Lasail, and depleted-Lasail patterns (this
(Fig. 5b; Shervais, 1982). Compositions falling beneath the study). (b) Tholeiitic Alley and Boninitic Alley patterns in compar-
main positive correlation trend in Fig. 5a testify to magnetite ison to the Geotimes and Lasail ranges (this study) and low- and
fractionation, and therefore they are assigned to felsic sub- high-silica boninite glass compositions from (Kusano et al., 2017).
units and excluded from Fig. 5b. The Zr–Ti diagram also (c) Lasail, Tholeiitic Alley, and Boninitic Alley compositions in
comparison to V2 type I and II groups of Godard et al. (2003).
shows whether such evolved lavas belong to the Tholeiitic
Alley or Geotimes suites, as Ti-bearing magnetite fraction-
ation occurs at lower Zr concentration in Tholeiitic Alley
relative to Geotimes (Alabaster et al., 1982; Pearce et al., ley”. A subset of these Phase 2 compositions depart the linear
1981). Figure 5a further provides a simple discrimination of Zr–Ti trend at the upper limit of the Boninitic Alley array, at
many of the Geotimes and Boninitic Alley samples on the lower Zr concentration than for Tholeiitic Alley.
basis of the absolute concentrations of their incompatible el- The Ti–V diagram (Fig. 5b) is particularly effective at dis-
ements. The compositions of Phase 1 and Phase 2 gabbros criminating units within the Semail extrusive suite and within
and plagiogranites from the length of the ophiolite are also ophiolites in general (Gilgen et al., 2014, 2016; Pearce, 2014;
plotted for comparison in Fig. 5a (de Graaff et al., 2019; Shervais, 1982). This discrimination is partly based on the
Haase et al., 2016). The Zr–Ti evolution of the Phase 1 in- redox-sensitive compatibility of V relative to Ti in the man-
trusives closely matches that of “Felsic Geotimes”, whereas tle source: source oxidation decreases V compatibility and
most of the Phase 2 compositions overlap with “Felsic Al- thus increases V/Ti ratios in its partial melts (Mallmann and

Figure 5. Geochemical unit-discrimination diagrams (anhydrous compositions). Circles: this study. Squares: published analyses re-examined
as unknowns (asterisked if assigned differently to the publications; Gilgen et al., 2014, 2016; Kusano et al., 2017). (a) Whole-rock Ti–Zr after
Alabaster et al. (1982). (b) Whole-rock Ti–V and Ti/V ratios (grey lines) after Shervais (1982). (c) Whole-rock Y–Cr after Pearce (1980).
(d) Zr–Zr/Y after Alabaster et al. (1982). (e) Clinopyroxene median Mg# (molar Mg/(Mg + Fe) ×100) vs. Ti (atoms per formula unit) for
a subset of samples. (f) Clinopyroxene lower and upper Mg# quintile medians with corresponding Ti for the same dataset as in (e). Dashed
lines in (a)–(d): previous whole-rock unit fields based on stratigraphically defined data for Geotimes (Alabaster et al., 1982; Belgrano and
Diamond, 2019; Einaudi et al., 2003; Godard et al., 2006; Kusano et al., 2012), Lasail (Alabaster et al., 1982; Belgrano and Diamond, 2019;
Godard et al., 2006; Kusano et al., 2012), Tholeiitic Alley (Alabaster et al., 1982; Kusano et al., 2014), and Boninitic Alley (Ishikawa et al.,
2002; Kusano et al., 2014). Volcanic glasses (Kusano et al., 2017) and Semail intrusives (excluding wehrlites; de Graaff et al., 2019; Haase
et al., 2016). Dashed lines in (e) and (d): previous clinopyroxene fields (Alabaster et al., 1982; Belgrano and Diamond, 2019; Gilgen et al.,
2014; Kusano et al., 2012). Coloured fields: this study.
O’Neill, 2009; Shervais, 1982). This combines with increas- Consequently, the occasional samples affected by incipient
ing V/Ti ratios at higher partial-melt degrees and Ti de- alteration of these types have been excluded from Fig. 5b.
pletion upwards through the Semail extrusive sequence to The Cr–Y diagram (Fig. 5c; Pearce, 1980) allows
trace an anti-clockwise progression in Fig. 5b from Geotimes incompatible-element depletion (monitored by Y concentra-
through to Boninitic Alley, with each unit falling along radi- tion) to be considered a function of magmatic differentiation
ating V/Ti trends. Apart from the aforementioned issue with (monitored by Cr depletion). Accordingly, it is particularly
magnetite fraction (affected samples have been screened out useful in distinguishing primitive Geotimes from Lasail sam-
with Fig. 5a), V appears to be slightly mobile during epi- ples which fall in the overlapping area of Fig. 5b. For these
dosite and pumpellyite alteration (e.g. Gilgen et al., 2016). samples, the incompatible-element depletion that is diagnos-

tic of Lasail (Alabaster et al., 1982; Godard et al., 2006; Ku- 2016). Compositional zonation and intra-sample variation
sano et al., 2012) allows for mostly unambiguous sample as- among the clinopyroxenes are also useful for discriminat-
signment to either a Lasail or Geotimes fractionation path at ing between units whose medians fall in overlapping zones.
respectively lower or higher Y content for a given Cr con- Steep trends of dispersion at high Mg# are diagnostic of La-
tent (Fig. 5c). As chromium is rapidly depleted during differ- sail lavas (Gilgen et al., 2016). To show this in a reproducible
entiation of wet melts (Pearce, 1980), its concentration lies and representative way, we plot the median of the upper and
below the XRF detection limit (∼ 4 µg g−1 ) for much of our lower quintiles of Mg# with their corresponding Ti values for
sample set. Concentrations below the detection limit are not the totality of EMP measurements on each sample and join
plotted in Fig. 5c and therefore the Alley and Geotimes fields them with a line (Fig. 5f). Core-to-rim zonation is indicated
in fact extend to lower Cr and higher Y contents than those by an arrow where these trends were petrographically clear
indicated by the coloured fields. and consistent.
Following our addition of many recent analyses, the unit
fields of Alabaster et al. (1982) and Gilgen et al. (2014) now 4.5 Interpretation of transitional compositions
mostly overlap in the Zr/Y–Zr diagram. The exceptions are
for particularly depleted Boninitic Alley lavas and particu- The rather large fields of the four main volcanic units in
larly evolved Geotimes lavas (Fig. 5d). Boninitic Alley and Fig. 5 reflect the true range of compositions within each
“Depleted Lasail” also have characteristically steeper trends unit, as these fields encompass samples taken from clearly
than the other units in this diagram. stratigraphically defined positions. However, where samples
with intermediate geochemistry coincide with an intermedi-
4.4 Geochemical discrimination of Lasail from ate stratigraphic position, i.e. where they outcrop between
Tholeiitic Alley two clearly identifiable units, these samples are probably
transitional. Such lavas are important as they demonstrate
The expanded dataset of analyses now available for the Se- temporal overlaps between eruptive episodes. Transitional
mail volcanics show that the Lasail and Tholeiitic Alley units lavas between Geotimes and Lasail have been widely re-
cannot be straightforwardly discriminated at the regional ported (A’Shaikh et al., 2005; Alabaster et al., 1982; Bel-
scale based on whole-rock geochemistry alone. In Fig. 5a–d, grano and Diamond, 2019; Kusano et al., 2012). A subset of
for example, the fields for Lasail and Tholeiitic Alley overlap our samples is also geochemically and stratigraphically inter-
to a large extent. Godard et al. (2003) speculated that their mediate between Geotimes and Lasail and thus assigned to
units V2-I and V2-II may be equivalent to Lasail and Al- a “Transitional Geotimes–Lasail” group. Similarly, samples
ley, respectively, and they distinguished them based on their with geochemistry intermediate between Tholeiitic Alley and
contrasting incompatible-element patterns and related ratios. Boninitic Alley occasionally occur at the top of Tholeiitic
However, the differences between Lasail and Tholeiitic Alley Alley and in areas where both Tholeiitic Alley and Boninitic
are not apparent when the data of Kusano et al. (2012, 2014) Alley are present, as also reported by Kusano et al. (2014).
are plotted in the Godard et al. (2003) diagrams (Belgrano We assign these intermediate samples to a Transitional Alley
and Diamond, 2019). In fact, V2-I rather appears to corre- group. Transitional lavas between Geotimes and Alley have
spond to the Tholeiitic Alley unit, as suggested by the along- so far not been described. However, the lowermost Tholeiitic
strike continuity of V2-I described by Godard et al. (2003). Alley lavas overlying Lasail and Geotimes in the north of the
Lavas directly comparable to V2-II were not recovered dur- western Fizh block consistently have slightly lower V/Ti ra-
ing this study nor by Kusano et al. (2014). However, as noted tios than typical Alley as well as less-depleted incompatible-
by Kusano et al. (2014), the spoon-shaped incompatible- element patterns in Fig. 4 (e.g. samples YV15-21, TB2-41A,
element patterns of V2-II fit those of an evolved Boninitic TB2-46A). Though marginally assigned to the Tholeiitic Al-
or “Transitional Alley” lava (Fig. 4c). ley unit on the basis of the literature fields, these lavas ap-
The only reliable way to geochemically discriminate La- pear to represent an intermediate stage between Geotimes
sail and Tholeiitic Alley identified thus far is through their and Alley and can be traced in the RTP geomagnetic map
divergent clinopyroxene compositions (Fig. 5e, f; Alabaster into the footwall of the Mandoos VMS deposit. These tran-
et al., 1982; Belgrano and Diamond, 2019; Gilgen et al., sitional units also tend to have intermediate field characteris-
2014, 2016). Fortunately, the outcrop appearance and strati- tics; however, these differences are subtle and generally not
graphic associations of these two units are typically distinct mappable without prior knowledge and geochemical refer-
from one another, so they could generally be discriminated ence points.
during this study by combining field observations with their
whole-rock XRF composition. In other cases, clinopyrox- 4.6 Interpretation of Depleted Lasail compositions
ene analyses were necessary for assignment. In Fig. 5e, the
median Mg# and Ti concentrations in each sample are plot- Certain rare lavas exhibit Lasail-like clinopyroxene composi-
ted relative to the previously published fields (Alabaster et tions but have highly depleted low-silica boninite whole-rock
al., 1982; Belgrano and Diamond, 2019; Gilgen et al., 2014, (spilite) major-element and immobile-element compositions

(Figs. 4a, 5, 6). No fresh glasses were recovered from these

lavas, so the spilite compositions in Fig. 6 are potentially al-
tered; however, the reasonable correlation of Si and Mg in
Fig. 6a suggests mostly limited mobilisation of these ele-
ments. Depleted Lasail’s immobile element compositions are
comparable to Boninitic Alley (Figs. 4a, 5) but with some-
what deeper negative Nb–Ta anomalies and an absence of
spoon-shaped light rare-earth element enrichment (Fig. 4a).
Gilgen et al. (2016) sampled similar lavas and their dyke
equivalents (pillow lava sample RAM010 was remeasured
with ICP-MS for this study), assigning them to Lasail on the
basis of their clinopyroxene compositions. Belgrano and Dia-
mond (2019) also reported similar lavas within a lens of “Ax-
ial Lasail” intercalated within Geotimes near Wadi Hawarim.
In our figures, these samples are distinguished from nor-
mal Lasail as Depleted Lasail. The presence of Depleted
Lasail lavas has been confirmed by sampling in the Hilti,
Fizh, and Aswad blocks. Stratigraphically, Depleted Lasail
lavas occur within the axial Geotimes sequence near Wadi
Hawarim (sample TB3-04B; Belgrano and Diamond, 2019),
directly overlying Geotimes and underlying Tholeiitic Alley
near Wadi Bidi (sample TB5-27A), and atop a thick accumu-
lation of normal Lasail lavas just south of the Mandoos VMS
deposit (samples RAM005, RAM010, TB2-40D). The ob-
servation of these lavas beneath the less-depleted Tholeiitic
Alley sequence marks a departure from the otherwise consis-
tent trend of depletion upwards through the volcanic stratig-
raphy. Further work is required to understand the significance
of these early, depleted lavas. In the interim, the newly col-
lected Depleted Lasail lavas are differentiated in our figures
but are not incorporated into the expanded geochemical fields
for the normal Lasail unit.
4.7 How boninitic is the “Boninitic Alley” unit?
Boninites are defined by the IUGS (International Union

of Geological Sciences; Le Bas, 2000) as volcanic
rocks with whole-rock SiO2 > 52 wt %, MgO > 8 wt %, and
TiO2 < 0.5 wt % after normalisation to 100 wt %. The def- Figure 6. Boninite classification diagrams based on whole-
inition of boninite-series lavas has been incrementally ex- rock major-element oxide concentrations normalised to 100 wt %.
panded to include fractionation series which pass through (a) SiO2 vs. MgO with rock fields after Pearce and Reagan (2019),
SHMB meaning siliceous high-magnesium basalt, HMA meaning
the IUGS fields, extending from low-Si compositions to
high-magnesium andesite, D meaning dacite. (b) MgO vs. TiO2
high-magnesium andesites (HMA) at < 8 wt % MgO (Fig. 6;
with boninite/low-Ti basalt field adapted to log scale after Pearce
Pearce and Reagan, 2019; Pearce and Robinson, 2010; Rea- and Reagan (2019), and original BADR–boninite series divider
gan et al., 2017). Boninites can be further divided into low-Si (dashed grey line) after Pearce and Robinson (2010). Filled circles:
boninite (LSB) and high-Si boninite (HSB) series, with a di- volcanic glasses (this study). Inverted triangles: volcanic glasses
viding composition along an olivine fractionation line which (Kusano et al., 2017). Unfilled circles: Boninitic Alley and De-
passes through SiO2 = 57 wt % at MgO = 8 wt % (Pearce and pleted Lasail spilites (this study). Unfilled squares: Boninitic Alley
Reagan, 2019). and Depleted Lasail spilites (Belgrano and Diamond, 2019; Gilgen
To test whether the Semail Boninitic Alley unit is really et al., 2014, 2016). West Pacific boninite analyses from the north
boninitic, we consider the major-element compositions of Tonga Trench (Falloon et al., 2007; Falloon and Crawford, 1991),
unaltered Boninitic Alley/UV2 volcanic glasses reported by Izu–Bonin forearc (IODP Expedition 786; Murton et al., 1992), and
Bonin forearc (IODP Expedition 352; Reagan et al., 2017).
Kusano et al. (2017), which fall along the same immobile-
element trends as the Boninitic Alley spilites in Fig. 5. In ad-
dition, the major-element compositions of several volcanic

