RESEARCH ARTICLE Pangea Migration

10.1029/2020TC006585
Xavier Le Pichon1 , Mark Jellinek2 , Adrian Lenardic3 , A. M. Celal Şengör4 , and
Key Points: Caner İmren5
• P late tectonics of the stationary 1
Collège de France, Paris, France, 2Department of Earth, Ocean, and Atmospheric Sciences, University of British
hemispheric equatorial Pangea was
radically different from the present Columbia, Vancouver, BC, Canada, 3Department of Earth Science, Rice University, Houston, TX, USA, 4Maden
dispersed continental stage Fakültesi, Jeoloji Bölümü ve Avrasya Yerbilimleri Enstitüsü, İstanbul Teknik Üniversitesi, İstanbul, Turkey, 5Maden
• Hemispheric surrounding slab Fakültesi, Jeofizik Bölümü, İstanbul Teknik Üniversitesi, İstanbul, Turkey
curtain led to fixity with respect to
the mantle causing thermal isolation
and a warm underlying upper
mantle
Abstract We confirm the proposition of Le Pichon et al (2019, https://doi.org/10.1029/2018TC005445)
• Consequences are sea level control, that Pangea was ringed by a hemispheric subduction girdle from its formation 400 Ma to its dispersal
Pangea migration to equator and 100 Ma. We quantify the northward migration, that we attribute to True Polar Wander (TPW), of its
associated oscillations, fracturing,
flood basalts, and finally, breakup
axis of symmetry, between 400 and 150 Ma, from southern latitudes to the equatorial zone. The spatial
stabilizing within the equatorial zone of the axis of symmetry in a fixed position with respect to the lower
mantle was marked by alternating clockwise and counterclockwise oscillations between 250 and 100 Ma
Correspondence to:
that we relate to tectonic events. A subduction girdle is predicted to set up lateral temperature gradients
A. M. C. Şengör,
sengor@itu.edu.tr from the relatively warm sub-Pangean mantle to a cooler sub-oceanic mantle. Over time, this effect acts
to destabilize the Pangea landmass and its associated subduction girdle. Quantitatively, a scaling theory
Citation:
for the stability of the subduction girdle against the mantle overturn constrains the maximum magnitude
Le Pichon, X., Jellinek, M., Lenardic, of sub-Pangean warming before breakup to be order 100 °C, consistent with constraints on the Pacific-
A., Şengör, A. M. C., & İmren, C. Atlantic oceanic crustal thickness differences. Our predictions are in line with the recent analyses of the
(2021). Pangea migration. Tectonics, Jurassic-Cretaceous climate change and with the existing models for potential driving forces for a TPW
40, e2020TC006585. https://doi.
org/10.1029/2020TC006585 oscillation of Pangea across the equator. The timing and intensity of the predicted sub-Pangean warming
potentially contributed to the enigmatically large Siberian Traps and CAMP flood basalts at 250 Ma and
Received 20 OCT 2020 201 Ma, respectively.
Accepted 4 MAY 2021

1. Introduction
Le Pichon et al. (2019, called Le Pichon for short) corroborated a much earlier proposal by Le Pichon and
Huchon (1983, 1984) that Pangea and Panthalassa each occupied a whole hemisphere surrounded by a sub-
duction zone girdle following a great circle (Figure 1). This configuration, thus, had an axis of symmetry.
These authors further proposed that Pangea was stationary or moved very little with respect to the mantle
and consequently, defined an absolute reference frame which is simply determined by the position of this
axis of symmetry. Pangea first formed in the southern latitudes but migrated steadily northward until its
axis of symmetry was within the equatorial plane 250 Ma. But, we first need to define what we call Pangea
in this study. Restricting the name of Pangea to the continuous landmass that resulted from the amalgama-
tion of all continental material, as is often done, is fraught with difficulties. For example, one might argue
that Laurasia was not achieved when the North Atlantic break was initiated as the Khangai/Khantey ocean
was not yet closed. Instead, in this study, we focus on the hemispheric subduction girdle. We consider that
Pangea existed when all continental material was contained within the hemispheric subduction girdle. This
is because the hemispheric constraint implied that continental material covered 80% of the surface of the
hemisphere and the remaining 20% of oceanic space, mostly collected in one ocean called the Palaeo-Te-
thys, could be considered to be intracontinental. Further, the presence of a surrounding curtain of slabs
down to at least 400 km depth led to stationarity, thermal isolation and associated sub-Pangean mantle
warming (Lenardic et al., 2011). Thus, the whole hemisphere had very peculiar kinematic, thermal, and
dynamic constraints as discussed by Le Pichon. During Pangea time, two hemispheres, one continental and
the other oceanic, were separated by a subduction girdle that prevented mixing in the underlying mantle
between the two hemispheres. The system was dynamically stable when the axis of symmetry was within
the equatorial plane. An explanation for the formation of such a system may be given by the proposal
© 2021. American Geophysical Union. by Zhong et al. (2007) that the supercontinent assembly may result from degree 1 planform convection.
All Rights Reserved. The supercontinent assembly, then, leads to degree-2 planform convection with antipodal upwelling and

LE PICHON ET AL. 1 of 18
 Tectonics 10.1029/2020TC006585

consequent migration through the True Polar Wander (TPW) to the equa-
torial location, which is its position of stability, as the associated geoid of
order-2 has then its negatives centered on the poles and its positives on
the equator.

Pangea and its associated effects on the whole Earth system, may well be
unique amongst supercontinents. For this reason, we propose to call a
hemispheric continent such as Pangea, a Buridanian continent, because
the common term “supercontinent” does not necessarily imply a single
continent (e.g., Gondwana-Land was a supercontinent, but it shared the
earth with other continents at different times). We, thus, call Buridanian a
continent, such as Pangea, that is hemispheric, separated from the oppo-
site hemispheric ocean by a subduction girdle and which is in a position
of dynamic equilibrium when the axis of symmetry lies in the equatorial
plane. The subduction girdle great circle then passes through the poles.
The term Buridanian refers to the earth model of Jean Buridan (1300–
1358), twice the rector magnificus of the University of Paris in the 14th
century. Buridan assumed an earth with one major continent occupying
one hemisphere and a major ocean occupying the other. His model was
the only dynamic earth model proposed in the Middle Ages until James
Hutton's model in the 18th century (see Duhem, 1958; Şengör, 2003).

With this definition of Pangea, the beginning and end of this superconti-
nent coincide, respectively, with the formation and disruption of the sub-
duction girdle. Although one may argue that the reconstructions before
320 Ma are too uncertain to establish the existence of the subduction gir-
dle as an established fact, it is still the best explanation for the assembly of
Laurasia and Gondwana-Land and 400 Ma may be considered as the time
at which all existing continental material was contained within a single
hemispheric subduction girdle. From then on, the mantle within this
Figure 1. Lambert Azimuthal Equal Area projections with pole of hemisphere was thermally isolated from the mantle below Panthalassa.
projection that best fits the Pangea hemisphere. The upper hemisphere It is often stated that Pangea did not exist after 180 Ma because that was
is projected unto the equatorial plane of projection within the first great the beginning of the first large North Atlantic break. But, the subduction
circle. The hidden hemisphere is shown distorted beyond the great girdle was still intact until about 100 Ma (see 120 Ma reconstruction in
circle (thick red line). Symbols (⊕) in red show locations of north and
south poles. Global plate reconstructions from Müller et al. (2016) for
Figure 1). It is only after 100 Ma that the hemispheric subduction girdle
220 to 120 Ma (the color in the oceans indicates the age, red is youngest) lost its identity and, as a consequence, the global mantle was well mixed
and from Domeier and Torsvik (2014) for 360 Ma to 260 Ma. The main again. This is why we consider 400 Ma and 100 Ma as the beginning and
subduction zones in these reconstructions are shown in blue. Note end of Pangea. An important observation is that Pangea, which is the lat-
how the peripheral subduction zone is essentially continuous and est supercontinent, is also the only one that can be reconstructed without
fits well the Pangea hemisphere, especially for the most reliable pre-
240 Ma reconstructions. Note also how this great circle goes through the
large uncertainties, at least for the Mesozoic (post-250 Ma) period, as the
geographic poles between 260 and 100 Ma. Adapted from Le Pichon. reconstructions are more uncertain prior to 250 Ma. Pangea is, conse-
Although there is still no consensus on some aspects of the Paleozoic quently, the best one to test theories for the formation, evolution, and
reconstructions for which Le Pichon proposed an alternate version (see longevity of supercontinents.
their Figure 9), these differences would not affect the main conclusions
of the Pangea formation stage that we discuss here. The reader should In this context, this study focuses on the 60° northward migration of Pan-
be aware however of the significantly larger degree of uncertainty of the gea at a steady rate of 0.4°/Myr from the southern latitudes from its for-
Paleozoic reconstructions.
mation 400 Ma in the early Devonian to 250 Ma at the end of Permian,
when it entered the equatorial zone, and then to its stabilization in the
equatorial zone between 250 and 100 Ma. This is clearly seen in Figure 2,
where we have plotted the latitude of the axis of symmetry of the hemispheric continent between 400 and
100 Ma. This migration to the equatorial zone where Pangea stabilized indicates that Pangea reached its
equilibrium position when its axis of symmetry belonged to the equatorial plane. The northward migration
is often attributed to plate tectonic processes (e.g., Steinberger & Torsvik, 2008). However, we do not con-
sider likely that this hemispheric piece of the upper mantle and lithosphere moved at a steady velocity of
about 40 mm/yr during 150 Myr without even distorting the hemispheric shape of the subduction girdle.

LE PICHON ET AL. 2 of 18
 Tectonics 10.1029/2020TC006585

Figure 2. Latitude of projection of the axis of symmetry of the Pangea hemisphere versus time obtained from a set of reconstructions defined in Figure 1
between 400 and 100 Ma at 20 Myr interval. The counterclockwise (CCW) and clockwise (CW) rotations that affected Pangea during its migration are after
Torsvik et al. (2012). The dates of the Siberian and Camp traps coincide with the beginning and end of the first large CCW rotation of Pangea. The dashed
black line shows in a schematic and qualitative fashion the long-term evolution of global sea level. There is a consensus to consider that a high sea level existed
400 Ma, a low sea level between 260 and 180 Ma, and high sea level again in Upper Cretaceous near 100 Ma so that long-term sea level trend was decreasing
between 400 and 260 Ma and rising between 260 and 100 Ma (Miller et al., 2005; Snedden & Liu, 2010).

