2002 - Besse and Courtillot

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. B11, 2300, doi:10.
1029/2000JB000050, 2002
Correction published 10 October 2003
Apparent and true polar wander and the geometry

of the geomagnetic field over the last 200 Myr
Jean Besse and Vincent Courtillot
Laboratoire de Paléomagnétisme, UMR 7577, Institut de Physique du Globe de Paris, Paris, France
Received 8 November 2000; revised 15 January 2002; accepted 20 January 2002; published 15 November 2002.
[1] We have constructed new apparent polar wander paths (APWPs) for major plates over
the last 200 Myr. Updated kinematic models and selected paleomagnetic data allowed
us to construct a master APWP. A persistent quadrupole moment on the order of 3% of
the dipole over the last 200 Myr is suggested. Paleomagnetic and hot spot APW are
compared, and a new determination of ‘‘true polar wander’’ (TPW) is derived. Under the
hypothesis of fixed Atlantic and Indian hot spots, we confirm that TPW is episodic,
with periods of (quasi) standstill alternating with periods of faster TPW (in the Cretaceous).
The typical duration of these periods is on the order of a few tens of millions of years with
wander rates during fast tracks on the order of 30 to 50 km/Myr. A total TPW of some 30
is suggested for the last 200 Myr. We find no convincing evidence for episodes of
superfast TPW such as proposed recently by a number of authors. Comparison over the
last 130 Myr of TPW deduced from hot spot tracks and paleomagnetic data in the
Indo-Atlantic hemisphere with an independent determination for the Pacific plate supports
the idea that, to first order, TPW is a truly global feature of Earth dynamics. Comparison
with numerical modeling estimates of TPW shows that all current models still fail to some
extent to account for the observed values of TPW velocity and for the succession of
standstills and tracks which is observed. INDEX TERMS: 1527 Geomagnetism and
Paleomagnetism: Paleomagnetism applied to geologic processes; 3040 Marine Geology and Geophysics: Plate
tectonics (8150, 8155, 8157, 8158); 8120 Tectonophysics: Dynamics of lithosphere and mantle—general;
KEYWORDS: Paleomagnetism, polar wander, Earth rotation, TPW
Citation: Besse, J., and V. Courtillot, Apparent and true polar wander and the geometry of the geomagnetic field over the last 200 Myr,
J. Geophys. Res., 107(B11), 2300, doi:10.1029/2000JB000050, 2002.
1. Introduction hereinafter referred to as BC91] to blend the two approaches