glasses collected during this study are plotted for assign- termined the magnetic mineralogy of a subset of these sam-
ment into the boninite series or tholeiitic basalt–andesite– ples.
dacite–rhyolite (BADR) or high magnesium andesite fields.
Boninites from the west Pacific are also plotted for compari- 5.1 Bulk magnetic property results
son with the Semail examples.
The Boninitic Alley glasses all fall within the LSB and Aeromagnetic anomalies (represented here by the Batinah
HMA fields (Fig. 6). Tholeiitic Alley follows a fractiona- RTP map) are principally caused by differences in magneti-
tion path from basalt through the high-magnesium andesite sation between adjacent rock bodies. These differences are
field into dacite. The compositional difference between the controlled by the different magnetic susceptibilities, i.e. in-
two Alley units is slight in Fig. 6a, and is more clearly duced magnetisation, of the rocks, as well as by differences
defined by greater incompatible-element depletion (includ- in NRM (Blakely, 1995; Clark, 1997, 2014). A basic assump-
ing Ti), higher V/Ti ratios, and higher Cr in the boninites tion of the RTP transformation carried out on the Batinah
(Figs. 4b, 5b, c, 6b). Spilitised samples assigned to Boninitic survey is that source magnetisation is parallel to the local
Alley via Fig. 5 are also plotted for comparison in Fig. 6. geomagnetic field, which may not be the case for rock bod-
Though these altered compositions should be treated with ies with strong remanent magnetism (Blakely, 1995; Clark,
caution, the majority plot along similar trends to the fresh 2014). The Königsberger ratio (Q) of remanent-to-induced
glasses, suggesting that the higher MgO content is not simply magnetisation is therefore a useful parameter for assessing
due to spilitisation. The Boninitic Alley series may therefore the potential for artefacts or “remanence effects” in the RTP
extend to more magnesian compositions than those recorded map data used in this study (Clark, 1997). For rock bodies
by the glasses (e.g. to ∼ 17 wt % MgO). with Q 1, induced magnetisation dominates. Thus, in the
The trace-element patterns of Boninitic Alley samples absence of anisotropy of susceptibility, RTP processing ac-
(as classified by Fig. 5) have highly depleted incompatible- curately centres anomalies over their geographic sources. For
element trends in Fig. 4b. Of these, seven tend towards bodies with Q 1, magnetism is dominated by remanence.
monotonic compatibility-controlled depletion with negative In this case, and if the NRM direction is different from the
Nb–Ta and positive Th anomalies. The remaining sample field direction, remanence effects may be introduced into the
(TB2-42A, Wadi Zab’in, Fizh block) has a classic boni- RTP map, leading to inaccurate estimates of the magneti-
nite spoon-shaped MORB-normalised pattern enriched in sation and geographic location of anomalies (Clark, 1997,
Th, Nb, Ta, and light rare-earth elements. A similar range 2014).
in compositions was noted by Kusano et al. (2017), who Figure 7 shows that Geotimes and Tholeiitic Alley lavas
subdivided Boninitic Alley/UV2 into “low-Si” and “high- have generally higher susceptibilities and remanences than
Si” groups with monotonically depleted and spoon-shaped Lasail and Boninitic Alley lavas. Further, Q is on average
MORB-normalised patterns, respectively (Fig. 4b). These greater than one for all units, with Boninitic Alley having the
differences were explained by the addition of sediment melts largest spread in values. Although averaging NRM and Q on
to the “high-Si” group, supported by attendant shifts in Nd a unit basis is properly achieved by incorporating the vector
and Hf isotopic compositions (Kusano et al., 2017). directions of these properties for each sample (Clark, 2014),
Taken together, the compositional similarities between the the medians of our un-oriented sample set give a sound idea
Boninitic Alley glasses, Boninitic Alley spilites, and the west of the potential for remanence effects in each unit (Fig. 7).
Pacific boninites in Figs. 4–6 indicate an LSB-series protolith Comparison of these results with the chemical and min-
for the Boninitic Alley spilites and demonstrate the effective- eralogical compositions of the lavas allows the origin of the
ness of the immobile-element fields in Fig. 5 for discriminat- magnetic properties to be identified (primary vs. secondary).
ing altered boninites. This in turn permits assessment of the reliability of the aero-
magnetic data as an aid in geological mapping. The magnetic
properties of volcanic rocks are principally controlled by
the partitioning of Fe between strongly magnetic (ferromag-
5 Interpretation of the aeromagnetic survey netic) oxide phases and weakly magnetic (paramagnetic) sili-
cate phases (Clark, 1997). Therefore, it is instructive to com-
Aeromagnetic interpretation is greatly clarified by under- pare measured magnetic susceptibilities with the whole-rock
standing the magnetic petrology of the surveyed units (Clark, proportion of Mg relative to Fe (expressed as Mg# meaning
1997). To aid interpretation of the Batinah aeromagnetic map molar Mg/(Mg + Fetotal ) ×100). Figure 8 shows this com-
(Isles and Witham, 1993), we measured the magnetic suscep- parison, including the measurements of Geotimes lavas from
tibility (K) and the natural remanent magnetisation (NRM) Wadi Shaffan by Einaudi et al. (2003), which we have recal-
intensity of samples that had been assigned to a specific vol- culated to mass-normalised susceptibility using the reported
canic unit by either a clear stratigraphic situation or by geo- densities.
chemical criteria. To identify the origin and establish the po- Between Mg# values of 45 and 80 (Fig. 8), which are
tential reliability of these magnetic properties, we further de- typical of Geotimes and Tholeiitic Alley, susceptibility in-

Figure 7. Statistical summary of bulk magnetic measurements of

hand specimens of the Semail volcanic units carried out for this
study. Bold lines: medians. Coloured boxes: interquartile ranges.
Whiskers: minimum–maximum. Felsic subunits are grouped with
parent units. Numbers in coloured boxes show number of sam-
ples analysed. (a) Magnetic susceptibility (both Magnon and KT-5
measurements grouped). (b) Natural remanent magnetism (NRM), Figure 8. Magnetic susceptibility vs. whole rock Mg# (molar
un-oriented samples. (c) Königsberger ratios (Q) of remanent-to- Mg/(Mg + Fetotal ) ×100) for the four main Semail volcanic units.
induced magnetisation calculated for each sample with un-oriented Circles and diamonds: this study. Squares: Geotimes/V1 sam-
NRM, a density of 2.67 g cm−3 (mean for Geotimes spilites; ples from Einaudi et al. (2003). Outlying analyses: fresh boni-
Einaudi et al., 2003) and the local geomagnetic field for 1992 nite TB3-01A; highly carbonated Geotimes TB3-20I (Loss on igni-
(42 900 nT; Thébault et al., 2015). tion = 16 wt %) from vicinity of late carbonate vein; high-Si Transi-
tional Alley TB3-15C (70 wt % SiO2 ) and Tholeiitic Alley TB2-34
(84 wt % SiO2 ).
creases along a scattered array with decreasing Mg#. Above
Mg# ∼ 80, the primitive lavas of the Lasail and Boninitic
Alley units mostly have susceptibilities close to zero. One 5.2 Magnetic mineralogical results
particularly fresh high-Mg# boninite (TB3-01A) has low but
significant susceptibility, continuing along the extension of As shown above, the Semail lavas mostly display increas-
the Geotimes–Alley trend. From this array we can infer that, ing magnetic susceptibility with decreasing Mg#, as would
for the Semail spilites, Fe is primarily incorporated into para- be expected for fresh lava suites (e.g. Vogt and Johnson,
magnetic silicates in rocks with Mg# > 80, whereas in rocks 1973), even though the rocks are pervasively hydrothermally
with Mg# between 45 and 80 or in fresh lavas, Fe is signifi- altered. The ferromagnetic mineralogy of our samples eluci-
cantly incorporated into ferromagnetic oxides. dates how these apparently primary magnetic characteristics
At Mg# < 45 magnetic susceptibility is scattered towards could have persisted through spilite alteration. This mineral-
lower values. This could be due to magmatic fractionation of ogy was determined by high-temperature susceptibility (KT )
titanomagnetite from evolved magmas (e.g. high-Si outliers and low-temperature magnetic experiments, for which four
in Fig. 8). However, apart from these two extreme samples, representative examples are given in Fig. 9. Plots for each
there is no clear link between low magnetic susceptibility sample, grouped by unit and mineralogical interpretation, are
and the felsic units shown to have fractionated magnetite in shown in Figs. S2–9 in the Supplement.
Fig. 5a. An alternative explanation is that for many of these Figure 9a shows a simple KT curve with magnetic un-
low-Mg# samples (e.g. from Einaudi et al., 2003), relatively blocking upon heating corresponding with a dK/dT mini-
intense hydrothermal alteration (e.g. incipient epidotisation) mum at 493 ◦ C. This unblocking temperature (Tub ) is typ-
has either leached Mg or sequestered the available Fe into ical of titanomagnetite with Ti substitution depressing Tub
paramagnetic silicates or weakly magnetic oxides (e.g. epi- from stoichiometric magnetite’s Curie temperature of 580 ◦ C
dote, hematite). Intense calcite alteration around late carbon- (Dunlop and Özdemir, 1997). The weak K upon cooling in-
ate veins also appears to result in destruction of the magnetic dicates that this titanomagnetite was oxidised to weakly mag-
minerals (TB3-15C; Fig. 8). netic (Ti) hematite above 600 ◦ C during the experiment. The
curve in Fig. 9b is similar, except unblocking occurs rapidly
at a Tub of ∼ 574 ◦ C, almost equivalent to the Curie temper-
ature of stoichiometric magnetite.

Figure 9. Magnetic mineralogy of the Semail volcanics. (a–d) High-temperature magnetic susceptibility (K) of four representative samples
with the first derivative of the heating curve (temperatures marked at dK/dT minima/maxima). (a) Titanomagnetite in relatively fresh
Boninitic Alley (TB3-01A). (b) Magnetite in Tholeiitic Alley (TB2-33C). (c) A mixture of maghemite and magnetite in Felsic Alley (TB3-
25E). (d) Maghemite in Geotimes (TB3-07C). (e–h) Low-temperature FC and ZFC remanences and RT-IRM during cooling and warming
of the same samples as in a–d. Asterisked “transition” in (e) possibly caused by unintentional movement of the sample. (i) Frequency of
ferromagnetic mineral occurrences (excluding weakly magnetic hematite) deduced from high- and low-T experiments for the entire sample
set (n = 36) and within each unit (nGeotimes = 14, nLasail = 6, nThol. Alley = 9, nBon. Alley = 11). Where multiple magnetic minerals are
detected in a single sample, all the minerals contribute to the unit count.

The KT curves in Fig. 9c and d are more complex, with 1997). Maghemite “bumps” and inversions comparable to
an increase in K upon heating from room temperature to those in Fig. 9d occur in 47 % of the total sample set and in
∼ 350 ◦ C punctuated with a step at 130–40 ◦ C. In Fig. 9c, this 86 % of Geotimes samples. Given the often significant pro-
is followed by a drop in K at around 409 ◦ C, but the majority portion of magnetism lost during these inversions, we con-
of magnetisation persists until unblocking at ∼ 576 ◦ C. The clude, in agreement with Perrin et al. (1994), that maghemite
initial increase in K from room temperature in Fig. 9c and d is a prevalent carrier of magnetism in the Semail lavas, and
is a common feature of maghemite and has been explained in particular in the Geotimes unit.
as the thermally prompted relaxation of lattice stresses at
the contact between maghemite rims and magnetite cores 5.3 Implications for aeromagnetic interpretation
(“maghemite bump”; Kontny and Grothaus, 2017; Liu et al.,
2004; Velzen and Zijderveld, 1992). Maghemite (γ -Fe2 O3 ) Figure 7 demonstrates that, typically, highly magnetised ar-
is a typical oxidation product of magnetite which retains eas in the aeromagnetic map should correlate with the oc-
the cubic spinel structure and much of the magnetism of its currence of Geotimes and Tholeiitic Alley lavas, whereas
precursor magnetite (Clark, 1997). We interpret the marked weakly magnetic zones should correlate with Lasail and
drops in K at 300–450 ◦ C (Fig. 9c, d) as the structural inver- Boninitic Alley. Figure 8 demonstrates that this correlation
sion of maghemite to hematite during heating (Dunlop and is connected to the characteristically high Mg# of Lasail
Özdemir, 1997). and Boninitic Alley in comparison to Geotimes and Tholei-
The low-temperature magnetic behaviour of these samples itic Alley. As these differences are also clearly visible on
supports our high-T interpretations (Fig. 9e–h). The absence a unit basis (Fig. 8), the high Mg# of Lasail and Boninitic
of a Verwey transition at ∼ 120 K for TB3-01A (Fig. 9e) is Alley must be inherited from their protoliths, as supported
consistent with Ti substitution in magnetite (Moskowitz et by these units’ high Cr, abundant olivine (phenocrysts or
al., 1998). Contrastingly, a marked Verwey transition occurs pseudomorphs; Kusano et al., 2012, 2014), and magnesian
at ∼ 110 K in TB2-33C (Fig. 9f), supporting the interpre- clinopyroxene compositions. The mixed, variably oxidised
tation of stoichiometric magnetite from Fig. 9b. For TB3- ferromagnetic mineralogy of the Semail lavas may explain
25E, a weak Verwey transition is detectable (Fig. 9g). This much of the scatter in Fig. 8. However, though this scatter
is consistent with the mixture of magnetite and maghemite exists for each unit (Fig. 7), at the spatial resolution of the
deduced from the high-T data, as the Verwey transition aeromagnetic survey (∼ 80 m), the average character of each
is suppressed by partial oxidation of magnetite and is ab- unit is more likely to be represented. The common occur-
sent in maghemite (Dunlop and Özdemir, 1997). For KT rence of relict titanomagnetite in Alley, as well as oxidised,
curves featuring drops in K at ∼ 330 ◦ C, similar to that but nevertheless magnetic maghemite in Geotimes (Fig. 9),
in Fig. 9d, which we attribute to maghemite inversion, the provides a mineralogical explanation for how these primary
presence of monoclinic pyrrhotite (with TCurie = 320 ◦ C) is properties could have been partly preserved through spilite
difficult to rule out. However, upon cooling through ∼ 30– alteration. These conclusions agree with the qualitative dif-
35 K, monoclinic pyrrhotite should undergo a characteristic ferences observed with field magnets (described in Sect. 3),
“Besnus” transition leading to a loss of remanence (Dunlop and they attest to the usefulness of the relationships shown in
and Özdemir, 1997; Rochette et al., 1990). Such a transition Fig. 7 for aeromagnetic mapping.
was not detected for any of the Semail lavas, which strongly However, Königsberger ratios greater than unity for the
supports the interpretation that maghemite is present in sam- Semail extrusives (Fig. 7) indicate the potential for signifi-
ples like TB3-07C (Fig. 9d, h). cant remanence effects on the RTP data (Clark, 1997; Flint
A summary of ferromagnetic mineral occurrences de- et al., 1999). As the ophiolite was formed during a period of
duced from the high- and low-T data (excluding weakly normal geomagnetic polarity (Perrin et al., 1994), the extru-
magnetic hematite) is given for the entire sample set (n = 38) sive sequence should not be complicated by naturally oppos-
in Fig. 9i. Multiple ferromagnetic phases often occur within ing primary remanence directions. This is supported by mea-
a single sample (e.g. Fig. 9c). Titanomagnetite (e.g. Fig. 9a, surements from the northern extrusives, whose characteristic
e) occurs in 37 % of our samples but is particularly preva- remanence directions can be reconciled with each other by
lent in Tholeiitic Alley (55 %), in agreement with Perrin et tectonic rotations of < 90◦ (Perrin et al., 1994, 2000). Nev-
al. (1994). The good positive correlation with Zr in Fig. 5a ertheless, these minor block rotations, as well as any syn-
(prior to magnetite fractionation) shows that Ti is rather im- volcanic tilting or later remanence resetting during obduc-
mobile during spilitisation. These titanomagnetites therefore tion and rotation (Feinberg et al., 1999; Morris et al., 2016)
appear to be primary magmatic phases. Almost stoichio- could all lead to inconsistent RTP anomalies. We accordingly
metric magnetite, as evidenced by Tub = 570–580 ◦ C (e.g. acknowledge the limitations of the RTP map for remote geo-
Fig. 9b, f), is present in 30 %–60 % of samples from each logical mapping. To mitigate the influence of remanence ef-
unit. This magnetite could either be interpreted as hydrother- fects on our interpretation, each structural block was consid-
mal or as the Ti-poor phase of magmatic titanomagnetites ered separately in terms of the RTP character of each unit,
which have unmixed during cooling (Dunlop and Özdemir, and precedence was given to field-based constraints. In this