We prefer to explore the possibility that the Devonian to Permian northward migration of the axis of rota-
tion toward the equatorial plane was due to the TPW. Figure 2 shows the remarkable 60° of latitude motion
at a steady latitudinal velocity of about 40 mm/yr ending rather abruptly when it reached the equatorial
zone with oscillations about the axis of symmetry of the hemisphere. (Note that for the reconstructions we
used, the velocity would be significantly larger, 50–60 mm/yr, as their authors assume an eastward com-
ponent of motion). The rate of migration slowed and eventually, reversed near 180 Ma during Toarcian as
the axis of symmetry of Pangea reached a steady-state position within the equatorial plane. The peripheral
subduction “girdle” from then on coincided with a polar great circle. The stabilization in the equatorial
plane between 250 and 100 Ma, from end of Permian to Albian, ended with the demise of Pangea at 100 Ma
when the subduction girdle that had begun to dislocate 180 Ma lost its identity and global thermal mixing
of the mantle was, ultimately, re-established. This stabilization was accompanied by successive clockwise
(CW) and counterclockwise (CCW) rotations of diminishing amplitudes attributed to theTPW with respect
to the axis of symmetry of Pangea (Steinberger & Torsvik, 2008; Torsvik et al., 2012). These rotations are
consistent with a modeling study that investigated the oscillation of Pangea across the equator in terms of
an elastic restoring force arising in response to memory of Earth's rotational bulge (Creveling et al., 2012).
Pangea northward migration is, thus, closely related to its identity of Buridanian continent that can only be
dynamically stable within the equatorial zone.

LE PICHON ET AL. 3 of 18
 Tectonics 10.1029/2020TC006585

As mentioned above, the stationarity of Pangea lead to thermal isolation

and sub-Pangean warming (Jellinek et al., 2020; Lenardic et al., 2011).
These authors have proposed that, as a consequence, the long-term sea
level change would record this evolution. The assembled continent stage
would correspond to maximum thermal isolation of the mantle of the
two hemispheres and to low sea level, whereas the dispersed continent
would correspond to a globally well-mixed mantle and to high sea level.
Long-term sea level should thus, be an important marker of the state of
isolation or thermal mixing of the global mantle. This is verified in Fig-
ure 2 by the schematic qualitative sea level curve and will be used as a
confirmation of the predicted evolution of the global mantle from isola-
tion to good mixing.

In the following: (a) We corroborate the validity of the Pangea reference

frame proposed by Le Pichon. (b) We describe and discuss the northward
migration of Pangea to the equatorial zone. (c) We discuss the effect of
the hemispheric subduction girdle in the context of Pangea stability
and relate its formation and dispersal to long-term sea level evolution.
Figure 3. Late Permian Pangea (modified from Şengör & Atayman, 2009, Through a scaling analysis of the mechanics acting to stabilize Pangea
Figure 15B). Key to symbols: Lines with teeth are subduction zones against breakup, we argue that the effects of the subduction girdle on the
(teeth on the upper plate), double lines are spreading centers, lines with evolution of the mantle thermal structure lead to the eventual breakup of
hachures are normal fault belts (usually delimiting rifts), upsidedown v's
Pangea. (d) Finally, we argue that equatorial rotations of Pangea prior to
are flood basalts (exposed), pluses are flood basalts (covered by younger
deposits), half arrows show strike-slip motion. In northern Siberia, data final breakup, coupled to girdle induced thermal effects, initiated local-
are insufficient to decide on the tectonic regime. ized episodes of continental volcanism.

2. Confirmation That Pangea Provides an Absolute Reference Frame

The six reconstructions of the supercontinent Pangea shown in Figure 1 are widely used reconstructions by
Müller et al. (2016) from 220 to 120 Ma and by Domeier and Torsvik (2014) from 360 to 260 Ma. The largest
deviations of the fits of Müller et al. (2016) with respect to a great circle reach 700–800 km and probably
are of the same order as the uncertainties of the reconstructions. Their authors had no a priori bias for a
hemispheric configuration of Pangea as they do not even mention it. Note that the fact that Pangea fits a
hemisphere is a property of all recent reconstructions and was already demonstrated on the 200 Ma recon-
struction published by Le Pichon and Huchon (1984) based on Norton and Sclater (1979). This is because
the adjustment of Gondwana-Land and Laurasia during their collision imposes this hemispheric shape,
best explained by the constraint of a hemispheric subduction girdle. Indeed, the most striking character of
the reconstructions in Figure 1 is that Pangea was surrounded by a peripheral, essentially continuous, sub-
duction zone coinciding with a great circle (see in particular, the 220 to 180 Ma reconstructions). Of course,
the reconstruction of former subduction zones is model dependent. But, other recent reconstructions have
such a subduction girdle over most, if not all of the periphery of Pangea. For example, Stampfli et al. (2013)
and Scotese (2017) have such a hemispheric subduction girdle over more than 80% of the periphery.

This is also the case for the Late Permian Pangea of Şengör and Atayman (2009) and Şengör et al. (2018)
shown in Figure 3, when it already acquired its Pangea A2 geometry. With the exception of the northern
Siberian Arctic shelf, the entire supercontinent is surrounded by a subduction girdle and we now briefly
argue why this choice is made using Figure 3 as a guide. If we begin in the northeast, the Tuva-Mongol
Fragment was flanked by a subduction zone dipping toward it, along which the Khangai-Khantey accre-
tionary prism accumulated since the Neoproterozoic and finally ended during the early Cretaceous (Şengör
et al., 2018). The same subduction zone can be followed into Southeast Asia to connect with the Loei Oro-
gen in Thailand, where subduction was active since the latter half of the Paleozoic (Şengör, 1984). Then
the major Gondwanide orogenic belt of Du Toit (1937) girdles Gondwana-Land and continues into the
margin of North America at least since the Devonian. The Tasmanides had orogeny since the beginning of
the Paleozoic and with changing plate boundary configurations, it still continues in the Southwest Pacific
Island arc systems that now connect with the western Pacific Island arc systems.

LE PICHON ET AL. 4 of 18
 Tectonics 10.1029/2020TC006585

In western North America, the earliest evidence for orogeny is the Devonian Antler event that indicates
the presence of an offshore island arc with a marginal basin separating it from the main continent the
contents of which are now found in the Roberts Mountain Allochthon (Madrid et al., 1992). A second
deformational episode was the Sonoma event that greatly resembled the Antler and lasted well into the Tri-
assic (Caravaca et al., 2018), and generated the Golconda Allochthon and the Kaipato volcanics. After the
Sonoma, the interpretations become multifarious because of the “terrane” methodology that assumes that
every coherent continental block had to have been separated from the main continent by an ocean. That
is manifestly not true. In southwestern United States, subduction-related shortening had already com-
menced in the latest Permian and that regime evolved northwards (Burchfiel et al., 1992; Dickinson, 2004;
Gabrielse & Yorath, 1991). The various independent terrane interpretations (e.g., Nokleberg et al., 2000)
are indefensible, simply because everything east of the Cache Creek suture are directly related to the North
American Craton (see Gabrielse & Yorath, 1991). Therefore, the Slide Mountain Ocean or Angayuchem
Oceans depicted in Nokleberg et al. (2000) do not exist. Nokleberg ignores the enormous amount of strike-
slip faulting in the Canadian Cordillera throughout the Phanerozoic that greatly intermingled many units
(see especially Dickinson, 2009). The Stikine arc is also a native American arc originally lying east of the
Cache Creek suture (see Wernicke, 1988). Thus, an east-dipping subduction zone existed outboard of the
Cache Creek, not dissimilar to the earlier Sonoma arc farther south. It is likely that they formed a single
entity that later disrupted. The Wrangellia and the Alexander “terranes” were claimed to be “oceanic pla-
teaux,” but the underlying Hasen Creek formation has coarse clastics bearing quartz (for references, see
Greene et al., 2005). Therefore, Wrangellia also could not have been “out in the Ocean.” It was most likely
a fringing arc that later localized a flood basalt event, much like the Columbia basalts of the Miocene to
early Pliocene times. There is no justification for placing the Canadian Cordilleran “terranes” out into the
ocean. The various units, now strike-slip shuffled, were continental margin and fringing island-arc type
environments, much like the eastern margin of Asia today. The entire Cordillera, from British Columbia to
Patagonia shows clearly that subduction was at the continental margin since at least the Jurassic (Kirsch
et al., 2016). Before that, down into the Triassic, again, subduction was at the continental margin or not
very far away (Dickinson, 2004, 2009). In the Andes, subduction was active from at least the early Paleozoic
(e.g., Lucassen & Franz, 2005); the medial Devonian to early Carboniferous Chanic orogeny in Argentina
followed an episode of extension (Ariza et al., 2018), but continued in the late Carboniferous with sub-
duction dipping beneath the continent (García-Sansegundo et al., 2014). Subduction-related magmatism
is known from the northern Andes from the Permian onwards (Spikings et al., 2019). The Transantarctic
mountains are a part of the Gondwanide orogen. The Ross Mountains have evidence for early Paleozoic
orogeny, and the entire Transantarctic mountains show evidence of subduction-related magmatism in
the Antarctic Peninsula, Thurston Island, Marie Byrd Land, and Northern Victoria Land from where the
subduction zone continued into the Lachlan Belt in Australia (Milne & Millar, 1991). In eastern Austral-
ia, including the Tasmanides, subduction-controlled orogeny was continuous from the Cambrian to the
present with almost always continent-ward dipping subduction (Glen et al., 2016). This completes our tour
around Pangea from about the Devonian to the Cretaceous, except the northern Margin of Siberia, where
the Omolon arc has an as yet uncertain western connection (see Şengör & Natal'in, 1996). We, conse-
quently, maintain that there is a consensus for a continuous or nearly continuous hemispheric subduction
girdle, although complex plate tectonic patterns, very difficult to reconstruct without ambiguity, may have
affected it locally.