and to use then available paleomagnetic data from North
[2] Analysis of the fossil magnetization preserved in America (NAM), Africa (AFR), Eurasia (EUR) and India
rocks is the basis for constraining such diverse geophysical (IND) and kinematic models from the Indian, central
problems as dynamo generation in the Earth’s core, plate Atlantic and North Atlantic oceans into a single ‘‘synthetic’’
kinematics and paleogeographic reconstructions, and mantle APWP that could next be transferred to any desired plate.
dynamics leading to true polar wander (TPW). The second On the basis of a selection of 111 poles, the paths were
half of the twentieth century saw the advent and consol- defined in 20 Myr windows extending back to 200 Ma. This
idation of plate tectonics: paleomagnetic measurements on first-order analysis provided a reasonable check of the
lava and sediments coming mostly from continental areas or consistency of individual plate paleomagnetic data and
from hot spot volcanics demonstrated continental drift and kinematic models with the rigid plates and geocentric dipole
could be blended into apparent polar wander paths (APWP); hypotheses. It also filled gaps in the available APWPs from
oceanographic exploration led to the discovery of seafloor individual plates. However, the study suffered from a num-
spreading related magnetic anomalies and transform faults, ber of limitations in space and time resolution and was not
allowing the construction of kinematic models for each truly global since a number of major plates (South America-
individual ocean basin. APWPs and kinematic models were SAM, Australia-AUS, Antarctica-ANT and the Pacific-PAC)
developed from the ’60s to the ’80s under the key assump- were not included.
tion that, when averaged over a sufficient amount of time, in [3] Slight departures from a purely centered dipole field
excess of a few thousand years, the Earth’s magnetic field had been noted as early as 1970, when Wilson [1970] argued
could be described accurately by an axial centered dipole. for a far-sided and right-handed distribution of virtual geo-
Of course, both types of largely ocean- and continent-based magnetic poles during the Cenozoic. A number of analyses
data could not be independent. Over a decade ago, we confirmed that when averaged over the last few million years,
proposed [Besse, 1986; Besse and Courtillot, 1988, 1991, an axial quadrupole component is detectable, with an ampli-
Copyright 2002 by the American Geophysical Union. tude on the order of 3 to 6% of the axial dipole [e.g., Coupland
0148-0227/02/2000JB000050$09.00 and Van der Voo, 1980; Merrill and McElhinny, 1983;
EPM 6-1
EPM 6- 2 BESSE AND COURTILLOT: APPARENT AND TRUE POLAR WANDER
Constable and Parker, 1988; Schneider and Kent, 1990b; Paleomagnetic Database (GPMD) of McElhinny and Lock
Quidelleur et al., 1994; Carlut and Courtillot, 1998; Johnson [1995]. We have used the June 1999 version (GPMD V3.3).
and Constable, 1997]. Livermore et al. [1983, 1984] tried to In the process of using and searching the database, we have
extract such a quadrupolar term in a worldwide paleomag- uncovered a few errors and inconsistencies. The more
netic database going back 200 Ma. Besse and Courtillot frequent occurrences were older data later re-used in a larger
[1991] argued that, for the period prior to 5 Ma going to study or an update without any notice of this redundancy in
200 Ma, a significant quadrupolar term could not be extracted the database. In order to avoid confusion in the database, and
unequivocally from the data available at that time. to allow others to easily check our computations, we have
[4] On the other hand, comparison of paleomagnetic and preferred not to alter the database version we started with,
hot spot APW led Besse and Courtillot [1991] to infer that but rather send a list of the problems to the managers of the
significant true polar wander, amounting to more than 20, database for future revisions, equally accessible to all. We
had occurred in an episodic, irregular way in the last 200 have, however, checked that these errors and inconsistencies
Myr. Recent advances in seismic tomography and dynamic were sufficiently few and small that their integration in the
modeling have raised fascinating new questions on the master APWPs resulted in changes of mean pole positions
existence and role of slab avalanches, plume events and always less than 1, or in mean age less than 1 Myr. As a
the viscosity structure of the deep mantle [e.g., Machetel result, these changes would hardly alter the Figures.
and Weber, 1991; Ricard et al., 1993; Tackley et al., 1994; [8] The key question is which selection criteria are
Steinberger and O’Connell, 1997; Bunge et al., 1998; applied to sift the base and extract a robust high quality
Evans, 1998; Richards et al., 1999; Greff-Lefftz, 2001; M. subset that allows one to address a given problem. As was
Greff-Lefftz and P. Bunge, personal communication, 2000] already the case in BC91, we have decided to retain data
(see, e.g., review by Tackley [2000]). Reliable measures of from both magmatic and sedimentary rocks. In a similar
TPW are required to constrain these models. recent study of TPW, Prévot et al. [2000] chose to retain
[5] Paper BC91 was based on a database compiled and only magmatic rocks, which they believe to carry a more
updated in our institute [Besse, 1986], prior to availability reliable magnetization, namely a thermoremanent magnet-
of the global paleomagnetic computerized database of ization (TRM). The question of the relative reliabilities of
McElhinny and Lock [1995]. In its June 1999 version, sedimentary and volcanic magnetizations is an old one in
the GPMD includes all paleomagnetic data published up to paleomagnetism. For instance, deepwater sediments can
the end of 1997. As will be seen below, and still retaining have significant compaction errors [e.g., Celaya and Clem-
the selection criteria used in BC91, the total amount of ent, 1988; Tarduno, 1990] but there are none in our data set.
reliable data has been increased by a factor of more than 2. We believe that there can be as many problems with
Also, significant methodological progress, most notably volcanic rocks as with sediments (e.g., remagnetizations
regarding satellite altimetry data analysis, has led to much versus inclination shallowing, too short versus too long
improved description of fracture zones and continent/ocean integration time for recording of magnetization) and in any
boundaries and hence better reconstructions of past plate case it is interesting to compare results obtained in either
motions. Improvement in normalization and correlation of case, as will be done below.
timescales has resulted in data sets such as the Global [9] For the sake of comparison, and because we believe
Seafloor Anomaly Chart [Royer et al., 1992] in which the the end result to be sufficiently reliable, we have retained a
marine geophysics community has integrated most avail- selection procedure similar to BC91. Applying selection
able data into a single coherent synthesis. criteria automatically, i.e., writing selection filters, is quite
[6] As a consequence of publication of these large sets of straightforward with a properly computerized database. The
new data, we have felt it useful to update our BC91 paper a BC91 criteria were the following: (1) at least 6 sites and 36
decade after its publication to try and evaluate whether samples per study; (2) a 95% confidence interval less than
improved plate reconstructions, APWPs, constraints on 10 in the Cenozoic and 15 in the Mesozoic; (3) evidence
field geometry and mantle dynamics could be obtained. In for successful alternating field and/or thermal demagnet-
order to decrease the risk of generating artifacts due to ization (i.e., Demagnetization Code equal to or larger than 2
limited site distribution [see, e.g., Quidelleur et al., 1994], in the McElhinny and Lock [1995] terminology); (4) dating
we have updated the NAM, EUR, AFR and IND data sets uncertainties less than 15 Myr; and (5) absence of remagne-
and have added data from AUS, ANT, SAM and GRE tization ensured by the fact that differences between mag-
(Greenland). We have also included some data from ODP/ netization and stratigraphic ages should be less than 5 Myr
DSDP drilling sites and two skewness poles coming from and by rejecting poles with a negative fold test or negative
analyses of marine magnetic anomaly profiles from the reversal test.
Indian Ocean (potential problems when using this type of [10] These simple criteria are of course to a large degree
data are discussed below). This results in much improved arbitrary and are based on our experience in trying to
geographical coverage both in latitude and longitude (Fig- eliminate most problematic studies. This is discussed in
ure 1). The case of the Pacific plate and data is left to the BC91 and compared to different selection criteria used by
discussion section. other authors. Part of the justification comes a posteriori
from a comparative examination of the results.
[11] One of the key aspects of our selection process has
2. Extended Database been an attempt to identify mobile zones from which data
2.1. Paleomagnetic Data From Continents should be excluded from what should in principle be rigid
[7] Faster paleomagnetic data search and treatment is now plates. This is important not only for plate reconstructions,
made easy due to the availability of the computerized Global but also for field geometry analysis: Carlut and Courtillot
BESSE AND COURTILLOT: APPARENT AND TRUE POLAR WANDER EPM 6-3
Figure 1. (a) Master apparent polar wander path for Africa from the Present back to 200 Ma, with
associated ellipses of 95% confidence in shaded gray (averages every 10 Myr, with a 20 Myr sliding
window); actual mean ages for each time window are indicated. (b) Distribution of individual
paleomagnetic poles and sampling sites; dots: all poles used in this study to construct the master APWP,
transferred to South African coordinates according to their age; stars: site locations reconstructed to
appropriate time in South African frame (solid star, Northern Hemisphere, open star, Southern
Hemisphere). Histograms from the Present back to 200 Ma for (c) number of data, (d) confidence
intervals at the 95% level, and (e) Fisher’s concentration parameter K.
EPM 6- 4 BESSE AND COURTILLOT: APPARENT AND TRUE POLAR WANDER
[1998] noted, for instance, in a study of the mean geo- declination of the original remanence of the block, on the
magnetic field over the last 5 Myr that inclusion of tectoni- present field direction (inclination and declination) and on
cally rotated Ethiopian data produced a very erroneous the azimuth of the magnetized body (i.e., generally of the
mean field model. Some tectonic areas, such as mountain ridge where the crustal block was generated). A great circle
ranges (Himalayas, Alpine ranges) or rift zones (Afar in of possible pole positions is deduced from the skewness
Africa), are of course rather obvious and sites from these value in the hypothesis of a dipolar centered geomagnetic
regions were eliminated. Others zones, such as the Colorado field. The intersections of a set of such semicircles derived
plateau in North America or the African rift south of the from the same anomaly (chron) at distant locations on the
Afar depression are less straightforward. And actually, rocks same plate in principle allow the determination of a virtual
often outcrop simply because of some amount of tectonic geomagnetic pole [e.g., Schouten and Cande, 1976; Gordon
deformation. Our task is then to test whether the extent of and Cox, 1980].
deformation requires that the data be rejected or not. A [14] The analysis of the skewness of marine magnetic
major source of error is linked to local tectonic rotations anomalies generated by seafloor spreading is therefore a
about vertical axes: crustal deformation at a large scale or at way of obtaining pole positions for oceanic plates, for
depth may lead to small block rotations and many paleo- which other kinds of data may be lacking. The method
magnetic studies fail to take into account the plunge of fold has been extensively applied by R. Gordon and colleagues
axes (the resulting error may exceed 15). Detailed analysis [e.g., Petronotis and Gordon, 1999] to the case of the
of individual poles and comparison with synthesized Pacific plate. Because of problems in connecting that plate
APWPs may allow one to spot remaining outliers with a to the others through a plate circuit, we have not included it
previously undetected tectonic origin. In some cases, we in the set used for construction of synthesized APWPs; we
have used another method: some poles that where suspected return to this question in the final section of this paper. We
to have undergone rotations without significant poleward note, however, that the theoretical potential of the skewness
motion have been integrated using a method derived from method for the determination of VGPs from purely oceanic
McFadden and McElhinny [1988]. We treated these poles plates is unfortunately hampered by strong limitations due
using only inclination data, i.e., as small circle constraints, to the parasitic effect of neighboring sources and/or to the
as we also did for data coming from DSDP Legs (see tectonic tilt of blocks generated close to the spreading
below). These poles (11 studies in western North America, center, leading to strong artifacts: Cande [1976] discusses
one in South America and one in Europe) are marked in the problems of ‘‘anomalous skewness’’, which are further
Table 1. elaborated on by Petronotis et al. [1992] and Dyment et al.
[12] Another major problem is to determine to what [1994].
extent large plates, such as Eurasia or Africa, can be [15] In the case of the Indian Ocean, careful analysis
considered as rigid. The ocean-based kinematic syntheses [Dyment et al., 1994] of conjugate marine magnetic anoma-
of Müller et al. [1993] imply that Africa and South America lies from the Carlsberg ridge and the Wharton Basin, and
need to be subdivided into a small number of rigid subplates the Central and South-East Indian ridges, leads to the
that have undergone some (slight) amount of differential conclusion that anomalous skewness becomes negligible
rotation since the time of breakup. Recent surveys of South when the spreading rate is faster than 50 km/Myr. The
Asian data [Cogné et al., 1999; Yang and Besse, 2001] have determination of skewness from anomaly 34 (84 Ma), i.e.,
emphasized differences between the Cretaceous segment of the time when India began its northward motion, up to
the Chinese APWP and the synthesized APWP of BC91 for anomaly 21 (47 Ma), when a sudden velocity drop resulted
Eurasia at the same period. Cogné et al. [1999] concluded from the India-Asia collision [Patriat and Achache, 1984] is
that ‘‘rigid’’ Eurasia may have actually undergone some possible because the spreading rate is far above this critical
amount of internal deformation, possibly due to the relative value. Dyment et al. [1994] found that anomalies 21, 24 and
motions of three rigid subplates. Therefore, and contrary to 29 were suitable for the determination of reasonably accu-
our analysis in BC91 where China was assumed to have rate paleolatitudes; they found good agreement between the
been rigidly attached to Eurasia (hence Europe) since the great circle of possible pole positions for these anomalies
Cretaceous, we have excluded Chinese data from the and the BC91 APWP.
present compilation. This final data set comprises 221 poles [16] However, Dyment et al. [1994] did not publish such
(Tables 1a – 1g). great circles for anomalies 25 and 26 which are prominent in
the Indian Ocean. Torcq [1997] and F. Torcq and J. Besse
2.2. Oceanic Data (manuscript in preparation, 2002) filled this gap, by analyzing
2.2.1. Data From Skewness of Marine Magnetic profiles from the Somali, Madagascar, Arabia, Crozet, and
Anomalies Central Indian basins. Anomalies 25 and 26 (61.5– 57.2 Ma
[13] The skewness of a (usually marine) magnetic anom- following the Harland et al. [1989] timescale) are easy to
aly is the phase shift required to restore the shape of the recognize and their shapes are well determined. Anticipat-
measured anomaly to the one which would have been ing results given later in this paper, the intersections of
observed, had the anomaly been created at the geomagnetic skewness derived circles for the Indian and African plates
pole [Schouten and McCamy, 1972]. The anomaly is then yield a pole in excellent agreement with data from other
‘‘undistorted’’, i.e., has the same symmetry elements as the continents. There are unfortunately no other sets of usable
magnetic source body. This is particularly useful for accu- skewness data from the plates involved in our synthesis.
rate determination of the location of the center or edges of 2.2.2. Oceanic Data From Boreholes
individual crustal blocks corresponding to a given chron. [17] Paleolatitudes can also be inferred from inclination
The value of skewness depends on the inclination and measurements performed on unoriented cores [e.g., Cox and
Table 1a. Selected Paleomagnetic Data for Antarctica Used in This Study Extracted From the GPMD V3.3 Databasea
Rock Name Age Dage Plat Plong Dp Dm Slat Slong B N Dc Tests Rock Type RNO Year
McMurdo Volcanics 1.5 3 86.4 81.2 13.4 13.7 77.5 161 14 212 3 Ro extrusives 1187 1983
McMurdo Volcanics 2 4 87.3 317.3 6.3 6.3 78 165.4 47 450 3 Rc extrusives 398 1988
Pensacola Mountains, Dufek Intrusion 172 8 60 43 10.3 11.5 82.6 50 30 91 2 Rb intrusives, gabbro 2489 1972
Nash Hills-Pagano Nunatek Intrusives 175 16 41.2 55.2 5.3 5.3 82.8 88.6 8 37 4 Rc intrusives, granite, 1107 1987
baked sediments
Northern Prince Albert Mountains, 177 3 47.8 45.5 5.5 5.5 74.5 162 15 227 3 M intrusives, dolerites 7079 1993
Ferrar dolerites
Northern Prince Albert Mountains, 177 3 78.8 0.9 5.9 5.9 75.4 161.3 7 98 3 M intrusives, dolerites 7080 1993
Ferrar dolerites
Northern Prince Albert Mountains, 177 3 45 40 3 4 78 162 46 83 2 Co, M intrusives, dolerites 3592 1962
Ferrar dolerites
Queen Alexandra Range, 193 30 52 24.1 6.8 7.7 84 165 14 84 2 M extrusives 3231 1971
Mount Falla Lavas
Queen Alexandra Range, 193 30 52.5 23.9 8.4 9.5 84 165 7 42 2 M intrusives 3232 1971
Jurassic Intrusions
Queen Alexandra Range, 193 30 44.1 51.5 8.8 11 84 165 12 72 2 M extrusives 3230 1971
Storm Peak Lavas
a
From McElhinny and Lock [1995]. Column headings are rock name, where sampled, formation or unit; age, mean age of magnetization, in Ma; Dage: maximum age of magnetization minus minimum age (i.e., error
is Dage/2); Plat, Plong, VGP latitude (N) and longitude (E); Dp, Dm, semiaxes of the 95% level confidence ellipse (in degrees) for the VGP; Slat, Slong, site latitude and longitude; B, number of sites; N, number of
samples; Dc, demagnetization code, or quality; tests, N, no test; M, rock magnetic test; G, F, R, conglomerate, fold, and reversal tests, respectively; O, a,b, c, +, significance level of tests; RNO, reference number in the
GPMD; year, year of publication.
Table 1b. Selected Paleomagnetic Data for India Used in This Study Extracted From the GPMD V3.3 Databasea
Sonhat Sill 61.5 7 37 285 2.1 3.2 23 82 11 62 3 N intrusives, dolerite 2892 1974
Deccan Traps, 11 sections combined 62.5 5 34.3 261.4 3.9 3.9 18.5 76.5 181 1070 3 Ro extrusives, basalts 2874 1973
Deccan Traps, Jabalpur 62.5 5 48 286 3 6 23.1 80.8 8 93 2 N extrusives, basalts 2915 1971
Deccan Traps, Aurangabad 62.5 5 33 287 5 7.3 19.8 76.5 25 142 3 N extrusives 2967 1972
Mt. Pavagarh Traps 64 8 39.2 285.6 4 6.8 22.5 73.5 16 88 3 N extrusives, basalts, rhyolites 2755 1974
Deccan dike swarms 65.5 5 37.2 280.5 9.7 9.7 21.5 74.3 11 67 3 Rc intrusives, dolerites 8106 1996
Deccan Traps – Nagpur to Bombay traverse 65.5 5 38.4 282.4 6.1 6.1 20 75 16 119 4 Rc extrusives, basalts 5727 1991
Deccan Traps, Malwa Plateau 65.5 5 36.3 270.4 11.4 17.4 22.5 75.8 13 76 2 Rb extrusives, basalts 2975 1971
BESSE AND COURTILLOT: APPARENT AND TRUE POLAR WANDER
Deccan Traps, Mahabaleshwar 65.5 5 40 276 7.4 7.4 17.9 73.6 28 190 2 Rc, M extrusives, basalts 2693 1972
Deccan traps, Dhar region 65.5 5 29 293 5.3 8.3 22.4 75.4 6 37 3 Rc, M extrusives 67 1981
Deccan Traps 66 4 32.6 290.8 3.8 5.9 20 76.5 21 110 4 Rc, M extrusives, basalts 564 1986
Central Kerala dikes 69 2 34.6 274 11.8 15.5 9.7 76.7 6 39 3 Ro intrusives, dolerites 7150 1994
Rajmahal Traps combined 116 2 7.5 296.5 3 3.5 24.5 87.5 48 294 3 Ro extrusives, basalts 3000 1971
a
See Table 1a footnote.
EPM
6-5
EPM
6- 6
Table 1c. Selected Paleomagnetic Data for Australia Used in This Study Extracted From the GPMD V3.3 Databasea
Newer Volcanics 2.5 5 86.6 266.3 1.9 1.9 38.5 143.5 46 133 2 Ra extrusives 1897 1971
Port Campbell Limestone, 12 4 77.2 303.5 4.2 4.2 38.7 143.1 30 48 3 N sediments, limestone 139 1985
Glenample Formation
Springfield Basin, red clays 14 18 80.4 270.5 8.4 8.4 32.5 138.5 6 37 2 Rc sediments, red clays 1976 1976
New South Wales, Nandewar, 17.5 5 79.4 246 4.2 4.2 32 150 51 149 2 Ro extrusives, basalts 1924 1974
Warrumbungle Volcanoes
Tweed and Main Range Volcanos 22.5 5 77.4 290.9 4.1 4.1 30 150 75 278 2 Ro extrusives, basalts 1925 1974
combined
Queensland, lavas and dikes 23.5 3 78 260 8 12 27 152.2 12 68 3 Rc extrusives, intrusives 1858 1966
Liverpool, Springsure Volcanoes and 30 10 68.9 272.4 4.3 4.3 35 150 52 162 2 Ro extrusives, basalts 1926 1974
older volcanics combined
Victoria, Browns Creek Formation 37 4 65.5 292.5 2.5 2.5 38.8 143.4 33 66 4 Rb, M sediments, red-brown clays 7097 1994
Barrington Volcano, Nerriga Province, 50 20 68.5 310.9 5.2 5.2 35 150 46 143 2 Ro extrusives, basalts 1927 1974
older volcanics combined
SW Queensland, Morney Profile, 50 30 58.8 298 3.8 3.8 27 141.5 37 2 Rb sediments, weathered 1972 1978
Eromanga Basin
Sydney Basin, Mogo Hill Basalt 57.5 3 40.6 310.2 8.6 8.6 33.2 151.1 9 44 3 M extrusives 241 1981
New England, New South Wales, 60 20 59.2 297.2 9.1 9.9 30.5 151.5 7 53 4 Ro sediments, weathered 1964 1988
weathered profile
Mt. Dromedary Intrusive Complex 95 10 56 318 9 9 36.3 150 22 55 3 Co, M intrusives 1848 1963
Western Australia, Bunbury Basalt 97.5 15 49 341 10 10 33.4 115.6 5 54 2 Ro, M extrusives, basalts 1932 1976
Tasmania, Cygnet Alkaline Complex 105 10 50 338 10 10 43 147 15 45 2 Co intrusives, alkali-syenite 1973 1962
Sydney Basin, Prospect dolerite 168 10 53 359.6 6.4 6.4 33.8 150.9 10 59 4 N intrusives, dolerite 84 1982
Tasmanian Dolerite 174 16 50.7 354.5 5.2 5.2 42 147.5 21 42 2 N intrusives, dolerite 1960 1977
New South Wales, Garrawilla Volcanics 197 20 46.1 355.2 10 10 31 150 14 36 2 N extrusives 1938 1976
and Noombi extrusives
a
Table 1d. Selected Paleomagnetic Data for South America Used in This Study Extracted From the GPMD V3.3 Databasea
Rock Name Age Dage Plat Plong Dp Dm Slat Slong B N Dc Tests Rock Type RN- Year
O
Fernando de Noronha Volcanics, Brazil 2.5 2.5 87 164.8 3.3 6.4 3.8 32.4 22 88 2 Rc extrusives, intrusives 3400 1967
Basalts, Quixaba Formation, Brazil 3.5 1.5 80.3 234 2.9 2.9 3.9 32.4 7 61 3 M extrusives, basalts 553 1986
Lipiyoc Formation, Argentinab 8.5 1 85.7 260.5 7.9 7.9 22.5 66.5 17 159 4 Ro, M extrusives, ignimbrites 7818 1996
Remedios, Sao Jose Formations, Brazil 9.5 1.5 84.5 136 7.5 7.5 3.9 32.4 7 45 3 Ro, M extrusives, basalts 555 1986
Rio Azul Sediments, Argentina 11 4 80.2 76 2.8 4.7 30.7 68.8 89 267 3 Rb sediments 8129 1990
Neuquen-Mendoza Lavas, Argentina 14 28 87.3 165.9 9.7 10.4 37.5 70 22 198 2 Rc, M extrusives, basalts 3319 1968
Miocene Basalt, Argentina 16.5 23 85.8 179.1 10.9 14.8 37.5 70 6 42 3 Ro, M extrusives, basalts 3145 1970
Volcanic Hills, Argentina 75.5 19 70.2 224.5 12.2 12.2 33 65 12 36 3 Ro extrusives, basalts 94 1983
Intrusives, Cabo de Santo Agostinho, Brazil 92 14 87.6 135.1 4.5 4.5 8.4 35 9 100 2 N intrusives 651 1980
Volcanics and red beds, Sierra de Los Condores, 111 28 84.2 270.6 4.7 4.7 32.2 64.1 8 78 2 Co extrusives, sediments, 2238 1976
Argentina redbeds
Maranhao Basin Intrusives, Brazil 118 12 83.6 261 1.9 1.9 6.5 42 21 190 2 N intrusives 611 1979
Cerro Barcino Formation, Argentina 119 13 84.9 0.8 5.6 5.6 43.5 69 11 66 4 F+ sediments, sandstones, 7498 1994
siltstones
Serra Geral Basalts, Brazil 119 10 84.6 295.4 3.7 3.7 29 50 37 260 2 Rb extrusives, basalts 2387 1976
Serra Geral Formation, Brazil 119 10 78.2 234.1 5.7 5.7 26 53 30 74 2 Rc, Co, M extrusives, basalts, 3599 1962
intrusives, diabases
Vulcanitas Cerro Colorado Formation Argentina 121 12 81 194 13.4 13.4 32 64 6 86 2 N extrusives, intrusives, 2700 1972
sediments, redbeds
El Salto-Almafuerte Lavas, Argentina 124 10 72 205 6.5 6.5 32.2 64.2 15 65 2 Rc extrusives 2003 1978
Serra Geral Formation, Brazil 128 15 83.5 280.5 2.2 3.6 29 50 79 261 2 Rb extrusives, basalts, 7675 1983
andesites, rhyolites
Serra Geral Formation Younger Group, 131 8 82 218 7.8 7.8 26 52 28 85 3 Ro, M extrusives, basalts 6280 1990
Brazil, Uruguay
Serra Geral Formation Main Group, 136 7 85 288 1.1 1.1 26 52 287 850 3 Ro, M extrusives, basalts 6279 1990
Brazil, Uruguay
Maranhao Basalts, Brazil 175 4 85.3 82.5 6.9 6.9 6.4 47.4 15 121 2 N extrusives, basalts 610 1979
a
b
Declinations are suspect, due to possible local tectonic rotations, and inclinations only are used (see text).
EPM
6-7
EPM
Table 1e. Selected Paleomagnetic Data for Africa Used in This Study Extracted From the GPMD V3.3 Databasea
Kenya and Tanzania, East African Volcanics 1 2 88.7 104 3.2 3.2 0 36 54 155 2 Rc, M extrusives 2234 1976
Libya, Haruj Assuad Volcanics 1 2 83 171 5 8.5 27.8 17.3 14 78 2 Rc extrusives, basalts 2611 1974
6- 8
Northern Nigeria, Jos Plateau newer basalts 1 2 66.9 242.2 2.4 4.8 9.6 8.8 10 97 3 Rb extrusives, basalts 6134 1991
Canary Islands, Gran Canaria and Tenerife basalts 1 1 80.2 140 3 5.2 28.1 16.5 11 43 2 N extrusives, basalts 2097 1978
Madagascar, Volcanics, Itasy 1 2 77.7 267.2 9.3 14.5 19.1 46.7 8 56 2 N extrusives 2330 1971
Sao Tome Volcanics 2 2 86.4 199.4 2.4 4.8 0 6.5 49 164 2 Rc extrusives, basalts 2509 1972
Tanzania, Ngorongoro Caldera 2.5 1 81 62 6 12 3.2 35.5 20 102 2 Ro, M extrusives 2944 1971
Combined Gran Canaria and Tenerife younger basalts 2.5 2.5 82.9 131.9 2.4 4.1 28.1 16.5 24 128 2 Ro extrusives 2099 1978
Madagascar, Plio-Pleistocene Volcanics combined 2.5 5 82.9 255.5 4 6.6 16.5 47.6 28 171 2 Rc extrusives 2329 1971
Kenya and Tanzania, East African Volcanics 3.5 3 86.5 147.6 2.3 2.3 0 36 102 255 2 Rb, M extrusives 2235 1976
Spain, Canary Islands, Late Tertiary basalts 3.5 1.5 83.8 126.8 3.8 6.3 28.1 16.5 13 85 2 Ro extrusives 2098 1978
Pliocene Volcanics, Canary Islands and Madeira 3.5 1.5 82.6 128.8 3.2 3.2 28 16 72 214 2 Rb, M extrusives, basalts 2517 1973
Volcanics, Kenya 4 6 83.9 296.6 3.9 3.9 1.5 36 161 240 2 Ro extrusives, basalts, welded tuff 796 1979
Spain, Lanzarote, Canary Islands, Famara Volcanics 7.5 2.5 87.5 178.2 5.4 8.5 29.2 13.5 17 108 4 M extrusives, basalts 7644 1995
Canary Islands, Basalts Series II, Fuerteventura 8 6 77.8 146.2 3.2 5.8 28.6 14.1 10 51 3 N extrusives 756 1979
Kenya, Ngorora Formation 11.5 3 85.7 255.8 1.9 3.8 1 35.5 104 312 4 Rb sediments 8133 1990
Libya Volcanics, Jebel Soda 11.5 3 78.4 196.1 7.4 7.4 28.7 15.6 12 138 2 Rc extrusives, basalts 2521 1973
Libya Volcanics, Jebel Soda 11.5 3 69 184 6.8 6.8 28.8 15.5 9 57 2 Ro extrusives, alkali basalts 2625 1975
Kenya and Tanzania, East African Volcanics 12 2 86.5 186.6 6.1 6.1 0 36 22 161 2 Rc extrusives 2236 1976
Miocene Volcanics, Canary Islands 13 8 81.9 114.4 3.5 3.5 28 16 99 291 2 Rb, M extrusives, basalts 2519 1973
Volcanics, Kenya 13.5 3 80.1 34.2 8.9 8.9 1.6 35.9 14 56 2 N extrusives, nephelinites, welded tuff. 797 1979
Santa Antao Volcanics, Cape Verde 14 9 84.5 168.2 3.7 6.9 17.1 25.1 40 120 2 Rc extrusives, intrusives 3263 1968
Sao Nicolau Volcanics, Cape Verde 14 9 86.8 124.7 4.2 7.8 16.6 24.3 12 36 2 Ro extrusives 3265 1968
Sao Tiago Volcanics, Cape Verde 14 9 82.3 178.9 2.8 5.3 15.1 23.6 30 93 2 Rc extrusives, intrusives 3268 1968
Sao Vicente Volcanics, Cape Verde 14 9 83 87.1 3.7 6.9 16.8 25 46 143 2 Rc extrusives, intrusives 3264 1968
Kenya, Turkana lavas 18.5 9 84.6 163.3 2.3 2.3 0 36 62 109 2 Rb, M extrusives 2237 1976
Algeria, Massif de Cavallo 19 6 86.8 22.9 2.2 3.3 32 5 13 51 2 N extrusives 2791 1969
Egypt, Cairo region basalts 19.5 7 66 167 2.3 2.3 30 31 11 132 4 F+ extrusives, basalts 8110 1995
Egypt, Cairo region basalts 19.5 7 76 111 3 3 30 31 18 216 4 F+ extrusives, basalts 8111 1995
Principe Volcanics 24 4 82.8 96.6 4.5 9 1.5 7.5 25 78 2 Rc extrusives, basalts 2512 1972
Ethiopian Traps, Lima Limo 30 1 75.5 207.5 3.5 6.9 13.2 37.9 20 302 4 Rc extrusives, basalts 6691 1999
Ethiopian Traps, Wegel Tena 30 1 81.1 226.4 4.3 8.6 11.5 39.2 20 158 4 Rc extrusives, basalts 6691 1999
Ethiopia Southern Plateau Volcanics 34 8 75.1 170.3 6.4 6.4 9.1 41 22 92 2 Rc extrusives 2764 1974
Egypt, Iron ores combined, Baharia Oasis 37 4 83.5 138.6 7 7 28.2 28.9 9 109 3 Ro sediments, iron ores 769 1981
Egypt, Basalts, Wadi Abu Tereifiya 44.5 11 69.4 189.4 3.2 6.1 30 32.1 6 111 3 N extrusives, basalt 1 1979
Ethiopian Trap Series 50 30 80.8 167.7 4.3 4.3 9.3 39 20 52 2 Ro extrusives, basalts 3025 1970
Egypt, Nubian Sandstone Combined 72.5 15 81.8 222.7 3.3 3.3 24.5 33.5 23 255 3 Rb sediments, sandstones 766 1981
Madagascar, Volcanics Combined 74 12 63.5 219.6 4.1 4.1 21 47.3 30 211 2 N extrusives 2338 1971
Sudan, Northern Volcanic Field 80.5 5 55.9 277.8 11.3 11.3 19 33.3 6 54 3 M extrusives, basalts, latites, phonolites 1179 1989
Madagascar, Volcanics, Massif d’Androy 82 16 65 252 8 8 24.3 46 7 36 2 N extrusives 2789 1969
Madagascar, Volcanics Combined 89.5 3 69.1 240.1 4.9 4.9 19 44.9 33 170 2 N extrusives 2339 1971
South Africa, Cretaceous Kimberlites 1 90.5 19 64.1 226.1 5.2 5.2 29 26 14 118 4 M intrusives, kimberlite 5983 1989
Egypt, Wadi Natash Volcanics 93 14 69.3 258.1 5.8 5.8 24.4 34.3 15 342 3 N extrusives, basalts, andesites 765 1981
Mozambique, Lupata Series Volcanics 111 4 61.8 259.5 3.2 4.5 16.7 34.2 7 61 2 F+ extrusives, trachytes, phonolites, kenyites 3586 1963
Namibia, Kaoko lavas 122 19 48.3 266.6 2.5 3.9 20 14 40 118 3 Ro, M extrusives, basalts 2580 1975
Zimbabwe, Mateke Hills Complexes 173 8 58.7 260.5 7 9.4 21.8 31.2 6 49 3 N intrusives, granite, granophyre, gabbro 3452 1964
a
See Table 1a footnote for explanation.
Table 1f. Selected Paleomagnetic Data for Europe Used in This Study Extracted From the GPMD V3.3 Databasea
River Volga Sediments 1 2 81 227 4.3 5.9 47 47 8 67 2 M sediments, clays, sandstones 3883 1979
Massif Central Lavas 2 3 4 77.5 187.5 12 12 44.6 3.5 10 73 3 Rc extrusives, basalts 3055 1970
Coiron Lavas 6 10 80.4 142.9 7.2 7.2 45 4 36 220 2 Ro extrusives, basalts 3050 1970
Suevites, Nordlinger Ries 15 2 77.5 146.3 1.6 2.1 49.9 10.5 12 111 2 N metamorphics, tuffaceous rocks 3472 1965
Vogelsberg Volcanics 16.5 3 85.1 200.9 6.6 8.3 50.5 9.4 31 93 4 Rc, M extrusives, basalts 6264 1990
Rheinland Pflaz Volcanics 22.5 13 70 108 6 8 50.5 7.5 22 53 2 Ro extrusives, basalts, trachyte, phonolite 3603 1962
Lausitz Volcanics 22.5 13 74.6 120.5 6.5 8.2 51 14.7 24 148 2 Rc, M extrusives, basalts, phonolites 3304 1967
Tertiary Volcanics, northern Bavaria 28.5 25 78 140 7.2 7.2 50.1 11.4 22 316 2 Rc, M extrusives 2287 1977
Hocheifel Tertiary Volcanics 34 22 80.8 182 4.2 4.2 50.3 7 47 351 3 Rb extrusives, andesites, basalts 781 1984
Scottish Tertiary dikes 46.5 23 73.4 196.8 4 5 55 4 21 84 2 Ro, Co intrusives, dykes 3456 1966
Tertiary dikes 53 6 77 213.5 10.9 14.2 53.6 2.9 11 54 2 Co intrusives 3182 1969
Faeroe Island Volcanics 57 3 77 161 2 2 62 7 253 1809 2 Ro extrusives 3208 1970
Northern Ireland, Antrim Basalts 57.5 15 69.6 162.9 5 7 55 6 25 225 2 Co extrusives, basalts 3026 1970
Northern England, Skye Lavas 57.5 15 71.5 165.2 2.8 3.8 57.4 6.3 90 344 2 N extrusives, basalts 2506 1972
Northern England, 58.5 13 75 240 5.5 5.5 54.5 2 10 83 3 Co intrusives 2870 1974
Cleveland-Armathwaite dike
U.K., Rhum and Canna igneous rocks 59 2 81.4 181.9 3.2 3.9 57 6.5 109 453 2 Rc, M extrusives, intrusives, dykes 68 1981
Scotland, dike swarm, Skye 59 2 82.5 158 2.1 2.5 57.1 5.9 409 1636 2 C+, Ra intrusives, dykes 75 1982
U.K. Ardnamurchan igneous complex 59.5 3 77 175 3.3 4.2 56.7 6.2 62 484 3 M intrusives 146 1984
Scotland, Mull lavas 62 2 72.2 168.3 3 4.1 56.4 6.1 78 492 2 M intrusives, basalts 1979 1977
Aix-en-Provence Sediments, Franceb 71.5 23 73 156 10 10 43.5 5.5 10 96 3 F+, Rc, M sediments, sandstones, limestones 6218 1989
Northern Ireland, Antrim basalts 71.5 15 70.9 125.8 12.1 13.6 55.1 6.4 19 79 3 M extrusives, basalts 2493 1972
Germany, Limestones, Munster Basin 92.5 9 75.8 181.1 3.8 3.8 52 8 9 191 3 Fo, M sediments, limestones 758 1979
Combined sills and dikes, Spitsbergen 108 7.5 61.8 222.9 6.8 8.2 77.5 17 13 150 4 Rc, M intrusives 1493 1985
France, Berriasian limestones 142 3 74.1 183.1 3.1 4.2 44.4 4.2 1 163 4 R-, M limestone, marls 617 1985
Norway, Hinlopenstretet Dolerites 144 5 66 200 7 8 79 20 9 157 4 M intrusives, dolerites 6220 1989
Switzerland and France, Blue limestones 155 2.5 77.2 149 4.6 6.5 47.3 7.2 24 204 3 F+, Ro, M sediments, limestones 427 1984
France, Terres Noires 158 3 77.6 129.7 7.1 10.1 44.5 4.3 5 40 3 F+ sediments 8204 1992
Switzerland, France, 159 2 76.5 142.5 4.8 6.7 46.5 6 16 298 4 F+, Ro, M sediments, ferriferous oolites 6501 1991
Jura Mountains Oolites
Germany, limestones 155 3 72.7 125.7 4.7 6.3 49 11 12 252 2 Ro, M sediments, limestones 2289 1977
Poland, Oxfordian sediments 155 2 70.1 147 4.2 5.7 52.9 18 6 61 4 M sediments, limestones 3119 1988
Poland, limestones, 158 7 72.3 150.4 7.7 10.6 50.3 19.5 8 55 4 Ro, M sediments, limestones 1802 1987
Krakow-Czestochowa Upland
Alsace, France, Jurassic sediments 168 4 63.1 120.1 6.1 8.7 48.7 7.5 7 47 4 F+, Ro, M sediments, limestones 792 1988
Sweden, Scanian basalts 181 13 69 102 10 11 55.5 14 21 170 4 Rc, M extrusives, basalts 7078 1993
France, Thouars and Airvault 183 9 70.1 102.6 5.1 5.1 48 0.2 14 114 2 Rb sediments 3225 1987
Toarcian stratotypes
Yorkshire, U.K., Liassic sediments 192 6 76.9 134.7 2.5 2.5 54.6 0.8 29 185 3 M sediments 701 1982
Bretagne, France, Kerforne dikes 198 20 61.3 78.8 10.2 10.2 48.3 4.5 7 76 4 M intrusives, dolerites 7128 1993
Paris Basin sedimentsb 202 13 51.3 105 2.8 4.2 49.7 4 16 496 4 Rb, M sediments, limestones 7820 1996
a
EPM
b
Declinations are suspect, due to possible local tectonic rotations, and inclinations only are used (see text).
6-9
Table 1g. Selected Paleomagnetic Data for North America Used in This Study Extracted From the GPMD V3.3 Databasea
EPM
Alberta, Canada and Montana, Tills and Paleosols 1 2 80.4 119.5 8.2 10.6 48.8 113.5 22 250 4 Rc sediments, tills, paleosols 7663 1995
Alberta, Canada and Montana, Kennedy Drift 1 2 80.4 119.5 8.2 10.6 48.8 113.5 22 237 3 Rc sediments, tills, paleosols 7774 1995
Wyoming and Montana, Clinker deposits 1 2 81 158 6 7.8 46 105 17 150 3 F+, Ro, M sediments, baked 1227 1984
Northwest Territories, Canada Katherine Creek sedimentb 1.5 3 77.8 123.7 7.9 8.8 65 127.6 8 100 4 Ro sediments, tills 8031 1996
6 - 10
SW Wyoming, Leucite Hills Volcanicsb 1.5 1 83.2 117.8 6.1 8.5 41.6 109 39 232 2 Rc intrusives 1265 1980
Colorado, Lake City Calderab 23 2 76.2 210 11.2 13.6 38 107.3 17 128 3 C+, Rc, Fo, M extrusives, intrusives 356 1986
San Luis Basin, Colorado, Conejos and Hinsdaleb 26 6 79.7 162.6 6.9 9.7 37.2 105.6 23 168 4 Rc extrusives, basalts, 8164 1997
andesites, dacites
Colorado, Volcanics, San Juan Mountainsb 30.5 9 85.2 304.3 8 10.7 37.5 106.5 18 164 2 Rc extrusives, intrusives 2556 1974
Labrador, Canada, Mistastin Lake Impact Structure 38 8 85.5 117.7 3.4 4 55.9 63.4 10 73 3 N metamorphics, impact 3235 1969
melt rocks
Wyoming, East Fork, Washakia Basin sedimentsb 44 12 83.9 144.2 8.8 11.7 43 109.5 10 85 3 Rc, Fo sediments 1138 1986
Wyoming, Rattlesnake Hills Volcanicsb 46 6 82.6 151.9 12.3 12.3 42.8 107.3 9 64 2 N extrusives 2209 1977
Wyoming, Absaroka basaltsb 46.5 5 83.5 177.4 10.1 10.1 44.5 110 19 91 2 Rc, M extrusives 2014 1977
Virginia, Monterey Intrusions 47 4 87.6 45.9 12 12 38.4 79.6 6 36 3 Rc, M intrusives, felsite 2471 1974
Montana, Robinson Anticline Intrusive Complex 50.5 5 77.1 145.8 4.7 6.1 46.2 111.5 16 93 3 Rc, M intrusives, syenite, 458 1988
trachyte sills
Montana, Highwood Mountains intrusions 51 4 81.2 167.3 7.1 7.1 47.4 110.6 29 220 2 N intrusives 1262 1980
Wasatch and Green River Formationsb 51 4 77.8 128.4 4.8 6.9 41.6 110.4 25 129 4 Rb sediments, redbeds, shales, 8194 1997
limestones
Montana, Bearpaw Mountains intrusions 52 4 80.5 198.4 5.8 5.8 48.2 109.7 18 160 2 Ro intrusives 1261 1980
Montana, Paleocene igneous rocks 63 8 81.8 181.4 5.4 5.4 47.6 108.9 36 311 3 Rc intrusives 293 1983
Alberta, Canada, Edmonton Group 63.5 1 72 183 10.3 11.6 51.9 112.9 20 60 3 N sediments 1729 1985
Montana, Alkalic intrusions 64 6 80.5 185.1 5.6 5.6 47.5 109 33 284 2 Ro intrusives 1264 1980
Montana, Boulder Batholith 75 10 73 249 12 13 46 112.5 27 314 2 Ro, M intrusives 2835 1973
Montana, Adel Mountains Volcanics 76 10 83.4 200.9 6.5 7.7 47.5 111.9 26 193 4 F+, G+, M extrusives, intrusives 6165 1991
Montana, Maudlow Formation 79.5 11 69.8 207.8 9.8 9.8 46.1 111.1 11 55 4 F+, Rc sediments, volcaniclastic 6223 1989
Montana, Elkhorn Mountains Volcanics 80 6 80.3 189.5 9.6 9.6 46.1 112 15 112 4 F+, Rc extrusives, andesites, 6191 1991
sediments, volcanic
Arkansas Alkalic Intrusives 94 12 74.1 192.5 5.7 5.7 34.3 92.5 20 147 4 M intrusives 401 1988
New England Intrusions 102 3 76.6 167.5 5.3 5.3 43.6 71.6 16 106 4 M intrusives, syenite 8096 1996
New England Intrusions 112 2 74.5 195.2 3.8 3.8 43.5 71 27 157 4 M intrusives, syenite, 8095 1996
rhyolite, gabbro
Quebec, Canada, Monteregian Hills intrusives 120 10 69.9 188.7 3.7 5 45.5 73 16 74 2 Rc intrusives 3282 1968
Quebec, Canada, Monteregian intrusives 120 10 71.3 189.5 2.7 3.6 45.3 72.8 32 147 2 Ro intrusives 3178 1969
Quebec, Canada, Monteregian intrusives 120 10 72.4 191 2.8 3.7 45.3 73.2 70 760 3 Ra intrusives 1610 1979
Notre Dame Bay, Newfoundland, Canada, dikes 122 20 73.9 201 5.3 6.8 49.5 55.1 15 86 3 N intrusives, lamprophyre 5942 1981
Southern Maine intrusions 123 5 72.2 198.9 3.3 3.3 43.3 70.7 41 236 4 R+, M intrusives, gabbro, 7837 1996
granodiorite, diorite
Quebec, Canada, Mount Megantic intrusions 130 20 75.2 181.4 4.2 5.9 42.4 71.2 44 101 3 Rb, Co intrusives, gabbro, granite, 1724 1985
syenite
Notre Dame Bay, Newfoundland, Canada, dikes 135 30 67 212 4.3 5.6 49.5 55.5 10 68 3 N intrusives, lamprophyre 1611 1979
New York, Ithaca kimberlites 142 8 58 203.1 3.8 3.8 42.5 76.5 7 48 4 Rb, C+, M intrusives, kimberlite 6871 1993
Louisiana, Winfield Salt Dome Cap Rock 152 11 76.2 120.5 2.8 4.9 31.9 92.6 100 4 Ro, M sediments, anhydrite 7674 1993
Wyoming, Twin Creek Formationb 168 21 83 286 12 15 43.2 110.5 8 60 4 F*+ sediments, limestones 6010 1990
Vermont, New Hampshire, White Mountains Plutons 169 16 88.4 82.1 6.1 6.1 44 71.5 10 50 4 N intrusives, granite, syenite, 6018 1990
diorite
North Carolina, northwest dikes 180 30 52.6 60.7 4.9 9.8 36 79.8 11 77 3 N intrusives 1517 1987
North Carolina, north-south dikes 180 30 69.6 47.1 5.5 9.2 36 79.8 15 100 3 Ro intrusives 1518 1987
Vermont, New Hampshire, White Mountain Volcanics 180 20 85.5 124.5 4.5 6 44 71 12 130 2 Rb extrusives 3434 1966
New Brunswick, Canada, Caraquet dike 191 5 74.1 114 7.6 11.4 46.8 66 8 36 3 M intrusives, dolerite 1661 1981
BESSE AND COURTILLOT: APPARENT AND TRUE POLAR WANDER EPM 6 - 11
1987
1987
1966
1981
1968
1982
1982
1972
1968
1961
1989
1994
1990
1986
Year
Gordon, 1984]. Despite the huge number of cores which
have been drilled in recent decades in the various steps of
international drilling projects, only a few reliable data can
RNO
1517
1518
3434
3280
1661
1477
1477
3546
2461
3610
5963
7235
6020
1232
be used for the determination of paleolatitudes. Peirce
[1976, 1978] provided early compilations of these data,
and selected the most reliable studies. One of the main
intrusives, metamorphics,
problems encountered in oceanic cores is the absence of
extrusives, intrusives
extrusives, intrusives,
orientation: declination is not available, unless one uses the
extrusives, intrusives
extrusives, basalts
sediments, red beds,
sediments, red beds
sediments, redbeds,
baked sediments
Rock Type
intrusives, dolerite
intrusives, diabase
extrusives, basalts
present Earth’s field to reorient the core. Inclination alone