fashion, aeromagnetic inferences could be made between ref- be attributed to its highly variable grade of hydrothermal
erence points and under cover on a case-by-case basis. alteration. In the prominent accumulations of boninites in
the Aswad, Rajmi, and Daris areas, the sampled boninites
5.4 Observed reduced-to-pole anomalies are mostly altered to greenschist-facies and are weakly mag-
netic (Figs. 3d, 7). Contrastingly, a relatively fresh sample
The alternation between strongly and weakly magnetic units (TB3-01A; Fig. 9a) from the eastern Fizh block retains sig-
upwards through the Semail volcanostratigraphy suggests nificant magnetic susceptibility as well as its primary titano-
that uniformly dipping extrusive sections should be straight- magnetite (Fig. 9a). This sample, and presumably other fresh
forward to interpret in aeromagnetic data. Good examples boninites, is sited within a NNW–SSE-oriented positive RTP
of this exist along the east-dipping western Fizh and Hilti anomaly which extends for ∼ 30 km along strike. This incon-
blocks. Nevertheless, over the mapping area, a variety of sistency makes the inference of boninites from aeromagnetic
RTP anomaly characters are observed for each volcanic unit data alone challenging; and accordingly, we have confirmed
(Fig. 10). all the mapped boninite accumulations by field observations
The Geotimes unit typically corresponds to relatively posi- or sampling.
tive but patchy RTP anomalies between 0.5 and 2 µT. This in-
consistency makes Geotimes challenging to infer in many ar- 5.4.1 The Batinah complex
eas. Fortunately, Geotimes is usually well-exposed at the sur-
face and accurately delineated by the existing regional map Inferences under cover at the top of the section in the Sarami
set. A number of factors may explain this patchiness. Firstly, and Haylayn blocks are complicated by the discontinuous
Geotimes has generally undergone higher-grade spilitic alter- presence of magnetic Triassic volcanics tectonically em-
ation than the overlying units (Alabaster and Pearce, 1985), placed over this part of the ophiolite as part of the Batinah
and it is locally altered to weakly magnetic epidosite (Gilgen complex (BRGM, 1986b; Woodcock and Robertson, 1982).
et al., 2016). Intercalated lenses of the weakly magnetic Given the shallow dip of the top of ophiolite in this area
Lasail unit within Geotimes are also occasionally present, (Shelton, 1990), the magnetic anomalies situated over the
and can explain some of the patchiness around wadis Hatta, Batinah complex volcanics could also be influenced by the
Ashar, Fizh (Fizh block), and Ghuzayn (Alabaster et al., underlying Semail volcanics. Where exposed, the faulted
1982; Belgrano and Diamond, 2019; Kusano et al., 2012; contact between the ophiolite and the Batinah complex is
Umino et al., 1990). Commonly, however, patches of weak readily picked in the field and is well-defined by the regional
RTP magnetism apparently wholly confined to the Geo- maps (BRGM, 1986a, b). However, north-east of this contact,
times unit do not correspond to weakly magnetic or anoma- or where it disappears under cover, there is some uncertainty
lously altered lavas as confirmed in the field (e.g. along the in interpretation. To avoid confusion, we have delineated our
Hilti block). One possible explanation for this is that shal- bedrock interpretation of the overthrust Batinah complex in
low, weakly magnetic intrusions exist beneath the volcanics. Fig. 10 and we have not attempted to interpret the ophiolite
Another explanation, which better explains the widespread beneath the faulted contact.
distribution of this patchiness, is that magnetic remanence
effects (as permitted by Königsberger ratios above unity; 5.4.2 Repeated ophiolite blocks visible in the
Fig. 7c) locally have a significant influence on the aeromag- aeromagnetic map
netic anomalies related to the Geotimes unit.
Exposures of Lasail lavas greater than ∼ 100 m wide A positive gravity anomaly suggests that a repeated ophio-
(comparable to the spatial resolution of the aeromagnetic sur- lite block exists underneath the Batinah plane to the north-
vey) consistently coincide with relatively weak RTP anoma- east of the Sarami block (Shelton, 1990). Isolated outcrops
lies between 0 and 0.5 µT. The consistency of the Lasail of serpentinite and listvenite around the anomaly may be the
unit’s weak RTP anomalies is presumably due to the unit’s surface expressions of this block, or may simply be large olis-
consistently weaker magnetism compared to other units toliths (BRGM, 1986b). This repeated block is well-resolved
(Fig. 7a, b). as an arcuate 2–2.5 µT positive RTP anomaly 30–40 km long
Tholeiitic Alley typically corresponds to consistent, strat- and 5–8 km wide in Fig. 10 (Anomaly no. 3). The intensity
iform, and positive anomalies between 1 and 2.5 µT, which and dimensions of the magnetic anomaly resemble those of
are straightforward to interpret and to trace between refer- the east-dipping Alley sequences in the Hilti and eastern Fizh
ence points. There are some areas of Tholeiitic Alley lavas blocks. However, with limited outcrop available for confir-
which correspond to weak RTP magnetism, but this is typi- mation, it is also feasible that the geophysical anomaly cor-
cally where Tholeiitic Alley is anomalously thin (e.g. Aswad responds to a sheet of serpentinised magnetite-bearing peri-
block) or cut by major faults (e.g. Haylayn block). dotite. Less-clearly resolved, but nevertheless positive, mag-
Although the measured magnetism of the Boninitic Al- netic anomalies also extend as strips to the north (Anomaly
ley unit is rather weak (Fig. 7a, b), it has an inconsistent no. 1) and south (Anomaly no. 2) of the exposed eastern Fizh
RTP character ranging from 0 to 2.5 µT. This can possibly block lavas. These anomalies indicate that the eastern Fizh

Figure 10. Reduced-to-pole (RTP) magnetic anomalies for the Batinah coast and Yanqul area (Isles and Witham, 1993), marked with our
interpretations of the block boundaries, the extent of volcanic bedrock, structurally controlled re- or de-magnetised features, the potentially
magnetic Batinah complex modified after BRGM (1986a, b) and Woodcock and Robertson (1982), and possible repetitions or extensions of
ophiolite blocks (numbered) visible in the RTP map, supported by the gravity interpretations of Shelton (1990). The Batinah RTP map and
remagnetised features are viewable as separate layers in the supplementary Geospatial PDF. The Batinah RTP base map is reproduced with
permission from the Public Authority for Mining, Sultanate of Oman, who reserve all rights.

repetition probably continues for some 80 km along strike ping. To illustrate the relative confidence in which the var-
and is open to extension along the eastern coast of the U.A.E. ious areas are mapped, Fig. 11 shows the locations of these
previous maps, the 279 lava samples, 9 dyke or sill samples,
5.4.3 Block boundaries and fault zones in the and also the sites where we mapped the units using GPS-
aeromagnetic map equipped digital tablets.
Our sampling and observation density is highest in the
The aeromagnetic map reveals weak and even negative mag- north of the mapping area. This is partly due to the wider,
netic anomalies at the boundaries between all the mapped more tectonically complex sequence in the north but also
tectonic blocks (Fig. 10). The boundary zones between due to the availability of samples and observations collected
the Fizh, Hilti, and Sarami blocks are complex and oc- during parallel projects in this area. Between these sampling
cur across multiple structures which isolate smaller blocks points, the certainty of our unit identifications depends on
between the main blocks (e.g. between Wadi Bargah and the field and aeromagnetic characteristics of the units, on the
Jizi, also south of Wadi Ahin). These boundary zones along local structural complexity, and on the coverage by previous
Wadi Jizi, Wadi Ahin, and Wadi Bani Suq are occupied maps. For example, our fewer reference samples in the south-
by sediment-dominated melange (Robertson and Woodcock, ern blocks are somewhat offset by the excellent coverage of
1983a), which explains the weak RTP magnetism of these the JICA (2000, 2002) maps in those areas.
zones. As Geotimes is mostly well defined in the regional map
The aeromagnetic map also reveals strips of weak and set and typically readily recognisable in the field, this unit
slightly negative RTP anomalies up to kilometres wide co- is delineated with high confidence in our map. Similarly, the
inciding with major fault zones within the blocks (Fig. 10). generally consistent field appearance, stratigraphic position,
Where both sides of the fault consist of igneous rocks (e.g. and magnetic character of the Lasail unit allowed for rela-
Wadi Bargah and north and south of Wadi Hawqayn; Fig. 10) tively unambiguous mapping on the surface and inference of
these anomalies must be explained by de- or remagnetisa- its presence below cover.
tion around the fault planes. The approximate spatial extents Collectively, the Alley group lavas were relatively straight-
of these tectonically controlled weak aeromagnetic anoma- forward to delineate on the basis of their distinctive field
lies are marked in Fig. 10. The slightly negative RTP values characteristics and generally positive RTP anomalies. Where
within many of these anomalies cannot be achieved purely boninites are thickly accumulated and altered to pale,
by magnetic mineral destruction. Rather, they imply that a greenschist-facies spilites at their basal contact with Tholei-
secondary magnetic remanence that is not parallel to the ge- itic Alley (e.g. Aswad, Fizh, Daris), this contact is easy to
omagnetic field has been imparted to the surrounding vol- follow in the field and in satellite imagery. However, where
canics. Fluid-related remagnetisation of the lower crustal and Boninitic Alley occurs as a thin cap on the sequence (e.g.
mantle sections of the ophiolite during obduction and rota- Hilti, Sarami blocks), it often proved difficult to distinguish
tion has previously been documented throughout the ophi- from Tholeiitic Alley without geochemical analysis and care-
olite (Feinberg et al., 1999; Morris et al., 2016; Usui and ful field observations. It is therefore possible that further oc-
Yamazaki, 2010). The structurally controlled remagnetisa- currences of Boninitic Alley may be present as thin cappings
tion outlined in Fig. 10 indicates that the fluids responsible in addition to the mapped layers. Consequently, areas where
for resetting magnetism deeper in the section may have lo- boninites are suspected, but are untested by sampling, are
cally transgressed into the upper crust. Field evidence for marked on the map as “Undifferentiated Boninitic & Tholei-
this hydrothermal resetting is compelling, as the idiomor- itic Alley”.
phic hematite in the obduction-related quartz–hematite veins Tabular intrusions intruded along faults, as well as epidote-
(Q0 ) is often strongly magnetic (described in Sect. 3.6). cemented fault-breccias, were taken as evidence for “syn-
As pure hematite is only very weakly magnetic, this mag- magmatic” deformation, similarly to Reuber (1988). Unlike
netism presumably results from pseudomorphic replacement chlorite, epidote is not part of the overprinting Q0 quartz–
of hematite by magnetite (mushketovisation) which post- hematite vein assemblage, and pervasive epidote alteration
dates the emplacement of the quartz–hematite veins. in the upper crust occurred mostly during Tholeiitic Alley-
phase magmatism (Gilgen et al., 2016).
The complete map of the volcanic units is presented in
6 Map construction and presentation Fig. 12. For detailed use, the reader is referred to the Geospa-
tial PDF provided in the Supplement to this article. The
Previously published volcanostratigraphic sections and geo- Geospatial PDF format allows the georeferenced input and
logical maps from 22 publications were used as a basis for final map layers to be viewed in Adobe Acrobat® and di-
mapping the Semail volcanic units. The suite of 190 lava rectly imported into common geospatial mapping software
samples that we assigned to volcanostratigraphic units by (e.g. Avenza MapPublisher® ).
geochemistry were combined with 89 previously published
sample locations to provide reference points for our map-

Figure 11. Summary of datasets used to construct the final map in addition to the aeromagnetic survey in Fig. 10. Red circles: field mapping
locations (this study). Reference samples include triangles: volcanic glasses (this study); filled diamonds: dykes and sills (this study); unfilled
diamonds: dyke samples from Adachi and Miyashita (2003); squares: lava and volcanic glass samples from Gilgen et al. (2014, 2016),
Kusano et al. (2017), and MacLeod et al. (2013); unfilled squares: P1 and P2 intrusive samples from Haase et al. (2016) and de Graaff et
al. (2019). Numbered outlines: coverage of previous publications used to support our mapping: (1) BRGM (1993b), (2) BRGM (1993a), (3)
BME (1987a), (4) BME (1987c), (5) BME (1987b), (6) BRGM (1986b), (7) BRGM (1986a) and local maps, (8) Reuber (1988), (9) Kusano
et al. (2012), (10) Umino et al. (2003), (11) Haymon et al. (1989), (12) A’Shaikh et al. (2005), (13) Alabaster et al. (1980), (14) Robertson
and Woodcock (1983), (15) Alabaster and Pearce (1985), (16) Ernewein et al. (1988), (17) JICA (2000), (18) Umino (2012), (19) Kusano et
al. (2014), (20) JICA (2002), (21) Reuber et al. (1991), and (22) Adachi and Miyashita (2003).

Figure 12. New map of volcanic units in the Semail ophiolite (this study). Strong colours: upper-crustal units in outcrop. Pale conjugate
colours: upper-crustal units inferred as bedrock underlying sedimentary cover. Lower crustal and mantle rocks (copied and simplified from
the regional map set) are shown for context. See Sect. 1.4 for changes versus previous maps and details on bedrock inferences under
sediments. VMS deposit locations after Gilgen et al. (2014). The georeferenced multi-layer version of this figure, together with the mapping
inputs in Fig. 11, is provided in the supplementary Geospatial PDF.

7 Discussion of newly mapped features turbed sections yield the closest approximations of volumet-
ric proportions from areal proportions. For this reason, the
7.1 Tectonic omissions of Geotimes in the Haylayn tectonically imbricated volcanics on the western side of the
block ophiolite are excluded from the area calculations in Table 2.
The north-eastern side of the northern ophiolite (i.e. the Bati-
The Geotimes unit is present above the SDC throughout nah coast) dips relatively consistently to the east. From this
most of the map area (Fig. 12). However, in the Haylayn strip we calculated the bedrock areas of each upper-crustal
block (in the region of Wadi Wadiyah), Geotimes occurs unit within the mapping area (Table 2). Due to tectonic com-
only discontinuously and the volcanic section is almost in- plications in the Aswad, eastern Fizh, Sarami, and Haylayn
variably bounded above and below by faulting. These faults blocks, however, we additionally calculated these areas based
correspond to strip-like negative anomalies in the RTP map on the relatively intact upper-crustal exposures along the
(Fig. 13). Along these faults, Tholeiitic Alley lavas are com- western Fizh block (Wadi Fayd to Wadi Kabiyat) and Hilti
monly juxtaposed against the SDC, and Geotimes is either block (Wadi Bargah to Wadi Ahin). These Fizh and Hilti
absent or is only a few hundred metres thick, as previously areal proportions are the most representative available for the
noted by Lippard et al. (1986). Rather than viewing the con- pre-obduction state of the Semail crust (Table 2).
tact between Alley and the SDC as conformable, as previ- Interestingly, the Batinah coast and Fizh–Hilti spatial sub-
ously mapped by BRGM (1986a) and Juteau et al. (1988), we sets are relatively consistent with each other. In both cases,
attribute the missing Geotimes lavas in the Haylayn block to our inferred bedrock maps indicate that Tholeiitic Alley is
faulting (Fig. 13). This conclusion is supported by the neg- the most areally extensive volcanic unit. If the SDC, which is
ative RTP anomalies along this contact as well as by our largely comagmatic with Geotimes (Miyashita et al., 2003;
observation, ∼ 500 m to the north of Wadi Wadiyah, of a Pearce et al., 1981), is considered with the volcanics and
fault zone several metres wide hosting a Q0 quartz–hematite upper-crustal intrusions to make up an upper-crustal total,
vein marked by sub-horizontal slickensides. This zone oc- the proportion of axial upper crust (SDC + Geotimes; Phase
curs within the weak RTP anomaly and between outcrops 1) is ∼ 54 vol %, whereas Phase 2 off- or post-axial lavas
of the SDC and Tholeiitic Alley (confirmed by sampling; and intrusions constitute ∼ 46 vol %. In the Fizh–Hilti sub-
Fig. 13). The tectonically disturbed character of the contact set, 24 vol % of the volcanics or 15 vol % of the upper crust
between the SDC and the volcanics in the adjacent Sarami is made up of boninite-series lavas.
block is also demonstrated by a major fault zone several me- These significant volumes of Phase 2 lavas imply that their
tres wide at the base of Geotimes in Wadi Shaffan (Einaudi et generation must have depleted considerable volumes of man-
al., 2003). The presence of such faults precludes simple com- tle following the axial-spreading-stage (Geotimes) melt ex-
parisons between the southern and northern blocks based on traction. Such extensive Phase 2 melting and magmatic over-
volcanostratigraphy. printing has previously been described in the mantle sec-
tion (Arai et al., 2006; Python and Ceuleneer, 2003). Plu-
7.2 Proportions of the upper-crustal units tonic evidence for Phase 2 magmatism has also been doc-
umented throughout the ophiolite (Adachi and Miyashita,
Early descriptions of the Semail volcanic stratigraphy 2003; Goodenough et al., 2010; de Graaff et al., 2019; Haase
showed Geotimes with a relatively consistent thickness of 1– et al., 2016; Juteau et al., 1988; Rollinson, 2009; Tsuchiya et
1.75 km, overlain by Lasail or Alley sequences ∼ 0.75 km al., 2013; Yamasaki et al., 2006). The volumetric significance
in thickness (Alabaster et al., 1982). In the Fizh block, Reu- of this late plutonism is poorly constrained, though Phase 2
ber (1988) contrastingly indicated that the “V2” lavas could intrusives apparently constitute about half of the crustal ex-
attain thicknesses of up to 1.5 km, whereas Geotimes can posures in the U.A.E. portion of the ophiolite (Goodenough
be as thin as 0.5 km. Our mapping indicates that complete et al., 2010, 2014), which is comparable to the proportions
tectonically undisturbed volcanic sections between the SDC of Phase 1 to Phase 2 lavas documented here. It follows from
and the post-Alley Suhaylah Formation are rare. The strati- these proportions that many of the major tectonic and mag-
graphic columns in Fig. 12 are constructed from dip mea- matic features in the mantle and lower crustal sections are
surements along some of these complete sections (sections feasibly related to Phase 2 magmatism. For example, ∼ 14
B, D, E) and along other instructive sections. These new mantle diapirs have been mapped by Nicolas et al. (2000).
sections are more consistent with Reuber (1988), indicating Although troctolitic cumulates at the head of the Maqsad
that undisturbed Geotimes can be as thin as 0.3 km, whereas mantle diapir suggest it formed during the MORB-like ax-
Tholeiitic Alley is up to 1 km thick, and Boninitic Alley lo- ial phase, the same is not clear for many of the other diapirs
cally up to 2 km thick. (Python and Ceuleneer, 2003).
In addition to these stratigraphies, the inferred areas of The unit proportions documented in Table 2 are the first es-
buried bedrock in our map permit realistic estimates to be timates of the composition of proto-arc upper crust based on
made of the areal proportions of the different upper-crustal a large mapped area (2012 km2 ). The proportion of boninites
units. Uniformly dipping but otherwise tectonically undis- in the Semail ophiolite is comparable to the exposures of