Le Pichon argued that the peripheral subduction girdle around Pangea implied stationarity and that, conse-
quently, Pangea defines an absolute reference frame. Between 260 and 100 Ma, this great circle defined by
the subduction girdle passed through the geographic poles. Thus, the Mesozoic Pangea hemisphere had an
axis of symmetry in the equatorial plane. The present geoid also has an axis of symmetry in the equatorial
plane which coincides with the axis of symmetry of the two antipodal LLSVP's at the base of the mantle
(we adopt the value of 0°N and 10°E given by Dziewonski et al., 2010, for the position of this axis). Le
Pichon and Huchon (1984) and Le Pichon proposed that this axis of symmetry coincided with the axis of
symmetry of Pangea 100 Myr ago and that, as a consequence, we can relate the Pangea absolute reference
frame to the Present. In this reference frame, South America is the only continent that has remained adja-
cent to the absolute position of the subduction girdle since Pangea time. Africa, on the other hand, before
opening the South Atlantic, was not situated close to the axis of symmetry of Pangea and, consequently,

LE PICHON ET AL. 5 of 18
 Tectonics 10.1029/2020TC006585

did not sit squarely on top of the so-called Africa LLSVP as it does to-
day (see Figure 3b of Le Pichon). As discussed at length by Le Pichon,
this disagrees with the proposal first made by Burke and Torsvik (2004)
and Torsvik et al. (2006), that Africa has remained over the LLSVP for
long geologic times. Quoting Le Pichon, “the main problem posed by
the “zero-longitude” Africa change, as it is often called, is that it implies
no movement for the LLSVP as well as Africa. But the equatorial axis of
symmetry that goes presently somewhere through the Gulf of Guinea,
near the center of Africa, was situated quite far from it in Pangea time.
This would imply that Pangea had an equatorial axis of symmetry quite
far from the equatorial axis of rotation of the lower mantle, which does
not make sense.”

We do not claim, however, that the peripheral subduction zone during

Pangea time maintained its position along a fixed great circle to a few
tens of kilometers all the time. We earlier mentioned that deviations of
Figure 4. Lambert Azimuthal Equal Area projections with pole of up to 800 km exist even for the best controlled pre-250 Ma reconstruc-
projection at 0°N, 10°E (Pangea reference frame) of reconstructions by
Müller et al. (2019). The red great circle indicates the position predicted
tions in Figure 1. Further, plate tectonic processes, such as formation and
for the subduction girdle in the Pangea reference frame. Present contours destruction of marginal basins, were active along this subduction girdle.
for the whole world in black and 20, 40, 60, and 80 Ma for South America The question, then, is how far the girdle could locally depart from the
and Africa in color from red for 80 Ma to blue for 20 Ma. South America hemispheric great circle. We next give new arguments that confirm that
migrates along the great circle while Africa migrates to the NE. the part of the subduction girdle next to South America has remained
stationary within 1,000 km, that is one tenth of the radius of the hemi-
sphere, with respect to the deep mantle since at least 250 Ma as predicted
by the Pangea reference frame. We propose that 1,000 km is a reasonable estimate of the possible amount
of local motion of the subduction girdle away from the average great circle. Note that this implies that the
shortening that occurred during this period within the Andes must have been smaller than about 1,000 km
in agreement with recent geologic estimates (e.g., Schepers et al., 2017).

(a) Williams et al. (2015) pointed out that published absolute motion models often imply unreasonably high
rates of paleo-trench migration (TM) and Net Lithospheric Rotation (NLR) and that the best models, in this
respect, were the Van der Meer et al. (2010) slab model and the NLR models, both models being actually
quite close. Tetley et al. (2019) demonstrated further that they could devise absolute motion models with
low TM and NLR rates that fit reasonably well with major hot spots migration lines younger than 60–80 Ma.
These models are close to the Pangea reference frame for the last 80 Myr. This is demonstrated in Figure 4
with reconstructions of Müller et al. (2019), who used the method of Tetley et al. (2019) to search for the
absolute reference frame. Their new absolute motions models produce both low NLR and TM, and also fit
reasonably well with major hot spots migration lines younger than 60–80 Ma, as stated above. Notice how,
in Figure 4, South America remained close to the subduction zone, as it moved approximately parallel to it,
about 20° CCW, at an average rate of 0.25°/Myr or about 30 mm/yr (see the directions of absolute motion
in Figure 7 of Müller et al., 2019). (See also Figure 1 of Chen et al., 2019 which shows a distance of 800–
1,200 km from the vertical of the 60–80 Ma slab to the trench. Note that, as the slab, because of its rigidity
does not plunge vertically below the trench, only a part of this distance may be due to relative westward
migration of the trench.) The only part of the subduction zone where the absolute motion of South Amer-
ica was toward the trench is the northwestern equatorial portion where the trench is oriented NE/SW, as
confirmed by the recent interpretation of Chen et al. (2019). We conclude that the Pangea reference frame
is compatible within about 1,000 km for the last 80 Myr with the information coming from hot spot tracks.
Africa, on the other hand moved northeastward, 20° to the north but also 20° to the east. This large eastward
shift in the longitude of Africa, during the last 80 Myr, is in disagreement with the so-called zero-longitude
Africa concept.

This result is important because there is a consensus that hot spot trails younger than 60–80 Ma are fair-
ly reliable indicators of absolute motion, in contrast with absolute motions derived from 80 to 130 Ma
hot spot trails which are fewer and more uncertain. This is first because, prior to 83.5 Ma, one loses a
well-established kinematic link between the Pacific hot spots and the Indian and Atlantic ones. Second,

LE PICHON ET AL. 6 of 18
 Tectonics 10.1029/2020TC006585

Figure 5. (a) Same projection as in Figure 4 for reconstructions at 140 (black) and 180 Ma (red) of Van der Meer
et al. (2010) based on positions of slabs in the mantle. The frame of reference is offset 8° westward from the Pangea one
indicated by the red circle, within the range of errors of this method. (b) Global longitude correction curve (blue) versus
time for reference frame based on positions of slabs in the mantle with respect to Torsvik et al. (2008) reference frame
after Figure 3 of Van der Meer et al. (2010). The corresponding correction curve for the Pangea reference frame (red
circles) is shown for comparison. See text.

because as hot spot trails diverge increasingly with increasing age, complex geodynamic models of mantle
flow are used to predict the deflection of the plumes that created the hot spots and reconcile the trails,
the so-called moving hot spot models (Doubrovine et al., 2012; Steinberger et al., 2004, see discussion in
Le Pichon and Tetley et al., 2019). Williams et al. (2015) and Tetley et al. (2019) showed, indeed, that 80
to 130 Ma absolute motions derived from hot spots diverged increasingly from those fitting the slab and
NLR method.

(b) Van der Meer et al. (2018) confirmed the relationship established by Van der Meer et al. (2010) between
the depth of slabs in the mantle and the age at which they left the surface. Both studies determined the ages
of slab remnants identified by tomography in the mantle by tying them to the age of the orogeny presumed
to have caused their formation (top of slab, end of subduction; bottom of slab, initiation of subduction).
Slabs now at the core-mantle boundary (CMB) are 220–290 Ma and estimates for ages at a given depth
are accurate to better than 40 Myr compared to a travel time of 200–220 Myr from surface to CMB. The
resulting subduction speed in the lower mantle is inferred to lie between 8 and 12 mm/yr (Van der Meer
et al., 2010, 2018). Of course, this work adopted geological scenarios and tomographic interpretations that
are non-unique. An examination by one of us (A.M.C.S.) of the Van der Meer et al. (2018) geological cor-
relations identified several serious difficulties in their interpretations, in particular, concerning ophiolite
spreading and obduction ages in Greece and Turkey. These reservations are not critical for this study, as the
consistency in the majority of the data used by both studies demonstrate that it is possible to relate a depth
in the lower mantle to an average geological age. It is worth noting, however, that the actual age of some of
the slabs at this level may differ by as much as 25 Myr from this average for the 2010 paper of Van der Meer
(their Figure 1) and up to 40 Myr for the 2018 paper (their Figure 99a). Consequently, taking their age-depth
relationship to provide an age constraint with a 40 Myr maximum uncertainty, we investigate the validity of
the Pangea reference frame for ages older than 80 Ma. Van der Meer et al. (2010) used the information com-
ing from the lower mantle slabs to test the paleolongitudes of the reconstructions of Torsvik et al. (2008)
and found that they were systematically offset westward by 18°, between 140 and 180 Ma, with respect to
the slab information. The offset linearly decreases with time after 140 Ma. Figure 5a shows the Van der Meer
et al. (2010) reconstructions for 140 and 180 Ma after correction of the 18° offset within our Pangea frame
of reference. Although the agreement with the Pangea frame is not perfect (we tested that a 6°–8° eastward
shift of the Van der Meer reconstruction would match our Pangea frame), it is within the uncertainties of
this slab method discussed in Van der Meer et al. (2010). These include uncertainties in the definition of the

LE PICHON ET AL. 7 of 18
 Tectonics 10.1029/2020TC006585

depth of the tomographic signal, especially in the lower mantle, in dating

the dates of beginning and end of subduction and finally in relating a de-
tached slab to the orogeny that produced it. These uncertainties are min-
imized where the imaged slab is attached to an active subduction zone.

Figure 5b shows the global longitude correction against time obtained

by Van der Meer et al. (2010) with respect to the Torsvik et al. (2008) re-
constructions. These reconstructions are very close to those of Müller
et al. (2016) who adopted the same reference frame. We show in Figure 5b
the global longitude correction for our Pangea reference frame, taking the
axis of symmetry of the present geoid at 0°N, 10°E, as explained above.
Both curves agree for ages smaller than 200 Ma with a maximum devia-
tion of 6° which is within the error bounds of both methods. We consid-
er the agreement of these two completely independent methods to be a
strong argument in favor of their validity. It validates, in particular, the
stationarity of South America, which is a major consequence of the sta-
tionarity of Pangea, as explained above, as it confirms that South Amer-
ica, between the Present and 200 Ma ago, moved longitudinally by less
than 700 km in its northern part and less than 400–500 km in its central
portion.

Both curves show a large westward shift of the reference frame with re-
spect to the Torsvik et al. (2008) reference frame, starting at 60 Ma and
extending to 120 Ma. The shift is a maximum of 18° for Van der Meer
et al. (2010) frame based on subduction slabs and 24° for the Pangea
frame based on stationarity of Pangea. Le Pichon discussed the cause of
this 18°–24° difference with the Torsvik et al. (2008) reference frame. For
times younger than about 70 Ma, Torsvik et al. (2008) used the Indian-At-
Figure 6. S wave tomography (red, low velocity; blue, high) at five lantic hot spot frame that results in Africa migrating eastward by only 7°
different depths from Cottaar and Lekic (2016) with same Pangea pole during the last 70 Myr versus 13° for the Pangea frame. For times older
of projection as Le Pichon (Figure 3). The red great circle indicates the from those well constrained by hot spot trails (70 Ma), Torsvik et al. (2008)
expected location of the subduction girdle in the Pangea reference frame.
adopted, for ages older than 130 Ma, the corrected paleomagnetic model
The reconstructed contours of continents after Müller et al. (2019) are
shown for the expected ages of slabs at this depth (see text). of Steinberger and Torsvik (2008) that assumes no longitudinal motion of
Africa, with a transition between the two models. Thus, in their model,
during the first phase of opening of the South Atlantic between 130 and
70 Ma, Africa is essentially fixed and it is South America that moves west-
ward. The slab reference frame as well as our Pangea frame are not compatible with the existence of such
a large westward motion of South America between 70 and 120 Ma. Finally, we note that the 200–250 Ma
part of the Global Longitude correction proposed by Van der Meer et al. (2010) relies on slabs assumed to
be older than 220 Ma that are now situated on the CMB. We suggest that the identification of these slabs is
problematic and that the decrease in longitude correction, that they lead to and that the Pangea reference
frame does not have, is not valid. We conclude that the slab information used by Van der Meer et al. (2010)
confirms the validity of the Pangea frame.

