Content uploaded by W. Spakman
Author content
All content in this area was uploaded by W. Spakman on Nov 09, 2018
Content may be subject to copyright.
Greater India Basin hypothesis and a two-stage
Cenozoic collision between India and Asia
Douwe J. J. van Hinsbergena,b,1, Peter C. Lippertc,d, Guillaume Dupont-Nivete,f,g, Nadine McQuarrieh,
Pavel V. Doubrovinea,b, Wim Spakmani, and Trond H. Torsvika,b,j,k
aPhysics of Geological Processes, University of Oslo, Sem Sælands vei 24, NO-0316 Oslo, Norway; bCenter for Advanced Study, Norwegian Academy of
Science and Letters, Drammensveien 78, 0271 Oslo, Norway; cDepartment of Geosciences, University of Arizona, Tucson, AZ 85721; dDepartment of Earth
and Planetary Sciences, University of California, Santa Cruz, CA 95064; eGéosciences Rennes, Unité Mixte de Recherche 6118, Université de Rennes 1,
Campus de Beaulieu, 35042 Rennes Cedex, France; fPaleomagnetic Laboratory Fort Hoofddijk, Department of Earth Sciences, University of Utrecht,
Budapestlaan 17, 3584 CD, Utrecht, The Netherlands; gKey Laboratory of Orogenic Belts and Crustal Evolution, Ministry of Education, Peking University,
Beijing 100871, China; hDepartment of Geology and Planetary Science, University of Pittsburgh, Pittsburgh, PA 15260; iDepartment of Earth Sciences,
University of Utrecht, Budapestlaan 4, 3584 CD, Utrecht, The Netherlands; jCenter for Geodynamics, Geological Survey of Norway, Leiv Eirikssons vei 39,
7491 Trondheim, Norway; and kSchool of Geosciences, University of the Witwatersrand, WITS 2050, Johannesburg, South Africa
Edited by B. Clark Burchfiel, Massachusetts Institute of Technology, Cambridge, MA, and approved March 29, 2012 (received for review October 19, 2011)
Cenozoic convergence between the Indian and Asian plates pro-
duced the archetypical continental collision zone comprising the
Himalaya mountain belt and the Tibetan Plateau. How and where
India–Asia convergence was accommodated after collision at or be-
fore 52 Ma remains a long-standing controversy. Since 52 Ma, the
two plates have converged up to 3,600 35 km, yet the upper
crustal shortening documented from the geological record of Asia
and the Himalaya is up to approximately 2,350-km less. Here we
show that the discrepancy between the convergence and the
shortening can be explained by subduction of highly extended
continental and oceanic Indian lithosphere within the Himalaya be-
tween approximately 50 and 25 Ma. Paleomagnetic data show that
this extended continental and oceanic “Greater India”promontory
resulted from 2,675 700 km of North–South extension between
120 and 70 Ma, accommodated between the Tibetan Himalaya and
cratonic India. We suggest that the approximately 50 Ma “India”–
Asia collision was a collision of a Tibetan-Himalayan microconti-
nent with Asia, followed by subduction of the largely oceanic
Greater India Basin along a subduction zone at the location of
the Greater Himalaya. The “hard”India–Asia collision with thicker
and contiguous Indian continental lithosphere occurred around
25–20 Ma. This hard collision is coincident with far-field deforma-
tion in central Asia and rapid exhumation of Greater Himalaya crys-
talline rocks, and may be linked to intensification of the Asian
monsoon system. This two-stage collision between India and Asia
is also reflected in the deep mantle remnants of subduction imaged
with seismic tomography.
continent–continent collision ∣mantle tomography ∣plate reconstructions ∣
Cretaceous
The present geological boundary between India and Asia is
marked by the Indus–Yarlung suture zone, which contains de-
formed remnants of the ancient Neotethys Ocean (1, 2) (Fig. 1).
North of the Indus–Yarlung suture is the southernmost continen-
tal fragment of Asia, the Lhasa block. South of the suture lies the
Himalaya, composed of (meta)sedimentary rocks that were
scraped off now-subducted Indian continental crust and mantle
lithosphere and thrust southward over India during collision. The
highest structural unit of the Himalaya is overlain by fragments of
oceanic lithosphere (ophiolites).
We apply the common term Greater India to refer to the part
of the Indian plate that has been subducted underneath Tibet
since the onset of Cenozoic continental collision. A 52 Ma mini-
mum age of collision between northernmost Greater India and
the Lhasa block is constrained by 52 Ma sedimentary rocks in
the northern, “Tibetan”Himalaya that include detritus from
the Lhasa block (3). This collision age is consistent with indepen-
dent paleomagnetic evidence for overlapping paleolatitudes for
the Tibetan Himalaya and the Lhasa blocks at 48.66.2Ma
(Fig. 2; SI Text) as well as with an abrupt decrease in India–Asia
convergence rates beginning at 55–50 Ma, as demonstrated by
India–Asia plate circuits (e.g., ref. 4). Structural (5) and strati-
graphic (6) data show that ophiolites were emplaced over the
Tibetan Himalaya in the latest Cretaceous (approximately
70–65 Ma), well before the Tibetan Himalaya–Lhasa collision.