allows only the determination of small circles of equal
paleocolatitude on which the pole must lie. Another problem
extrusives
diabase
intrusives
intrusives
intrusives
intrusives
is the underestimation of inclination due to arithmetic

averaging (but see Enkin and Watson [1996]). In order to
overcome this problem, Peirce [1976] used an unpublished
method of Cox, which also gives an estimate of the standard
G+, M
Tests
error (SM). We have selected those data corresponding to

C+
C+
Ro
Rb
M
M
N
N
N
N
N
N
values of SM less than 5. A minimum amount of 36 samples

in a study was used, as was the case for terrestrial paleo-
magnetic poles.
Dc
3
3
2
3
2
3
3
2
2
3
4
3
[18] In addition, as regards data postdating the Peirce

[1976, 1978] compilations, we have used the paleomagnetic
100
130
375
313
77
36
40
80
80
78
49
80
53
89
studies of oceanic cores drilled during campaigns of the

N
Ocean Drilling (ODP) and Deep Sea Drilling (DSDP)

Projects in the Indian (Wharton Basin and Ninety-East
15
12
25
20
20
78
29
16
11
11
B
Ridge) and South Atlantic (South American plate) Oceans.

We used the Leg 73 paleolatitudes published by Tauxe et al.
79.8
79.8
65.4
81.5
81.5
76.3
72.7
77.5
74.3
Slong
71
66
75
73
75
[1983], and the Leg 115 data of Schneider and Kent [1990a]
and Vandamme and Courtillot [1990]. The latter studies give
paleolatitudes for both the Indian and African plates. All
40.5
46.8
40.2
40.5
40.5
34.3
34.3
41.5
44.9
Slat
42
36
36
44
39
these results are listed in Table 2b. The distribution of these

data, altogether representing 19 determinations from 0 to
10.5
12.6
11.4
Dm
1.5
3.9
9.8
5.4
9.2
9.3
9.3
3.7
6.5
4
6
120 Ma, is also plotted in Figure 1a.

10.5
2.3. Combining Data Types

7.6
1.5
3.9
5.4
4.9
5.5
4.5
4.9
4.9
7.2
3.7
4.5
Dp
[19] The space and time distributions of selected paleo-

Plong
104.5
124.5
115.1
magnetic data are shown in map (Figures 1a and 1b) and

88.6
94.5
60.7
47.1
96.1
96.1
87.5
73.1
108
104
114
histogram (Figures 1c – 1e) form. A total number of 242

independent data have finally been retained. This final data
62.2
74.1
55.5
52.6
69.6
85.5
66.1
66.1
65.5
65.5
Plat
62
63
68
73
set comprises 221 ‘‘classical’’ poles, two poles determined

using marine magnetic anomaly skewness data and 19
Dage
14
30
10
20
30
20
15
14
10
‘‘inclination only’’ data from oceanic cores. The total

3
5
5
8
8
number of data is thus increased by more than 2 with

respect to the BC91 database, which comprised only 111
Age
208
191
201
206
180
195
200
180
180
194
194
198
Pennsylvania, Culpeper Basin intrusives and baked sediments 201

191
poles. The map (Figure 1b) shows all poles (as dots) and the
corresponding sites (as stars) in (South) African coordi-
nates, with sites restored to their relative positions at the
appropriate time. We see that geographical coverage is quite
Vermont, New Hampshire, White Mountain Volcanics
reasonable; it is much improved over BC91. The histogram

New Jersey, Passaic Formation, Preakness Basalt
New Jersey, Newark Basin Hettangian red beds
(Figure 1c) shows that a large number of data (more than

Nova Scotia, Canada, North Mountain Basalt
30) is available for the 0– 30 and 50– 80 Ma time windows.

By contrast, the 140– 150 Ma time window contains only

South Carolina Piedmont, Diabase dikes
South Carolina Piedmont, Diabase dikes
New Brunswick, Canada, Caraquet dike
eight studies. The mean number of data in each 10 Myr time

Rock Name
window is on the order of 15. As a matter of comparison,

North Carolina, north-south dikes
Combined Pennsylvania diabase

North Carolina, northwest dikes
the corresponding mean for the recent Prévot et al. [2000]

Newark Supergroup Volcanics
Connecticut Valley Volcanics
study is on the order of 4, and there are less than 5 data in

Table 1g. (continued)
twelve 10 Myr time windows.

See Table 1f.
Newark Group
3. Kinematic Models
[20] Despite the significant improvement in the paleomag-
netic database, the number of data is still rather low if one
a
b
wants to construct the APWP of a single plate based on its

EPM 6 - 12 BESSE AND COURTILLOT: APPARENT AND TRUE POLAR WANDER
Table 2a. Pole Positions Deduced From Skewness of Marine motions at the time of each anomaly between 0 and
Anomalies 24 and 25 in the Indian Oceana 130 Ma, by simply subtracting the absolute motion of
Mean Age Anomaly l(N) f(E) ND 95%CI K (South) Africa with respect to hot spots from the motion
of other plates (Table 3). The finite poles of rotation which
57.2 25 76.8 224.7 17 6.9 30.7
61.5 26 76.4 200.6 14 6.2 48.2 are not given by Müller et al. [1993] are those of Müller and
a
Definitions: l(N), pole latitude; f(E) pole longitude; ND, number of
Roest [1992] for the motion between Europe and North
determinations; 95%CI, confidence interval; K, precision parameter. America, and those of Srivastava and Tapscott [1986] for
Greenland and North America. They were combined with
the North America versus South Africa motion to obtain
own paleomagnetic poles only. As in BC91, we therefore take total motion with respect to South Africa. The Arabia to
advantage of the fact that the APWPs from individual plates Africa rotation poles are those of Le Pichon and Gaulier
cannot be independent one from the other and should be [1988]. For periods earlier than 130 Ma, the South America
related through plate kinematic models. And indeed, accurate to South Africa kinematic parameters are those of Nürnberg
up-to-date models are available for the Indian, North Atlan- and Müller [1991], the Australia to East-Antarctica to South
tic, central Atlantic, and South Atlantic Oceans, allowing one Africa rotation poles are those of Royer and Sandwell
to relate the seven selected plates over the entire period of [1989], using respectively a recomputed fit and a fit after
interest here with reasonable confidence. A possible motion Lawver and Scotese [1987]. Prior to 130 Ma, the Mada-
between Eurasia and North America [Van der Voo, 1993] may gascar to South Africa motion uses the parameters from the
constitute a source of error during Jurassic times. These Global Isochron Chart of Royer et al. [1992], with a final fit
models can be used to transfer all paleomagnetic data to a from Lawver and Scotese [1987].
common, single reference frame, arbitrarily taken here to be [23] For each plate, paleomagnetic or skewness poles and
the southern part of the African plate. small circles based on DSDP/ODP inclination data were
[21] Müller et al. [1993] have published a new set of data, transferred onto (South) Africa. For this, we interpolated a
in which plate motions with respect to hot spots are Eulerian pole of rotation between the two published finite
computed based on combined Atlantic and Indian Ocean rotation poles with ages bracketing the estimated age
hot spot tracks between the Present and 130 Ma. The initial respectively of the pole and drill hole or crustal block
relative plate motion parameters are those quoted in the corresponding to the individual datum. Errors incurred in
database of Royer et al. [1992], and are computed at any such reconstructions are unlikely to exceed 2 (see Molnar
given age for the same anomaly. We have taken advantage and Stock [1987], discussion by Besse and Courtillot [1988,
of these compilations, both because of their unifying fea- 1991], and Acton and Gordon [1994]). We must be cautious
tures, and because one of the aims of this paper is indeed to in case the timescales used by different authors are not
compute relative motions between the paleomagnetic and identical and therefore require some discussion. McElhinny
hot spot reference frames. Müller et al. [1993] extensively and Lock [1995] used the Harland et al. [1989] timescale,
discuss the implications of their plate motion models. Two whereas Müller et al. [1993] used that of Kent and Grad-
important features of their reconstructions are the internal stein [1986]. We have compared the differences in the ages
deformation of the African and South American plates of each particular chron or anomaly, and of geological stage
during breakup phases in the Cretaceous, and a new model boundaries, when the two different timescales are used.
of the final breakup of Gondwana, in which India rifted Between 120 Ma and the Present, the corresponding mean
away from Antarctica at the same time it rifted away from age differences are generally less than 1 Myr, with a
Australia (i.e., between chrons M10 and M11). maximum of 1.4 Myr between 55 and 170 Ma. Between
[22] On the basis of the Royer et al. [1992] and Müller et 120 and 130 Ma (i.e., from the Valanginian to the early
al. [1993] kinematics, we have recomputed the relative Aptian), the difference reaches 4 to 7.5 Ma, but this is a time
Table 2b. Selected Data From ODP and DSDP Legsa

Mean Age Sl (N) Sf (E) Colatitude NC 95%CI Site Plate Rock Type Author
3.3 26.1 5.1 121.7 83 4.2 519 AFR s Tauxe et al. [1983]
15 20.1 61.5 82 49 4.1 222 IND s Pierce [1976]
18 26.1 10.3 123.7 98 3.1 521 AFR s Tauxe et al. [1983]
26 26.1 5.1 121.2 55 3.1 522 AFR s Tauxe et al. [1983]
27.5 28 36.3 104.7 42 4.4 17 AMS s Pierce [1976]
28 26.1 5.1 122.9 100 1.5 522 AFR s Tauxe et al. [1983]
32.5 28.5 2.2 123.8 65 4.4 523 AFR s Tauxe et al. [1983]
40.3 28.5 2.2 132.3 82 2.7 523 AFR s Tauxe et al. [1983]
45.5 28.5 2.2 133.5 102 3.8 523 AFR s Tauxe et al. [1983]
47 4.2 73.4 102.3 52 1.8 713 IND v Vandamme and Courtillot [1990]
54.7 24.9 87.4 141.9 49 3 253 IND s Pierce [1976]
57.5 19.2 99.3 119.7 38 4.8 212 AUS s Pierce [1976]
63.9 29.5 3.5 135.9 257 2.6 524 AFR s Tauxe et al. [1983]
66 7.5 59 115.2 56 4.2 707 AFR v Vandamme and Courtillot [1990]
104.5 16.1 116.3 116.7 61 3.9 260 AUS s Pierce [1976]
114.5 23.3 111 135.9 44 4.7 263 AUS s Pierce [1976]
114.5 23.3 111 132 106 2.9 263 AUS s Pierce [1976]
a
Sl(N), site latitude; Sf(E) site longitude; colatitude, paleomagnetic colatitude; NC, number of samples; 95%CI, confidence interval; site, number,
plate to which datum belongs; rock type, s, sediment ; v, volcanics; author, reference.
Table 3. Finite Rotation Poles for Gondwanaa