Table 2. Areal extents and inferred volumetric proportions of each Semail volcanic unit based on the bedrock map in Fig. 12 for both the
Batinah coast and the structurally intact western Fizh and Hilti blocks. Proportions are given both as fractions of the volcanics and of the
upper crust (including upper-crustal intrusions and the sheeted dike complex, SDC).
Batinah coast Fizh & Hilti blocks Summary

Unit Bedrock Vol % of Vol % of Bedrock Vol % of Vol % of Rock Vol % of upper-
area (km2 ) upper crust volcanics area (km2 ) upper crust volcanics types crust (Fizh & Hilti)
Boninitic Alley 120 6.0 11 50 15 24 Boninite-series 15
Undifferentiated Alley 148 7.4 13 2 0.6 1.0 Off- and post-axial
Tholeiitic Alley 416 21 37 82 24 40 tholeiite lavas and 31
Lasail 76 3.8 6.7 14 4.1 6.8 intrusions
Geotimes 374 19 33 58 17 28 Phase 1 axial
54
SDC 708 35 – 124 37 – lavas and dykes
Upper-crustal intrusions 170 8.4 – 9.0 2.7 –
Total volcanic rocks 1134 56 206 61
Total upper crust 2012 339
Vol %: volume fraction assumed equivalent to areal fraction
boninitic upper crust estimated for the southern Mariana in the southern ophiolite blocks (de Graaff et al., 2019; Haase
forearc by submarine surveying (20 area %–30 area %; Rea- et al., 2016), structural arguments for outcrop bias in the
gan et al., 2013). U.A.E. (Ambrose and Searle, 2018), and by the synchronous
Although the unit proportions documented in the present ages for prograde metamorphism in the southern and north-
study are representative for a significant length along the ern metamorphic soles (Guilmette et al., 2018).
palaeo-spreading axis (over ∼ 150 km), any large-scale vari- Owing to the tectonic omissions of Geotimes in the Sarami
ations perpendicular to this axis are difficult to estimate and Haylayn blocks, our new map is somewhat equivocal
across the 30–60 km wide Semail ophiolite. The comparable on the issue of ophiolite-scale variations in subduction influ-
units and geochemistry between the NE (Batinah coast) and ence. However, it is at least clear that subduction-influenced
SW (Yanqul) flanks of the ophiolite, which are separated by Lasail and Alley volcanism occurred over the length of the
∼ 30 km perpendicular to the spreading axis, suggests some mapping area. In fact, some of the ophiolite’s thickest ac-
continuity of volcanostratigraphy at this scale. However, flat- cumulations of Felsic Alley and Boninitic Alley lavas oc-
lying ophiolites (e.g. Mirdita), the IBM arc basement and the cur in the Haylayn and Wuqbah blocks (Figs. 12, 13), indi-
recently recognised Matthew–Hunter proto-arc show signif- cating that, at least during Phase 2 magmatism, subduction-
icant variations in upper-crustal architecture with distance influenced volcanism was comparably significant throughout
from the trench (Dilek et al., 2008; Hickey-Vargas et al., the mapping area.
2018; Patriat et al., 2019; Reagan et al., 2017). The upper-
crustal unit proportions documented herein are therefore not 7.4 Lateral continuity of the Alley unit
necessarily representative of an entire proto-arc crustal tract,
but rather represent the best estimate currently available. On the basis of the Suhaylah Village section, it has long been
held that Geotimes locally tops the volcanic sequence, be-
7.3 Along-ophiolite differences in subduction-zone ing directly overlain by Suhaylah Formation sediments, and
influence therefore that the Alley strata are discontinuous (Fleet and
Robertson, 1980; Lippard et al., 1986). However, the pres-
It has previously been proposed that a gradient of subduction- ence of Geotimes at the top of this section was apparently
zone influence on magma generation existed from SE to NW not confirmed by geochemical analyses. In contrast, our field
along the Semail ophiolite. This proposal is based on Cr con- observations, whole-rock analyses, and clinopyroxene anal-
tents in mantle spinel (Python et al., 2008), abundant Phase yses reveal that the uppermost ∼ 300 m of the Suhaylah sec-
2 intrusives in the U.A.E. (Goodenough et al., 2010) versus tion is in fact composed of Tholeiitic Alley pillow basalts and
abundant troctolites in the south-eastern blocks (Python and andesitic massive flows (Fig. 14).
Ceuleneer, 2003), and an apparent paucity of Lasail/Alley A similar locality with Geotimes topping the section was
volcanics in the south-eastern ophiolite blocks. The existence indicated in upper Wadi Ahin (map by BME, 1987b). How-
of this gradient has since been challenged by the apparently ever, our sampling (TB4-23A) and field observations show
uniform hydrous influence on axial magmatism (MacLeod et that Boninitic Alley lavas are present there, and that they are
al., 2013), the identification of numerous Phase 2 intrusions actually faulted against, rather than overlain by, sediments on

Figure 13. Mapping example: tectonic omissions of Geotimes in the Haylayn block near the boninite-hosted Daris-3A VMS deposit. (a) Dis-
tribution of volcanic units according to the pre-existing regional geological map (BRGM, 1986a). (b) RTP aeromagnetic survey (Isles and
Witham, 1993) with inferred bedrock contacts and faults (this study). (c) Revised map (this study) with sample location symbols as in Fig. 11.
Coordinate grid: UTM 40 N in kilometres.
both sides of the wadi. Additional occurrences of Suhaylah

sediments directly overlying Geotimes had been mapped be-
tween Wadi Hayl and Wadi Kabiyat, ∼ 6 km north of Suhay-
lah in the Fizh block (BME, 1987c). Our field observations,
as well as those of Robertson and Woodcock (1983), rather
indicate that this sequence is disrupted by normal faulting,
which has removed much of the section. Our sampling fur-
ther shows that the uppermost unit along this section is
Boninitic Alley (samples LD10-1461, LD10-1464). No other
examples of Suhaylah sediments directly overlying Geotimes
are indicated in the regional map set or were found during the
present study. We thus conclude that Alley/V2-stage lavas are
continuously present throughout the northern ophiolite.
7.5 Lasail basaltic seamounts
The discrete accumulations of Lasail lavas that overlie Geo-

times have previously been likened to off-axis “seamounts”,
regularly spaced and associated with VMS deposits (Al-
abaster et al., 1982; Lippard et al., 1986; Pearce et al., 1981).
Prior to this study, these accumulations had not been mapped
out in detail. The newly mapped Lasail accumulations are
generally less than a few hundred metres thick, less than a
kilometre wide, and are irregularly distributed at intervals
of 10–20 km wherever the top of the Geotimes unit is well-
exposed. These findings thus support the seamount interpre-
Figure 14. Mapping example: the fault-bounded north-east-dipping
tation, but their frequency is slightly higher and their volu-
volcanic section at Suhaylah, modified after Robertson and Wood-
cock (1983) and BME (1987c), showing Tholeiitic Alley lavas (this
metric significance slightly lower than originally suggested
study) overlying a thin Geotimes section and underlying the type- by Alabaster et al. (1982).
locality Suhaylah sediments. Our samples analysed to prove unit Thick Lasail accumulations occur in the region of pro-
affiliations are numbered: (1) SP18-A1, (2) SP18-A3, (3) TB4-17J, posed spreading-axis segment boundaries, e.g. Wadi Hatta
(4) SP18-A8, and (5) SP18-A9. to Rajmi, and Ghuzayn, as noted by Alabaster et al. (1982)
and MacLeod and Rothery (1992), but the unit’s distribution
is not limited to these positions. Lasail also occurs within the
middle of intact structural blocks (e.g. Salahi area). More-

over, as also shown by Gilgen et al. (2014), there are no clear of the newly mapped Rajmi–Zab’in boninite accumulations
associations between these seamounts and VMS deposits. (the only orthopyroxene-bearing volcanic unit; Ishikawa et
The most significant of the Lasail accumulations occurs al., 2002), the possibility that some of these gabbronorites are
between wadis Ashar and Rajmi (Fig. 15) and does in fact co- the intrusive equivalent of the boninite lavas warrants consid-
incide with a proposed area of axial segmentation (MacLeod eration.
and Rothery, 1992; Reuber, 1988; Smewing, 1980). With The mantle and lower crustal section in the Rajmi–Zab’in
a maximum thickness of ∼ 1 km and a continuous base ∼ area (Fig. 16) is cut by numerous high-temperature mylonitic
15 km along strike (Figs. 12, 15), the weakly magnetic La- shear zones (BRGM, 1993b; Smewing, 1980; Takazawa et
sail lavas are well distinguished from the strongly magnetic al., 2003). Though cataclastic zones beneath the Moho and
overlying Alley lavas in the aeromagnetic survey, as con- Q0 quartz–hematite veins in the upper crust (BRGM, 1993b)
firmed by sampling and field observations (Fig. 15). Isolated indicate that many of the faults in this area experienced
occurrences of Axial Lasail volcanism began synchronously post-magmatic brittle reactivation, the high-temperature fab-
with earliest Geotimes eruptions in this area (Alabaster et al., ric and sheeted intrusions emplaced along these fault zones
1982; Belgrano and Diamond, 2019) and the Geotimes unit demonstrate the syn-magmatic timing of the original struc-
itself is abnormally thin, being only a few hundred metres tures. Greater displacement of the axial sequence relative to
thick (section A in Fig. 12), suggesting short-lived axial vol- the Alley sequence along these syn-magmatic faults further
canism and a rapid development of subduction influence in suggests that part of this upper-crustal deformation occurred
this area. after axial magmatism but before or during Alley-stage mag-
matism.
7.6 Semail boninites The centres of the two most prominent boninite lava
outcrops at Rajmi–Zab’in are clearly fed by dykes intrud-
Analyses of volcanic glass demonstrate that boninites (LSB) ing along these syn-magmatic faults (Fig. 16), demonstrat-
sensu stricto are present throughout the northern ophiolite ing that boninite migration through the existing axial crust
(Fig. 6; Kusano et al., 2017). The well-defined high-V/Ti and was structurally controlled. Moreover, the mylonitic shear
high-Zr/Y fractionation trends followed by both Boninitic zones which underlie and connect with these syn-magmatic
Alley spilites and the boninitic glasses (Fig. 5) indicate that faults indicate that boninitic melts were structurally chan-
all of these lavas belong to the same unit, which in turn de- nelled from the level of the mantle. In fact, the major man-
rives from an LSB parental melt (Fig. 6). tle shear zone underlying the Wadi Zab’in boninite dyke
Prior to this study, little was known about the areal ex- and lava complex (Fig. 16) is well-studied and records the
tent and volcanic morphology of boninites in the ophiolite. passage of hydrous fluids and melts from the metamor-
Our mapping, though unlikely to have recorded all the mi- phic sole through the entire mantle sequence (Arai et al.,
nor boninite occurrences within the mapping area, shows 2006; Kanke and Takazawa, 2014; Takazawa et al., 2003).
that boninites occur in each of the mapped blocks. These Comparable shear-zone-controlled Phase 2 magmatism is
occurrences are discontinuous and have a highly variable widespread in the U.A.E. portion of the ophiolite (Goode-
stratigraphic thickness (0.1–2 km; sections in Fig. 12). In the nough et al., 2010; Styles et al., 2006). The structurally con-
Wuqbah, Hilti, and Sarami blocks, layers of Boninitic Alley trolled boninitic volcanism documented herein completes the
only up to ∼ 200 m thick discontinuously cap and locally in- spectacular mantle-to-seafloor section at Rajmi, constituting
terfinger with the uppermost Tholeiitic Alley lavas (e.g. Ku- a complete anatomy of a proto-arc magmatic system.
sano et al., 2014). Contrastingly, there are boninite accumu- Interestingly, the boninite-focusing fault zones between
lations 1.5–2 km thick with bases 2–5 km wide in the Aswad, Wadi Rajmi and Wadi Zab’in extend directly into the man-
Fizh, and Haylayn blocks (Fig. 12). tle (Fig. 16) and there they enclose a cluster of structurally
controlled podiform chromitite deposits around Wadi Rajmi
7.6.1 Structurally controlled boninitic volcanism (Boudier and Al-Rajhi, 2014; Rollinson, 2008). Chromites
within these deposits record formation from a series of melts
The best-preserved boninite accumulations can be found evolving from arc-like (comparable to Tholeiitic Alley) to
between Wadi Rajmi and Wadi Zab’in in the Fizh block highly oxidised boninitic compositions (Rollinson, 2008;
(Fig. 16). Here, three sub-parallel dyke swarms spaced at Rollinson and Adetunji, 2015). The short duration permit-
∼ 5 km fed a 2 km thick accumulation of Boninitic Alley ted for boninitic volcanism (∼ 0.5 Ma; Gilgen et al., 2016,
lavas. Ishikawa et al. (2002) noted the boninitic character of and references therein) together with the close geochemi-
these dyke swarms, which apparently show mutually intru- cal and structural association between these chromite de-
sive relationships with orthopyroxene-series (gabbronorite) posits and the thick overlying boninite accumulations sup-
intrusions (Umino et al., 1990). Together with the shallowly ports the long-held association between hydrous, boninitic
SSE-dipping SDC along this segment, the presence of these melts and podiform chromitite formation (Matveev and Ball-
gabbronorites has previously been linked to axial segmenta- haus, 2002; Rollinson, 2008). The mantle sections beneath
tion (MacLeod and Rothery, 1992; Reuber, 1988). In light the other significant boninite accumulations (e.g. Aswad,