Paleomagnetic data suggest that these ophiolites formed at equa-
torial paleolatitudes (7).
Motion between continents that border the modern oceans
is quantified through time using plate reconstructions based
on marine magnetic anomalies. The Eurasia-North America-
Africa-India plate circuit demonstrates 2;860 30 and 3;600
35 km of post-52 Ma India–Asia convergence for the western and
eastern Himalayan syntaxes, respectively (4). It has long been re-
cognized that the amount of upper crustal shortening since 52 Ma
reconstructed from the geology of the Himalaya and Asia ac-
counts for only approximately 30–50% of this total convergence
(8) (Fig. 1). Previously proposed solutions for this “shortening
deficit”include major unrecognized shortening in Siberia (8, 9),
a>1;000-km eastward extrusion of Indochina from an original
position within Tibet that would lead to >2;000 km of Cenozoic
intra-Asian shortening (10–12), or a much younger Tibetan Hi-
malaya–Lhasa collision (13). However, geologically recon-
structed shortening of approximately 1,050–600 km (from west
to east) within and north of the Pamir and the Tibetan plateau
without invoking major Indochina extrusion (14) is consistent
with 1;100 500 km of paleomagnetically constrained conver-
gence between the Indus–Yarlung suture and Eurasia since
approximately 50 Ma (15). The excess convergence should there-
fore mostly have been accommodated in the Himalaya, to the
south of the Indus–Yarlung suture.
The Himalaya consists of upper continental crust that was
decoupled from now-subducted Greater India (1, 2), which is
generally considered to have formed the contiguous margin of
northern India (2, 16–18). The Himalaya is divided into three
tectonostratigraphic zones (Fig. 1). From north to south, these
include the non- to low-metamorphic grade sedimentary rocks
of the Tibetan Himalaya, separated by the South Tibetan detach-
ment from the igneous and high-grade metamorphic Greater
Himalaya, which overlies the low-grade, internally thrusted
Lesser Himalaya along the Main Central thrust (1, 2) (Fig. 1).
Author contributions: D.J.J.v.H. designed research; D.J.J.v.H., P.C.L., G.D.-N., N.M., P.V.D.,
W.S., and T.H.T. performed research; D.J.J.v.H. and P.C.L. analyzed data; and D.J.J.v.H.,
P.C.L., G.D.-N., and N.M. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1To whom correspondence should be addressed. E-mail: d.v.hinsbergen@fys.uio.no.
This article contains supporting information online at www.pnas.org/lookup/suppl/
doi:10.1073/pnas.1117262109/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1117262109 PNAS ∣May 15, 2012 ∣vol. 109 ∣no. 20 ∣7659–7664
EARTH, ATMOSPHERIC,
AND PLANETARY SCIENCES
Composite balanced cross-sections across the Himalaya docu-
mented approximately 500–900 km of shortening (19, 20). These
estimates are the sum of (i) shortening in the Tibetan Himalaya
(since the Paleogene); (ii) the amount of overlap between
the Greater Himalaya and the Lesser Himalaya along the Main
Central thrust that was largely established between approximately
25–20 and 15–10 Ma; and (iii) the amount of shortening within
the Lesser Himalaya since approximately 15–10 Ma (2, 19, 20).
The main uncertainty in the Himalayan shortening estimates
are associated with the Greater Himalaya. Intense Miocene de-
formation has effectively overprinted any older structures in the
Greater Himalaya, but well-dated prograde mineral growth and
magmatism in the Greater Himalaya shows evidence for burial
and heating between 45 and 25 Ma (21–23). The amount of short-
ening accommodated within or below the Greater Himalaya prior
to the Miocene therefore remains geologically unconstrained, but
could be considerable. Detrital zircon studies of the Lesser,
Greater, and Tibetan Himalaya suggest that their Neoproterozoic
to lower Paleozoic (Cambrian-Ordovician) stratigraphies are si-
milar (16, 24–26), suggesting that the net effect of post-Ordovi-
cian tectonics within the Himalaya was limited to perhaps several
hundreds of kilometers of shortening.
The amount of India–Asia convergence since 52 Ma (up to
3;600 35 km, ref. 4) exceeds the estimated total crustal short-
ening within Asia and the Himalaya by up to 2,350 km (14, 19, 20)
(Fig. 1). In this paper, we will consider several possible explana-
tions of the observed discrepancy and propose a tectonic scenario
that can reconcile the disparate estimates and identify the struc-
ture(s) that accommodated the ‘missing’convergence.