NW Africa onto South Africa Arabia onto South Africa
An Age, Ma l(N) f(E) w, deg Source Age, Ma l(N) f(E) w, deg Source
A34 84.0 0.0 0.0 0.0 4.7 32.75 22.64 1.89 (2)
100.0 17.3 172.8 5 (1) 13 32.15 22.59 5.36 (2)
110.0 15.7 174.0 .9 (1) 30 32.11 22.57 7.36 (2)
M0 118.7 18.1 173.6 1.1 (1)
M10 130.0 29.8 15.2 1.6 (1)
East Antarctica to South Africa India to South Africa Australia to South Africa
An Age, Ma l(N) f(E) w, deg Source l(N) f(E) w, deg Source l(N) f(E) w, deg Source
5 10.4 7.9 49.5 1.5 (1) 23.7 146.7 4.6 (1) 11.9 130.4 6.6 (1)
6 20.5 10.8 48.0 2.8 (1) 29.6 156.2 7.6 (1) 14.5 133.5 11.76 (1)
13 35.5 14.1 50.7 5.5 (1) 21.4 139.3 15.3 (1) 13.3 130.6 20.16 (1)
18 42.7 15.5 44.8 7.2 (1) 23.6 138.1 17.9 (1) 15.9 130.8 22.76 (1)
21 50.3 10.3 42.9 8.8 (1) 23.7 144.1 22.1 (1) 16.0 127.6 23.76 (1)
25 58.6 5.3 39.1 10.3 (1) 23.3 152.7 28.6 (1) 15.9 122.7 24.16 (1)
31 68.5 1.1 41.6 11.8 (1) 20.5 160.8 42.4 (1) 11.4 118.2 25.66 (1)
33y 73.6 1.9 41.4 13.5 (1) 22.2 162.3 45.7 (1) 13.5 115.2 26.36 (1)
33o 80.2 4.6 39.7 16.0 (1) 23.9 163.2 48.8 (1) 16.3 110.5 27.16 (1)
34 84.0 2.6 38.1 17.9 (1) 22.9 161.2 51.0 (1) 14.7 106.3 27.86 (1)
90.0 3.9 33.2 22.1 (1) 22.5 156.4 53.0 (1) 15.7 97.8 29.06 (1)
100.0 5.6 29.0 29.3 (1) 21.4 149.3 56.8 (1) 18.0 85.0 31.86 (1)
110.0 6.5 26.1 36.6 (1) 21.4 149.3 56.8 (1) 20.5 73.8 35.36 (1)
M0 118.7 7.1 24.4 43.1 (1) 21.7 148.4 57.0 (1) 22.0 65.7 39.26 (1)
M10 130.0 11.9 21.0 50.4 (1) 22.3 143.1 61.0 (1) 28.2 56.9 43.56 (1)
M16 140.0 7.0 26.9 50.70 (3) 21.3 139.0 64.7 (4) 21.7 63.8 46.9 (5)
M21 150.0 4.7 29.0 52.84 (3) 21.6 135.2 66.6 (4) 19.2 63.7 49.7 (5)
FIT 165.0 7.8 31.4 58.0 (5) 26.9 133.4 67.8 (4) 22.6 62.7 55.2 (5)
Madagascar to South Africa
An Age, Ma l(N) f(E) w, deg Source
M0 118.7 3.5 76.3 0.9 (1)
M10 130.0 3.4 79.2 6.9 (1)
M16 141.9 5.40 76.20 8.32 (6)
M21 149.9 4.00 71.40 11.32 (6)
FIT 165.0 3.41 81.70 19.73 (7)
a
Recomputed following (1) Müller et al. [1993]; (2) Le Pichon and Gaulier [1988]; (3) Royer et al. [1988]; (4) India to South Africa motion is India to East
Antarctica [Royer and Sandwell, 1989] plus East Antarctica to South Africa [Royer et al., 1988]; (5) Australia to South Africa motion is Australia to East
Antarctica [Royer and Sandwell, 1989] plus East Antarctica to South Africa [Royer et al., 1988]; (6) Royer et al. [1988]; and (7) Lawyer and Scotese [1987].
window with no sedimentary data. From 135 Ma back to the done by iteratively finding a direction such that the Fisher
earliest seafloor spreading anomalies at 160 Ma, the differ- average of pole directions and the points on the small circle
ences are between 0 and 2.8 Myr, averaging 1 Myr in which are closest to this direction are statistically the same).
absolute value. Prior to 160 Ma, only geological age assign- [25] In order to give some feeling for the quality of
ments of formations from which paleomagnetic data are individual means, four examples showing the poles and
available are used and not plate kinematic parameters. For sites on an appropriate plate reconstruction are given for 20,
our purpose, these differences are smaller than the uncer- 60, 90, and 200 Myr in Figure 2. Thirty-nine individual data
tainties in age determinations in most paleomagnetic stud- are available for the 20 Myr reconstruction (Figure 2a), with
ies, and often less than the error due to reassigning an age to good longitudinal and latitudinal coverage. The correspond-
an anomaly when the exact time feature used in pointing the ing pole distribution is Fisherian and the 95% confidence
anomaly is not accurately known. We have therefore used interval is less than 3. Data come from most major plates.
ages as given by the original authors, regardless of the The 60 Myr reconstruction (Figure 2b) involves 46 data; the
timescale they used. However, as mentioned above in the distribution is approximately Fisherian with a 95% confi-
discussion of the new paleomagnetic database, we have dence interval of 3, although several poles streak E-W, east
scaled the ages of all DSDP/ODP sites using the Harland et of (65N, 270E). There are no data from South America.
al. [1989] timescale. The 90 Myr reconstruction (Figure 3a) contains only 13
data with a 95% confidence interval of 2.9, but geo-
graphical coverage is quite good, with all major landmasses
4. New Synthesized APWPs except Antarctica providing compatible data. There are still
4.1. African APWP: Old and New 20 data available for the 200 Myr reconstruction (Figure 3b),
[24] All data and sites are shown in Figures 1a and 1b in again with most plates (except South America) in their
(South) African coordinates, together with average poles Pangea configuration represented, a good Fisherian pole
(Table 4) calculated every 10 Myr with a 20 Myr window, distribution and a 95% confidence interval of about 4.
using a method derived from McFadden and McElhinny Altogether, the A95 values range from 2 to 7, with only a
[1988] in order to combine poles and small circles (this is slight degradation as one goes back in time (Figure 1d). The
Table 4. Master Apparent Polar Wander Path for the Past 200 Myr Calculated for a 20 Myr Sliding Window Every 10 Myra
South Africa South America India Australia
Window Age N l (N) f (E) A95 K l (N) f (E) A95 K l (N) f (E) A95 K l (N) f (E) A95 K
0 3.1 30 86.5 171.6 2.6 105.3 86.3 172 2.6 105.1 86.1 174.8 2.6 105.2 86.1 174.8 2.6 105.2
10 8.3 54 86.0 160.8 2.0 95.8 85.4 162.5 2.0 94.4 85.0 168.1 2.0 94.2 85.0 168.1 2.0 94.2
20 18.9 38 85.4 151.9 2.7 75.9 84.0 154.8 2.7 76.2 83.3 164.2 2.7 75.6 83.3 164.2 2.7 75.6
30 29.5 23 85.1 162.2 3.8 65.6 82.8 158.1 3.8 66.2 81.5 169.2 3.8 65.4 81.5 169.2 3.8 65.4
40 40.0 24 84.3 172.4 3.3 82.1 81.3 162.4 3.3 81.9 79.5 174.4 3.2 85.5 79.1 175.1 3.2 87.5
50 52.2 31 84.7 174.7 3.4 60.1 80.9 164.4 3.4 59.4 77.9 179.3 3.4 58.2 76.3 178.0 3.4 59.7
60 59.7 45 84.7 217.6 2.8 57.4 81.1 190.5 2.9 56.1 75.9 196.8 2.9 54.7 74.4 191.3 2.8 57.2
70 67.3 34 83.8 241.6 3.2 60.3 80.3 204.3 3.2 61.0 74.2 204.8 3.2 61.2 72.9 197.9 3.2 61.7
80 77.9 14 84.7 275.8 6.0 45.1 81.4 206.1 5.9 47.2 74.7 207.4 5.9 47.0 73.1 202.2 6.0 45.7
90 90.0 13 85.0 320.8 5.3 61.7 82.2 202.1 5.2 65.2 75.5 207.4 5.1 65.9 72.9 203.8 5.1 67.2
100 97.6 12 85.5 9.7 6.8 42.0 81.7 180.1 6.7 43.0 76.6 195.8 6.7 43.1 73.0 194.4 6.7 43.0
110 113.6 17 77.7 20.2 4.1 76.7 80.0 183.6 4.2 74.8 75.1 193.8 4.2 75.3 71.3 194.5 4.2 75.0
120 119.1 20 76.4 17.3 2.3 209.6 78.2 189.4 2.4 182.9 73.1 193.9 2.4 184.3 69.3 196.0 2.4 183.7
130 126.4 14 75.3 14.0 3.2 154.5 75.8 192.9 2.8 205.5 70.6 193.0 2.8 205.6 66.8 196.5 2.8 205.8
140 136.8 7 72.4 5.8 6.5 87.4 73.8 197.6 6.0 103.2 68.3 194.2 6.0 103.4 64.6 198.4 6.0 103.4
150 151.6 10 67.0 26.6 6.8 50.8 75.0 159.9 6.6 54.3 73.6 167.7 6.6 54.5 68.7 175.2 6.6 54.4
160 162.3 15 62.9 31.6 5.0 58.5 72.5 144.0 5.0 59.7 73.7 149.7 5.0 59.7 68.4 161.3 5.0 59.7
170 173.4 21 56.9 39.3 6.0 28.8 69.7 112.5 6.7 23.6 75.5 110.1 6.7 23.6 70.9 131.8 6.7 23.6
180 178.8 18 53.2 45.7 5.4 41.3 65.5 95.9 5.6 39.7 73.0 83.4 5.6 39.7 69.9 109.2 5.6 39.7
190 189.7 23 54.5 45.0 4.2 52.9 65.3 98.4 4.2 52.9 72.6 86.8 4.2 52.9 69.3 111.8 4.2 52.9
200 196.7 19 58.2 46.9 4.3 61.6 63.2 106.0 4.3 61.6 69.8 95.6 4.3 61.6 66.1 117.7 4.3 61.6
Antarctica Europe North America Greenland
Window Age N l (N) f (E) A95 K l (N) f (E) A95 K l (N) f (E) A95 K l (N) f (E) A95 K
0 3.1 30 86.5 171.6 2.6 105.3 86.3 172 2.6 105.1 86.1 174.8 2.6 105.2 86.1 174.8 2.6 105.2
10 8.3 54 86.0 160.8 2.0 95.8 85.4 162.5 2.0 94.4 85.0 168.1 2.0 94.2 85.0 168.1 2.0 94.2
20 18.9 38 85.4 151.9 2.7 75.9 84.0 154.8 2.7 76.2 83.3 164.2 2.7 75.6 83.3 164.2 2.7 75.6
30 29.5 23 85.1 162.2 3.8 65.6 82.8 158.1 3.8 66.2 81.5 169.2 3.8 65.4 81.5 169.2 3.8 65.4
40 40.0 24 84.3 172.4 3.3 82.1 81.3 162.4 3.3 81.9 79.5 174.4 3.2 85.5 79.1 175.1 3.2 87.5
50 52.2 31 84.7 174.7 3.4 60.1 80.9 164.4 3.4 59.4 77.9 179.3 3.4 58.2 76.3 178.0 3.4 59.7
60 59.7 45 84.7 217.6 2.8 57.4 81.1 190.5 2.9 56.1 75.9 196.8 2.9 54.7 74.4 191.3 2.8 57.2
70 67.3 34 83.8 241.6 3.2 60.3 80.3 204.3 3.2 61.0 74.2 204.8 3.2 61.2 72.9 197.9 3.2 61.7
80 77.9 14 84.7 275.8 6.0 45.1 81.4 206.1 5.9 47.2 74.7 207.4 5.9 47.0 73.1 202.2 6.0 45.7
90 90.0 13 85.0 320.8 5.3 61.7 82.2 202.1 5.2 65.2 75.5 207.4 5.1 65.9 72.9 203.8 5.1 67.2
100 97.6 12 85.5 9.7 6.8 42.0 81.7 180.1 6.7 43.0 76.6 195.8 6.7 43.1 73.0 194.4 6.7 43.0
110 113.6 17 77.7 20.2 4.1 76.7 80.0 183.6 4.2 74.8 75.1 193.8 4.2 75.3 71.3 194.5 4.2 75.0
120 119.1 20 76.4 17.3 2.3 209.6 78.2 189.4 2.4 182.9 73.1 193.9 2.4 184.3 69.3 196.0 2.4 183.7
130 126.4 14 75.3 14.0 3.2 154.5 75.8 192.9 2.8 205.5 70.6 193.0 2.8 205.6 66.8 196.5 2.8 205.8
140 136.8 7 72.4 5.8 6.5 87.4 73.8 197.6 6.0 103.2 68.3 194.2 6.0 103.4 64.6 198.4 6.0 103.4
150 151.6 10 67.0 26.6 6.8 50.8 75.0 159.9 6.6 54.3 73.6 167.7 6.6 54.5 68.7 175.2 6.6 54.4
160 162.3 15 62.9 31.6 5.0 58.5 72.5 144.0 5.0 59.7 73.7 149.7 5.0 59.7 68.4 161.3 5.0 59.7
170 173.4 21 56.9 39.3 6.0 28.8 69.7 112.5 6.7 23.6 75.5 110.1 6.7 23.6 70.9 131.8 6.7 23.6
180 178.8 18 53.2 45.7 5.4 41.3 65.5 95.9 5.6 39.7 73.0 83.4 5.6 39.7 69.9 109.2 5.6 39.7
190 189.7 23 54.5 45.0 4.2 52.9 65.3 98.4 4.2 52.9 72.6 86.8 4.2 52.9 69.3 111.8 4.2 52.9
200 196.7 19 58.2 46.9 4.3 61.6 63.2 106.0 4.3 61.6 69.8 95.6 4.3 61.6 66.1 117.7 4.3 61.6
a
Window, age of the center of window; age, mean age computed from the data; N, number of studies; l, f, latitude and longitude of mean VGP; A95
uncertainty at the 95% confidence level; K, Fisher’s precision parameter.
overall mean A95 is 2.9 ± 0.8 (1s) for the new synthesized the 95% probability level when the test parameter is
APWP, significantly less than the value of 4.1 ± 1.2 found positive. Four apparently discrepant poles deserve some
by BC91; the decreasep(4.1/2.9 = 1.4) is not much less than comment. Two poles appear to be significantly different
the maximum value ( 2.2 = 1.5) expected from increasing from the coeval synthesized pole: those for NAM in the 0 –
the number of data. 10 Myr window and EUR at 70 Ma (60 –80 Myr window).
[26] There are enough data that a significant synthetic For the NAM pole, there are only 6 data and the discrep-
APWP can be calculated with a time resolution increased ancy could for instance be due to a slightly larger quadru-
by a factor of 2, i.e., one average every 5 Myr with a 10 pole term, on the order of 10% of the axial dipole (for the
Myr time window (Table 5). This results in a mean A95 of last few million years, ranges of estimates are between 3 and
4.2 ± 1.9, i.e., similar to BC91 with a twofold increase in 8% [Schneider and Kent, 1988; McElhinny et al., 1996;
time resolution. Johnson and Constable, 1997; Carlut and Courtillot,
[27] We have plotted in Figure 4 the angular test param- 1998]). The pole from EUR is based on a combination of
eter gc – go [from McFadden and McElhinny, 1990], deter- only three studies, all being to some extent questionable.
mined for pairs of poles consisting of the calculated mean One study comes from the Aix-en-Provence series (south of
pole from each plate versus the corresponding overall mean France), in which rotations have subsequently been
(synthesized) pole for the same 20 Myr time window, as a revealed. The two other studies are from the Antrim and
function of age. Pole pairs are not significantly different at Mull lavas of the British Tertiary Igneous Province. Their
Figure 2. Plate reconstructions at (a) 20 and (b) 60 Ma in rectilinear projection, together with
transferred poles in equal-area projection (the South African plate is kept fixed in longitude). The
transferred poles and site locations are also shown as open circles and solid stars, respectively, in
the projections. Insets: polar view of the synthetic African APWP (small linked dots) with all data for the
appropriate time window (larger open circles).
ages in the database (around 70 Ma) are now known to be assigned age may easily account for the observed discrep-
too old by some 10 Myr when compared with more recent ancy. For two poles, there is a problem of rounding the age
estimates based on a combination of new radiometric ages at a time of significant change. Such is the case for the 50–
and the magnetic polarity timescale [Saunders et al., 1997]. 60 Ma Australian pole and for the 160 –170 Ma Greenland
Indeed, a large part of these lavas erupted during reversed pole. The effect of the slight differences in ages of observed
chron C26r dated at 59 ± 1 Myr. Such an error in the versus predicted poles illustrates difficulties linked to the
Figure 3. Plate reconstructions at (a) 90 and (b) 200 Ma. Same legend as Figure 2.
sparseness of the database and can lead to misleading APWP computed with the BC91 database but with the
tectonic or field-geometry interpretations. For instance the updated kinematics used in the present paper (BC91’). The
calculated 54.4 ± 5.2 Myr mean pole from AUS data only 0– 140 Ma and 120 – 200 Ma segments are shown separately
can be compared either with either the 55 or 59 Myr overall in Figure 5 to avoid too much overlap and lack of legibility. We
mean (synthesized) pole, in one case with a discrepancy, in first see that improvement in the kinematic models results in
the other with none. A similar case is also shown for GRE at insignificant changes from the Present back to 160 Ma. On the
160 –170 Ma. Note again (as in Figure 1d) that there is no other hand, the 170 (actually 175) and 180 (actually 178– 179)
sign of degradation with age. Ma poles are significantly displaced. This is essentially due to
[28] In Figure 5, we compare three synthesized APWPs in a different Gondwana fit (see section 4, APWP for Africa).
(south) African coordinates: triangles are for the original [29] Now, comparing BC91 and BC01, we see that the
BC91 curve, dots for the new BC01 curve, and stars for an increased database has resulted in a systematic 2 shift of
Table 5. Synthetic Apparent Polar Wander Path for the Past 200 Myr Calculated for a 10 Myr Sliding Window Every 5 Myra
South Africa South America India Australia
Window Age N l (N) f (E) A95 K l (N) f(E) A95 K l (N) f (E) A95 K l (N) f (E) A95 K
0 2.1 25 86.6 182.2 3.0 96.2 86.8 176.3 3.0 96.0 87.0 197.1 3.0 95.7 87.5 204.0 3.0 95.1
5 3.1 30 86.2 176.9 2.6 105.5 86.5 168.0 2.6 104.8 86.8 197.1 2.6 106.2 87.5 207.6 2.6 105.1
10 11.9 21 85.0 170.7 3.1 107.4 85.5 139.9 3.1 107.3 85.8 231.1 3.1 104.0 85.9 277.2 3.2 100.9
15 14.8 24 84.2 170.9 3.2 85.1 84.8 136.7 3.2 84.2 84.9 231.2 3.2 86.8 84.9 279.3 3.2 89.0
20 19.6 16 81.7 165.7 4.5 69.3 82.1 131.8 4.6 66.4 83.4 221.9 4.5 70.0 84.1 273.0 4.5 69.3
25 26.0 14 83.4 184.5 5.4 55.8 84.4 126.1 5.3 59.2 81.3 264.7 5.7 50.8 78.6 294.7 5.7 50.8
30 30.2 13 79.0 207.7 5.4 63.3 83.5 163.8 5.3 64.7 74.9 263.6 5.0 72.3 73.0 282.8 5.0 72.4
35 33.7 12 78.9 201.7 4.5 96.1 82.9 152.1 4.7 89.4 75.1 266.3 4.5 96.6 72.9 285.3 4.5 96.5
40 39.0 8 77.3 191.6 7.2 63.3 80.0 139.5 7.3 62.0 74.9 264.4 7.2 63.7 73.2 283.7 7.1 64.7
45 46.4 12 79.1 200.6 5.2 73.0 82.0 126.5 5.3 72.6 70.6 277.8 5.5 67.1 70.2 294.6 5.4 69.1
50 49.9 17 78.7 204.1 4.2 74.5 82.2 127.7 4.2 75.5 67.7 278.5 4.5 66.1 69.5 295.7 4.1 76.4
55 55.0 22 76.7 214.7 4.2 56.1 82.7 145.3 4.1 57.5 60.7 277.7 4.8 43.6 67.2 294.2 4.1 58.0
60 60.7 24 73.9 224.0 4.3 48.7 82.4 168.5 4.3 48.9 50.8 277.5 4.6 42.5 64.3 292.7 4.3 48.8
65 64.0 24 71.6 234.2 3.6 67.8 82.0 193.7 3.6 68.3 42.4 279.4 4.0 56.1 60.6 293.1 3.6 67.9
70 68.3 15 71.1 237.4 4.7 67.2 82.3 199.0 4.8 65.0 36.8 281.0 4.5 74.5 59.5 295.4 4.8 65.8
75 75.4 10 72.5 232.4 7.4 44.9 83.6 170.6 7.2 46.4 33.6 282.1 7.7 41.2 62.1 300.6 7.3 45.4
80 79.1 9 68.3 251.9 7.0 56.1 82.3 228.4 6.9 57.5 25.7 287.5 7.1 54.2 54.8 303.5 6.9 58.1
85 83.4 5 66.9 253.1 9.8 61.7 82.5 229.5 10.2 57.7 22.6 289.1 9.4 67.2 54.4 305.4 10.0 59.4
90 91.4 8 66.7 242.7 3.4 259.8 84.8 175.8 3.6 232.6 21.9 289.2 3.9 200.0 59.9 309.4 3.9 203.5
95 94.1 8 66.5 246.0 6.6 72.2 86.2 178.5 6.5 74.0 20.5 291.7 7.0 63.9 59.4 313.6 6.6 71.1
100 100.2 6 68.8 254.1 11.0 39.6 87.8 29.2 11.5 35.9 19.9 298.0 10.9 40.1 59.2 327.5 11.9 33.8
105 104.6 4 62.9 251.5 25.6 15.4 88.0 179.8 24.6 16.6 14.9 295.0 25.6 15.4 58.6 322.7 24.3 17.0
110 110.8 7 58.0 261.7 8.3 57.1 85.1 264.2 8.7 52.2 8.3 297.5 8.3 56.9 53.8 326.2 8.9 49.9
115 116.4 13 56.3 260.9 2.7 235.5 86.4 261.2 2.4 313.9 6.8 297.2 2.7 240.5 54.1 330.5 2.3 342.0
120 120.1 13 51.8 260.9 3.2 163.9 83.9 238.5 3.1 178.2 2.4 296.1 3.2 166.2 52.9 327.9 2.9 204.6
125 122.8 10 50.6 260.7 3.4 197.5 83.4 233.0 3.4 197.7 0.9 296.4 3.7 173.9 53.1 328.6 3.8 166.4
130 130.9 4 49.5 265.8 7.4 156.9 82.6 252.6 7.2 163.0 3 301.5 8.0 132.3 50.2 335.6 8.2 127.3
135 134.3 5 49.5 264.6 5.6 189.2 83.2 246.5 5.5 192.6 3.5 301.6 5.9 167.5 49.6 335.4 6.2 152.3
140 139.3 5 43.6 265.9 8.1 90.1 77.5 238.2 8.1 90.1 10 300.7 7.9 93.8 44.1 328.5 8.2 88.4
145 142.0 3 40.3 263.2 11.9 108.1 74.7 227.2 11.9 108.1 12.4 297.9 11.7 111.7 43.1 323.2 12.0 107.0
150 153.9 4 59.1 255.7 6.3 216.2 85.8 61.6 6.3 216.2 5.7 308.3 6.0 231.9 50.6 352.4 6.1 231.1
155 155.7 7 58.2 258.9 3.8 253.8 87.3 39.6 3.8 253.8 4.1 309.8 3.9 244.6 48.5 351.8 3.9 245.7
160 158.9 5 53.8 257.9 7.1 117.6 88.2 189.6 7.1 117.6 1.1 307.9 6.4 145.8 47.9 346.6 6.4 144.1
165 168.0 8 52.1 260.5 9.8 33.7 86.6 228.5 9.8 33.7 0.7 310.7 9.8 33.7 44.8 347.3 9.8 33.7
170 170.2 13 56.5 259.1 6.8 37.8 88.9 41.9 6.8 37.8 3.4 312.4 6.8 37.8 46.2 353.3 6.8 37.8
175 176.7 13 65.7 260.8 7.2 33.8 79.7 32.2 7.2 33.8 10.6 318.3 7.2 33.8 45.5 6.6 7.2 33.8
180 178.4 12 66.8 263.9 7.6 33.4 78.4 26.6 7.6 33.4 10.9 319.9 7.6 33.4 44.2 8.0 7.6 33.4
185 182.4 9 69.5 261.4 7.6 46.5 75.9 32.6 7.6 46.5 13.6 320.6 7.6 46.5 44.9 11.9 7.6 46.5
190 191.3 13 62.9 258.9 5.5 58.4 82.5 37.0 5.5 58.4 8.8 316.0 5.5 58.4 46.4 2.6 5.5 58.4
195 194.4 14 61.7 257.9 4.9 67.9 83.7 41.6 4.9 67.9 8.0 314.9 4.9 67.9 46.9 0.8 4.9 67.9
200 198.9 8 63.9 244.6 5.7 94.9 78.8 70.6 5.7 94.9 13.7 311.7 5.7 94.9 52.6 5.8 5.7 94.9
Window Age N l (N) f(E) A95 K l (N) f (E) A95 K l (N) f(E) A95 K l (N) f(E) A95 K
0 2.1 25 86.8 178.8 3.0 96.0 86.7 178.7 3.0 96.1 86.5 180.7 3.0 96.2 86.5 180.7 3.0 96.2
5 3.1 30 86.5 171.6 2.6 105.3 86.3 172.0 2.6 105.1 86.1 174.8 2.6 105.2 86.1 174.8 2.6 105.2
10 11.9 21 85.9 151.8 3.1 107.3 85.0 155.7 3.1 107.6 84.6 164.4 3.1 107.7 84.6 164.4 3.1 107.7
15 14.8 24 85.2 151.0 3.2 85.1 84.2 154.9 3.2 84.3 83.6 163.0 3.2 84.2 83.6 163.0 3.2 84.2
20 19.6 16 82.8 146.7 4.5 67.9 81.4 149.7 4.5 67.8 81.0 156.2 4.5 68.3 81.0 156.2 4.5 68.3
25 26.0 14 85.8 152.1 5.4 57.5 83.8 153.2 5.3 58.7 82.8 165.7 5.3 57.9 82.8 165.7 5.3 57.9
30 30.2 13 83.5 197.0 5.2 66.1 81.6 183.4 5.3 64.4 79.6 187.9 5.4 63.2 79.6 187.9 5.4 63.2
35 33.7 12 83.6 185.3 4.6 91.6 81.2 173.4 4.6 90.4 79.3 180.4 4.6 92.4 79.3 180.4 4.6 92.4
40 39.0 8 81.8 167.2 7.3 62.7 78.8 160.2 7.3 62.2 77.3 167.7 7.3 62.4 77.2 168.0 7.3 61.8
45 46.4 12 84.8 154.9 5.3 71.2 81.1 150.4 5.3 72.7 79.6 167.9 5.2 73.5 78.7 169.9 5.1 76.3
50 49.9 17 85.1 156.4 4.2 76.0 81.3 151.9 4.2 75.3 79.3 170.3 4.2 74.4 78.0 171.9 4.2 73.9
55 55.0 22 85.3 181.7 4.1 58.1 81.4 168.3 4.2 57.0 77.9 183.4 4.2 55.9 76.1 180.9 4.2 56.9
60 60.7 24 84.3 211.5 4.3 48.9 80.5 188.9 4.3 48.8 75.4 195.5 4.3 48.7 73.8 189.3 4.3 48.9
65 64.0 24 82.7 240.6 3.6 68.1 79.8 209.5 3.6 68.0 73.5 207.3 3.6 67.9 72.6 200.1 3.6 68.4
70 68.3 15 82.8 251.7 4.8 65.2 80.0 213.2 4.8 64.5 73.4 209.7 4.8 64.9 72.5 202.3 4.8 65.8
75 75.4 10 86.7 245.7 7.3 45.8 81.3 188.6 7.2 46.9 75.7 197.6 7.2 46.8 73.7 192.0 7.3 46.0
80 79.1 9 80.9 292.0 6.9 58.1 81.0 232.5 6.9 57.0 73.5 221.4 6.9 57.0 72.5 216.0 6.9 56.8
85 83.4 5 81.0 298.2 10.2 56.9 81.1 230.5 10.3 56.3 73.5 220.5 10.2 56.7 72.1 216.1 10.2 57.6
90 91.4 8 87.6 324.9 3.9 203.4 80.8 185.8 3.5 245.6 75.3 196.4 3.5 247.1 71.9 195.5 3.5 251.7
95 94.1 8 86.3 356.4 6.6 72.2 81.8 183.4 6.4 76.5 76.4 197.7 6.4 76.3 73.0 195.8 6.4 76.3
100 100.2 6 80.3 32.6 12.0 33.4 85.1 144.5 11.5 36.3 81.7 198.3 11.5 36.3 78.1 190.4 11.4 36.5
105 104.6 4 82.4 23.7 24.2 17.1 81.2 173.5 24.7 16.5 76.8 191.5 24.6 16.5 73.0 191.3 24.7 16.5
110 110.8 7 78.1 9.2 8.9 49.1 82.1 203.8 8.7 51.9 75.4 208.8 8.7 51.7 72.5 206.1 8.7 51.8
115 116.4 13 76.5 19.6 2.2 346.5 79.7 186.2 2.5 278.4 74.7 194.6 2.5 281.3 70.9 195.5 2.5 280.3
120 120.1 13 77.0 11.9 2.9 205.9 76.5 193.5 3.1 178.5 71.2 194.2 3.1 178.6 67.4 197.2 3.1 178.3
125 122.8 10 76.8 14.2 3.8 165.7 75.7 192.4 3.5 193.5 70.5 192.5 3.5 193.6 66.7 196.1 3.5 193.8
130 130.9 4 71.9 18.8 7.8 139.7 77.4 195.2 7.0 172.8 71.9 196.5 7.0 173.8 68.2 198.7 7.0 174.3
Table 5. (continued)
Window Age N l (N) f (E) A95 K l (N) f(E) A95 K l (N) f (E) A95 K l (N) f (E) A95 K
135 134.3 5 71.8 18.5 6.0 165.6 76.8 190. 5.5 194.5 71.8 192.3 5.5 195.1 67.9 195.3 5.5 195.7
140 139.3 5 70.9 358 8.0 92.3 72.6 203.1 8.1 90.0 66.7 197.2 8.1 90.0 63.2 201.6 8.1 90.0
145 142 3 72.0 346.8 11.9 109.1 68.7 201.9 11.7 111.3 63.1 193.9 11.7 111.2 59.4 199.8 11.7 111.2
150 153.9 4 63.9 41.5 6.1 231.0 73.2 134.3 6.1 227.2 75.7 141.6 6.1 226.9 70.4 155.2 6.1 226.6
155 155.7 7 62.9 37.2 3.9 245.6 74.3 137.4 3.5 296.2 76.2 146.8 3.5 295.5 70.9 159.1 3.5 295.2
160 158.9 5 65.2 31.1 6.4 144.1 71.9 149.7 7.1 117.9 72.4 154.1 7.1 117.7 67.2 164.8 7.1 117.8
165 168 8 62.8 26.5 9.8 33.7 70.6 149.5 9.7 34.0 71.2 151.9 9.7 34.0 65.9 163.2 9.7 34.0
170 170.2 13 60.6 34.8 6.8 37.8 70.1 129.9 7.7 30.4 73.5 131.5 7.7 30.4 68.4 147.4 7.7 30.4
175 176.7 13 53.2 45.1 7.2 33.8 66.0 96.4 7.4 32.4 73.4 84.5 7.4 32.4 70.3 110.4 7.4 32.4
180 178.4 12 51.6 44.6 7.6 33.4 66.0 91.7 7.6 33.4 73.8 77.9 7.6 33.4 71.1 105.2 7.6 33.4
185 182.4 9 50.1 48.5 7.6 46.5 63.7 87.7 7.6 46.5 71.7 71.2 7.6 46.5 69.6 97.9 7.6 46.5
190 191.3 13 55.9 43.1 5.5 58.4 65.9 102.5 5.5 58.4 72.8 92.8 5.5 58.4 69.2 116.8 5.5 58.4
195 194.4 14 57.2 42.3 4.9 67.9 65.8 105.8 4.9 67.9 72.5 97.2 4.9 67.9 68.6 120.2 4.9 67.9
200 198.9 8 58.0 54.4 5.7 94.9 59.4 103.5 5.7 94.9 66.4 90.5 5.7 94.9 63.0 111.9 5.7 94.9
a
Same as Table 4.
the 10 to 60 Ma poles, and 4 shift of the 80 to 110 Ma not distinguishable from zero in any individual 10 or 20
poles. The 120 and 130 Ma poles agree, but the two paths Myr time frame (except at 52 and 59 Myr), and in any case
diverge prior to 140 Ma, with a maximum difference of 14 the effect is small and on the order of contributions from
at 170 Ma. Actually, when one takes into account the A95 other sources of paleomagnetic uncertainty. With the reso-
uncertainties, only the 173– 175 and 178 Ma poles fail to lution we are able to achieve, these quadrupolar effects
intersect. Altogether, the new BC01 African APWP returns remain negligible and the geocentric axial dipole (GAD)
on itself rather than forming an open loop in the 90– 200 Ma hypothesis is clearly satisfactory on the timescales of
interval. interest to our study (10 to 200 Myr).
4.2. Checking the Geocentric Axial Dipole Hypothesis 4.3. Description of Individual Plate APWPs
[30] The comparison between the 10 Myr and 5 Myr 4.3.1. Africa
resolution APWPs is best described in the case of the plate [31] We have already illustrated the main characteristics
with the fastest polar wander rates, i.e., India. This is done of the BC01 synthetic APWP in South African coordinates
in a later section. However, before comparing the new and (using the 20 Myr sliding window) (Figures 1 and 5). The
old versions of the synthetic APWP for other plates, we path can be described in terms of a succession of track
address the question of dipolarity of the past geomagnetic segments and standstills, i.e., of episodic polar wander [e.g.,
field. As recalled above, a significant quadrupole, on the Briden, 1967; Gordon and Cox, 1980; Cox and Hart, 1986;
order of a few percent of the axial dipole, has been Irving and Irving, 1982]. Poles tend to cluster at 10– 20 Ma,
identified in the mean geomagnetic field for the last 5 90– 100 Ma, 180 – 190 Ma and a sharp hairpin occurs at
Myr [e.g., Johnson and Constable, 1997; Carlut and 140 Ma. These relative standstills and cusps separate tracks
Courtillot, 1998]. In Figure 6, we have plotted all the poles with rather fast and regular polar wander: 20– 90 Ma, 100–
from our database referred to a common site longitude. In 140 Ma, 140– 180 Ma.
each 20 Myr time window, all sites are related to a common 4.3.2. Europe and North America
longitude (taken to be 0), and all such frames are stacked [32] The synthetic APWPs in European and North Amer-
from 0 to 200 Ma. When a mean pole is computed in each ican coordinates share the same major features, related
time window, a jagged path results (Figure 7). Poles tend to through the slow opening of the North Atlantic Ocean at
follow an erratic course, yet remaining most of the time in the end of the Cretaceous (Figure 8a). A track from 10 to 50
the hemisphere opposite to the one centered on the reference Ma is preceded by a more complex path between 130– 140
meridian. Of course, the angular distance from the pole is and 50 Ma, with a general change in trend. A previously
small (on the order of 2 on average) and the uncertainties undetected loop occurs between 50 and 110 Ma. It can in a
are such that most 95% confidence intervals include the way be taken to represent a standstill, but some of the
geographical pole. When a grand 200 Myr average is features in the loop appear to be resolved by individual
computed, the mean pole lies at l = 88.6N, f = 176.5E mean poles. In particular, the 52 and 59 Ma averages are
(A95 = 1.2), i.e., on the far side of the reference meridian, statistically distinct. A problem comes from the fact that
1.4 ± 1.2 away from the geographical pole. If taken at face much of the loop overlaps in time with the Cretaceous Long
value, this would imply a far-sided effect due to a persistent Normal Superchron, during which oceanic spreading rates
quadrupole on the order of 3% (±2%) of the dipole, i.e., are assumed to be constant for lack of observable chrons.
somewhat smaller than the 5 Myr average. This translates to This introduces a decrease in the temporal resolution of the
a maximum effect of 3 on inclination at the equator, and kinematic models which are used to transfer poles.
translates into a paleolatitude error of 1.5 at most when the [33] There are distinct cusps at 140 and 180 –190 Ma, as
axial dipole hypothesis is used. We conclude that there are was the case for the African path. The overall shape of the
indications in favor of a persistent quadrupole on the 200 Eurasian APWP between, say, 50 and 150 Ma confirms the
Myr timescale, with an amplitude on the order of half the loop which was one of the new, previously ill-recognized
one obtained for the last 5 Myr. However, these values are features uncovered in BC91. The loop appears a bit smaller
Figure 4. The McFadden and McElhinny [1990] test applied to each calculated mean pole for each
plate versus the corresponding overall mean (synthetic) pole for the same 20 Myr time window as a
function of age (the vertical axis corresponds to the calculated test parameter gc – go). If negative (gray
shaded zone), pole pairs are considered distinct at the 95% probability level. Two poles (AUS near 55 Ma
and GRE near 165 Ma) fall in between rounded ages and can be assigned to either one of two synthetic
poles (at 50 and 60 Ma and 160 and 170 Ma, respectively); hence two test points linked by a straight line
segment are shown.
and may have a more complex structure than previously der Voo [1990] emphasized that a pole transfer using the
recognized. The timing of the so-called Cretaceous standstill Bullard et al. [1965] fit led to North American and European
[Westphal et al., 1986; Besse and Courtillot, 1991] appears to APWP segments in better agreement during the Paleozoic
be somewhat later: 60– 110 Ma rather than 70– 130 Ma. This and part of the Mesozoic. The problem is to understand when
loop was a significant feature used to analyze paleomagnetic and how the change in configuration between the reconstruc-
data from mobile zones and large continental blocks in Asia tion based on the oldest seafloor data [see Royer et al., 1992]
[e.g., Besse and Courtillot, 1988, 1991; Enkin et al., 1992]. on the one hand, and on the Bullard et al. [1965] fit on the
Consequences in the changes in shape of the Eurasian APWP other, was achieved. Our European database allows one to
for such analyses are not explored further in this paper. compute separate 185 and 190 Ma mean poles with three and
[34] Because most data on which the Jurassic segment of five studies, respectively. The comparison between these two
the NAM APWP could be based carried some amount of poles and a version of our master APWP from which Euro-
uncertainty (tectonic rotations for data from the SW United pean data have been removed is inconclusive at the 95%
States, remagnetization on intrusions from the NE United level, leaving this question still open, as already concluded
States), Courtillot et al. [1994] proposed to transfer their by Van der Voo [1993].
available data from other continents to generate synthetic 4.3.3. South America
NAM APWPs. They discussed the effects of data selection [35] As is well known, the SAM APWP is remarkable in
and kinematic reconstructions. They concluded that the high that it displays very little polar wander (Figure 9a). Most
latitude APWP of Irving and Irving [1982] was to a large poles have latitudes higher than 80: South America has
extent vindicated, though with much reduced uncertainties at basically remained in the same position with respect to the
the 95% confidence level. On the other hand, the lower geographical poles (i.e., spin axis) in the last 200 Myr.
latitude paleomagnetic Euler pole-based path of May and However, the path is ordered and shows small but resolv-
Butler [1986], though on the edge of the confidence intervals, able features, such as a track from 200 to 180 Ma, a change
was not supported by the synthesized path. It is interesting to in direction near 180 Ma, a track from 180 to 130– 140 Ma
compare the path for North America from the present study and little motion since (with a real standstill between 50 Ma
with the earlier attempts by Courtillot et al. [1994]. Basically, and the Present).
the new path (Figure 8b and Table 4) vindicates the con- 4.3.4. Australia, India, and Antarctica
clusions based on BC91. The 130 –200 Ma segment of the [36] It is particularly informative to draw the three
NAM APWP remains at rather constant latitudes on the order APWPs on Figure 9b. In BC91, the Indian APWP was
of 70– 75N, with the mean A95 reduced from 7 to 5. shown only back to 120 Ma, and only two paleomagnetic
Reliable poles are now obtained at 140 and 170 Ma, due to poles for that plate (Deccan and Rajmahal Traps) were
the increased number of data, which were not sufficient in available; Australia and Antarctica had not been included
BC91. The main difference lies with the 200 Ma pole. Van in our previous analysis.
Figure 5. Comparison of three synthesized APWPs in South African coordinates with their associated
95% ellipses of confidence: triangles are for BC91 [Besse and Courtillot, 1991], dots for BC01 (this
paper), and stars for BC91’ (BC91 data transferred using the kinematics used in the present paper). (a)
From Present to 140 Ma and (b) from 120 to 200 Ma. The ellipses of confidence for the BC91’ path are
not figured for more clarity. In Figure 5b, the age numbers in italic refers to the BC91’path. Ages are in
Ma. Equal-area projection.
[37] The Indian path is the most elongated one, corre- poles (and for that reason, not used in constructing the
sponding to the fastest average apparent polar wander veloc- synthetic path) and very accurately dated falls only 1 away
ity. Three phases (tracks) are clearly recognized from 180 to from the 67 Ma synthetic pole (Table 4), attesting to the
130, 130 to 80– 70 and 80– 70 Ma to the Present. These are validity of the data selection process, to the quality of the
separated by a standstill (i.e., poles are in the joint intersection selected paleomagnetic data and plate kinematic parameters.
of their 95% confidence intervals) at 200– 170 Ma, another [38] Because of its fast polar wander rate, the synthetic
one at 140 –110 Ma and a directional change at 80– 70 Ma. APWP in Indian plate coordinates is the one for which the
The classical rotation and northward motion of India are influence of improving time resolution from 20 Myr to 10
clearly outlined; the fast velocity (150 km/Myr) peaking Myr moving averages is best displayed. This is shown in
between 70 and 50 Ma, and the slowdown after the India- Figure 10, where only the 95% confidence circles for the
Asia collision at 50 Ma are also seen. The 66 Ma Deccan 10 Myr APWP have been included for clarity. The mean
Traps pole [Vandamme et al., 1991], which is based on a very poles are labeled with the mean age of the data they are
large number of data already used in our Indian selection of actually based on (Tables 4 and 5). The excellent agree-
Ocean ridge. After that time, Australia and India belong to