Figure 15. Mapping example: part of a Lasail basaltic accumulation in the northern Fizh block near the Tholeiitic Alley-hosted Mandoos
VMS deposit (Gilgen et al., 2014). (a) Distribution of volcanic units according to the pre-existing regional geological map (BRGM, 1993b).
(b) RTP aeromagnetic survey (Isles and Witham, 1993) with inferred bedrock contacts and faults (this study). (c) Revised map (this study)
with sample location symbols as in Fig. 11. Coordinate grid: UTM 40 N in kilometres.
Daris) may therefore be attractive areas for future chromite fractionation path, supporting the hypothesis that these in-
exploration. trusives formed by fractional crystallisation of a Geotimes
melt (Haase et al., 2016). The Phase 2 intrusive compositions
7.7 Relationships between the volcanic units and the rather mostly follow the Tholeiitic Alley fractionation path,
intrusive phases extending to far lower Ti/Zr ratios than recorded for the La-
sail lava unit, supporting the intrusive–extrusive equivalence
The intrusive rocks of the Semail ophiolite have been subdi- of the Phase 2 intrusive and Tholeiitic Alley suites.
vided into two main groups, with the axial “High-Level” and Interestingly, a subset of Phase 2 intrusive composi-
post-axial “Late Intrusive” groupings of Lippard et al. (1986) tions are continuous with the upper Boninitic Alley field
respectively coming under the Phase 1 and Phase 2 groupings (Fig. 5a), suggesting that these intrusives evolved from de-
of Goodenough et al. (2014). It is generally accepted that the pleted, boninite-series melts. The intrusions emplaced into
Phase 1 intrusions are related to the axial SDC and Geotimes Boninitic Alley lavas in the Rajmi area (Fig. 15) are also
unit (Goodenough et al., 2010; Lippard et al., 1986); how- presumably boninite series, as these boninites are the last-
ever, the relationship between the Phase 2 intrusives and the erupted lavas in this area.
various Phase 2 volcanic units is less clear. This uncertainty
has persisted because prior to this study, the distributions and 7.8 Relationships between faulting, dyke swarms,
relative proportions of the Phase 2 volcanic units were un- boninites, and sulfide ore deposits
clear.
Several lines of evidence suggest that the majority of the A genetic association between late dyke swarms, normal
Phase 2 intrusives are comagmatic with the Tholeiitic Alley faulting, volcaniclastic breccias, boninitic volcanism, and
lavas. Firstly, our map shows that the Alley suite is volumet- VMS deposits has previously been noted for the “Bowl-
rically far more significant than the Lasail lavas (Table 2), ing Alley” fault zone between Wadi Bani Umar and Wadi
suggesting that Alley should have a more significant pro- Kabiyat (Fizh block) and the nearby Aarja and Bayda VMS
portion of intrusive equivalents. Secondly, the Phase 2 in- deposits (Gilgen et al., 2014; Haymon et al., 1989; Smewing
trusive complexes characteristically span a range of com- et al., 1977). Early mapping viewed the fault zone as an east-
positions throughout the ophiolite (gabbro–diorite–tonalite– facing half graben that along its southern reach swings SW
trondhjemite), often within single complexes (Lippard et al., into Wadi Jizi (Smewing et al., 1977). In contrast, our map-
1986). The equivalent compositional range (basalt to rhyo- ping reveals unusual repetition of the Geotimes unit to the
lite) is characteristic of the Tholeiitic Alley lava suite, but east of Tholeiitic Alley, which is suggestive of a full graben
not of Lasail, which is predominantly comprised of primitive running N–S from at least present-day Wadi Fizh to Wadi
basalts. Bargah (Fig. 17). The continuation of this rift-like feature
This evidence is supported by the Zr–Ti trends of the south of Wadi Jizi is marked by a strip of Boninitic Alley
different lavas and intrusives (Fig. 5a). The Phase 1 intru- lavas, Felsic Alley lavas, and Suhaylah sediments bounded
sive compositions fall along and continue from the Geotimes to the E and W by faults, with Geotimes lavas repeated

Figure 16. Boninites capping the volcanic succession in the Wadi Rajmi and Wadi Zab’in region and their structural connection to previ-
ously mapped boninite dykes, ultramafic and orthopyroxene-series intrusives (Adachi and Miyashita, 2003; BRGM, 1993b; Umino et al.,
1990; Usui and Yamazaki, 2010), mantle shear zones (Boudier and Al-Rajhi, 2014; BRGM, 1993b; Takazawa et al., 2003), and podiform
chromitites (Boudier and Al-Rajhi, 2014; BRGM, 1993b; Rollinson, 2008). The boninite-hosted Safwa VMS deposit is also shown (Gilgen
et al., 2014). Coordinate grid: UTM 40 N in kilometres. Sample symbology as in Fig. 11. Strong colours denote outcrops; pale conjugate
colours denote inferred bedrock beneath gravels.
to the E of the Alley lavas. During ophiolite emplacement, and adjacent to the faults and grabens bounding the de-
the rift was dismembered by faulting parallel to Wadi Jizi posits, and the grabens are filled with blocky volcaniclas-
(marked by the Zabyat conglomerates; Robertson and Wood- tics and boninitic lavas. The boninite-hosted Safwa VMS
cock, 1983) and by NNE–SSW faulting along the western deposit is also located between the two major W–E-striking
edge of the eastern Fizh block. Several further corridors of in- fault-localised boninite dyke swarms between Wadi Fayd and
trusions run parallel to the Bowling Alley axis, which we in- Wadi Zab’in (Fig. 16).
terpret as related to secondary rift structures (Fig. 17). Thus, The extensional tectonics associated with these late dyke
the entire axis-parallel rift is ∼ 30 km long and it was the lo- swarms therefore appears to be favourable both for VMS
cus of VMS mineralisation from the end of Geotimes (Lasail mineralisation, as pointed out by Smewing et al. (1977) and
deposit) to Boninitic Alley volcanism (Aarja deposit; Smew- Haymon et al. (1989), as well as boninite extrusion, as sim-
ing et al., 1977; Gilgen et al., 2014), indicating that exten- ilarly noted for the “infill” boninites of the Troodos ophio-
sional tectonics and vigorous hydrothermal activity occurred lite (Cameron, 1985; Osozawa et al., 2012). Further work is
in this area throughout axial volcanic Phase 1 and later Phase needed to resolve the specific timing and genetic associations
2. between these features.
Our mapping indicates that a similar lithotectonic associ-
ation exists for the boninite-hosted Rakah and Hayl al Safil
VMS deposits in the Yanqul area (Gilgen et al., 2014). Here,
Alley-stage “sheeted” dyke swarms (confirmed by analy-
sis of dyke samples TB3-25H and TB4-20L) run parallel

8 Comparison with the Izu–Bonin–Mariana proto-arc

sequence
The Semail ophiolite has been proposed as analogous to the

Eocene Izu–Bonin–Mariana (IBM) proto-arc crust on the ba-
sis of their similar structure and their similar 1–2 Myr pro-
gression from axial basaltic to boninitic magmatism (Bel-
grano and Diamond, 2019; MacLeod et al., 2013; Reagan et
al., 2019; Rioux et al., 2016; Rollinson and Adetunji, 2015).
To facilitate comparison of the mapped Semail sequence with
the IBM proto-arc record, we plot the updated Semail vol-
canic unit fields from Fig. 5 together with the published com-
positions of IBM lavas, as well as Tonga Trench boninites
and MORB glasses, in Fig. 18.
The biggest geochemical discrepancy between the Semail
and IBM volcanic sequences is the absence of a high-Ti/V
“moist” MORB-like equivalent to Geotimes in the IBM (Bel-
grano and Diamond, 2019; MacLeod et al., 2013). The high-
est Ti/V forearc basalts recovered from the IBM derive from
the Amami-Sankaku basin rear-arc (Unit 1; Hickey-Vargas
et al., 2018) and from the Bonin forearc (DeBari et al.,
1999). Originally thought to be fragments of pre-subduction-
initiation MORB, these Bonin forearc basalts were later re-
classified as forearc (proto-arc) basalts (Reagan et al., 2010).
These high-Ti/V proto-arc basalts overlap with the Geotimes
and MORB arrays in Fig. 18a, c, and d, but Geotimes follows
a still-higher-Ti/V trend in Fig. 18b. No direct equivalent
for the Geotimes unit has thus been identified in the IBM
proto-arc sequence. A pre-axial episode of melt extraction,
inferred for the IBM source from trace elements and Hf–Nd
isotopes (Li et al., 2019; Shervais et al., 2019; Yogodzinski et
al., 2018) but so far not for the Semail ophiolite, may explain
this discrepancy (Belgrano and Diamond, 2019). This dif-
ference in pre-axial mantle source history implies that when
subduction initiation occurs above normal, depleted MORB
mantle as appears to have been the case for the Semail ophi-
olite (Belgrano and Diamond, 2019; Godard et al., 2006), the
first proto-arc lavas to erupt may have a composition closer
to MORB than forearc basalts from the IBM type locality
(Reagan et al., 2010).
Recent drilling in the IBM forearc identified a unit of prim-
itive basalts directly overlying the axial volcanics (P-FAB;
Shervais et al., 2018). The low-Ti primitive major-element
composition of P-FAB is closely comparable to Lasail. The
trace-element composition of P-FAB mostly overlaps with
Lasail but is not an exact equivalent. However, the compo-
sitional relationship between P-FAB and the axial forearc
Figure 17. Interpreted location of the “Bowling Alley” rift, mod-
basalts which they overlie is closely comparable to the re-
ified after Alabaster and Pearce (1985), BME (1987a), Robertson
lationship between Lasail and Geotimes, with both P-FAB
and Woodcock (1983), and Smewing et al. (1977). The rift is ge-
netically associated with numerous VMS deposits, minor copper and Lasail being slightly more primitive, depleted, and with
showings and areas of disseminated sulfide mineralisation (Haymon lower Ti/V ratios than their underlying axial lavas (Fig. 18).
et al., 1989; Smewing et al., 1977). We therefore hypothesise from the known lateral discontinu-
ity of the Lasail unit that P-FAB may also be laterally dis-
continuous atop the IBM axial crust.

Figure 18. Comparison of the expanded Semail volcanic unit fields with Izu–Bonin–Mariana (IBM) proto-arc volcanic rocks, north Tonga
Trench boninites and MORB glasses. West Pacific boninite analyses from N. Tonga Trench and N. Lau basin (Falloon et al., 2007; Falloon
and Crawford, 1991), Izu–Bonin forearc (IODP Expedition 786; Murton et al., 1992), and Bonin forearc (IODP Expedition 352; Shervais et
al., 2019). Normal Forearc Basalt (N-FAB) and Primitive Forearc Basalt (P-FAB) from Shervais et al. (2019) and other IBM forearc basalts
(FAB) from Reagan et al. (2010) and Ishizuka et al. (2011). High Ti/V forearc basalts from DeBari et al. (1999) and Hickey-Vargas et
al. (2018). MORB glasses from Jenner and O’Neill (2012).
Tholeiitic Alley is the closest geochemical equivalent to and Tonga Trench boninites are highly depleted and scat-
the normal IBM forearc basalts, especially in terms of Ti/V tered, but mostly overlap with Boninitic Alley. The most
(Fig. 18). The key difference between these two units is that notable exception to this similarity are the higher Zr con-
in the IBM these low Ti/V forearc basalts make up the axial tents in the IBM boninites, shifting them to higher Zr ar-
crust, whereas in the Semail ophiolite they are erupted on top rays in Fig. 18a and d. Positive Zr–Hf anomalies are ubiq-
of the axial crust. uitous in the IBM boninites and have been attributed to con-
Boninitic Alley is depleted in incompatible elements and tamination by slab melts originating from an amphibolite-
enriched in fluid-mobile elements similarly to the type Bonin bearing slab (Pearce et al., 1992; Reagan et al., 2017). The
forearc boninites (Ishikawa et al., 2002; Kusano et al., 2014). absence of these anomalies in Boninitic Alley suggests dif-
In terms of major elements, the Semail ophiolite lacks an fering boninite-contaminating slab components between the
HSB suite, and Boninitic Alley follows a slightly shallower Semail and IBM settings.
lower-Si fractionation trend on the MgO–SiO2 diagram than Overall, the comparable structure, chronology, and pro-
the IBM and Tonga Trench boninites (Fig. 6a). In terms of gressive depletion of the Semail and IBM magmatic se-
maximum MgO contents, Boninitic Alley is more similar to quences supports the proto-arc interpretation for the gene-
the subduction-initiation-related IBM boninites (Murton et sis of the Semail ophiolite. However, this short evaluation
al., 1992; Reagan et al., 2017; Shervais et al., 2019) than to suggests that the pre-axial mantle source and contaminating
the remarkably magnesian plume- and slab-window-related slab melts were probably different between the two settings
north Tonga Trench boninites (Falloon et al., 2007, 2008). (see also Belgrano and Diamond, 2019; Kusano et al., 2017).
In terms of the trace elements plotted in Fig. 18, the IBM The Semail ophiolite further lacks a suite of true HSB, which

are common in the IBM. The Semail–IBM analogy is thus inforce the strongly subduction-influenced character of the
far from perfect. However, complex tectonic and magmatic ophiolite as a whole (e.g. de Graaff et al., 2019; Goodenough
histories appear to be common, or even necessary, precondi- et al., 2010; Haase et al., 2016; MacLeod et al., 2013; Pearce
tions for initiating subduction (Patriat et al., 2019; Stern and et al., 1981) and caution against unconfirmed assumptions
Gerya, 2017). Compositional differences between different that geological features in the lower-crustal or mantle se-
proto-arc volcanic sequences and the IBM type locality are quences of the ophiolite (e.g. mantle diapirs) must necessar-
therefore perhaps to be expected, and these differences rep- ily reflect processes related to an oceanic spreading axis.
resent a promising line of research to describe the variety of Obduction-related remagnetisation is common around ma-
situations in which subduction may initiate. jor fault zones. Moreover, quartz–hematite and quartz–
carbonate veins with significant chloritic and carbonate alter-
ation haloes cut the entire crustal sequence and the anticlinal
9 Conclusions structure that warps the ophiolite. These veins clearly post-
date the ophiolite’s crust-formation stage but nevertheless
We have presented the methodology and final product of a they are often sited in reactivated syn-magmatic faults. Care
5-year project remapping the volcanic units in the Semail is therefore necessary when interpreting the significance of
ophiolite. Our key conclusions can be summarised into three chloritised hydrothermal features where clear cross-cutting
categories: relationships are unavailable.
Mapping methodology. Previous regional mapping of the Previously proposed associations between the ends of
ophiolite relied mainly on field observations and early satel- spreading-ridge segments, VMS deposits, and off-axis Lasail
lite imagery. With these methods, resolution of the various seamounts (Alabaster et al., 1982) are less convincing in light
units within the “V2” stage extrusives was impractical at a of the new map. Boninite accumulations, late dyke swarms,
regional scale. Our study has benefitted from rapid and rou- and extensional tectonics are rather more commonly associ-
tine whole-rock XRF and EMP mineral analyses. These tech- ated with economic VMS deposits.
niques have enabled large numbers of sample analyses to be Accretion of proto-arc crust. Despite representing appar-
used to identify units over several hundred square kilometres. ently differing modes of subduction initiation (Arculus et al.,
The bulk-magnetic properties of submarine lava suites 2015; Guilmette et al., 2018), the close comparability be-
are strongly controlled by major-element composition and tween the Semail and IBM magmatic sequences is supported
can be well preserved despite overprinting by sub-seafloor by our new mapping and sampling, as well as by recently col-
spilitic alteration. It should therefore be feasible to distin- lected datasets from both settings (Belgrano and Diamond,
guish volcanic units using aeromagnetic data in other ophi- 2019; MacLeod et al., 2013; Reagan et al., 2017; Shervais
olite or greenstone-belt settings. However, strong remanent et al., 2019; Whattam and Stern, 2011). However, subtle
magnetism in submarine volcanics leads to inconsistencies compositional differences between stratigraphically equiva-
in RTP data, and therefore less susceptible data transforma- lent lava units in each setting suggest contrasting pre-axial
tions should be used if possible. histories and slab-contaminations between the two settings.
Combining magnetic petrology and geochemistry allows The two most substantial axial (Geotimes) and post-axial
aeromagnetic data to be usefully correlated with specific (Tholeiitic Alley) volcanic episodes were somewhat vari-
units in volcanic terranes. Our confidence in these interpre- able in thickness but almost continuous along the length of
tations was sufficient to map specific units beneath sedimen- the palaeo-spreading axis within the ophiolite (∼ 150 km).
tary cover, thereby significantly expanding our bedrock map Contrastingly, the off-axis Lasail and final Boninitic Alley
and the prospective area for VMS deposit exploration. episodes were discontinuous and variable in thickness. Over-
Upper-crustal features of the Semail ophiolite. Of the two all, ∼ 50 vol % of the Semail upper crust was accreted after
lava units traditionally ascribed to Phase 2 magmatism, Alley the main axial stage.
is volumetrically far more significant than Lasail. Boninitic melts in the ophiolite were locally channelled
The traditional Alley unit is systematically divisible into along lithosphere-scale shear-zone networks which appar-
two main units at the ophiolite scale: lavas belonging to a vol- ently extend from the metamorphic sole to the palaeo-
umetrically dominant tholeiitic series versus lavas belonging seafloor. Podiform chromitite deposits formed along these
to a subordinate low-Si boninitic series. same melt channels in association with the ascending proto-
Incipient boninitic volcanism (“Depleted Lasail”) locally arc basaltic and boninitic melts (Boudier and Al-Rajhi, 2014;
occurred both during and directly following the Geotimes ax- Rollinson, 2008), concurrently with VMS mineralisation on
ial stage as part of the Lasail unit (see also Belgrano and Di- the seafloor (Gilgen et al., 2014). A complete cross section
amond, 2019), but before the eruption of the Alley tholeiites through this system is accessible in the Wadi Rajmi–Zab’in
and boninites. area of the Fizh block.
Alley/Phase 2 volcanism is continuously present through- The new map represents significant progress compared to
out the northern ophiolite and was volumetrically compara- the previous two-unit stratigraphy and it should prove useful
ble to axial Geotimes/Phase 1 volcanism. These findings re- for further research and mineral exploration in the ophiolite.