Following Le Pichon, we investigate further this conclusion, using the distribution of S wave-derived tomo-
graphic highs (Cottaar & Lekic, 2016) at five different levels in the mantle. In Figure 6, we have added the
presumed ages of the slabs at the origin of the anomalies using an average curve through the age-depth re-
lation of Van der Meer et al. (2018). The Cottaar and Lekic (2016) tomographic solutions shown in Figure 6
come from a cluster analysis of five global models given for five depth slices about 300 km thick, which is
equivalent to averaging over about 30 Myr. They are compared to the most recent reconstructions of Müller
et al. (2019) used above, as well as to the predicted position of the subduction girdle in the Pangea frame.

Figure 6 confirms that the Müller et al. (2019) frame is close to the Pangea frame at 70 Ma, but is progres-
sively shifted eastward, up to 24°–28°, 240 Ma. We checked that the reconstructions of Müller et al. (2016)
are shifted by the same amounts. On the other hand, as already pointed out by Le Pichon, the position of the
subduction girdle in the Pangea frame broadly coincides with a belt of high velocity anomalies.

LE PICHON ET AL. 8 of 18
 Tectonics 10.1029/2020TC006585

Figure 7. Time series of two-dimensional numerical simulations of supercontinental formation with a subduction
girdle and subsequent breakup. The cartoon is based on numerical models, scaling analysis, and laboratory tank
experiments discussed more fully in Lenardic et al. (2011). The formation of a subduction girdle inhibits lateral
convective thermal mixing and leads to isolation of the oceanic and sub-supercontinental mantle domains. Whereas
the oceanic potential temperature declines as a result of this process (red curve), a volume of mantle beneath the
supercontinent warms depending on the longevity of this arrangement (red shading in simulations). For plausible
conditions during Pangea formation and breakup, the maximum lateral temperature difference that can be sustained
across the subduction girdle is order 100 °C (see text).

To conclude, the “slab method” agrees with the Pangea reference frame for the last 180 Myr. The increasing
difference with the reference frames of Müller et al. (2016, 2019) for times older than 80 Ma results from
the choice they made, in the absence of reliable indications from hot spot trails, to use as initial models the
frames of Torsvik et al. (2008). These are based on the paleomagnetic absolute frame of Steinberger and
Torsvik (2008) in which no longitudinal motion of Africa is assumed.

(c) Van der Meer et al. (2018) identified two slabs, now parts of South America trench system, that are con-
tinuous to (or close to) the CMB. The positions of the slabs in the mantle confirm that there has been no
large migration of this subduction system since about 200 Ma. These slabs originate in active trenches and
can be followed to the lowermost mantle. They call the first the Brasilia slab (their Figure 22, 0–2400 km,
0–200–173 Ma). This is the Peruvian slab of Fukao and Obayashi (2013) who mapped it to a depth of
1,200 km. This slab (astride the northwestern coast of South America at the 1,100 km level in Figure 6) can
be followed downward in Figure 6, except at the 1,900 km level. Figure 6 and Figure 22 of the 2018 paper,
indicate that this slab does not extend beyond a horizontal distant of 1,000 km from the trench (see also
Chen et al., 2019). The second slab, that they call Cocos, is depicted through a profile along 17°N below the
Southern Caribbean (their Figure 33, 0–2,550 km, 0–170–160 Ma). It is the Central America slab of Fukao
and Obayashi (2013) that mapped it to a depth of 2,200 km. The corresponding positive anomaly (west of
the northern Gulf of Mexico at the 1,100 km level in Figure 6) can be followed downward in Figure 6 from
1,100 to 2,700 km. This Cocos profile leads to the same conclusion as for the Brazil slab. The positions of
both slabs are best fit by the Pangea frame and exclude the possibility of larger than 1,000 km westward
migration of South America since Middle Jurassic. We conclude that the Pangea frame of reference is valid
since at least 240 Ma and that, consequently, Pangea was stationary with respect to the deep mantle between
240 and 100 Ma within the limits that we have defined.

3. The TPW of Pangea Revisited

Between 400 and 250 Ma, Pangea migrated northward by about 50°–60° to stabilize within the equatorial
zone during Triassic (Le Pichon). We explore the attribution of this northward migration of Pangea to the
TPW in a way that is consistent with our observational data and with calculations by Creveling et al. (2012).
We first determine the variation of latitude of Pangea with time. This is usually done by estimating the
center of mass of continents (Steinberger & Torsvik, 2008). However, the 20% of the surface of the Pangea
hemisphere occupied by the oceanic surface of the Tethyan realm was entirely situated within the eastern
equatorial area (Figure 1). Consequently, the distribution of continents within Pangea was asymmetric,
with more continental area to the west than to the east. Thus, the center of mass of continents did not

LE PICHON ET AL. 9 of 18
 Tectonics 10.1029/2020TC006585

coincide with the axis of symmetry of the Pangea hemisphere surrounded by the great circle subduction
girdle. Here, we use the hemispheric geometry to define its axis of symmetry in a simple and a fairly accu-
rate way by fitting a great circle to its periphery. We have tested that the resolution of the latitude is better
than 5°.

Figure 2 shows the evolution of the latitude of the axis of symmetry with time from the beginning of the
assembly of Pangea in early Devonian 400 Ma to its demise through breakup in middle Cretaceous 100 Ma.
The latitude is obtained from the reconstructions of Figure 1 derived from paleomagnetic data of Stein-
berger and Torsvik (2008). As discussed earlier, Figure 2 shows that the latitude first increased steadily at
0.4°/Myr between about 400 and 250 Ma. Then, the axis entered the equatorial zone. The northward motion
slowed and eventually reversed near 180 Ma after the latitude had slightly overshot 5°N to slowly return to
the equator. Steinberger and Torsvik (2008) and Torsvik et al. (2012) showed that this increase was accom-
panied by a large (18°–22.5°) CCW rotation of Pangea about its axis during the last part of the ascent (250–
220 Ma). The CCW rotation was followed by a CW one of similar amplitude during the return to the equator
(195–160 Ma). A second set of similar but twice smaller rotations occurred between 160 and 100 Ma. These
authors considered the rotations as TPW and corrected for them. As a consequence, the reconstructions we
use have been corrected for these rotations. Since these corrections were made with respect to the center
of mass of continents rather than the axis of symmetry of Pangea, the actual overshoot of latitude above
the equator is somewhat larger than shown in Figure 2. To summarize, the CCW rotation appeared when
the northward motion of Pangea began slowing down; it reversed to CW rotation when southward motion
started, to finally reverse to CCW once more when southward motion decreased to change to northward
(see Figure 2).

Creveling et al. (2012) also interpreted this set of successive rotations as TPW and considered them to be a
part of an oscillation that extended from 250 to 100 Ma, driven as a result of restoring TPW-induced elastic
(membrane) stresses in the lithosphere forming Earth's equatorial bulge. Note that, although Creveling
et al. (2012) actually fit the proposed extension of rotation to 50 Ma, based on Torsvik et al. (2012, Figure 22),
we do not consider the evolution between 100 and 50 Ma. This is because there is no significant coherent
rotation during this time period. The small CCW rotation detected by Torsvik et al. (2012) between 100 and
50 Ma is due to a large CCW rotation of Africa and a very large CCW rotation from India that closed Teth-
ys while South America and North America moved CW and Eurasia stayed still. Pangea, as a Buridanian
continent, did not exist after 100 Ma. We adopt the interpretation of Creveling et al. (2012) for 250–100 Ma
and consider further this adjustment to be the last phase of a large 50°–60° TPW that brought the Pangea
hemisphere from the southern latitudes to the equatorial zone at a steady rate of 0.4°/Myr between 400 and
250 Ma, as Pangea assembled and progressively acquired its stable configuration.

As noted by Creveling et al. (2012), a phase of TPW that brings Pangea to the equator can be initiated by
an internal mantle load associated with time variable mantle convection. Creveling et al. (2012) suggested
that the formation of a non-equatorial super-swell, associated with a warm mantle plume, could be a po-
tential cause. However, the formation of Pangea itself could lead to a similar effect. The initial formation of
a subduction girdle (Figure 1) would lead to significant mass anomalies which, together with sub-Pangean
thermal isolation (Jellinek et al., 2020; Lenardic et al., 2011), could provide the type of internal mantle
loading that Creveling et al. (2012) argued could drive a TPW event that would bring Pangea to the equator.
Critically, it could lead to both Pangea and its subduction girdle moving toward the equator. The continued
existence of a subduction girdle, particularly after Pangea stabilized at the equator, could lead to the even-
tual demise of Pangea itself. We turn to this issue in the next section.

4. The Subduction Girdle and Pangea Stability

A subduction girdle around Pangea is connected to the slab curtains that could have isolated the sub-
Pangean mantle from the effects of whole mantle convective thermal mixing. As discussed above, there
is evidence that slabs penetrate within the lower mantle (Fukao & Obayashi, 2013), where they sink near
vertically, although more slowly than in the upper mantle (Van der Meer et al., 2010, 2018). Van der Meer
et al. (2018) interpret the absence of “long-term stagnation (>60 Myr)” of slabs in the upper mantle as
demonstrating that there is “a single layer convective system” with a slab deceleration zone between 660 and

LE PICHON ET AL. 10 of 18
 Tectonics 10.1029/2020TC006585

1,500 km. Alternatively, one could interpret their data as reflecting the stagnation of a varying proportion
of slabs at three different levels (410, 660, and 1,000–1,300 km) during times corresponding to a few tens
Myr as shown by Fukao and Obayashi (2013). In any case, these data indicate that (a) most slabs reach the
410 km discontinuity within 10–20 Myr; (b) most of them enter the lower mantle before 70 Myr; and (c)
slabs reach the CMB about 250 ± 30 Myr after their initial subduction. The decrease in the rate of descent by
an average factor of 5 for long continuous slabs implies a corresponding shortening (and related thickening)
by about a factor of 5 (Van der Meer et al., 2018). We can conclude that an essentially continuous slab cur-
tain could be established around Pangea down to 410 km in a geologically short time, but below this depth
the progression was much slower and more complex with quite probable discontinuities, as confirmed by
Figure 6.