the same plate.
[41] The Antarctic APWP (Figure 9b) again displays
similar features prior to 70 Ma, with two standstills and
two tracks, but since then, apparent polar wander is almost
negligible, much like South America. The shapes of the
three APWPs shown in Figure 9b are easily related one to
the other as a function of the opening of ocean basins
between plate pairs. The relative orientations of the relative
velocity vectors and the respective rates of polar wander and
plate motions combine in such a way that the respective
influences of and relationships between the various param-
eters are readily apparent.
4.3.5. All Paths
[42] The succession of tracks, standstills, loops, and
directional changes for all APWPs shows similarities and
differences between the various continents. All paths begin
with slow motion or standstill between 200 and 170– 180
Ma and then turn into tracks which last until 140 Ma. A
major re-organization occurs at 140 Ma, with AFR, EUR,
NAM and SAM changing track direction, and IND, AUS,
ANT entering a standstill. IND, AUS, ANT and SAM
undergo a change at 110 Ma, but the other three plates do
Figure 6. All poles from our database referred to a not. AFR, EUR, and NAM have a track ending near 100
common site longitude. Open star and 95% confidence Ma, and enter a standstill or loop. A 80 Myr change is
circle: grand 200 Ma average. Inclination alone allows only common to IND, AUS and ANT; one at 40– 50 Ma to AFR,
the determination of small circles of equal paleo-colatitude EUR and NAM. Altogether, the 180– 140 Ma track is the
on which the pole must lie. main feature which stands out to be common to all plates.
[43] A common succession of a fast (220 – 200 Ma, not
documented in this paper) phase followed by a general
ment between the two curves is obvious. Of course, some
features are more readily apparent in the higher resolution
curve, such as the slowdown at 55 Ma, compatible with an
early collision age for India versus Asia [e.g., Patriat and
Achache, 1984; Jaeger et al., 1989], or the sharper reorien-
tation in the track at 75– 80 Ma. The 10 Myr resolution pole
at 100 Ma (nominal age 100.2 ± 4.1 Ma, with 6 data) and the
corresponding 20 Ma resolution pole (nominal age 97.6 ± 5.8
Ma, with 12 data) are both well within the intersection of
their respective 95% confidence circles (10.9 and 7.3). The
small loop in the 10 Myr APWP based on four mean poles
between 130 and 142 Ma is not resolved given A95 between
6 and 12; all are within less than 5 from the 20 Ma
resolution pole at 140 Ma (nominal age 136.8 ± 5.4 Ma)
which has an A95 of 6.4. At the older end of the curves, the
10 Myr resolution poles at 194 and 199 Ma and the 20 Myr
resolution pole at 196 Ma are fully compatible.
[39] In conclusion, the 20 Myr resolution curve captures
all the essential features which can be found in the 10 Myr
curve, and it is not expected that limiting our discussion to
the former would run the risk of having missed any
significant feature. The high-resolution curve also confirms
the validity of the rough classification of APW periods into
tracks and standstills.
[40] The Australian APWP (Figure 9b) displays almost
the same track times and orientations, standstills and track Figure 7. Mean poles computed in successive 20 Myr
changes as that for India until 70 Ma. Motion since 70 Ma is time windows, every 10 Myr, using the common site
similar to that indicated by India (i.e., mostly northward longitude poles of Figure 6. Full radial scale is 10 only.
motion follows after counterclockwise rotation) except it is Notice the small angular distance of each mean from the
three times slower between 80 and 50 Ma, and then pole (on the order of 2 on average); 95% confidence circles
accelerates a bit from 40 Ma onwards, in relation with the are not shown, for clarity; all but two include the
southward jump of the eastern branch of the South Indian geographical pole.
assumption of hot spot fixity. This is a long debated subject