However, maps are inevitably interpretations of reality and Science and Engineering, University of Minnesota (IRM Visiting
so the present contribution is provided in an editable format Fellowship).
with the hope that it may be updated in the future as new data
become available.
Review statement. This paper was edited by Andrea Di Muro and
reviewed by Kathryn Goodenough and Yuki Kusano.
Data availability. All new geochemical and rock magnetic data
are openly available at https://doi.org/10.1594/PANGAEA.899794
(Belgrano et al., 2019).
References
A’Shaikh, D., Miyashita, S., and Matsueda, H.: The petrological and
Supplement. The Supplement includes the complete set of mag- geochemical characteristics of an ophiolite volcanic suite from
netic mineralogical figures (Figs. S1–S9), tables of criteria used the Ghayth area of Oman, J. Miner. Petrol. Sci., 100, 202–220,
to assign each sample to a unit (Table S1), previously published https://doi.org/10.2465/jmps.100.202, 2005.
sample locations used for mapping (Table S2), analyses of standard Adachi, Y. and Miyashita, S.: Geology and petrology of
materials (Table S3–4), and a Geospatial PDF of the final vol- the plutonic complexes in the Wadi Fizh area: Multi-
canic map and input layers. The supplement related to this article ple magmatic events and segment structure in the north-
is available online at: https://doi.org/10.5194/se-10-1181-2019- ern Oman ophiolite, Geochem., Geophy. Geosy., 4, 8619,
supplement. https://doi.org/10.1029/2001GC000272, 2003.
Alabaster, T. and Pearce, J. A.: The interrelationship be-
tween magmatic and ore-forming hydrothermal pro-
Author contributions. The project was conceived by LWD and cesses in the Oman ophiolite, Econ. Geol., 80, 1–16,
SAG. Field mapping and geochemical analysis were led by TMB https://doi.org/10.2113/gsecongeo.80.1.1, 1985.
and YV. ARB and TMB measured and interpreted the rock mag- Alabaster, T., Pearce, J. A., Mallick, D. I. J., and Elboushi, I. M.:
netic properties. KAT provided assistance in the field and logisti- The volcanic stratigraphy and location of massive sulphide de-
cal support. TMB and LWD prepared the article with contributions posits in the Oman ophiolite, in: Proceedings of the International
from all co-authors. Ophiolite Symposium 1979, Cyprus, edited by: Panayiotou A.,
751–757, Cyprus Geological Survey Department, Nicosia, 1978.
Alabaster, T., Pearce, J. A., and Malpas, J.: The vol-
Competing interests. The authors declare that they have no conflict canic stratigraphy and petrogenesis of the Oman ophi-
of interest. olite complex, Contrib. Mineral. Petr., 81, 168–183,
https://doi.org/10.1007/BF00371294, 1982.
Ambrose, T. K. and Searle, M. P.: 3D structure of the
Acknowledgements. We gratefully acknowledge the support of the northern Oman – UAE ophiolite: widespread, short-lived,
Public Authority for Mining (PAM), Sultanate of Oman, in particu- supra-subduction zone magmatism, Tectonics, 38, 233–252,
lar Salim Omar Al-Ibraheem, Mohammed Al Araimi, Mohammed https://doi.org/10.1029/2018TC005038, 2018.
Al-Battashi, and Ali Al Hashmi. We thank Samuel Weber, Alan- Arai, S., Kadoshima, K., and Morishita, T.: Widespread arc-related
nah Brett, Sarah Pein, Nevena Novakovic, Lisa Richter (all Uni- melting in the mantle section of the northern Oman ophiolite as
versity of Bern) and Ali Al Hashmi (PAM) for assistance dur- inferred from detrital chromian spinels, J. Geol. Soc., 163, 869–
ing field mapping. We thank Lydia Zehnder (ETH Zurich) for her 879, https://doi.org/10.1144/0016-76492005-057, 2006.
help with numerous XRF measurements. Pierre Lanari and Elias Arculus, R. J., Ishizuka, O., Bogus, K. A., Gurnis, M., Hickey-
Kempf kindly facilitated our microprobe measurements. Daniel Pe- Vargas, R., Aljahdali, M. H., Bandini-Maeder, A. N., Barth, A.
ters and Thomas Pettke provided LA-ICP-MS support. Kathryn P., Brandl, P. A., Drab, L., do Monte Guerra, R., Hamada, M.,
Goodenough and Yuki Kusano are thanked for their insightful re- Jiang, F., Kanayama, K., Kender, S., Kusano, Y., Li, H., Loudin,
views and helpful comments, and Andrea di Muro is thanked for his L. C., Maffione, M., Marsaglia, K. M., McCarthy, A., Meffre, S.,
editorial handling. Rock magnetism measurements were performed Morris, A., Neuhaus, M., Savov, I. P., Sena, C., Tepley III, F. J.,
by Thomas M. Belgrano as a visiting fellow at the Institute for Rock van der Land, C., Yogodzinski, G. M., and Zhang, Z.: A record of
Magnetism (IRM) at the University of Minnesota, under the wel- spontaneous subduction initiation in the Izu–Bonin–Mariana arc,
come guidance of Dario Bilardello and Michael Jackson. The IRM Nat. Geosci., 8, 728, https://doi.org/10.1038/NGEO2515, 2015.
is a US national multi-user facility supported through the Instru- Ballhaus, C., Fonseca, R. O. C., Münker, C., Kirchen-
mentation and Facilities Program of the National Science Founda- baur, M., and Zirner, A.: Spheroidal textures in igneous
tion, Earth Sciences Division, and by funding from the University rocks – Textural consequences of H2 O saturation in
of Minnesota. basaltic melts, Geochim. Cosmochim. Acta, 167, 241–252,
https://doi.org/10.1016/j.gca.2015.07.029, 2015.
Belgrano, T. M. and Diamond, L. W.: Subduction-zone contribu-
tions to axial volcanism in the Oman-U.A.E. ophiolite, Litho-
Financial support. This research has been supported by the Uni-
sphere, 11, 399–411, https://doi.org/10.1130/L1045.1, 2019.
versity of Bern, by the Swiss National Science Foundation (grant
Belgrano, T. M., Diamond, L. W., Vogt, Y., Biedermann, A. R.,
no. 200020-169653 to Larryn W. Diamond) and by the College of
Gilgen, S. A., and Al-Tobi, K.: Geochemical compositions (XRF,

LA-ICP-MS, EMPA) and bulk magnetic properties of volcanic Mesozoic Gondwana, Gondwana Res., 11, 453–475,
rocks used for mapping in the northern Semail ophiolite (Oman), https://doi.org/10.1016/j.gr.2007.01.005, 2007.
PANGAEA, https://doi.org/10.1594/PANGAEA.899794, 2019. Dilek, Y., Furnes, H., and Shallo, M.: Geochemistry of the Juras-
Blakely, R. J.: Potential Theory in Gravity and Magnetic Applica- sic Mirdita Ophiolite (Albania) and the MORB to SSZ evolu-
tions, Cambridge University Press, Cambridge, 1995. tion of a marginal basin oceanic crust, Lithos, 100, 174–209,
BME (Bishimetal Exploration): Geological Map of Al Wasit https://doi.org/10.1016/j.lithos.2007.06.026, 2008.
(Scale:1 : 50 000: Sheet NG40-14 E-III), Ministry of Petroleum Dunlop, D. and Özdemir, Ö.: Rock Magnetism: Fundamentals and
and Minerals, Sultanate of Oman, 1987a. Frontiers, Cambridge University Press, Cambridge, 1997.
BME (Bishimetal Exploration): Geological Map of Mahab Duretz, T., Agard, P., Yamato, P., Ducassou, C., Burov, E. B., and
(Scale:1 : 50 000: Sheet NG40-14 F-IV), Ministry of Petroleum Gerya, T. V: Thermo-mechanical modeling of the obduction pro-
and Minerals, Sultanate of Oman, 1987b. cess based on the Oman Ophiolite case, Gondwana Res., 32, 1–
BME (Bishimetal Exploration): Geological Map of Wadi Bani 10, https://doi.org/10.1016/j.gr.2015.02.002, 2016.
Umar (Scale 1 : 50 000: Sheet NG40-14E-II), Ministry of Einaudi, F., Godard, M., Pezard, P., Cochemé, J. J., Coulon,
Petroleum and Minerals, Sultanate of Oman, 1987c. C., Brewer, T., and Harvey, P.: Magmatic cycles and for-
Boudier, F. and Al-Rajhi, A.: Structural control on chromitite de- mation of the upper oceanic crust at spreading centers:
posits in ophiolites: the Oman case, Geol. Soc. London Spec. Geochemical study of a continuous extrusive section in
Pub., 392, 263–277, https://doi.org/10.1144/sp392.14, 2014. the Oman ophiolite, Geochem. Geophy. Geosy., 4, 1–25,
BRGM (Bureau de Recherches Geìologique et Minères): Geologi- https://doi.org/10.1029/2002GC000362, 2003.
cal Map of As Suwayq (Scale 1 : 100 000: Sheet NF40-3A), Min- Ernewein, M., Pflumio, C., and Whitechurch, H.: The death of
istry of Petroleum and Minerals, Sultanate of Oman, 1986a. an accretion zone as evidenced by the magmatic history of
BRGM (Bureau de Recherches Geìologique et Minères): Geologi- the Sumail ophiolite (Oman), Tectonophysics, 151, 247–274,
cal Map of Yanqul (Scale 1 : 100 000: Sheet NF40-2C), Ministry https://doi.org/10.1016/0040-1951(88)90248-X, 1988.
of Petroleum and Minerals, Sultanate of Oman, 1986b. Falloon, T. J. and Crawford, A. J.: The petrogenesis of high-calcium
BRGM (Bureau de Recherches Geìologique et Minères): Geolog- boninite lavas dredged from the northern Tonga ridge, Earth
ical Map of Aswad (Scale 1 : 50 000: Sheet NG40-14B1/B2), Planet. Sc. Lett., 102, 375–394, https://doi.org/10.1016/0012-
Ministry of Petroleum and Minerals, Sultanate of Oman, 1993a. 821X(91)90030-L, 1991.
BRGM (Bureau de Recherches Geìologique et Minères): Geologi- Falloon, T. J., Danyushevsky, L. V, Crawford, T. J., Maas, R.,
cal Map of Fizh (Scale 1 : 50 000: Sheet NG40-14B4), Ministry Woodhead, J. D., Eggins, S. M., Bloomer, S. H., Wright, D. J.,
of Petroleum and Minerals, Sultanate of Oman, 1993b. Zlobin, S. K., and Stacey, A. R.: Multiple mantle plume com-
Cameron, W. E.: Petrology and origin of primitive lavas from the ponents involved in the petrogenesis of subduction-related lavas
Troodos ophiolite, Cyprus, Contrib. Mineral. Petr., 89, 239–255, from the northern termination of the Tonga Arc and northern
https://doi.org/10.1007/BF00379457, 1985. Lau Basin: Evidence from the geochemistry of arc and backarc
Clark, D. A.: Magnetic petrophysics and magnetic petrology: aids to submarine volcanics, Geochem. Geophy. Geosy., 8, Q09003,
geological interpretation of magnetic surveys, AGSO J. Austral. https://doi.org/10.1029/2007GC001619, 2007.
Geol. Geophys., 17, 83–103, 1997. Falloon, T. J., Danyushevsky, L. V, Crawford, A. J., Meffre, S.,
Clark, D. A.: Methods for determining remanent and total magneti- Woodhead, J. D., and Bloomer, S. H.: Boninites and Adakites
sations of magnetic sources – A review, Explor. Geophys., 45, from the Northern Termination of the Tonga Trench: Im-
271–304, https://doi.org/10.1071/EG14013, 2014. plications for Adakite Petrogenesis, J. Petrol., 49, 697–715,
Cotta, A. J. B. and Enzweiler, J.: Certificate of Analysis of the https://doi.org/10.1093/petrology/egm080, 2008.
Reference Material BRP-1 (Basalt Ribeirão Preto), Geostand. Feinberg, H., Horen, H., Michard, A., and Saddiqi, O.: Obduction-
Geoanal. Res., 32, 231–235, https://doi.org/10.1111/j.1751- related remagnetization at the base of an ophiolite: Paleomag-
908X.2008.00894.x, 2008. netism of the Samail nappe lower sequence and of its continental
DeBari, S. M., Taylor, B., Spencer, K., and Fujioka, K.: A trapped substratum, southeast Oman Mountains, J. Geophys. Res., 104,
Philippine Sea plate origin for MORB from the inner slope of 17703, https://doi.org/10.1029/1999JB900171, 1999.
the Izu-Bonin trench, Earth Planet. Sci. Lett., 174, 183–197, Fleet, A. J. and Robertson, A. H. F.: Ocean-ridge metalliferous and
https://doi.org/10.1016/S0012-821X(99)00252-6, 1999. pelagic sediments of the Semail Nappe, Oman, J. Geol. Soc., 137,
de Graaff, S. J., Goodenough, K. M., Klaver, M., Lissenberg, 403–422, https://doi.org/10.1144/gsjgs.137.4.0403, 1980.
C. J., Jansen, M. N., Millar, I., and Davies, G. R.: Ev- Flint, R. B., Webster, S. S., and Uppill, R. K.: Re-interpretation of
idence for a Moist to Wet Source Transition Throughout the 1992 airborne magnetic survey of the Batinah coast and Yan-
the Oman-UAE Ophiolite, and Implications for the Geo- qul area, Northern Oman mountains, Muscat, 1999.
dynamic History, Geochem. Geophy. Geosy., 20, 651–672, Gale, A., Dalton, C. A., Langmuir, C. H., Su, Y., and
https://doi.org/10.1029/2018GC007923, 2019. Schilling, J.-G.: The mean composition of ocean
Dilek, Y. and Furnes, H.: Lithos Structure and geo- ridge basalts, Geochem. Geophy. Geosy., 14, 489–518,
chemistry of Tethyan ophiolites and their petrogene- https://doi.org/10.1029/2012GC004334, 2013.
sis in subduction rollback systems, Lithos, 113, 1–20, Gilgen, S. A., Diamond, L. W., Mercolli, I., Al-Tobi, K., Maid-
https://doi.org/10.1016/j.lithos.2009.04.022, 2009. ment, D. W., Close, R., and Al-Towaya, A.: Volcanostrati-
Dilek, Y., Furnes, H., and Shallo, M.: Suprasubduc- graphic Controls on the Occurrence of Massive Sulfide Deposits
tion zone ophiolite formation along the periphery of in the Semail Ophiolite, Oman, Econ. Geol., 109, 1585–1610,
https://doi.org/10.2113/econgeo.109.6.1585, 2014.