The formation of a subduction girdle, and slab curtains, is predicted to affect the thermal structure between
the upper mantle oceanic and continental hemispheres (Jellinek et al., 2020; Lenardic et al., 2011) (Fig-
ure 7). These two hemispheres differed in lithospheric thickness. Panthalassa had a relatively thin oceanic
lithosphere, whereas most of the surface of Pangea was covered by a thick continental lithosphere. This
arrangement can lead to an asymmetry in the amount of heat flowing through the two hemispheres which,
on its own, can result in a warmer subcontinental mantle. However, if convective mixing can occur, then
thermal mantle domains between sub-oceanic and subcontinental regions will tend to homogenize pre-
venting the development of a lateral temperature gradient (Cooper et al., 2013; Lenardic et al., 2005). The
establishment of a near continuous subduction girdle around Pangea would mechanically hamper lateral
mixing in the upper mantle. This effect, in turn, would result in an increase in mantle temperature below
the thick, stagnant, and mostly continental lithosphere of the Pangea hemisphere and a corresponding de-
crease below the thinner oceanic lithosphere of the Panthalassa hemisphere, while the average temperature
of the whole mantle would have stayed constant (Lenardic et al., 2011). Note, however, that the presence
of the Tethyan realm, with its subduction and accretionary zones, over close to 20% of the surface of the
Pangea hemisphere enabled ventilation of heat within Pangea, reducing the aspect ratio of the continental
masses and decreasing the resulting heating of the underlying upper mantle (Le Pichon).

Initial support of sub-Pangean upper mantle isolation, and associated warming, came from enhanced mid-
ocean ridge magmatism and crustal production with Pangea breakup and unusually smooth seafloor in
Jurassic seafloor in the Atlantic (Brandl et al., 2013; Kelemen & Holbrook, 1995; Whittaker et al., 2008).
The hypothesis also comes with the following added predictions: (a) a global enhancing of Earth's cooling
rate during Pangea breakup (the lateral mantle temperature gradient would act as an additional transient
plate-driving force [Jellinek & Lenardic, 2009]); (b) a cooling rate, associated with the Atlantic basin, being
greater than that of the Pacific due to enhanced temperature of the sub-Pangean mantle prior to breakup;
(c) a decaying mantle potential temperature trend over time that should be observable in decreasing oceanic
crustal thickness versus age from the Atlantic. An analysis of global oceanic seismic data reached conclu-
sions in accord with these predictions (Van Avendonk et al., 2017).

Van Avendonk et al. (2017) compiled oceanic crustal thickness data that spanned ages from Pangea breakup
to the present. They connected crustal thickness to mantle temperature at the time of crust formation. The
results showed that the older crust from the Atlantic was unusually thick relative to the modern crust, con-
sistent with a relatively warm mantle below Pangea at the time of its breakup and a progressive dissipation
of the Pangea thermal anomaly (Lenardic, 2017). Van Avendonk et al. (2017) also used their proxy mantle
temperature versus age data to calculate a global mantle cooling rate from Pangea breakup to the present
and found it to be higher than canonical values of mantle cooling over geological time. The global coverage
of their compiled data set allowed them to breakout cooling rates associated with the oceanic lithosphere
from different basins. The analysis indicated that the cooling rate of the Atlantic was greater than that of
the Pacific and was the dominant contribution to the unusually high global cooling rate coming out of Pan-
gea breakup. These inferences are consistent with the hypothesis that slab curtains thermally isolated the
Pangean mantle which lead to the eventual breakup of Pangea and an enhanced period of mantle cooling
(Lenardic, 2017).

Slab curtain induced thermal isolation is also predicted to lead to a warming of the sub-Pangean mantle
and a cooling of the mantle not contained with the subduction girdle (Lenardic et al., 2011). This, in turn,
would contribute to a progressive fall in sea-level, although the effect, in detail, will be model-dependent

LE PICHON ET AL. 11 of 18
 Tectonics 10.1029/2020TC006585

(particularly as factors beyond continental rise will contribute to sea level

variations). Global sea-level reconstructions are contentious in terms of
absolute values, but the monotonic evolution of the upper mantle lateral
temperature variations, associated with the Pangea isolation, is broadly
consistent with the reconstructions of sea level change and continental
flooding (e.g., Miller et al., 2005). We do not discuss here the absolute
values of sea level, but are only interested in the long-term evolution on
which there is a consensus. As shown in the qualitative and voluntarily
schematic curve of Figure 2, before 400 Ma, during the dispersed conti-
nents stage that preceded Pangea, the sea level was high. Then sea level
decreased during the formation stage of Pangea between 400 and 250 Ma
and reached a minimum low between 250 and 180 Ma during the fully
Figure 8. Cartoon showing the problem definition that leads to the developed Pangea stage, followed by a new increase after 180 Ma as the
scaling analysis in the Appendix A and to Equation 1. To be a stable dispersal phase began to reach a new high near 100 Ma. We take this
arrangement, a subduction girdle with a specified perimeter, average
evolution to indicate that the global mixing started to increase 180 Ma
lithosphere thickness dlith and descending at speed Vp must cross the
thickness of the volume of warm, trapped mantle over a time scale when the disruption of the subduction girdle started and that it was fully
that is less than that for warm mantle to spread out. The resistance of realized 100 Ma when the subduction girdle had essentially collapsed.
the subduction girdle to the lateral spread of trapped warm mantle is This evolution in sea level rise is, thus, broadly consistent with the in-
enhanced by buttressing effects related to the viscous stress arising through ferred consequences of the formation and collapse of an upper mantle
the lithosphere bending through a radius Rbend. The fragmentation and
subduction girdle. Coupled climate and tectonic models also indicate that
spread of continental fragments down an expected dynamic topographic
bulge into peripheral subduction zones at speed Vc,spr augments the it is consistent with paleo-climate proxies, which show a shift from Ju-
destabilizing buoyancy effect of spreading warm mantle and can reduce rassic global climate cooling to Cretaceous global warming with Pangea
the resistance of the girdle. breakup – a full discussion of climatic constraints and modeling can be
found in Jellinek et al. (2020).

Finally, that there was a rapid phase of dispersal of the broken pieces of
Pangea, once the peripheral subduction girdle was overridden, is demon-
strated by the absolute frame models of Tetley et al. (2019) that all have a peak in trench migration rates
between 160 and 60 Ma. This time lapse corresponds to the rapid dispersal of Pangea when portions of the
former supercontinent actually overrode the peripheral subduction zone, driving them into compression,
and potentially to trench migration. Indeed, such a relatively rapid transition in subduction regime and
evolving breakdown of the subduction girdle with the continental fragmentation following >100 million
years of stability of this arrangement, is curious. The reconstructed timing of this transition suggests that
this planform became unstable-in response to the emergence and growth of large lateral temperature varia-
tions, to the uplift of the supercontinent in response to subcontinental warming, or both (Figure 8).

We can provide a simple scaling analysis to determine if the magnitude of a lateral temperature gradient
that leads to subduction girdle instability is consistent with observational constraints on the decrease in
potential temperature of the asthenosphere below the Pangea oceans since their initiation (Van Avendonk
et al., 2017). A gradual uplift of the continent and the concomitant growth of the upper mantle temperature
variations leads to lateral differences in hydrostatic pressure that act to drive Pangea disaggregation: The
continent will be drawn down a growing topographic bulge and the warm mantle will spread laterally over
the top cold mantle, dragging continental fragments toward peripheral subduction zones. However, these
driving forces are modulated by two retarding effects that depend on the volumes of warm mantle and the
subduction girdle and on the rate of subduction. First, the extent to which the warm mantle spreads later-
ally will be modulated by the speed at which the descending cold slab material draws this buoyant material
downward. That is, the buoyancy force per area per time (buoyancy flux) carried laterally by the spreading
warm material must be comparable to (or greater) than that which is carried by a descending slab girdle.
Second, the lateral spread of the warm mantle through descending slabs, as well as the breaking and subse-
quent sliding of continental fragments along a gravitational potential, are inhibited through the buttressing
effects related to the lithosphere bending viscously to form the peripheral subduction zone. Consideration
of these driving and retarding effects (Appendix A) leads to a scale for the threshold lateral upper mantle
temperature variation for girdle collapse ΔTcollapse:

LE PICHON ET AL. 12 of 18
 Tectonics 10.1029/2020TC006585

 Vol girdle  Vplate 

ΔTcollapse  ΔTlith 
(1) 1 
2 g ΔTlith H
 Vbend  Vcont   
 Vol warm   

Here, ΔTlith is the temperature difference across the lithosphere that drives
plate-scale mantle convection, the subduction girdle/warm mantle region
 
volume ratio Volgirdle /Vol warm indicates the relative strengths of the lat-
eral and vertical buoyancy fluxes carried by the spreading warm man-
tle and subducting slabs, Vbend  lith /  lith Rb is the rate of plate bend-
ing through a bending radius Rb = 200 km (Conrad & Hager, 1999) and
 
Vcont  Γ Δc ght L / 2 2  m dlithVplate is a scale for the speed of continental
breakup (yielding). For a L = 40,000 km circumference subduction girdle
(Figure 1) the speed of continental breakup depends on the mean dy-
namic topography ℎt and the corresponding gradient Γ of the topographic
bulge down which the continental fragments slide during breakup. This
gradient depends on the vertical and lateral structure and rheology of
the continental and oceanic lithosphere forming the Pangea great circle,
Figure 9. Predicted threshold lateral temperature variations from
Equation 1 as a function of a plausible range in subduction speeds before
which can be complex as implied in the discussion above. However, a
and soon after Pangea breakup (Jellinek et al., 2020). The black curve is the minimum value expected for homogeneous sub-Pangean warming and
prediction for the condition where the trapped mantle spreads faster than a constant lithosphere effective viscosity will scale with the radius of the
the descending slab girdle buoyancy flux carries this buoyancy flux away. Pangea great circle
 Γ min 2 ht / L  0.16. By contrast, a maximum val-
The blue curves show how this condition is modified as the topographic ue, implying potentially that the mean dynamic topography indicates the
gradient down which continental fragments slide Γ is increased from 0.1
(solid blue curve) to 1 (dashed blue curve). See text and Appendix A. In
(minimum) yield strength of the oceanic/continental lithosphere mix-
3 3
calculations, we take m  4000 (kg/m ), Δ c  3000 (kg/m ), H  660 ture
 Γ max  ht / ht  1. Assuming temperature variations are confined
(km), g = 9.8 (m/s2),   5  10 5 (1/°C), ht  1000 (m), and lith  1023 mostly to the low effective viscosity upper mantle with height H over the
(Pa-s). 300 million year time scale of Pangea formation and fragmentation (cf.
Jellinek et al., 2020; Semple & Lenardic, 2018), a lithosphere thickness
dl  200km, reasonable upper mantle physical properties (Appendix A),
mean continental uplift of order ht  1,000m, a global mean potential temperature difference across the
lithosphere ΔTlith  1350 °C, the requisite magnitude lateral temperature variation is between 85 °C and
90 °C (order 100 °C) for pre-breakup speeds Vplate  0.1  1 cm/yr and for 10−1 ≤ Γ ≤ 100. This result is gov-
erned mostly by where the warm upper mantle has sufficient buoyancy to spread faster than the descending
lithosphere sinks (black curve in Figure 9). Extending the range of plate speeds to include the potentially
higher rates during the initial phase of Pangea fragmentation (Jellinek et al., 2020) enables a more thorough
analysis of Equation 1. In particular, this threshold temperature difference is comparatively less for pre-
breakup spreading rates less than around 1 cm/yr where Vcont > Vbend and relatively larger for syn-breakup
spreading rates exceeding a few cm/yr where Vcont < Vbend. Crucially, over the full range of conditions we
explore, the order of magnitude for the threshold lateral temperature variation governing subduction girdle
stability remains 100 °C.