[e.g., Molnar and Atwater, 1973; Molnar and Stock, 1987;
Tarduno and Gee, 1995; Steinberger and O’Connell, 1998;
Di Venere and Kent, 1999; Koppers et al., 2001]. In their
analysis, Müller et al. [1993] find that motions between
individual hot spots in the Atlantic and Indian Oceans are
less than a few km/Myr, i.e., close to an order of magnitude
less than typical plate motion. These Indo-Atlantic hot spots
therefore do form a relatively coherent frame for that vast
part of the Earth’s lithosphere (and underlying mantle)
ranging from North America to India.
[46] The major improvement of the present analysis over
that in BC91, where Morgan’s [1983] hot spot tracks on the
African plate were used, is the use of the revised model of
Müller et al. [1993], at least for the last 130 Myr. This
model combines a number of advantages such as inclusion
of hot spot tracks from AFR, NAM, SAM, IND, and AUS
and the same updated plate motions by Royer et al. [1992]
that we used in transferring the paleomagnetic data from
one plate to the next in order to build the synthetic APWPs.
Therefore, the kinematics and chronologies are compatible,
potentially reducing significantly sources of errors that
could have affected the BC91 analysis. Unfortunately, only
two hot spot tracks are available (Tristan da Cunha and
Meteor/New England) between 90 and 130 Ma, implying
potentially larger uncertainties for that period.
Figure 8. (a) Europe and (b) North American synthetic
APWPs (data in Table 4). Each pole corresponds to a mean
pole computed by using a 20 Myr sliding window. Equal-
area projection. Ages are in Ma; the values shown are the
actual mean ages derived from the data in the appropriate
window.
standstill of Pangea at 200 –180 Ma is followed by the

breakup of Gondwana and the opening of the central
Atlantic Ocean. An animated sequence derived from the
APWP (www.ipgp.jussieu.fr/besse/tpw_ jgr01) shows a
sequence of general rotations around 160– 140 Ma. The
next phase of breakup of Gondwana results in the 110– 100
Ma events. The directional change of IND and AUS at 80
Ma does not show as a major event.
5. True Polar Wander

[44] As was previously done by Livermore et al. [1984],
Andrews [1985], ourselves (BC91), and recently Prévot et al.
[2000], we have attempted to combine apparent polar
wander based on paleomagnetic data (paleomagnetic refer-
ence frame) with motions of hot spots with respect to
lithospheric plates (hot spot reference frame) in order to
derive true polar wander (TPW). Given the relative motions
between the plates, the position of these plates with respect
to the rotation axis (provided by paleomagnetic results under
the GAD assumption) and the relative motions between the
plates and a reference frame attached to the mantle (if such a
frame can be found), TPW is defined as the motion of the
mantle with respect to the rotation axis. The latter is defined
by hot spots, which are assumed to form an array of fixed
points that provides the mantle reference frame. Figure 9. Same as Figure 8 for the (a) South American
[45] A number of the assumptions made to determine and (b) Indian, Australian, and Antarctic synthetic APWPs
TPW may of course be found erroneous, such as the (see Table 4).
Figure 10. Comparison of two synthetic Indian APWPs. Triangles correspond to mean poles computed
every 10 Myr using a 20 Myr sliding window (Table 4); dots correspond to mean poles computed every 5
Myr using a 10 Myr sliding window, and their associated 95% confidence ellipses are shown (Table 5).
The dashed ellipse is the confidence interval of the ill defined 104.6 Ma pole. Ages are in Ma. Equal-area
projection.
[47] The resulting TPW curve is shown in Figure 11, with resolution), the new BC01 curve of Figure 11a, and that
95% confidence circles, for the best (10 Myr) time reso- obtained by Prévot et al. [2000].
lution available (Figure 11a) and for the 20 Myr time [48] The first 11 points of BC01 in Figure 11a, corre-
resolution (Figure 11b) (also see Table 6). Points are shown sponding to the period 0 – 55 Ma are all in the same
every 5 Myr (10 Myr) and every other point is statistically quadrant, between 4 and 9 away from the present rotation
independent. Besse and Courtillot [1991] compared BC91 pole, with 95% uncertainties ranging from 2 to 7. They are
TPW to two previous determinations by Livermore et al. not statistically distinct from each other and therefore could
[1984] and Andrews [1985], and we refer the reader to their correspond to a standstill. A mean position can be calcu-
Figure 3. In Figure 12, we compare BC91 (20 Myr lated from all (102) data points in that time window: it is
the 116 Ma pole. The 95% confidence intervals of the 100

and 110 Ma poles are too large to outline with certainty the
details of the track which is followed. The track extends to
142 Ma, when it turns back on itself. The path back to
200 Ma is again rather jagged although there is a tendency
to drift away from the current pole. When 20 Myr time
windows are used (Figure 11b), the path becomes somewhat
smoother and a succession of a standstill at 160– 130 Ma, a
circular track from 130 to 70 Ma, and a standstill at 50– 10
Ma follow each other. The true polar wander rate between
130 and 70 Ma averages 30 km/Myr. The BC91 TPW curve
was somewhat smoother, with a large track between 50 and
120 Ma, in part because it was based on 20 Myr averages.
However, differences between the two are of second order,
given the uncertainties. TPW amplitude was smaller in
BC91, with the same general features (tracks and sandstill)
from Present back to 110 Ma, with maximum differences of
order 3 – 4 in position between 70 and 90 Ma. On the other
hand BC91 had a 120– 150 Ma loop (or cluster) up to 8
away from the new BC01 path.
[49] It is interesting to compare the new BC01 TPW
curve with that recently derived by Prévot et al. [2000]. The
main difference between the two studies lies in the fact that
Table 6a. True Polar Wander Path for 10 Myr Sliding Windowa
Age, Ma N l(N) f(E) A95 K
3.1 30 86.3 172.2 2.6 105.2
11.9 21 85.5 153.7 3.1 107.3
14.8 24 84.9 151 3.2 85.1
19.6 16 82.5 143.7 4.6 67.1
26 14 85.4 143.9 5.3 58.5
30.2 13 83.6 185.3 5.2 65.9
33.7 12 83.5 172.7 4.6 90.4
39 8 81.4 156.1 7.2 63
46.4 12 84.3 136.7 5.4 70.1
49.9 17 84.7 134.2 4.1 76.5
55 22 85.8 154.9 4.1 58.2
60.7 24 86 191.2 4.3 49
64 24 85.1 232 3.6 68.7
68.3 15 85.6 245.7 4.9 63.4
75.4 10 89.4 154.6 7.3 46.1
Figure 11. True polar wander paths (TPWP) deduced from 79.1 9 83.1 300.3 6.9 57.6
the hot spot model of Müller et al. [1993] going from the 83.4 5 82.1 303.7 9.7 63.3
Present back to 130 Ma, and that of Morgan [1983] going 91.4 8 85.8 303.7 3.5 247.6
from 130 back to 200 Ma, with associated 95% confidence 94.1 8 84.6 313 6.6 72.4
100.2 6 81.6 342.4 11.1 38.4
ellipses (shaded light gray). (a) For 10 Myr sliding window. 104.6 4 81.9 306 25.2 15.8
The dashed ellipse is the confidence interval of the 104.6 Ma 110.8 7 75.5 301.7 8.3 56.6
pole; (b) for 20 Myr sliding window. Ages are in Ma and 116.4 13 75 296.7 2.7 248.1
correspond to the actual mean age of the data in the 120.1 13 72 286.6 3.2 172.3
122.8 10 71.4 284.6 3.5 194.5
corresponding window. Data from Table 6. 130.9 4 69.1 292.4 7.6 148.9
134.3 5 69.3 292.6 5.7 180
139.3 5 63.7 286.2 8.1 89.5
142 3 61.9 279.1 12.3 100.9
found to lie at 85.1N, 153.3E (A95 = 1.5), significantly 153.9 4 74.1 310.7 6.6 196.1
displaced from the pole. The youngest pole at 5 Ma (mean 155.7 7 71.8 309.2 4.1 214.3
158.9 5 69 296.2 7.7 100.5
age 3.0 ± 2.4 Ma) is at 86.4N, 166.8E (A95 = 2.5), also 168 8 63.7 293.8 9.8 33.7
significantly displaced from the pole but identical to the 170.2 13 66.4 299.5 6.6 40.5
55 Ma overall mean. It seems that TPW may have been 176.7 13 68.7 318.6 7.1 34.9
negligible for an extended period of about 50 Myr, but 178.4 12 67.5 321.6 7.6 33.4
182.4 9 68.4 324.5 7.6 46.5
accelerated a few Myr ago, with a velocity on the order of 191.3 13 62.4 305.4 5.7 53.6
100 km/Myr. Actually, the TPW path zigzags around the 194.4 14 60.2 301.8 4.6 74.8
pole with no clear track as early as 75 Ma. There is another 198.9 8 63.5 296 5.6 100.4
clustering of data points from 79 to 104 Ma, then the pole a
Apparent polar wander path for global (Indo-Atlantic) hot spots for the
jumps to a different location, which is well constrained by last 200 Myr.
Table 6b. True Polar Wander Path for 20 Myr Sliding Windowa samples per site (versus our 6), the rest being similar to
Age, Ma N l(N) f(E) A95 K ours. Also they rejected data with potentially insufficient
averaging of secular variation (K > 100). Their final data-
3.1 30 86.3 172.2 2.6 105.2
8.3 54 85.7 161 2 95.5 base contains 118 poles, i.e., less than half of ours. The
18.9 38 85 149.5 2.7 76.4 TPW they obtain is in general similar to ours (Figure 12).
29.5 23 84.7 151.3 3.8 66.4 However, their 95% uncertainties (not drawn in their figure)
40 24 84.1 157 3.4 79.8 are very large because of the small number of data points in
52.2 31 85 151.7 3.4 60
59.7 45 86.4 196.3 2.8 58.8
most time windows. Like us, they find a succession of
67.3 34 86.4 229.4 3.2 60 episodes, with a long and faster track separating two stand-
77.9 14 87.1 291 6 45.4 still periods. The most intriguing result they obtain is a
90 13 83.7 308.6 4.9 73.6 phase of very fast wander culminating around 115 Ma,
97.6 12 83.7 309.5 6.9 40.7 when 20 of polar wander take place at velocities in excess
113.6 17 76.7 298 4.5 64.6
119.1 20 73.6 291.8 2.6 164 of 500 km/Myr, i.e., more than ten times faster than the
126.4 14 70.5 287 2.9 183.1 mean rate we obtain on the track from 130 to 60 Ma. This
136.8 7 66.2 286.3 6.3 93.1 result is based on two sets of individual poles dated around
151.6 10 69.4 297.5 5.6 74.7 100 and 120 Ma, and more specifically to the recordings
162.3 15 67.7 300.3 5.5 49.2
173.4 21 67.3 308.7 5.8 31.4 around 115 Ma provided by South African kimberlites (114
178.8 18 68.5 317.6 5.3 43.1 to 118 Ma). It so happens that our database provides a
189.7 23 63.8 309.1 4.5 46.1 sequence of rather well determined poles, particularly at
196.7 19 61.4 298.5 4 71.8 90 Ma (A95 = 4.9; N = 13 poles) and 119 Ma (A95 = 2.7;
a
Apparent polar wander path for global (Indo-Atlantic) hot spots for the N = 20 poles). The angular difference is 10.0 ± 5.5 and the
last 200 Myr. corresponding polar wander rate in this 28 Myr period is on
Prévot et al. use only paleomagnetic data from magmatic the order of 40 km/Myr. We can attempt the same compar-
rocks and restrict themselves to basalts and andesites, or ison between the 119 Ma pole and a 114 Ma pole (A95 =
intrusive rocks with a similar composition. Their selection 4.2): the angular difference is 3.2 ± 6.5 and is not
criteria were at least 10 sites (versus our 6) with at least 5 resolvable. We therefore find no support for the episode
evo
>
'
Figure 12. Comparison between three TPW Paths (95% confidence intervals not shown for clarity):
dots, BC01 (this paper); diamonds and dashed line, Prévot et al. [2000]; and triangles, BC91 [Besse and
Courtillot, 1991]. Crosses are at 10 intervals (see also Figure 11 and Table 6b).
of super fast TPW suggested by Prévot et al. [2000]. As