Gilgen, S. A., Diamond, L. W., and Mercolli, I.: Sub-seafloor JICA: Report on the cooperative mineral exploration in the South
epidosite alteration: Timing, depth and stratigraphic distribu- Batinah Coast area, Sultanate of Oman: consolidated report,
tion in the Semail ophiolite, Oman, Lithos, 260, 191–210, available at: http://open_jicareport.jica.go.jp/pdf/11597283_02.
https://doi.org/10.1016/j.lithos.2016.05.014, 2016. pdf (last access: 31 July 2018), 2000.
Godard, M., Dautria, J.-M., and Perrin, M.: Geochemical variability JICA: Report on the mineral exploration in the Yanqul-Ghuzayn
of the Oman ophiolite lavas: Relationship with spatial distribu- area Sultanate of Oman: Final report, available at: http:
tion and paleomagnetic directions, Geochem. Geophy. Geosy., 4, //open_jicareport.jica.go.jp/pdf/11678257_01.pdf (last access:
1–25, https://doi.org/10.1029/2002GC000452, 2003. 31 July 2018), 2002.
Godard, M., Bosch, D., and Einaudi, F.: A MORB source for low- Jochum, K. P., Nohl, U., Herwig, K., Lammel, E., Stoll, B., and
Ti magmatism in the Semail ophiolite, Chem. Geol., 234, 58–78, Hofmann, A. W.: GeoReM: A New Geochemical Database
https://doi.org/10.1016/j.chemgeo.2006.04.005, 2006. for Reference Materials and Isotopic Standards, Geostand.
Goodenough, K. M., Styles, M. T., Schofield, D., Thomas, R. J., Geoanal. Res., 29, 333–338, https://doi.org/10.1111/j.1751-
Crowley, Q. C., Lilly, R. M., McKervey, J., Stephenson, D., and 908X.2005.tb00904.x, 2005.
Carney, J. N.: Architecture of the Oman–UAE ophiolite: evi- Juteau, T., Ernewein, M., Reuber, I., Whitechurch, H., and
dence for a multi-phase magmatic history, Arab. J. Geosci., 3, Dahl, R.: Duality of magmatism in the plutonic sequence
439–458, https://doi.org/10.1007/s12517-010-0177-3, 2010. of the Sumail Nappe, Oman, Tectonophysics, 151, 107–135,
Goodenough, K. M., Thomas, R. J., Styles, M. T., Schofield, https://doi.org/10.1016/0040-1951(88)90243-0, 1988.
D. I., and MacLeod, C. J.: Records of ocean growth and de- Kane, J. S., Potts, P. J., Meisel, T., and Wiedenbeck, M.: In-
struction in the Oman-UAE ophiolite, Elements, 10, 109–114, ternational association of geoanalysts’ protocol for the cer-
https://doi.org/10.2113/gselements.10.2.109, 2014. tification of geological and environmental reference materi-
Guilmette, C., Smit, M. A., van Hinsbergen, D. J. J., Gürer, D., als: A supplement, Geostand. Geoanal. Res., 31, 285–288,
Corfu, F., Charette, B., Maffione, M., Rabeau, O., and Savard, https://doi.org/10.1111/j.1751-908X.2007.00869.x, 2007.
D.: Forced subduction initiation recorded in the sole and crust Kanke, N. and Takazawa, E.: A kilometre-scale highly refractory
of the Semail Ophiolite of Oman, Nat. Geosci., 11, 688–695, harzburgite zone in the mantle section of the northern Oman
https://doi.org/10.1038/s41561-018-0209-2, 2018. Ophiolite (Fizh Block): implications for flux melting of oceanic
Haase, K. M., Freund, S., Beier, C., Koepke, J., Erdmann, M., and lithospheric mantle, Geol. Soc. London Spec. Pub., 392, 229–
Hauff, F.: Constraints on the magmatic evolution of the oceanic 246, https://doi.org/10.1144/SP392.12, 2014.
crust from plagiogranite intrusions in the Oman ophiolite, Con- Kontny, A. and Grothaus, L.: Effects of shock pressure and temper-
trib. Miner. Petrol., 171, 1–16, https://doi.org/10.1007/s00410- ature on titanomagnetite from ICDP cores and target rocks of the
016-1261-9, 2016. El’gygytgyn impact structure, Russia, Stud. Geophys. Geod., 61,
Haymon, R. M., Koski, R. A., and Abrams, M. J.: 162–183, https://doi.org/10.1007/s11200-016-0819-3, 2017.
Hydrothermal discharge zones beneath massive sul- Kusano, Y., Adachi, Y., Miyashita, S., and Umino, S.: Lava accre-
fide deposits mapped in the Oman ophiolite, Ge- tion system around mid-ocean ridges: Volcanic stratigraphy in
ology, 17, 531–535, https://doi.org/10.1130/0091- the Wadi Fizh area, northern Oman ophiolite, Geochem. Geophy.
7613(1989)017<0531:HDZBMS>2.3.CO;2, 1989. Geosy., 13, Q05012, https://doi.org/10.1029/2011GC004006,
Hickey-Vargas, R., Yogodzinski, G. M., Ishizuka, O., McCarthy, 2012.
A., Bizimis, M., Kusano, Y., Savov, I. P., and Arculus, Kusano, Y., Hayashi, M., Adachi, Y., Umino, S., and Miyashita,
R.: Origin of depleted basalts during subduction initiation S.: Evolution of volcanism and magmatism during initial
and early development of the Izu-Bonin-Mariana island arc: arc stage: constraints on the tectonic setting of the Oman
Evidence from IODP expedition 351 site U1438, Amami- Ophiolite, Geol. Soc. London Spec. Pub., 392, 177–193,
Sankaku basin, Geochim. Cosmochim. Acta, 229, 85–111, https://doi.org/10.1144/SP392.9, 2014.
https://doi.org/10.1016/j.gca.2018.03.007, 2018. Kusano, Y., Umino, S., Shinjo, R., Ikei, A., Adachi, Y., Miyashita,
Ishikawa, T., Nagaishi, K., and Umino, S.: Boninitic vol- S., and Arai, S.: Contribution of slab-derived fluid and sedimen-
canism in the Oman ophiolite: Implications for ther- tary melt in the incipient arc magmas with development of the
mal condition during transition from spreading ridge to paleo-arc in the Oman Ophiolite, Chem. Geol., 449, 206–225,
arc, Geology, 30, 899–902, https://doi.org/10.1130/0091- https://doi.org/10.1016/j.chemgeo.2016.12.012, 2017.
7613(2002)030<0899:BVITOO>2.0.CO;2, 2002. Li, H.-Y., Taylor, R. N., Prytulak, J., Kirchenbaur, M., Shervais,
Ishizuka, O., Tani, K., Reagan, M. K., Kanayama, K., Umino, S., J. W., Ryan, J. G., Godard, M., Reagan, M. K., and Pearce,
Harigane, Y., Sakamoto, I., Miyajima, Y., Yuasa, M., and Dunk- J. A.: Radiogenic isotopes document the start of subduction
ley, D. J.: The timescales of subduction initiation and subsequent in the Western Pacific, Earth Planet. Sc. Lett., 518, 197–210,
evolution of an oceanic island arc, Earth Planet. Sc. Lett., 306, https://doi.org/10.1016/j.epsl.2019.04.041, 2019.
229–240, https://doi.org/10.1016/j.epsl.2011.04.006, 2011. Lippard, S. J., Shelton, A. W., and Gass, I. G.: The Ophiolite of
Isles, D. J. and Witham, W. J. A.: Batinah Coast region and Raki Northern Oman, Blackwell Scientific Publications Ltd, London,
Hayl As Safil area, Airborne Geophysical Project: Final report on Great Britain, 1986.
geological interpretation, Ministry of Petroleum and Minerals, Liu, Q., Banerjee, S. K., Jackson, M. J., Deng, C., Pan, Y.,
Sultanate of Oman, Muscat, 1993. and Rixiang, Z.: New insights into partial oxidation model
Jenner, F. E. and O’Neill, H. S. C.: Analysis of 60 elements in of magnetites and thermal alteration of magnetic mineralogy
616 ocean floor basaltic glasses, Geochem. Geophy. Geosy., 13, of the Chinese loess in air, Geophys. J. Int., 158, 506–514,
Q02005, https://doi.org/10.1029/2011GC004009, 2012. https://doi.org/10.1111/j.1365-246X.2004.02348.x, 2004.

MacLeod, C. J. and Rothery, D. A.: Ridge axial segmentation in tion and an extended episode of ridge subduction?, Lithosphere,
the Oman ophiolite: Evidence from along-strike variations in the 4, 497–510, https://doi.org/10.1130/L205.1, 2012.
sheeted dyke complex, Geol. Soc. London Spec. Pub., 60, 39–63, Patriat, M., Falloon, T., Danyushevsky, L., Collot, J., Jean, M. M.,
https://doi.org/10.1144/GSL.SP.1992.060.01.03, 1992. Hoernle, K., Hauff, F., Maas, R., Woodhead, J. D., and Feig, S.
MacLeod, C. J., Johan Lissenberg, C. and Bibby, L. E.: “Moist T.: Subduction initiation terranes exposed at the front of a 2 Ma
MORB” axial magmatism in the Oman ophiolite: The evi- volcanically-active subduction zone, Earth Planet. Sc. Lett., 508,
dence against a mid-ocean ridge origin, Geology, 41, 459–462, 30–40, https://doi.org/10.1016/j.epsl.2018.12.011, 2019.
https://doi.org/10.1130/G33904.1, 2013. Pearce, J. A.: Geochemical evidence for the genesis and eruptive
Mallmann, G. and O’Neill, H. S. C.: The crystal/melt partition- setting of lavas from Tethyan ophiolites, in: Proceedings of the
ing of V during mantle melting as a function of oxygen fu- International Ophiolite Symposium 1979, Cyprus 1979, 261–
gacity compared with some other elements (Al, P, Ca, Sc, 272, Ministry of Agriculture and Natural Resources, Cyprus,
Ti, Cr, Fe, Ga, Y, Zr and Nb), J. Petrol., 50, 1765–1794, 1980.
https://doi.org/10.1093/petrology/egp053, 2009. Pearce, J. A.: Immobile Element Fingerprint-
Matveev, S. and Ballhaus, C.: Role of water in the origin of podi- ing of Ophiolites, Elements, 10, 101–108,
form chromitite deposits, Earth Planet Sc. Lett., 203, 235–243, https://doi.org/10.2113/gselements.10.2.101, 2014.
https://doi.org/10.1016/S0012-821X(02)00860-9, 2002. Pearce, J. A. and Reagan, M. K.: Identification, classifica-
Meffre, S., Falloon, T. J., Crawford, T. J., Hoernle, K., Hauff, tion, and interpretation of boninites from Anthropocene to
F., Duncan, R. A., Bloomer, S. H., and Wright, D. J.: Basalts Eoarchean using Si-Mg-Ti systematics, Geosphere, 15, 1–11,
erupted along the Tongan fore arc during subduction initi- https://doi.org/10.1130/GES01661.1, 2019.
ation: Evidence from geochronology of dredged rocks from Pearce, J. A. and Robinson, P. T.: The Troodos ophi-
the Tonga fore arc and trench, Geochem. Geophy. Geosy., 13, olitic complex probably formed in a subduction initi-
https://doi.org/10.1029/2012GC004335, 2012. ation, slab edge setting, Gondwana Res., 18, 60–81,
Metcalf, R. V. and Shervais, J. W.: Suprasubduction-zone ophio- https://doi.org/10.1016/j.gr.2009.12.003, 2010.
lites: Is there really an ophiolite conundrum?, in: Special Paper Pearce, J. A., Alabaster, T., Shelton, A. W., and Searle, M. P.:
438: Ophiolites, Arcs, and Batholiths: A Tribute to Cliff Hopson, The Oman Ophiolite as a Cretaceous Arc-Basin Complex: Ev-
191–222, 2008. idence and Implications, Philos. T. R. Soc. S.-A., 300, 299–317,
Miyashita, S., Adachi, Y., and Umino, S.: Along-axis magmatic sys- https://doi.org/10.1098/rsta.1981.0066, 1981.
tem in the northern Oman ophiolite: Implications of composi- Pearce, J. A., van der Laan, S. R., Arculus, R. J., Murton, B. J.,
tional variation of the sheeted dike complex, Geochem. Geophy. Ishii, T., Peate, D. W., and Parkinson, I. J.: Boninite and Harzbur-
Geosy., 4, 8617, https://doi.org/10.1029/2001GC000235, 2003. gite from Leg 125 (Bonin-Mariana Forearc): A Case Study of
Moghadam, H. S., Stern, R. J., and Rahgoshay, M.: The Dehshir Magma Genesis during the Initial Stages of Subduction, Pro-
ophiolite (central Iran): Geochemical constraints on the origin ceedings of the Ocean Drilling Program, Sci. Res., 125, 623–659,
and evolution of the inner Zagros ophiolite belt, Bull. Geol. Soc. https://doi.org/10.2973/odp.proc.sr.125.172.1992, 1992.
Am., 122, 1516–1547, https://doi.org/10.1130/B30066.1, 2010. Perrin, M., Prevot, M., and Bruere, F.: Rotation of the Oman ophio-
Morgan, G. B. and London, D.: Effect of current density on the elec- lite and initial location of the ridge in the hotspot reference frame,
tron microprobe analysis of alkali aluminosilicate glasses, Am. Tectonophysics, 229, 31–42, https://doi.org/10.1016/0040-
Mineral., 70, 1131–1138, https://doi.org/10.2138/am.2005.1769, 1951(94)90004-3, 1994.
2005. Perrin, M., Plenier, G., Dautria, J. M., Cocuaud, E., and Prévot, M.:
Morris, A., Meyer, M., Anderson, M. W., and MacLeod, C. Rotation of the Semail ophiolite (Oman): Additional paleomag-
J.: Clockwise rotation of the entire Oman ophiolite occurred netic data from the volcanic sequence, Mar. Geophys. Res., 21,
in a suprasubduction zone setting, Geology, 44, 1055–1058, 181–194, https://doi.org/10.1023/A:1026738313805, 2000.
https://doi.org/10.1130/G38380.1, 2016. Peters, D. and Pettke, T.: Evaluation of Major to Ultra Trace Ele-
Moskowitz, B. M., Jackson, M., and Kissel, C.: Low-temperature ment Bulk Rock Chemical Analysis of Nanoparticulate Pressed
magnetic behavior of titanomagnetites, Earth Planet. Sc. Lett., Powder Pellets by LA-ICP-MS, Geostand. Geoanal. Res., 41, 5–
157, 141–149, https://doi.org/10.1016/S0012-821X(98)00033-8, 28, https://doi.org/10.1111/ggr.12125, 2016.
1998. Pflumio, C.: Evidences for Polyphased Oceanic Alteration of
Murton, B. J., Peete, D. W., Arculus, R. J., Pearce, J. A., and the Extrusive Sequence of the Semail Ophiolite from the
Van der Laan, S. R.: Trace-Element Geochemistry of Vol- Salahi Block (Northern Oman), Petr. Stru. G., 5, 313–351,
canic Rocks from Site 786: The Izu-Bonin Forearc, Proceed- https://doi.org/10.1007/978-94-011-3358-6_17, 1991.
ings of the Ocean Drilling Program, Sci. Res., 125, 211–235, Python, M. and Ceuleneer, G.: Nature and distribution of dykes
https://doi.org/10.2973/odp.proc.sr.125.133.1992, 1992. and related melt migration structures in the mantle section
Nicolas, A., Boudier, F., Ildefonse, B., and Ball, E.: Accretion of the Oman ophiolite, Geochem. Geophy. Geosy., 4, 8612,
of Oman and United Arab Emirates ophiolite – Discussion https://doi.org/10.1029/2002GC000354, 2003.
of a new structural map, Mar. Geophys. Res., 21, 147–180, Python, M., Ceuleneer, G., and Arai, S.: Chromian spinels in mafic-
https://doi.org/10.1023/A:1026769727917, 2000. ultramafic mantle dykes: Evidence for a two-stage melt produc-
Osozawa, S., Shinjo, R., Lo, C.-H., Jahn, B.-M., Hoang, N., Sasaki, tion during the evolution of the Oman ophiolite, Lithos, 106,
M., Ishikawa, K., Kano, H., Hoshi, H., Xenophontos, C., and 137–154, https://doi.org/10.1016/j.lithos.2008.07.001, 2008.
Wakabayashi, J.: Geochemistry and geochronology of the Troo- Reagan, M. K., Ishizuka, O., Stern, R. J., Kelley, K. A., Ohara, Y.,
dos ophiolite: An SSZ ophiolite generated by subduction initia- Blichert-Toft, J., Bloomer, S. H., Cash, J., Fryer, P., Hanan, B. B.,