5. Discussion and Concluding Remarks: Pangea Migration, Subduction Girdle

Instability, and Pre-Breakup Volcanism
The analysis of the previous section indicates that the existence and longevity of the subduction girdle and
inferred order 100 °C upper mantle temperature variations prior to breakup are self-consistent. The analysis
is designed to constrain the magnitude of lateral temperature variations that would lead to Pangea insta-
bility and eventual breakup. The results are in line with the conclusions drawn on the basis of geophysical
inferences for differing Pacific-Atlantic basin heatflows (Van Avendonk et al., 2017). Furthermore, they are
broadly consistent with the dynamical models for TPW oscillations discussed above (Creveling et al., 2012),
although our results suggest a more thorough modeling study is now warranted (Figure 2). In particular,
that the subduction girdle can persist as a mass anomaly over the order 100 million years time scales re-
quired for 100 °C lateral temperature to emerge makes it a plausible driving force for the TPW resolved

LE PICHON ET AL. 13 of 18
 Tectonics 10.1029/2020TC006585

in Figures 1 and 2. Future work might address the co-evolutions of the proposed Pangea girdle, TPW and
inevitable consequences for the deep mantle LLSVP structure.

The simple nature of the analysis leading to Equation 1 and Figure 9 does not allow for direct predictions
related to the sequence of events that would occur during a breakup phase. However, in combination with
the evidence of Pangea rotations discussed in the previous section, the analysis does allow for a qualitative
prediction regarding volcanic-tectonic events leading into final Pangea breakup that is a natural direction
for future work. Once stabilized near the equator, the mantle below Pangea would be predicted to be rel-
atively warm as a subduction girdle would have been in place for a geologically significant time period
(Lenardic, 2017). The warm, sub-Pangean mantle could facilitate continental volcanism if it was allowed
to rise, decompress, and melt, in turn. Although the breakup of Pangea would allow for this behavior (Jell-
inek et al., 2020), it is reasonable to expect a more localized mantle melting and surface volcanism before
breakup. Rotations of Pangea around the equator would generate complex tensional and compressional
stresses within the Pangean elastic lithosphere. Given that Pangea is formed from a heterogeneous amal-
gam of cratons and mobile belts, such stresses could have concentrated along intra-supercontinental lith-
osphere strength gradients, opening local rift zones where the stress regime is predominantly tensional
(Dunbar & Sawyer, 1989; Kusznir & Bott, 1977; Vink et al., 1984). With elevated mantle temperatures below,
this condition enhances a potential for continental volcanism. If this picture is correct, then the timing of
volcanic events should coincide with the timing of Pangea rotations.

The following observations support this idea. Figure 2 shows that there is a close relationship between the
latitudinal migration of Pangea and its oscillations around its axis between 250 and 100 Ma as Pangea stabi-
lized rotationally within the equatorial zone. The end of oscillations at 100 Ma coincides with the beginning
of the phase of rapid breakup of Pangea. The two largest flood basalts events, the 250 Ma Siberian traps and
the 201 Ma CAMP traps occurred at the beginning and the end of the first large CCW rotation (250–200 Ma,
Figure 2). The reader is referred to Le Pichon for a geological discussion of both the events and the refer-
ences to the geological literature. These authors pointed out that both magmatic events were preceded by
a phase of extension lasting several tens of Myr. Fracturing due to this extension could have facilitated the
ascent and melting of warm, sub-Pangean mantle, and volcanism at the surface. The extension preceding
the Siberian event occurred at a time of slowing down of the northward migration of Pangea, while the
CCW rotation was accelerating. The extension preceding the CAMP event occurred while the northward
motion of Pangea was reversing itself and immediately before the rotation changed from CCW to CW. These
protracted pre-eruptive periods of extension are plausibly related to the changes in elastic stress regimes
associated with the oscillations of the Pangean landmass across the equator.

It is worth noting that the idea that the Siberian traps and/or the CAMP event are connected to mantle
plumes is not ruled out by our principal hypothesis. Indeed, a long-lived subduction girdle and sub-Pangean
mantle plumes are far from exclusive. Sinking slabs, associated with a subduction girdle, can initiate mantle
plumes as a result of the return flow required by mass conservation and consequent boundary layer insta-
bility above the CMB (Jellinek et al., 2003; Lenardic et al., 2011). An interesting avenue for future work, to
deconvolve the contributions of plumes and upper mantle heating, would be to provide tighter constraints
on the eruptive temperature(s) of the Siberian traps and Camp volcanism, as well as variations in this tem-
perature in space and time. Eruptive temperatures higher than our inferred average sub-Pangean tempera-
ture would be consistent with a significant plume contribution.

A final implication of our study echoes a caution from Le Pichon: “The character of plate tectonics was radi-
cally different between the supercontinent and the dispersed continent stages. We believe then that it is unwise to
assume steady state and infer for example that the analysis of the last 80-Ma plate kinematics is representative
of plate kinematics prior to the Cretaceous revolution.” The analysis of this study reinforces that point as a su-
percontinent constrained to fit within a single hemisphere, with only 20% ocean surface, is a powerful, and
potentially unique, kinematic constraint. Connecting Pangea breakup to the effects of a subduction girdle
adds a dynamic component. The lateral thermal anomaly that develops in the mantle, due to a subduction
girdle, generates a transient plate-driving force that contributes to Pangea breakup and dissipates subse-
quently (Lenardic, 2017). This provides a dynamic mechanism to explain the observational inferences from
Van Avendonk et al. (2017) for an enhanced episode of mantle cooling relative to a secular mean. Thus,
not only are geologically recent plate kinematic inferences not extendable into the Cretaceous, but neither

LE PICHON ET AL. 14 of 18
 Tectonics 10.1029/2020TC006585

are mantle dynamic and associated planetary cooling inferences. A final link that highlights the pitfalls of
extending mean trends back in time, comes from the potential link between Pangea breakup and the unique
climatic signals of the Cretaceous period (Jellinek et al., 2020; Lee et al., 2013). It is worth noting that not all
supercontinents are predicted to have the same degree of effects on the Earth system that Pangea had given
its large area of coverage and its long-lived subduction girdle (Lenardic et al., 2005, 2011).

In conclusion, plate reconstructions corroborate the conclusion of Le Pichon that Pangea was essentially
stationary with respect to the mantle from 250 Ma to its breakup and defined an absolute reference frame
over this time period. Added analysis indicates that from 400 to 250 Ma Pangea migrated toward the equator
at a steady rate of 0.4°/Myr. We attribute this migration to TPW. Our analysis indicates that over the bulk
of its lifetime, Pangea was ringed by a nearly continuous subduction girdle. Through a scaling analysis, we
argued that the effects of the subduction girdle on the evolution of mantle thermal structure lead to the
eventual breakup of Pangea that was preceded by localized episodes of continental volcanism.

Appendix A: Subduction Girdle Collapse

A subduction girdle can become unstable in response to the lateral spreading of warm mantle below Pan-
gea, which is driven by lateral differences in hydrostatic pressure that increases with relative subcontinental
warming. The resilience of a subduction girdle against this destabilization depends on the flux of mantle
material carried by subducting slabs, which can otherwise draw laterally spreading warm mantle down-
wards. This retarding effect is augmented by the additional buttressing effect of the bent slabs forming the
peripheral subduction zone, which retard continental breakup and fragmentation. A Pangean subduction
girdle will remain intact if the time scale for its bending and descent through the upper mantle depth H is
less than the time scales for the spread of trapped sub-Pangean mantle and the fragmentation and sliding of
continental fragments into the peripheral subduction zone.

An additional consideration is that to spread laterally, the warm subcontinental mantle must deform and
penetrate the surrounding, relatively viscous lithospheric sheets forming the subducting girdle. Assuming
that the lithosphere thickness dlith is very small compared to the effective half width or radius of the trapped
mantle volume, the predominant retarding viscous force inhibiting both the spread of warm mantle and
the descent of the subduction girdle are proportional to an appropriate mean mantle viscosity, which is
negligibly influenced by the volume of relatively viscous subducting slab. Furthermore, we also neglect the
retarding buttressing effects related to viscous bending stresses arising as the girdle is deformed around the

laterally spreading sub-Pangean warm upper mantle

 
 L Tlith

 
 L Tlith
, where H > Rb and the appropri-
2
H Rb2
ate mean lithosphere temperature in the mid-upper mantle is higher than that near the surface implying a
lower effective lithosphere viscosity.