stated by Hargraves [1989, pp. 1851, 1852] in his analysis
of the dating of the 114 and 118 Ma kimberlites in South
Africa, ‘‘it should be pointed out that accurate radiometric
dating of these bodies is difficult’’ and ‘‘inconsistencies
between the sequence of apparent pole positions and the
available radiometric ages may in part be due to uncer-
tainties in the age.’’ The 114 Ma pole belongs to a 118–
142 Ma petrochemical group, and the 114 Ma pole
belongs to a distinct 83– 114 Ma group. According to
Hargraves [1989] ‘‘There is no clear pattern of migration
with age’’ and ‘‘if it is argued that identical poles mean
identical emplacement and magnetization age, then. . .the
detailed validity of some of the radiometric ages must be
suspect.’’ The remarkable hypothesis of Prévot et al. is
therefore hardly supported by such uncertain data.
Actually, Prévot et al. [2000] average their data in four
subsets (0 –80, 80– 110, 110– 140, and 140 – 200 Ma) in
order to reduce uncertainties. At that level, our two
determinations are in good agreement, but evidence for
any fine scale behavior or superfast episodes is lost.
6. Discussion
Figure 13. Comparison between the ‘‘Indo-Atlantic’’
[50] The 10 and 20 Myr resolution TPW curves of Figure (open dots and dashed line; this paper) and ‘‘Pacific’’
11 provide our current best estimates of true polar wander (squares and solid line) TPW Paths. The Pacific TPW is
over the last 200 Myr. We confirm earlier findings [e.g., based on the Pacific APWP of Petronotis and Gordon [1999]
Andrews, 1985; Besse and Courtillot, 1991] that true polar and the plate/hot spot kinematic model of Engebretson et al.
wander appears to be episodic in nature, with periods of [1985]. The arrow shows the mean polar motion between
(quasi) standstill alternating with periods of faster TPW. The 1900 and 1990 [after Hulot et al., 1996].
typical duration of these standstill periods in on the order of a
few tens of millions of years (50 Myr). Typical polar wander
rates during fast tracks are on the order of 30 km/Myr. Also, that sense, the TPW estimate derived in this paper could be
because of all the uncertainties in models of hot spot termed an ‘‘Indo-Atlantic’’ TPW. The ‘‘Pacific’’ and ‘‘Indo-
kinematics prior to 130 Ma (and even possibly prior to Atlantic’’ TPW curves are compared in Figure 13. This
90 Ma), we feel it is not safe to place too much weight on comparison is interesting because the data sets they are
behavior prior to that time. The major event since then is based on are entirely different and independent. Despite
therefore the end of the 130 to 60 Ma period of relatively some significant differences to which we return shortly, it is
fast polar wander, with a standstill (i.e., no or little TPW) worth emphasizing that the two curves are similar in shape
from 50 Ma (actually because of larger 95% confidence (tracks, amplitudes, azimuths), particularly the 300 longi-
circles, possibly 80– 50 Ma) to 10 Ma. However, uncertain, tude trending track from 130 to 70 Ma, though the two are
evidence for the fact that Earth emerged from that standstill offset (in the same general direction) by about 7. More
to enter a new period of fast polar wander in a different precisely, the confidence intervals intersect (though points
direction 10 Myr ago (3 Ma at the higher resolution; see are not in the intersection) near 125, 90, 60, 40 and 30 Ma.
previous section) is particularly interesting. That period The main differences occur near 80 and 70 Ma: the 82 and
would then still be going on. Actually, the inferred rate 65– 72 poles derived from Petronotis and Gordon [1999],
and azimuth of this recent phase of accelerated TPW are and our poles at 77 and 67 Ma (which have moderate
compatible with the historical values based on direct meas- uncertainties), are clearly distinct.
urements: between 1900 and 1990, the axis of rotation has [52] We emphasize that there are ongoing debates on the
been moving at a rate of 135 km/Myr toward eastern Canada validity of the data used by Petronotis and Gordon [1999]
(281E; arrow on Figure 13 from Hulot et al. [1996]). to construct their Pacific APWP, and also on the question of
[51] A legitimate concern regarding the above conclu- fixity of Pacific hot spots with respect to each other and to
sions is due to the fact that the analysis is not truly global, in the Indo-Atlantic hot spots. For instance, Tarduno and
that it fails to encompass the Pacific plate. Petronotis and Cottrell [1997] have determined the paleolatitude (based
Gordon [1999] have compiled an APW path (nine poles on inclination-only data from cores) of the 81 Ma old
from 125 to 26 Ma) for the Pacific plate, with four poles Detroit Seamount, which is part of the Emperor chain, not
based on the analysis of skewness of ocean crust magnetic far from its northern termination in the Kuril Trench. The
anomalies, three on seamount magnetic anomaly modeling paleolatitude (36.2 + 6.9/7.2) is distinct from that based
[Sager and Pringle, 1988] and two unspecified. Using the on the 81 Ma pole of the Pacific APWP [Sager and Pringle,
Pacific plate versus hot spot kinematic model of Engebret- 1988; Gordon, 1983] which is on the order of 20N.
son et al. [1985], we have determined a corresponding 125 Tarduno and Cottrell [1997] exclude the possibilities of
to 26 Ma ‘‘Pacific hot spot only’’ TPW curve (Figure 13). In inadequate sampling of secular variation, bias due to unre-
moved overprints or off-vertical drilling. They point out the (but not necessarily much more) of (the major) Pacific hot
uncertainties encountered when building an APWP for an spots are a distinct possibility.
oceanic plate, such as the Pacific, solely from inversions of [55] Di Venere and Kent [1999] have addressed the
magnetic surveys over seamounts and/or analysis of skew- problem of relative motion between the Pacific and Indo-
ness of marine magnetic profiles. Petronotis and Gordon Atlantic groups of hot spots. Though they demonstrate that
[1999] evaluated the quality of their own skewness data, some motion must have taken place between West and East
which they rank from A (best) to D (worst). The uncertainty Antarctica, based on geological observation and paleomag-
in the 73Ma mean pole based on A quality data is 3 to 4 netic data, they conclude that this motion cannot account for
times larger than that of the mean based on all data (which more than 20%, and possibly as little as 4% of the 14.5
they prefer). Yet the A quality mean pole is compatible, due offset between the observed and predicted positions of the
to its large uncertainty, with the coeval 72 Ma pole of Sager 65 Ma Suiko seamount on the Emperor continuation of the
and Pringle [1988]. It is therefore not clear that the con- Hawaiian hot spot track. They also discuss the integrity of
clusion of Petronotis and Gordon [1999] (namely, that the the Pacific plate and the role of missing plate boundaries
two are significantly different, when all data from A to D and errors in kinematic plate circuits, and find that they play
quality are used) can be accepted. a small role. DiVenere and Kent conclude that most of the
[53] Following several authors [e.g., Parker et al., 1987; apparent motion between the two main groups of hot
Parker, 1991], Di Venere and Kent [1999] argue that the spots is real, with an average drift of about 20 km/Myr
reliability of the Pacific paleopoles based on either model- since 65 Ma.
ing of seamount magnetic anomalies or determination of [56] Despite criticism on the quality of skewness data and
skewness of marine magnetic anomalies should be consid- seamount poles, which have strong bases, and increasing
ered suspect. They recall that both are prone to numerous evidence that there can have been 10 km/Myr motion
biases and could yield errors in excess of 10 in the position between major individual hot spots or hot spot groups, we
of the mean poles derived from them. find the first-order agreement between the Petronotis and
[54] Let us now review briefly the question of hot spot Gordon [1999] Pacific and our Indo-Atlantic TPW curves
fixity. Norton [1995] has suggested that the famous 43 Ma shown in Figure 13 quite remarkable. We therefore propose
Hawaiian bend was actually a ‘‘nonevent,’’ i.e., indicated a that our BC01 curve can be considered as a good first-order
change in motion of the Hawaiian hot spot with respect to estimate of global TPW and that the frame of reference
the mantle rather than a change in Pacific plate motion. based on the surface traces of the major hot spots deforms
Koppers et al. [2001] have tested the fixed hot spot hypoth- only slowly, slower than plates and plate boundaries move
esis for Pacific seamount trails. They use seamount locations with respect to each other. In that sense, these hot spots can
to first determine stage Euler poles, which they then test be used as a frame of reference for the underlying mantle,
against observed 40Ar/39Ar age progressions. The stages regardless of their dynamics and depth of origin. Hence
tested are 0 – 43, 43– 80 and 80– 100 Ma. Koppers et al. find TPW appears to be a truly global phenomenon, with tracks,
that the 0 – 43 Ma Hawaiian and Foundation seamount trail cusps, standstills, and more generally amplitudes and azi-
pair is the only one compatible with the fixed hot spot muths which are now reasonably well determined. Cottrell
hypothesis. The 0– 43 Ma Louisville/Hawaii, 43– 80 Ma and Tarduno’s [2000] proposal that TPW has not exceeded
Emperor/Louisville and 80 – 100 Ma Magellan/Musician 5 for the last 130 Myr is not vindicated.
trail pairs would all require relative motions of order (or in [57] We now turn to evidence for episodes of fast to
excess of) 10 km/Myr. The 43– 80 Ma Emperor/Line pair superfast TPW, which have been proposed by a number of
shows particularly large discrepancies, requiring motions of authors. With due caution and suggestions of alternate
at least 30 km/Myr. However, we may note that the notion explanations, Petronotis and Gordon [1999] see possible
of stage poles does require minimum knowledge on the ages fast polar wander at 80– 70 Ma. Sager and Koppers [2000]
of the seamount trails being fitted, and therefore the deter- reexamined the 130– 40 Ma segment of the Pacific APWP
mination of a geometrical stage pole is not strictly inde- based on poles derived from magnetic anomaly modeling of
40
pendent from age information. Koppers et al. also publish, Ar/39Ar dated seamounts (and only seamounts, with no
but do not discuss, stage poles going back to 140 Ma. skewness data) and concluded that a rapid TPW episode (300
Altogether, four out of these six stage poles correspond to to 1100 km/Myr) occurred in 2 to 5 Myr at about 84 Ma. The
durations of only 10 to 20 Myr and the notion of a stage, back and forth motion on Sager and Koppers’ Pacific APWP
with regard to dating uncertainties, becomes fuzzy. The between 73 and 117 Ma, and even more so on their TPW
authors emphasize that uncertainties in age progression may curve, are uncomfortable features that could well be linked
not entirely reflect true geological uncertainties, such as to problems in data significance and statistical robustness.
prolonged volcanic activity, rejuvenation by younger hot The 84 Ma motion, between the remote yet quasi-coeval
spots and uncertainties in location of hot spot (e.g., volcano poles 84W and 84E, in entirely canceled by previous motion
hysteresis or hopping, control by lithospheric faults and from 93 to 84(W) Ma, and subsequent motion from 84(E) to
structures). On the basis of careful analysis of 14 Pacific 73 Ma (i.e., the 73 and 93 Ma points are not far from each
seamount tracks, using updated age determinations and other). Taken at face value, Sager and Koppers 84W and
bathymetry, Clouard and Bonneville [2001] show that these 84E poles would lead to TPW velocities in excess of 500
Pacific seamounts are created by different processes, most km/Myr, whereas the 93 and 73 Ma poles lead to values on
being short-lived and certainly not related to deep-mantle the order of 50 km/Myr. Very recently, Cottrell and Tarduno
phenomena. Only the Hawaii and Louisville chains qualify [2000] have re-analyzed the data on which Sager and
as long-lived hot spots that can be robustly tested for fixity. Koppers base their 84 Ma episode of rapid TPW. In addition
We conclude that inter-hot spot motions of order 10 km/Myr to pointing out modeling or magnetization uncertainties in
the seamount data, they test the TPW episode against the [58] We noted above the suggestion by Prévot et al.
reference paleomagnetic record from the Umbrian Appe- [2000] of the existence of an episode of superfast polar
nines (Italy) and conclude that the episode must be an wander (500 km/Myr) at 114– 118 Ma but concluded it was
artifact of spurious seamount data. likely an artifact. Note that Sager and Koppers [2000] find
no evidence for the Prévot et al. [2000] event, and vice
versa. Tarduno and Smirnov [2001] argue, in a comment on
the Prévot et al. [2000] analysis, that reliable, selected
paleomagnetic data from mid-Cretaceous (90 – 125 Ma)
granites from North America disagree with either paleola-
titude predictions based on the hot spot reference frame or
with the superfast (stepwise) TPW event of Prévot et al.
[2000]. As correctly identified by Camps et al. [2001] in
their reply, Tarduno and Smirnov [2001] failed to indicate
that the three predicted ‘‘paleolatitudes’’ for North America
are not independent. In Figure 14, we have plotted the
‘‘latitudes’’ predicted over the last 130 Myr (i.e., the time
with better ocean kinematics and hot spot data) for test
points respectively in North America, South Africa and
India. Three latitudes are shown: ‘‘APW’’ is simply the
paleomagnetic latitude predicted by our synthetic APWP,
i.e., by the world paleomagnetic data set, when plate kine-
matics are integrated. We have seen that this was compatible
with the very data originating from each respective plate or
continent. Under the assumption of a GAD field, this is
simply the geographical latitude. ‘‘HS’’ is the ‘‘latitude’’
predicted in a reference frame where the (Indo-Atlantic) hot
spots would have remained fixed with respect to the rotation
axis (i.e., no TPW). ‘‘TPW’’ is the latitude predicted by the
TPW curve, i.e., motion of the point, as if it were attached to
the hot spot reference frame, with respect to the rotation
axis. The three latitudes (of which only the first, APW,
corresponds to the true geographical latitude) are respec-
tively linked to three rotations: APW is linked to 1, the
rotation of the plate with respect to the Earth’s rotation axis,
HS is linked to 2, the rotation of the plate with respect to
the hot spot reference frame, and TPW is linked to 3, the
rotation of the hot spot reference frame with respect to the
Earth’s rotation axis. The three motions are linked through
the simple equation (1): 1 = 2* 3. Therefore, any of the
three latitudes can be derived from the two others, and they
are by construction mutually consistent. In practice, we use
APW and HS to deduce TPW. Figure 14a confirms that the
three latitude estimates happen to be similar, and rather
Figure 14. (opposite) ‘‘Relative’’ latitudes calculated for