Hickey-Vargas, R., Ishii, T., Kimura, J.-I., Peate, D. W., Rowe, melt compositions, Contrib. Miner. Petrol., 156, 273–288,
M. C., and Woods, M.: Fore-arc basalts and subduction initiation https://doi.org/10.1007/s00410-008-0284-2, 2008.
in the Izu-Bonin-Mariana system, Geochem. Geophy. Geosy., 11, Rollinson, H.: New models for the genesis of plagiogran-
Q03X12, https://doi.org/10.1029/2009GC002871, 2010. ites in the Oman ophiolite, Lithos, 112, 603–614,
Reagan, M. K., McClelland, W. C., Girard, G., Goff, K. R., Peate, https://doi.org/10.1016/j.lithos.2009.06.006, 2009.
D. W., Ohara, Y., and Stern, R. J.: The geology of the southern Rollinson, H. and Adetunji, J.: The geochemistry and oxida-
Mariana fore-arc crust: Implications for the scale of Eocene vol- tion state of podiform chromitites from the mantle section of
canism in the western Pacific, Earth Planet. Sc. Lett., 380, 41–51, the Oman ophiolite: A review, Gondwana Res., 27, 543–554,
https://doi.org/10.1016/j.epsl.2013.08.013, 2013. https://doi.org/10.1016/j.gr.2013.07.013, 2015.
Reagan, M. K., Pearce, J. A., Petronotis, K., Almeev, R. R., Av- Searle, M. P., Lippard, S. J., Smewing, J. D., and Rex, D.
ery, A. J., Carvallo, C., Chapman, T., Christeson, G. L., Ferré, C.: Volcanic rocks beneath the Semail Ophiolite nappe in
E. C., Godard, M., Heaton, D. E., Kirchenbaur, M., Kurz, W., the northern Oman mountains and their significance in the
Kutterolf, S., Li, H., Li, Y., Michibayashi, K., Morgan, S., Nel- Mesozoic evolution of Tethys, J. Geol. Soc., 137, 589–604,
son, W. R., Prytulak, J., Python, M., Robertson, A. H. F., Ryan, https://doi.org/10.1144/gsjgs.137.5.0589, 1980.
J. G., Sager, W. W., Sakuyama, T., Shervais, J. W., Shimizu, K., Shafaii Moghadam, H., Corfu, F., and Stern, R. J.: U–Pb zircon
and Whattam, S. A.: Subduction initiation and ophiolite crust: ages of Late Cretaceous Nain–Dehshir ophiolites, central Iran, J.
new insights from IODP drilling, Int. Geol. Rev., 59, 1439–1450, Geol. Soc., 170, 175–184, https://doi.org/10.1144/jgs2012-066,
https://doi.org/10.1080/00206814.2016.1276482, 2017. 2013.
Reagan, M. K., Heaton, D. E., Schmitz, M. D., Pearce, J. A., Shelton, A. W.: The interpretation of gravity data in
Shervais, J. W., and Koppers, A. A. P.: Forearc ages reveal Oman: constraints on the ophiolite emplacement mech-
extensive short-lived and rapid seafloor spreading following anism, Geol. Soc. London Spec. Pub., 49, 459–471,
subduction initiation, Earth Planet. Sc. Lett., 506, 520–529, https://doi.org/10.1144/GSL.SP.1992.049.01.29, 1990.
https://doi.org/10.1016/j.epsl.2018.11.020, 2019. Shervais, J. W.: TiV plots and the petrogenesis of modern
Reuber, I.: Complexity of the crustal sequence in the north- and ophiolitic lavas, Earth Planet. Sc. Lett., 59, 101–118,
ern Oman ophiolite (Fizh and southern Aswad blocks): https://doi.org/10.1016/0012-821X(82)90120-0, 1982.
The effect of early slicing?, Tectonophysics, 151, 137–165, Shervais, J. W., Reagan, M., Haugen, E., Almeev, R., Pearce, J.,
https://doi.org/10.1016/0040-1951(88)90244-2, 1988. Prytulak, J., Ryan, J. G., Whattam, S., Godard, M., Chapman, T.,
Reuber, I., Nehlig, P., and Juteau, T.: Axial segmentation at Li, H., Kurz, W., Nelson, W. R., Heaton, D., Kirchenbaur, M.,
a fossil oceanic spreading centre in the Haylayn block (Se- Shimizu, K., Sakuyama, T., Li, Y., and Vetter, S. K.: Magmatic
mail nappe, Oman): off-axis mantle diapir and advancing ridge Response to Subduction Initiation, Part I: Forearc basalts of the
tip, J. Geodynam., 13, 253–278, https://doi.org/10.1016/0264- Izu-Bonin Arc from IODP Expedition 352, Geochem. Geophy.
3707(91)90041-C, 1991. Geosy., 20, 314–338, https://doi.org/10.1029/2018GC007731,
Rioux, M., Bowring, S., Kelemen, P., Gordon, S., Miller, R., and 2019.
Dudás, F.: Tectonic development of the Samail ophiolite: High- Smewing, J. D.: An Upper Cretaceous ridge-transform intersection
precision U-Pb zircon geochronology and Sm-Nd isotopic con- in the Oman ophiolite, in: Proceedings of the International Ophi-
straints on crustal growth and emplacement, J. Geophys. Res.- olite Symposium 1979, Cyprus, 1979, edited by: Panayiotou A.,
Sol. Ea., 118, 2085–2101, https://doi.org/10.1002/jgrb.50139, 407–413, Cyprus Geological Survey Department, Nicosia, 1980.
2013. Smewing, J. D., Simonian, K. O., Elboushi, I. M., and Gass, I. G.:
Rioux, M., Garber, J., Bauer, A., Bowring, S., Searle, M., Kelemen, Mineralized fault zone parallel to the Oman ophiolite spread-
P., and Hacker, B.: Synchronous formation of the metamorphic ing axis, Geology, 5, 534–538, https://doi.org/10.1130/0091-
sole and igneous crust of the Semail ophiolite?: New constraints 7613(1977)5<534:MFZPTT>2.0.CO;2, 1977.
on the tectonic evolution during ophiolite formation from high- Stern, R. J. and Bloomer, S. H.: Subduction zone
precision U – Pb zircon geochronology, Earth Planet. Sc. Lett., infancy: examples from the Eocene Izu-Bonin-
451, 185–195, https://doi.org/10.1016/j.epsl.2016.06.051, 2016. Mariana and Jurassic California arcs, Geol. Soc. Am.
Robertson, A. H. F. and Woodcock, N. H.: Genesis of the Batinah Bull., 104, 1621–1636, https://doi.org/10.1130/0016-
mélange above the Semail ophiolite, Oman, J. Struct. Geol., 5, 7606(1992)104<1621:SZIEFT>2.3.CO;2, 1992.
1–17, https://doi.org/10.1016/0191-8141(83)90003-2, 1983a. Stern, R. J. and Gerya, T.: Subduction initiation in na-
Robertson, A. H. F. and Woodcock, N. H.: Zabyat Formation, Se- ture and models: A review, Tectonophysics, 746, 173–198,
mail Nappe, Oman: sedimentation on to an emplacing ophiolite, https://doi.org/10.1016/j.tecto.2017.10.014, 2017.
Sedimentology, 30, 105–116, https://doi.org/10.1111/j.1365- Stern, R. J., Reagan, M., Ishizuka, O., Ohara, Y., and Whattam, S.:
3091.1983.tb00653.x, 1983b. To understand subduction initiation, study forearc crust: To un-
Rochette, P., Fillion, G., Mattéi, J. L., and Dekkers, M. J.: Magnetic derstand forearc crust, study ophiolites, Lithosphere, 4, 469–483,
transition at 30-34 Kelvin in pyrrhotite: insight into a widespread https://doi.org/10.1130/L183.1, 2012.
occurrence of this mineral in rocks, Earth Planet. Sc. Lett., Styles, M., Ellison, R., Arkley, S., Crowley, Q. G., Farrant, A., and
98, 319–328, https://doi.org/10.1016/0012-821X(90)90034-U, Goodenough, K. M.: The geology and geophysics of the United
1990. Arab Emirates?: Volume 2, Geology, Abu Dhabi, 2006.
Rollinson, H.: The geochemistry of mantle chromitites from Takazawa, E., Okayasu, T., and Satoh, K.: Geochemistry and ori-
the northern part of the Oman ophiolite: Inferred parental gin of the basal lherzolites from the northern Oman ophio-

lite (northern Fizh block), Geochem. Geophy. Geosy., 4, 1021, Usui, Y. and Yamazaki, S.: Salvaging primary rema-
https://doi.org/10.1029/2001GC000232, 2003. nence from hydrothermally altered oceanic gabbros in
Thébault, E., Finlay, C. C., Beggan, C. D., Alken, P., Aubert, J., Bar- the Oman ophiolite: A selective destructive demagne-
rois, O., Bertrand, F., Bondar, T., Boness, A., Brocco, L., Canet, tization approach, Phys. Earth Planet. In., 181, 1–11,
E., Chambodut, A., Chulliat, A., Coïsson, P., Civet, F., Du, A., https://doi.org/10.1016/j.pepi.2010.04.008, 2010.
Fournier, A., Fratter, I., Gillet, N., Hamilton, B., Hamoudi, M., Velzen, A. J. and Zijderveld, J. D. A.: A method to study alter-
Hulot, G., Jager, T., Korte, M., Kuang, W., Lalanne, X., Langlais, ations of magnetic minerals during thermal demagnetization ap-
B., Léger, J.-M., Lesur, V., Lowes, F. J., Macmillan, S., Mandea, plied to a fine-grained marine marl (Trubi formation, Sicily),
M., Manoj, C., Maus, S., Olsen, N., Petrov, V., Ridley, V., Rother, Geophys. J. Int., 110, 79–90, https://doi.org/10.1111/j.1365-
M., Sabaka, T. J., Saturnino, D., Schachtschneider, R., Sirol, O., 246X.1992.tb00715.x, 1992.
Tangborn, A., Thomson, A., Tøffner-Clausen, L., Vigneron, P., Vogt, P. R. and Johnson, G. L.: Magnetic telechemistry of oceanic
Wardinski, I., and Zvereva, T.: International Geomagnetic Refer- crust?, Nature, 245, 373–375, https://doi.org/10.1038/245373a0,
ence Field: the 12th generation, Earth Planet. Space, 67, 79–90, 1973.
https://doi.org/10.1186/s40623-015-0228-9, 2015. Whattam, S. A. and Stern, R. J.: The “subduction initiation rule”:
Tsuchiya, N., Shibata, T., Yoshikawa, M., Adachi, Y., Miyashita, A key for linking ophiolites, intra-oceanic forearcs, and sub-
S., Adachi, T., Nakano, N., and Osanai, Y.: Petrology duction initiation, Contrib. Miner. Petrol., 162, 1031–1045,
of Lasail plutonic complex, northern Oman ophiolite, https://doi.org/10.1007/s00410-011-0638-z, 2011.
Oman: An example of arc-like magmatism associated Woodcock, N. H. and Robertson, A. H. F.: The upper Bati-
with ophiolite detachment, Lithos, 156–159, 120–138, nah Complex, Oman: allochthonous sediment sheets above
https://doi.org/10.1016/j.lithos.2012.10.013, 2013. the Semail ophiolite, Can. J. Earth Sci., 19, 1635–1656,
Umino, S.: Emplacement mechanism of off-axis large submarine https://doi.org/10.1139/e82-140, 1982.
lava field from the Oman Ophiolite, J. Geophys. Res.-Sol. Ea., Yamasaki, T., Maeda, J., and Mizuta, T.: Geochemical evidence in
117, B11210, https://doi.org/10.1029/2012JB009198, 2012. clinopyroxenes from gabbroic sequence for two distinct magma-
Umino, S., Yanai, S., Jaman, A. R., Nakamura, Y., and Iiyama, J. T.: tisms in the Oman ophiolite, Earth Planet. Sc. Lett., 251, 52–65,
The transition from spreading to subduction: Evidence from the https://doi.org/10.1016/j.epsl.2006.08.027, 2006.
Semail Ophiolite, northern Oman Mountains, in: Oceanic Crustal Yogodzinski, G. M., Bizimis, M., Hickey-Vargas, R., McCarthy,
Analogues: Proceedings of the Symposium “Troodos 1987”, A., Hocking, B. D., Savov, I. P., Ishizuka, O., and Arcu-
375–384, Cyprus Geological Survey Department, Nicosia, 1990. lus, R.: Implications of Eocene-age Philippine Sea and fore-
Umino, S., Miyashita, S., Hotta, F., and Adachi, Y.: Along-strike arc basalts for initiation and early history of the Izu-Bonin-
variation of the sheeted dike complex in the Oman Ophio- Mariana arc, Geochim. Cosmochim. Acta, 228, 136–156,
lite: Insights into subaxial ridge segment structures and the https://doi.org/10.1016/j.gca.2018.02.047, 2018.
magma plumbing system, Geochem. Geophy. Geosy., 4, 8618,
https://doi.org/10.1029/2001GC000233, 2003.

Se 10 1181 2019

Uploaded by

Copyright:

Available Formats

Se 10 1181 2019

Uploaded by

Document Information

Original Description:

Original Title

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

Se 10 1181 2019

Uploaded by

Copyright:

Available Formats

Solid Earth, 10, 1181–1217, 2019

A revised map of volcanic units in the Oman ophiolite: insights into

Correspondence: Thomas M. Belgrano ([email protected])

Received: 30 March 2019 – Discussion started: 18 April 2019

Published by Copernicus Publications on behalf of the European Geosciences Union.

2013; Stern and Bloomer, 1992). These findings have led

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

1.1 Approach 1.3 Geology and petrogenesis of the Semail

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

supra-subduction MORB-mantle source, triggered by

3. Tholeiitic Alley. This is part of the original “Alley”

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

assigned to a unit on the basis of field relationships. We

4.3 Geochemical results

The Semail volcanostratigraphy records a progressive deple-

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

(Figs. 4a, 5, 6). No fresh glasses were recovered from these

4.7 How boninitic is the “Boninitic Alley” unit?

Boninites are defined by the IUGS (International Union

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

Figure 7. Statistical summary of bulk magnetic measurements of

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Batinah coast Fizh & Hilti blocks Summary

Vol %: volume fraction assumed equivalent to areal fraction

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

both sides of the wadi. Additional occurrences of Suhaylah

7.5 Lasail basaltic seamounts

The discrete accumulations of Lasail lavas that overlie Geo-

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

8 Comparison with the Izu–Bonin–Mariana proto-arc

The Semail ophiolite has been proposed as analogous to the

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/

www.solid-earth.net/10/1181/2019/ Solid Earth, 10, 1181–1217, 2019

Solid Earth, 10, 1181–1217, 2019 www.solid-earth.net/10/1181/2019/