Bearing in mind these simplifications, a subduction girdle will collapse only if the buoyancy forces driving
the lateral spread of warm mantle and continental fragmentation are larger than the buoyancy force driving
subduction, which can be augmented by an additional buttressing force associated with the lithosphere
bend to form the subduction zone. For a subduction girdle with circumference L and radius L/2π, girdle
resilience at any time is ensured only where the following condition is met (see text for notation):

  lith dVplate    L2  
L  g lith ΔTlith Hdlith 
(A1)
 2 Rb2    
  g   m ΔTlatVolh  Δc Γ ht  2  
 4  
    

Here, the LHS gives the lithosphere buoyancy and viscous bending forces that depend on the effective vis-
cosity of the lithosphere µlith at the top of the mantle and the plate speed Vp. The RHS gives the buoyancy
force driving the spread of the warm mantle and the requisite lateral differences in hydrostatic pressure
for breakup and Pangea fragmentation (yielding), which is expressed through the topographic gradient
Γ. After some algebra, a critical lateral temperature difference ΔTlat for collapse of the subduction girdle is
(cf. Equation 1, see text):

LE PICHON ET AL. 15 of 18
 Tectonics 10.1029/2020TC006585

 Volgirdle  Vplate 
ΔTcollapse  ΔTlith 
(A2) 1 
2 g ΔTl H
 Vbend  Vcont   
 Vol warm   

This critical temperature for girdle collapse depends strongly on the relative volumes of the descending
subduction girdle and spreading warm mantle. It also depends on the rate at which lithospheric bending
forces are supplied to inhibit the fragmentation of Pangea, compared to the rate at which continental frag-
mentation occurs.

Data Availability Statement

There are no new data. Data all come from publications: Le Pichon et al. (2019), Lenardic et al. (2011),
Müller et al. (2016, 2019), Domeier and Torsvik (2014), Miller et al. (2005), Torsvik et al. (2012), Van der
Meer et al. (2010), Cottaar and Lekic (2016), Şengör and Atayman (2009). They are listed in the references
and cited in the text.

Acknowledgments References
The authors thank Doug Van der
Meer and an anonymous reviewer for Ariza, J. P., Boedo, F. L., Sánchez, M. A., Christiansen, R., Pérez Lujan, S. B., Vujovich, G. I., & Martínez, P. (2018). Structural setting of
detailed reviews. Both the Associate the Chanic orogen (Upper Devonian) at central-western Argentina from remote sensing and aeromagnetic data. Implications in the
Editor and the Editor-in-Chief Taylor evolution of the proto-pacific margin of Gondwana. Journal of South American Earth Sciences, 88, 352–366. https://doi.org/10.1016/j.
Schildgen have been extremely helpful jsames.2018.08.021
during the review process, especially Brandl, P. A., Regelous, M., Beier, C., & Haase, K. M. (2013). High mantle temperatures following rifting caused by continental insulation.
under the near-impossible pandemic Nature Geoscience, 6(5), 391–394. https://doi.org/10.1038/ngeo1758
circumstances, and we are accordingly Burchfiel, B. C., Lipman, P. W., & Zoback, M. L. (1992). The Cordilleran Orogen: Conterminous U.S. (The geology of North America Dec-
grateful. ade of North American geology, Vol. G3). The Geological Society of America. https://doi.org/10.1130/DNAG-GNA-G3
Burke, K., & Torsvik, T. H. (2004). Derivation of large igneous provinces of the past 200 million years from long-term heterogeneities in the
deep mantle. Earth and Planetary Science Letters, 227(3–4), 531–538. https://doi.org/10.1016/j.epsl.2004.09.015
Caravaca, G., Brayard, A., Vennin, E., Guiraud, M., Le Pourhiet, L., Grosjean, A.-S., et al. (2018). Controlling factors for differential sub-
sidence in the Sonoma Foreland Basin (Early Triassic, western USA). Geological Magazine, 155(6), 1305–1329. https://doi.org/10.1017/
S0016756817000164
Chen, Y.-W., Wu, J., & Suppe, J. (2019). Southward propagation of Nazca subduction along the Andes. Nature, 565, 441–447. https://doi.
org/10.1038/s41586-018-0860-1
Conrad, C. P., & Hager, B. H. (1999). Effects of plate bending and fault strength at subduction zones on plate dynamics. Journal of Geophys-
ical Research, 104, 17551–17571. https://doi.org/10.1029/1999jb900149
Cooper, C. M., Moresi, L.-N., & Lenardic, A. (2013). Effects of continental configuration on mantle heat loss. Geophysical Research Letters,
40(11), 2647–2651. https://doi.org/10.1002/grl.50547
Cottaar, S., & Lekic, V. (2016). Morphology of seismically slow lower-mantle structures. Geophysical Supplements to the Monthly Notices of
the Royal Astronomical Society, 207(2), 1122–1136. https://doi.org/10.1093/gji/ggw324
Creveling, J. R., Mitrovica, J. X., Chan, N.-H., Latychev, K., & Matsuyama, I. (2012). Mechanisms for oscillatory true polar wander. Nature,
491(7423), 244–248. https://doi.org/10.1038/nature11571
Dickinson, W. R. (2004). Evolution of the North American Cordillera. Annual Review of Earth and Planetary Sciences, 32, 13–45. https://
doi.org/10.1146/annurev.earth.32.101802.120257
Dickinson, W. R. (2009). Anatomy and global context of the North American Cordillera. Memoir 204. In S. M. Kay, V. A. Ramos, & W. R.
Dickinson (Eds.), Backbone of the Americas: Shallow subduction, plateau uplift, and ridge and terrane collision (pp. 1–29). Geological
Society of America. https://doi.org/10.1130/2009.1204(01)
Domeier, M., & Torsvik, T. H. (2014). Plate tectonics in the late Paleozoic. Geoscience Frontiers, 5, 303–350. https://doi.org/10.1016/
j.g.s.f.2014.01.00210.1016/j.gsf.2014.01.002
Doubrovine, P. V., Steinberger, B., & Torsvik, T. H. (2012). Absolute plate motions in a reference frame defined by moving hot spots in the
Pacific, Atlantic, and Indian oceans. Journal of Geophysical Research, 117, B09101. https://doi.org/10.1029/2011JB009072
Duhem, P. (1958). Le système du monde—Histoire des doctrines cosmologiques de Platon a Copernic (Cinquième Partie: La Physique
Parisienne au XIVe Siècle (suite), Vol. 9, p. 442). Hermann.
Dunbar, J. A., & Sawyer, D. S. (1989). How preexisting weaknesses control the style of continental breakup. Journal of Geophysical Re-
search, 94(B6), 7278–7292. https://doi.org/10.1029/JB094iB06p07278
Du Toit, A. L. (1937). Our wandering continents. Oliver and Boyd.
Dziewonski, A. M., Lekic, V., & Romanowicz, B. A. (2010). Mantle anchor structure: An argument for bottom up tectonics. Earth and
Planetary Science Letters, 299(1–2), 69–79. https://doi.org/10.1016/j.epsl.2010.08.013
Fukao, Y., & Obayashi, M. (2013). Subducted slabs stagnant above, penetrating through, and trapped below the 660 km discontinuity.
Journal of Geophysical Research: Solid Earth, 118(11), 5920–5938. https://doi.org/10.1002/2013JB010466
Gabrielse, H., & Yorath, C. J. (1992). Geology of the Cordilleran orogen in Canada (Geology of Canada Series, No.4, Vol. G2). The Geolog-
ical Society of America. https://doi.org/10.1130/DNAG-GNA-G2
García-Sansegundo, J., Farias, P., Heredia, N., Gallastegui, G., Charrier, R., Rubio-Ordóñez, A., & Cuesta, A. (2014). Structure of the An-
dean Palaeozoic basement in the Chilean coast at 31º 30' S: Geodynamic evolution of a subduction margin. Journal of Iberian Geology,
40(2), 293–308. https://doi.org/10.5209/rev_JIGE.2014.v40.n2.45300