three test points: (a) North America presently at 34.5N,
75W; (b) South Africa at 20S, 20E; and (c) India at 40N,
75E. In each case, three "latitudes" are given from Present
back to 130 Ma: APW is the paleolatitude predicted from the
synthetic APWP derived for that plate in the present paper;
HS is the paleolatitude in the (Indo-Atlantic) hot spot
reference frame, i.e., assuming no motion of the hot spots
with respect to the rotation axis [Müller et al., 1993]; TPW is
the ‘‘paleolatitude,’’ had the site remained fixed with respect
to the hot spot frame of reference, while that frame moved
with respect to the rotation axis, following our global TPW
curve. The three estimates are linked since the first rotation
(APW) is equal to the combination of the two others (HS and
TPW; see text). 20 Myr averaging windows are used, and
points are shown at their actual mean ages, with 2s
uncertainties derived from the 95% confidence intervals.
constant back to 80 Ma in North America. From 90 back to different Earth models. Interestingly, isoviscous mantle
130 Ma, the three estimates diverge. The observed paleo- models predict TPW rates much larger than observed, and
latitude (APW) remains constant, as has been long noted a significant viscosity increase in the lower mantle is
from NAM paleomagnetic data (and as is again emphasized required to stabilize the large scale pattern of convection
by Tarduno and Smirnov [2001]). However, the HS and and bring TPW rates closer to observed values. Richards
TPW estimates smoothly diverge in an opposite sense by et al. [1999] find occasional inertial interchanges of polar
roughly equal amounts. It so happens that they compensate axes with a duration of 20 Myr, due to avalanching in the
each other in equation (1). Therefore, contrary to what is lower mantle (though only one such event occurs in a 600
stated by Tarduno and Smirnov [2001], but in agreement Myr numerical run).
with Camps et al. [2001], there is no disagreement between [62] None of these early models actually feature litho-
the three curves. We also find no evidence, as stated above, spheric plates. The more recent study of M. Greff-Lefftz
for the sudden jerk in TPW near 115 Ma, which would need and P. Bunge (personal communication, 2000) and Greff-
to be compensated by an opposite, fortuitously coeval jerk Lefftz [2001] explores the respective and cumulative effects
in hot spot latitude. We note in passing that this agreement of lower mantle viscosity, upper to lower mantle phase
of the synthetic path with paleomagnetic data derived from transitions and heat flux from the core on TPW estimates,
granites somewhat alleviates worries on the quality of these combining 3-D spherical mantle circulation models with
data, as argued by Tarduno and Smirnov [2001]. solutions to the equations of conservation of angular
[59] Figures 14b and 14c display similar results for points momentum. M. Greff-Lefftz and P. Bunge (personal com-
on the African and Indian plates, in order to show that the munication, 2000) and Greff-Lefftz [2001] confirm that
respective behaviors of the three latitude estimates of course isoviscous mantle convection models predict TPW rates
depend on the distance and azimuth under which the APW going from 100 km/Myr (the maximum acceptable value
and TPW paths are seen. In the Indian case for instance, according to our study, applying to the current period) to
TPW results in very little change in latitude, at least back to 1000 km/Myr and more (i.e., unacceptable values for us, as
100 Ma, and the APW and HS latitudes are virtually long as the elusive superfast episodes are not confirmed).
identical. This was shown from paleomagnetic data of the Greff-Lefftz and Bunge find that phase transitions have
Reunion hot spot trace on the Indian and African plates little effect, and that combination of very high lower mantle
during ODP leg 115 [Schneider and Kent, 1990a; Van- viscosity (100 times that of the upper mantle, identical to
damme and Courtillot, 1990]. For the 100 –130 Ma period, the average they use for the lithosphere) with 12% bottom
the HS latitude of India becomes more northerly than the heating from the core [Davies, 1988] results in calculated
APW latitude, whereas the reverse holds for South Africa. TPW closest to that which we observe. ‘‘Because the
Such curves can be calculated for any point on any plate rotation axis can only change as fast at the Earth’s rotational
used in our database. They allow quick comparison of bulge relaxes by means of viscous flow,’’ inertial inter-
actual paleomagnetic data, predictions from the master change almost never happens. However, M. Greff-Lefftz
synthetic APW, and values predicted from the hot spot or and P. Bunge (personal communication, 2000) and Greff-
TPW curves (reference frames). Figure 14 serves to illus- Lefftz [2001] insist that their models strongly underestimate
trate that the characteristics of these three curves depend the true vigor of mantle convection, hence strongly under-
very much on the location to which they apply. estimate TPW rates.
[60] In conclusion, none of the several suggested super- [63] Therefore, all current modeling still fails to some
fast events is based on sufficiently robust sets of observa- (sometimes large) extent to account for the slow values of
tions. It remains reasonable to assume that many of these typical TPW velocity (30 – 100 km/Myr), and even more so
features correspond to erroneous individual data or other to account for the prolonged (50 Ma) periods with almost
sources of error. Only the recent phase of TPW since 10 Ma no TPW (standstills). Also they predict rather smooth
prompts us to accept that TPW velocities on the order of evolutions, rather than the alternating episodes which we
100 km/Myr can be maintained over periods of millions of feel we uncover from the data. The remarkable similarity
years, although we have no specific geodynamic explan- between TPW estimates for the Pacific plate and the rest of
ation for this event (which some authors associate with the world, which are based on completely different and
deglaciation and rebound). independent data sets, lends support to the idea that sig-
[61] Whether even faster velocities over shorter time- nificant TPW, on the order of 10 or more, occurred since
scales actually occurred cannot as yet be considered as a before the Cretaceous. The importance of the Pacific plate
strong constraint that should be modeled in numerical and severe limitations on presently available data from that
experiments. Studies of TPW and mantle dynamics have plate point to the need for many more direct (paleomagnetic
entered a new phase with the advent of flow models, where core) measurements as opposed to indirect/remote sensing
seismic tomography is used to infer 3-D maps of density determinations of magnetization direction (i.e., ‘‘skewness’’
heterogeneities that drive flow in the viscous mantle. For or ‘‘seamount’’ data). The possible links between episodes,
instance, Steinberger and O’Connell [1997] calculate the or major changes between TPW episodes, and either plate
degree 2 nonhydrostatic component of the geoid and derive motion and plate boundary reorganizations or avalanches,
inertia perturbations on Earth over the last 60 Myr. Their plumes and other major geodynamical events occurring in
results are in reasonable agreement with our earlier BC91 the lower mantle should remain the topic of fascinating,
TPW estimates, though slightly larger. However, their ongoing studies.
model has smooth and regular TPW changes, rather than
the episodic structure interrupted by standstills which we [64] Acknowledgments. Most computations, data handling and pro-
find. Richards et al. [1999] calculate polar motion for duction of diagrams were made using the Paleomac software package
kindly provided by J. P. Cogné http://www.ipgp.jussieu.fr./cogne). Part of Evidence for a rapid shift of the Pacific hotspots with respect to spin axis,
this project was funded by INSU-CNRS Intérieur de la Terre (INSU-IT). Geophys. Res. Lett., 10, 709 – 712, 1983.
We thank Stuart Gilder for comments, Randy Enkin and John Tarduno for Gordon, R., and A. Cox, Calculating paleomagnetic poles for oceanic
very careful and useful reviews, and Ted Evans for efficient suggestions as plates, Geophys. J. R. Astron. Soc., 63, 619 – 640, 1980.
Associate Editor. IPGP contribution NS 1759. Greff-Lefftz, M., Déformation et rotation de la Terre, Mémoire d’habilita-
tion, 120 pp., Univ. Paris 7, Paris, 2001.
Hargraves, R. B., Paleomagnetism of Mesozoic kimberlites in southern
References Africa and the Cretaceous apparent polar wander curve for Africa,
Acton, G. D., and R. G. Gordon, Paleomagnetic tests of Pacific plate J. Geophys. Res., 94, 1851 – 1866, 1989.
reconstructions and implications for motion between hotspots, Science, Harland, W. B., R. Armstrong, A. Cox, L. Craig, A. Smith, and D. Smith, A
263, 1246 – 1254, 1994. Geologic Time Scale, 263 pp., Cambridge Univ. Press, New York, 1989.
Andrews, J. A., True polar wander: An analysis of Cenozoic and Mesozoic Hulot, G., M. Le Huy, and J. L. Le Mouël, Influence of core flows on the
paleomagnetic poles, J. Geophys. Res., 90, 7737 – 7750, 1985. decade variations of the polar motion, Geophys. Astrophys. Fluid Dyn.,
Besse, J., Cinématique des plaques et dérive des p^
oles magnétiques: Evolu- 82, 35 – 67, 1996.
tion de la Téthys, collisions continentales et couplage manteau/noyau, Irving, E., and G. A. Irving, Apparent polar wander paths: Carboniferous
thèse d’etat, 380 pp., Univ. Paris 7, Paris, France, 1986. through Cenozoic and the assembly of Gondwana, Geophys. Surv., 5,
Besse, J., and V. Courtillot, Paleogeographic maps of the Indian Ocean 141 – 188, 1982.
bordering continents since the Upper Jurassic, J. Geophys. Res., 93, Jaeger, J. J., V. Courtillot, and P. Tapponnier, Paleontological view of the
11,791 – 11,808, 1988. ages of the Deccan traps, the Cretaceous/Tertiary boundary, and the India-
Besse, J., and V. Courtillot, Revised and synthetic polar wander paths of the Asia collision, Geology, 17, 316 – 319, 1989.
African, Eurasian, North American, and Indian plates and true polar Johnson, C., and C. G. Constable, The time-averaged geomagnetic field:
wander since 200 Ma, J. Geophys. Res., 96, 4029 – 4050, 1991. Global and regional databases for 0 – 5 Ma, Geophys. J. Int., 131, 643 –
Briden, J. C., Recurrent continental drift of Gondwanaland, Nature, 215, 666, 1997.
1334 – 1339, 1967. Kent, D. V., and F. M. Gradstein, A Jurassic to recent chronology, in The
Bullard, E. C., J. E. Everett, and A. G. Smith, A symposium on continental Geology of North America, vol. M, The Western Atlantic Region, edited
drift, IV, The fit of the continents around the Atlantic, Philos. Trans. by P. R. Vogt and B. E. Tucholke, pp. 45 – 50, Geol. Soc. of Am.,
R. Soc. London, 258, 41 – 51, 1965. Boulder, Colo., 1986.
Bunge, H.-P., M. A. Richards, C. Lithgow-Bertelloni, B. A. Romanowicz, Koppers, A. A. P., J. P. Morgan, J. W. Morgan, and H. Staudigel, Testing
and S. P. Grand, Time scales and heterogeneous structure in geodynamic the fixity of hot spots hypothesis using 40Ar/39Ar age progression along
Earth models, Science, 280, 91 – 95, 1998. seamount trail, Earth Planet Sci. Lett., 185, 237 – 252, 2001.
Camps, P., M. Prévot, M. Daignières, and P. Machetel, Comment on ‘‘Sta- Lawver, L. A., and C. Scotese, A revised reconstruction of Gondwanaland,
bility of the Earth with respect to the spin axis for the last 130 million in Gondwana Six, Structure, Tectonics and Geophysics, Geophys.
years’’ by J. A. Tarduno and A. Y. Smirnov, Earth Planet. Sci. Lett., 198, Monogr. Ser., vol. 40, edited by G. D. McKenzie, pp. 17 – 23, AGU,
529 – 532, 2001. Washington, D. C., 1987.
Cande, S., A paleomagnetic pole from the late Cretaceous marine magnetic Le Pichon, X., and J. M. Gaulier, The rotation of Arabia and the Levant
anomalies in the Pacific, Geophys. J. R. Astron. Soc., 44, 547 – 566, 1976. fault system, Tectonophysics, 153, 271 – 274, 1988.
Carlut, J., and V. Courtillot, How complex is the time-averaged geomag- Livermore, R. A., F. J. Vine, and A. G. Smith, Plate motions and the
netic field over the past 5 Myr?, Geophys. J. Int., 134, 527 – 544, 1998. geomagnetic field, I, Quaternary and late Tertiary, Geophys. J. R. Astron.
Celaya, M. A., and B. M. Clement, Inclination shallowing in deep sea Soc., 73, 153 – 171, 1983.
sediments from the North Atlantic, Geophys. Res. Lett., 15, 52 – 55, 1988. Livermore, R. A., F. J. Vine, and A. G. Smith, Plate motions and the
Clouard, V., and A. Bonneville, How many Pacific hotspots are fed by deep geomagnetic field, II, Jurassic to Tertiary, Geophys. J. R. Astron. Soc.,
mantle plumes?, Geology, 29, 695 – 698, 2001. 79, 939 – 961, 1984.
Cogné, J. P., N. Halim, Y. Chen, and V. Courtillot, Resolving the problem Machetel, P., and P. Weber, Intermittent layered convection in a model
of shallow magnetizations of Tertiary age in Asia: Insights from paleo- mantle with an endothermic phase change at 670 km, Nature, 350,
magnetic data from the Qiangtang, Kunlun, and Qaidam blocks (Tibet, 55 – 57, 1991.
China), and a new hypothesis, J. Geophys Res., 104, 17,715 – 17,734, May, S. R., and R. Butler, North American Jurassic apparent polar wander:
1999. Implications for plate motion, paleogeography, and cordilleran tectonics,
Constable, C. G., and R. L. Parker, Statistics of the geomagnetic secular J. Geophys. Res., 91, 11,519 – 11,544, 1986.
variation for the past 5 Myr, J. Geophys. Res., 93, 11,569 – 11,581, 1988. McElhinny, M. W., and J. Lock, Four IAGA databases released in one
Cottrell, R. D., and J. A. Tarduno, Late Cretaceous true polar wander: Not package, Eos Trans. AGU, 76, 266, 1995.
so fast, Science, 288, 2283a, 2000. McElhinny, M. W., P. L. McFadden, and R. T. Merrill, The time averaged
Coupland, D. H., and R. Van der Voo, Long-term nondipole components in paleomagnetic field 0 – 5 Ma, J. Geophys. Res., 101, 25,007 – 25,027,
the geomagnetic field during the last 130 Ma, J. Geophys. Res., 85, 1996.
3529 – 3548, 1980. McFadden, P. L., and M. W. McElhinny, The combined analysis of remag-
Courtillot, V., J. Besse, and H. Théveniaut, North American Jurassic appar- netization circle and direct observation in paleomagnetism, Earth Planet.
ent polar wander: The answer from other continents?, Phys. Earth Planet. Sci. Lett., 87, 161 – 172, 1988.
Inter., 82, 87 – 104, 1994. McFadden, P. L., and M. W. McElhinny, Classification of the reversal test
Cox, A., and R. G. Gordon, Paleolatitudes determined from paleomagnetic in paleomagnetism, Geophys. J. Int., 103, 725 – 729, 1990.
data from vertical cores, Rev. Geophys., 22, 47 – 72, 1984. Merrill, R. T., and M. W. McElhinny, The Earth’s Magnetic Field, 401 pp.,
Cox, A., and B. Hart, Plate Tectonics: How It Works, 392 pp., Blackwell Academic, San Diego, Calif., 1983.
Sci., Malden, Mass., 1986. Molnar, P., and T. Atwater, Relative motion of hot spots in the mantle,
Davies, G. F., Ocean bathymetry and mantle convection, 1, Large-scale Nature, 246, 288 – 291, 1973.
flow and hot spots, J. Geophys. Res., 93, 10,467 – 10,480, 1988. Molnar, P., and J. Stock, Relative motions of hotspots in the Pacific, Atlan-
Di Venere, V., and D. V. Kent, Are the Pacific and Indo-Atlantic hotspots tic and Indian Oceans since late Cretaceous time, Nature, 327, 587 – 591,
fixed? Testing the plate circuit through Antarctica, Earth Planet. Sci. 1987.
Lett., 170, 105 – 117, 1999. Morgan, W. J., Hotspot tracks and the early rifting of the Atlantic, Tecto-
Dyment, J., S. C. Cande, and J. Arkani-Hamed, Skewness of marine mag- nophysics, 94, 123 – 139, 1983.
netic anomalies created between 85 and 40 Ma in the Indian Ocean, Müller, R. D., and W. R. Roest, Fracture zones in the North Atlantic from
J. Geophys. Res., 99, 24,121 – 24,134, 1994. combined Geosat and Seasat data, J. Geophys. Res., 97, 3337 – 3350, 1992.
Engebretson, D. C., A. Cox, and R. G. Gordon, Relative motions between Müller, D. M., J. Y. Royer, and L. A. Lawver, Revised plate motions
oceanic and continental plates in the Pacific Basin, Spec. Pap. Geol. Soc. relative to the hotspots from combined Atlantic and Indian Ocean hotspot
Am., 206, 59 pp., 1985. tracks, Geology, 21, 275 – 278, 1993.
Enkin, R. J., and G. S. Watson, Statistical analysis of paleomagnetic in- Norton, I. O., Plate motion in the North Pacific: The 43 Ma nonevent,
clination data, J. Geophys. Int., 126, 495 – 504, 1996. Tectonics, 14, 1080 – 1094, 1995.
Enkin, R., Z. Yang, Y. Chen, and V. Courtillot, Paleomagnetic constraints Nürnberg, D., and R. D. Müller, The tectonic evolution of the South Atlan-
on the geodynamic history of the major blocks of China from the Permian tic from Late Jurassic to Present, Tectonophysics, 191, 27 – 53, 1991.
to the present, J. Geophys. Res., 97, 13,953 – 13,989, 1992. Parker, R. L., A theory of ideal bodies for seamount magnetism, J. Geo-
Evans, D. A., True polar wander, a supercontinental legacy, Earth Planet. phys. Res., 96, 16,101 – 16,112, 1991.
Sci. Lett., 157, 1 – 8, 1998. Parker, R. L., L. Shure, and L. A. Hildebrand, The application of inverse
Gordon, R., Late Cretaceous apparent polar wander of the Pacific plate: theory to seamount magnetism, Rev. Geophys., 25, 17 – 40, 1987.
Patriat, P., and J. Achache, India-Eurasia collision chronology has implica- edited by B. E. Tucholke and P. R. Vogt, pp. 379 – 404, Geol. Soc. of
tions for crustal shortening and driving mechanism of plates, Nature, 311, Am., Boulder, Colo., 1986.
615 – 621, 1984. Steinberger, B., and R. J. O’Connell, Changes of the Earth’s rotation axis
Peirce, J. W., Assessing the reliability of DSDP paleolatitudes, J. Geophys. owing to advection of mantle density heterogeneities, Nature, 387, 169 –
Res., 81, 4173 – 4187, 1976. 173, 1997.
Peirce, J. W., The northward motion of India since the Late Cretaceous, Steinberger, B., and R. J. O’Connell, Advection of plumes in mantle flow:
Geophys. J. R. Astron. Soc., 52, 277 – 311, 1978. Implications on hotspot motion, mantle viscosity and plume distribution,
Petronotis, K. E., and R. G. Gordon, A Maastrichtian paleomagnetic pole Geophys. J. Int., 132, 412 – 434, 1998.
for the Pacific plate from a skewness analysis of marine magnetic anom- Tackley, P., Mantle convection and plate tectonics: Toward an integrated
aly 32, Geophys. J. Int., 139, 227 – 247, 1999. physical and chemical theory, Science, 288, 2002 – 2007, 2000.
Petronotis, K. E., R. G. Gordon, and G. D. Acton, Determining paleomag- Tackley, P. J., D. J. Stevenson, G. A. Glatzmaier, and G. Schubert, Effects of
netic poles and anomalous skewness from marine magnetic anomaly multiple phase-transitions in a three-dimensional spherical model of con-
skewness data from a single plate, Geophys. J. Int., 109, 209 – 224, 1992. vection in Earth’s mantle, J. Geophys. Res., 99, 15,877 – 15,901, 1994.
Prévot, M., E. Mattern, P. Camps, and M. Daignières, Evidence for a 20 Tarduno, J. A., Absolute inclination values from deep sea sediments:
tilting of the Earth’s rotation axis 110 million years ago, Earth Planet. A reexamination of the Cretaceous Pacific record, Geophys. Res. Lett.,
Sci. Lett., 179, 517 – 528, 2000. 17, 101 – 104, 1990.
Quidelleur, X., J. P. Valet, V. Courtillot, and G. Hulot, Long-term geometry Tarduno, J. A., and R. D. Cottrell, Paleomagnetic evidence for motion of
of the geomagnetic field for the last five million years: An updated the Hawaiian hotspot during formation of the Emperor Seamounts, Earth
secular variation database, Geophys. Res. Lett., 21, 1639 – 1642, 1994. Planet. Sci. Lett., 153, 171 – 180, 1997.
Ricard, Y., M. Richards, and Y. A. Le Stunff, A geodynamical model of Tarduno, J. A., and J. Gee, Large-scale motion between Pacific and Atlantic
mantle density heterogeneity, J. Geophys. Res., 98, 21,895 – 21,909, hotspots, Nature, 378, 477 – 480, 1995.
1993. Tarduno, J. A., and A. Y. Smirnov, Stability of the Earth with respect to the
Richards, M., H. P. Bunge, Y. Ricard, and J. R. Baumgardner, Polar wan- spin axis for the last 130 million years, Earth Planet. Sci. Lett., 184,
dering in model convections models, Geophys. Res. Lett., 26, 1777 – 549 – 553, 2001.
1780, 1999. Tauxe, L., P. Tucker, N. Petersen, and J. L. LaBrecque, The magnetostrati-
Royer, J.-Y., and D. T. Sandwell, Evolution of the eastern India Ocean since graphy of Leg 73 sediments, Paleogeogr. Paleoclimatol. Paleoecol., 42,
the Late Cretaceous: Constraints from Geosat altimetry, J. Geophys. Res., 65 – 90, 1983.
94, 13,755 – 13,782, 1989. Torcq, F., Evolution et destruction de la Pangée du Carbonifère au Jurassi-
Royer, J.-Y., P. Patriat, H. Bergh, and C. R. Scotese, Evolution of the que, thèse de doctorat, 183 pp., Univ. Paris 7, Paris, France, 1997.
Southwest Indian Ridge from the Late Cretaceous (anomaly 34) to the Vandamme, D., and V. Courtillot, Palaeomagnetism of Leg 115 basement
middle Eocene (anomaly 20), Tectonophysics, 155, 235 – 260, 1988. rocks and latitudinal evolution of the Réunion hotspot, Proc. Ocean Drill.
Royer, J. Y., R. D. Müller, L. M. Gahagan, L. A. Lawver, C. L. Mayes, Program Sci. Results, 115, 111 – 117, 1990.
D. Nürnberg, and J. G. Sclater, A global isochron chart, Tech. Rep. 117, Vandamme, D., V. Courtillot, J. Besse, and R. Montigny, Paleomagnetism
Univ. of Tex. Inst. for Geophys., Austin, 1992. and age determinations of the Deccan traps (India): Results of a Nagpur-
Sager, W. W., and A. A. P. Koppers, Late Cretaceous polar wander of the Bombay traverse and review of earlier work, Rev. Geophys., 29, 159 –
Pacific plate: Evidence of a rapid true polar wander event, Science, 287, 190, 1991.
455 – 459, 2000. Van der Voo, R., Phanerozoic paleomagnetic poles from Europe and North
Sager, W. W., and S. Pringle, Mid Cretaceous to Early Tertiary apparent America and comparisons with continental reconstructions, Rev. Geo-
polar wander path of the Pacific plate, J. Geophys. Res., 93, 11,753 – phys., 28, 167 – 206, 1990.
11,771, 1988. Van der Voo, R., Paleomagnetism of the Atlantic, Tethys and Iapetus
Saunders, A. D., J. G. Fitton, A. C. Kerr, M. J. Norry, and R. W. Kent, The Oceans, 411 pp., Cambridge Univ. Press, New York, 1993.
North Atlantic igneous province, in Large Igneous Provinces, Geophys. Westphal, M., M. L. Bazhenov, J. P. Lauer, D. M. Pechersky, and J. C.
Monogr. Ser., vol. 100, edited by J. J. Mahoney and M. F. Coffin, pp. 45 – Sibuet, Paleomagnetic implications on the evolution of the Tethys belt
93, AGU, Washington, D.C., 1997. from the Atlantic Ocean to the Pamirs since the Triassic, Tectonophysics,
Schneider, D. A., and D. V. Kent, The paleomagnetic field from equatorial 123, 37 – 82, 1986.
deep sea sediments: Axial symmetry and polarity asymmetry, Science, Wilson, R. L., Permanent aspects of the Earth’s non-dipole magnetic field
242, 252 – 256, 1988. over Upper Tertiary times, Geophys. J. R. Astron. Soc., 19, 417 – 437,
Schneider, D. A., and D. V. Kent, Paleomagnetism of Leg 115 sediments: 1970.
Implications for Neogene magnetostratigraphy and paleolatitude of the Yang, Y., and J. Besse, New Mesozoic apparent polar wandering path for
Reunion Hotspot, Proc. Ocean Drill. Program Sci. Results, 115, 717 – south China: Tectonic consequences, J. Geophys. Res., 106, 8493 – 8520,
736, 1990a. 2001.
Schneider, D. A., and D. V. Kent, The time-averaged paleomagnetic field,
Rev. Geophys., 28, 71 – 96, 1990b.
Schouten, H., and S. C. Cande, Palaeomagnetic poles from marine mag-
netic anomalies, Geophys. J. R. Astron. Soc., 44, 567 – 575, 1976.
Schouten, H., and K. McCamy, Filtering marine magnetic anomalies,
J. Geophys. Res., 77, 7089 – 7099, 1972. J. Besse and V. Courtillot, Laboratoire de Paléomagnétisme, UMR 7577,
Srivastava, S. P., and C. R. Tapscott, Plate kinematics of the North Atlantic, Institut de Physique du Globe de Paris, 4 Place Jussieu, F-75252 Paris
in The Geology of North America, vol. M, The Western Atlantic Region, cedex 05, France. ([email protected]; [email protected])

2002 - Besse and Courtillot

Uploaded by

Copyright:

Available Formats

2002 - Besse and Courtillot

Uploaded by

Document Information

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

2002 - Besse and Courtillot

Uploaded by

Copyright:

Available Formats

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. B11, 2300, doi:10.

Correction published 10 October 2003

Apparent and true polar wander and the geometry

1. Introduction hereinafter referred to as BC91] to blend the two approaches

present Earth’s field to reorient the core. Inclination alone

is the underestimation of inclination due to arithmetic

error (SM). We have selected those data corresponding to

values of SM less than 5. A minimum amount of 36 samples

[18] In addition, as regards data postdating the Peirce

studies of oceanic cores drilled during campaigns of the

Ocean Drilling (ODP) and Deep Sea Drilling (DSDP)

Ridge) and South Atlantic (South American plate) Oceans.

these results are listed in Table 2b. The distribution of these

120 Ma, is also plotted in Figure 1a.

2.3. Combining Data Types

[19] The space and time distributions of selected paleo-

magnetic data are shown in map (Figures 1a and 1b) and

histogram (Figures 1c – 1e) form. A total number of 242

set comprises 221 ‘‘classical’’ poles, two poles determined

‘‘inclination only’’ data from oceanic cores. The total

number of data is thus increased by more than 2 with

Pennsylvania, Culpeper Basin intrusives and baked sediments 201

reasonable; it is much improved over BC91. The histogram

(Figure 1c) shows that a large number of data (more than

30) is available for the 0– 30 and 50– 80 Ma time windows.

By contrast, the 140– 150 Ma time window contains only

eight studies. The mean number of data in each 10 Myr time

window is on the order of 15. As a matter of comparison,

Combined Pennsylvania diabase

the corresponding mean for the recent Prévot et al. [2000]

study is on the order of 4, and there are less than 5 data in

twelve 10 Myr time windows.

wants to construct the APWP of a single plate based on its

Table 2b. Selected Data From ODP and DSDP Legsa

Table 3. Finite Rotation Poles for Gondwanaa

Ocean ridge. After that time, Australia and India belong to

assumption of hot spot fixity. This is a long debated subject

standstill of Pangea at 200 –180 Ma is followed by the

5. True Polar Wander

the 116 Ma pole. The 95% confidence intervals of the 100

of super fast TPW suggested by Prévot et al. [2000]. As

Figure 14. (opposite) ‘‘Relative’’ latitudes calculated for

You might also like