LE PICHON ET AL. 16 of 18
 Tectonics 10.1029/2020TC006585

Glen, R. A., Belousova, E., & Griffin, W. L. (2016). Different styles of modern and ancient non-collisional orogens and implications
for crustal growth: A Gondwanaland perspective. Canadian Journal of Earth Sciences, 53(11), 1372–1415. https://doi.org/10.1139/
cjes-2015-0229
Greene, A. R., Scoates, J. S., Weis, D., & Israel, S. (2005). Flood basalts of the Wrangellia Terrane, southwest Yukon: Implications for the
formation of oceanic plateaus, continental crust and Ni-Cu-PGE mineralization. In D. S. Emond, L. L. Lewis, & G. D. Bradshaw (Eds.),
Yukon exploration and geology (pp. 109–120). Yukon Geological Survey.
Jellinek, A. M., Gonnermann, H. M., & Richards, M. A. (2003). Plume capture by divergent plate motions: Implications for the distribution
of hotspots, geochemistry of mid-ocean ridge basalts, and estimates of the heat flux at the core–mantle boundary. Earth and Planetary
Science Letters, 205(3–4), 361–378. https://doi.org/10.1016/S0012-821X(02)01070-1
Jellinek, A. M., & Lenardic, A. (2009). Effects of spatially varying roof cooling on thermal convection at high Rayleigh number in a fluid
with a strongly temperature-dependent viscosity. Journal of Fluid Mechanics, 629, 109–137. https://doi.org/10.1017/S0022112009006260
Jellinek, A. M., Lenardic, A., & Pierrehumbert, R. T. (2020). Ice, Fire, or Fizzle: The Climate Footprint of Earth's Supercontinental Cycles.
Geochemistry, Geophysics, Geosystems, 21(2). https://doi.org/10.1029/2019GC008464
Kelemen, P. B., & Holbrook, W. S. (1995). Origin of thick, high-velocity igneous crust along the U.S. East Coast Margin. Journal of Geophys-
ical Research, 100(B6), 10077–10094. https://doi.org/10.1029/95JB00924
Kirsch, M., Paterson, S. R., Wobbe, F., Ardila, A. M. M., Clausen, B. L., & Alasino, P. H. (2016). Temporal histories of Cordilleran continen-
tal arcs: Testing models for magmatic episodicity. American Mineralogist, 101(10), 2133–2154. https://doi.org/10.2138/am-2016-5718
Kusznir, N. J., & Bott, M. H. P. (1977). Stress concentration in the upper lithosphere caused by underlying visco-elastic creep. Tectonophys-
ics, 43(3–4), 247–256. https://doi.org/10.1016/0040-1951(77)90119-6
Le Pichon, X., Engör, A. M. C., & İmren, C. (2019). Pangea and the lower mantle. Tectonics, 38(10), 3479–3504. https://doi.
org/10.1029/2018TC005445
Le Pichon, X., & Huchon, P. (1983). Pangée, géoïde et convection. Comptes rendus de l'Académie des Sciences, 296(II), 1313–1320.
Le Pichon, X., & Huchon, P. (1984). Geoid, Pangea and convection. Earth and Planetary Science Letters, 67(1), 123–135. https://doi.
org/10.1016/0012-821X(84)90044-X
Lee, C.-T. A., Shen, B., Slotnick, B. S., Liao, K., Dickens, G. R., Yokoyama, Y., et al. (2013). Continental arc-island arc fluctuations, growth
of crustal carbonates, and long-term climate change. Geosphere, 9(1), 21–36. https://doi.org/10.1130/GES00822.1
Lenardic, A. (2017). A supercontinental boost. Nature Geoscience, 10(1), 4–5. https://doi.org/10.1038/ngeo2862
Lenardic, A., Moresi, L., Jellinek, A., & Manga, M. (2005). Continental insulation, mantle cooling, and the surface area of oceans and
continents. Earth and Planetary Science Letters, 234(3–4), 317–333. https://doi.org/10.1016/j.epsl.2005.01.038
Lenardic, A., Moresi, L., Jellinek, A. M., O'Neill, C. J., Cooper, C. M., & Lee, C. T. (2011). Continents, supercontinents, mantle thermal mix-
ing, and mantle thermal isolation: Theory, numerical simulations, and laboratory experiments. Geochemistry, Geophysics, Geosystems,
12(10), Q10016. https://doi.org/10.1029/2011GC003663
Lucassen, F., & Franz, G. (2005). The early Palaeozoic orogen in the Central Andes: A non-collisional orogen comparable to the Cenozoic
high plateau? In A. R. M. Vaughan, P. T. Leat, & R. J. Pankhurst (Eds.), Terrane processes at the margins of Gondwana (No. 246, Vol.
246, pp. 257–273). Geological Society. https://doi.org/10.1144/GSL.SP.2005.246.01.09
Madrid, R. J., Poole, F. G., & Wrucke, C. T. (1992). Rocks of the Antler orogeny—The Roberts Mountains allochthon. In B. C. Burchfiel,
P. W. Lipman, & M. L. Zoback (Eds.), The Cordilleran orogen: Conterminous U.S. (Geology of North America, Vol. G-3, pp. 28–34).
Geological Society of America.
Miller, K. G., Kominz, M. A., Browning, J. V., Wright, J. D., Mountain, G. S., Katz, M. E., et al. (2005). The Phanerozoic record of global
sea-level change. Science, 310(5752), 1293–1298. https://doi.org/10.1126/science.1116412
Milne, A. J., & Millar, I. L. (1991). Mid-Palaeozoic basement of eastern Graham Land and its relation to the Pacific margin of Gondwana.
In M. R. A. Thomson, J. A. Crame, & J. W. Thomson (Eds.), Geological evolution of Antarctica (Proceedings of the Fifth International
Symposium on Antarctic Earth Sciences, pp. 335–339). Cambridge University Press.
Müller, R. D., Seton, M., Zahirovic, S., Williams, S. E., Matthews, K. J., Wright, N. M., et al. (2016). Ocean basin evolution and global-scale
plate reorganization events since Pangea breakup. Annual Review of Earth and Planetary Sciences, 44, 107–138. https://doi.org/10.1146/
annurev-earth-0600115-01221110.1146/annurev-earth-060115-012211
Müller, R. D., Zahirovic, S., Williams, S. E., Cannon, J., Seton, M., Bower, D. J., et al. (2019). A global plate model including lithospheric
deformation along major rifts and orogens since the Triassic. Tectonics, 38(6), 1884–1907. https://doi.org/10.1029/2018TC005462
Nokleberg, W. J., Parfenov, L. M., Monger, J. W. H., Norton, I. O., Khanchuk, A. I., Stone, D. B., et al. (2000). Phanerozoic tectonic evolution
of the Circum North Pacific (p. 122). USGS, Professional Paper 1626.
Norton, I. O., & Sclater, J. G. (1979). A model for the evolution of the Indian Ocean and the breakup of Gondwanaland. Journal of Geophys-
ical Research, 84(B12), 6803–6830. https://doi.org/10.1029/JB084iB12p06803
Schepers, G., Van Hinsbergen, D. J. J., Spakman, W., Kosters, M. E., Boschman, L. M., & McQuarrie, N. (2017). South-American plate
advance and forced Andean trench retreat as drivers for transient flat subduction episodes. Nature Communications, 8, 15249. https://
doi.org/10.1038/ncomms15249
Scotese, C. (2017). Plate Tectonics @ Night by C.R. Scotese (PALEOMAP Project Report 112171A). https://doi.org/10.13140/
RG.2.2.14706.91846
Semple, A. G., & Lenardic, A. (2018). Plug flow in the Earth's asthenosphere. Earth and Planetary Science Letters, 496, 29–36. https://doi.
org/10.1016/j.epsl.2018.05.030
Şengör, A. M. C. (1984). The Cimmeride orogenic system and the tectonics of Eurasia (No. 195). Geological Society of America Special Paper.
https://doi.org/10.1130/SPE195-p1
Şengör, A. M. C. (2003). The large wavelength deformations of the lithosphere: Materials for a history of the evolution of thought from the
earliest times to plate tectonics (Vol. 196). Geological Society of America.
Şengör, A. M. C., & Atayman, S. (2009). The Permian extinction and the Tethys: An exercise in global geology (Vol. 448, x+96). Geological
Society of America Special Paper. https://doi.org/10.1130/2009.2448
Şengör, A. M. C., & Natal'in, B. A. (1996). Palaeotectonics of Asia: Fragments of a synthesis. In A. Yin, M. Harrison, & M. (Eds.), The tec-
tonic evolution of Asia, Rubey Colloquium (pp. 486–640). Cambridge University Press.

of the continental lithosphere of Central Asia between ∼750 and ∼130 Ma Ago. Annual Review of Earth and Planetary Sciences, 46,
Şengör, A. M. C., Natal'in, B. A., Sunal, G., & van der Voo, R. (2018). The tectonics of the Altaids: Crustal growth during the construction

439–494. https://doi.org/10.1146/annurev-earth-060313-054826
Snedden, J. W., & Liu, C. (2010). A compilation of Phanerozoic sea-level change, coastal onlaps and recommended sequence designations
(Search and discovery, Article ID. 40594). Online H-Journal for E&P Geoscientists.

LE PICHON ET AL. 17 of 18
 Tectonics 10.1029/2020TC006585

Spikings, R. A., Cochrane, R., Vallejo, C., Villagomez, D., van der Lelij, R., Paul, A., & Winkler, W. (2019). Latest Triassic to early Cretaceous
tectonics of the northern Andes: Geochronology, geochemistry, isotopic tracing, and thermochronology. In B. K. Horton, & A. Folguera
(Eds.), Andean tectonics (pp. 173–208). Elsevier. https://doi.org/10.1016/B978-0-12-816009-1.00009-5
Stampfli, G. M., Hochard, C., Vérard, C., Wilhem, C., & vonRaumer, J. (2013). The formation of Pangea. Tectonophysics, 593, 1–19. https://
doi.org/10.1016/j.tecto.2013.02.037
Steinberger, B., Sutherland, R., & O'Connell, R. J. (2004). Prediction of Emperor-Hawaii seamount locations from a revised model of global
plate motion and mantle flow. Nature, 430(6996), 167–173. https://doi.org/10.1038/nature02660
Steinberger, B., & Torsvik, T. H. (2008). Absolute plate motions and true polar wander in the absence of hotspot tracks. Nature, 452(7187),
620–623. https://doi.org/10.1038/nature06824.PMID/18385737
Tetley, M. G., Williams, S. E., Gurnis, M., Flament, N., & Müller, R. D. (2019). Constraining absolute plate motions since the Triassic. Jour-
nal of Geophysical Research: Solid Earth, 124(7), 7231–7258. https://doi.org/10.1029/2019JB017442
Torsvik, T. H., Müller, R. D., Van der Voo, R., Steinberger, B., & Gaina, C. (2008). Global plate motion frames: Toward a unified model.
Reviews of Geophysics, 46(3), RG3004. https://doi.org/10.1029/2007rg000227
Torsvik, T. H., Smethurst, M. A., Burke, K., & Steinberger, B. (2006). Large igneous provinces generated from the margins of the large
low-velocity provinces in the deep mantle. Geophysical Journal International, 167(3), 1447–1460. https://doi.org/10.1111/j.13656-
246X.2006.03158.x10.1111/j.1365-246x.2006.03158.x
Torsvik, T. H., Van der Voo, R., Preeden, U., Mac Niocaill, C., Steinberger, B., Doubrovine, P. V., et al. (2012). Phanerozoic polar wander,
palaeogeography and dynamics. Earth-Science Reviews, 114(3–4), 325–368. https://doi.org/10.1016/j.earscirev.2012.06.007
Van Avendonk, H. J. A., Davis, J. K., Harding, J. L., & Lawver, L. A. (2017). Decrease in oceanic crustal thickness since the breakup of
Pangaea. Nature Geoscience, 10(1), 58–61. https://doi.org/10.1038/ngeo2849
Van der Meer, D. G., Spakman, W., Van Hinsbergen, D. J. J., Amaru, M. L., & Torsvik, T. H. (2010). Towards absolute plate motions con-
strained by lower-mantle slab remnants. Nature Geoscience, 3(1), 36–40. https://doi.org/10.1038/ngeo708
Van der Meer, D. G., van Hinsbergen, D. J. J., & Spakman, W. (2018). Atlas of the underworld: Slab remnants in the mantle, their sinking
history, and a new outlook on lower mantle viscosity. Tectonophysics, 723, 309–448. https://doi.org/10.1016/j.tecto.2017.10.004
Vink, G. E., Morgan, W. J., & Zhao, W.-L. (1984). Preferential rifting of continents: A source of displaced terranes. Journal of Geophysical
Research, 89(B12), 10072–10076. https://doi.org/10.1029/JB089iB12p10072
Wernicke, B., & Klepacki, D. W. (1988). Escape hypothesis for the Stikine block. Geology, 16(5), 461–464. https://doi.
org/10.1130/0091-7613(1988)016<0461:EHFTSB>2.3.CO;2
Whittaker, J. M., Müller, R. D., Roest, W. R., Wessel, P., & Smith, W. H. F. (2008). How supercontinents and superoceans affect seafloor
roughness. Nature, 456(7224), 938–941. https://doi.org/10.1038/nature07573
Williams, S., Flament, N., Dietmar Müller, R., & Butterworth, N. (2015). Absolute plate motions since 130 Ma constrained by subduction
zone kinematics. Earth and Planetary Science Letters, 418, 66–77. https://doi.org/10.1016/j.epsl.2015.02.026
Zhong, S., Zhang, N., Li, Z. X., & Roberts, J. H. (2007). Supercontinent cycles, true polar wander, and very long wave-length mantle
convection. Earth and Planetary Science Letters, 261(3–4), 551–564. https://doi.org/10.1016/j.epsl.2007.0704910.1016/j.epsl.2007.07.049

LE PICHON ET AL. 18 of 18