Size of Greater India Through Time
The only available technique to quantify past motions between
continental blocks that are not connected through a passive mar-
gin to oceanic basins is paleomagnetism. A wealth of paleomag-
netic data from the Ordovician, Triassic, and lower Cretaceous
rocks of the Tibetan Himalaya consistently demonstrates minor
net North–South (N-S) motion of the Tibetan Himalaya relative
to India since these times (SI Text). The youngest rocks that
demonstrate a relatively small Greater India are approximately
120 Mya old and show a net convergence between India and
the Tibetan Himalaya of 2.15.5°, or 233 877 km (Fig. 2).
After adding the modern width of Greater India (i.e., the Hima-
laya, approximately 250 km N-S), these data suggest that Greater
India in early Cretaceous time was not larger than approximately
900 km (SI Text). This conclusion is consistent with pre-Cretac-
eous Gondwana plate reconstructions (17, 27) as well as with the
notion that the Paleozoic and older stratigraphies of the Hima-
layan zones are correlative, suggesting that during their deposi-
tion Greater India formed a contiguous continental margin (16,
24–26). By assuming that Greater India remained only several
hundreds of kilometers wide until collision with Asia, Aitchison
et al. (13, 28) demonstrated that the leading edge of Greater India
would have passed the equator approximately 55 Ma ago, and sug-
gested that the 55–50 Ma collision record (what is widely regarded
as the Tibetan Himalaya–Asia collision) resulted from just the ob-
duction of ophiolites onto the leading, Tibetan-Himalayan edge of
Greater India. By also assuming negligible Cenozoic intra-Asian
shortening, their model proposed that the India–Asia collision oc-
curred at approximately 34 Ma. Paleomagnetic evidence from
upper Cretaceous (approximately 68 Ma) and Paleocene (approxi-
mately 59 Ma) rocks from the Tibetan Himalaya (29–31) and cra-
tonic India, however, demonstrate that during the late Cretaceous
to Paleocene, the Tibetan Himalaya was separated by 22.03.0°
of latitude (2;442 333 km N-S) from India (Fig. 2). These pa-
leomagnetic data pass both fold and reversal paleomagnetic field
tests at high confidence (SI Text), demonstrating their prefolding,
primary magnetic acquisition. This much larger size of Greater In-
dia at the time of collision than during the early Cretaceous results
in a paleomagnetically determined collision age of 48.66.2Ma
(Fig. 2; SI Text), consistent with the 52 Ma age of first arrival of
Lhasa-derived sediments in the Tibetan Himalaya (3). This large
dimension of Greater India is also consistent with the position of
the southern Asian margin at the time of collision based on re-
storation of intra-Asian deformation (14). Therefore, paleomag-
netic data show that between 118 and 68 Ma, Greater India
became extended and the Tibetan Himalaya drifted 24.16.3°
(2;675 699 km N-S) northward relative to cratonic India (Fig. 2),
followed by convergence of a similar magnitude after collision.
Cretaceous Greater Indian N-S Extension
Cretaceous extension within Greater India has previously
been inferred from sedimentary facies changes in the Tibetan
Himalaya (17, 18) and from lower Cretaceous (140–100 Ma)
alkali-basaltic volcaniclastic sediments with a geochemistry inter-
preted to record intracontinental rifting (32, 33). The 2;675
700 km of N-S extension between 118 and 68 Ma, inferred from
paleomagnetic data (Fig. 2), requires minimum extension rates of
40–67 mm∕y. Such rates are typical for midoceanic ridges (pre-
drift continental extension rates rarely exceed 20 mm∕y, ref. 34),
and the magnitude of extension is an order of magnitude larger
than that of typical extended continental margins (34). Thus, at
least one oceanic basin—the Greater India Basin(s) (GIB)—
must have formed between the Tibetan Himalaya and cratonic
Fig. 1. Tectonic map of the India–Asia collision zone. Bars represent the
amount of post-50 Ma India–Asia convergence, and the amount of intra-Asian
(14), Himalayan (19, 20), and missing shortening along the collision zone. We
calculated the Himalayan shortening deficit using a reconstruction of Asian
deformation (14) embedded in global plate circuits (4, 34). Our maximum es-
timate of undocumented convergence uses (i) plate convergence estimates
since 50 Ma, 3;600 35 km for the eastern Himalayan syntaxis (4); (ii) approxi-
mately 600 km of shortening reconstructed in eastern Tibet since 50 Ma
(14); and (iii) up to 650 km of shortening reconstructed from the eastern
Himalaya (51), resulting in a 2,350 km deficit. Since approximately 25–20 Ma,
the total amount of plate convergence is up to 1,300–1,000 km in the eastern
Himalaya (4). Shortening since approximately 25 Ma in the Himalaya, along
the main central thrust (MCT) and in the Lesser Himalaya (LH), is approxi-
mately 400–700 km (20), and in Tibet, approximately 200–300 km (14), leaving
a modest shortening deficit of several hundreds of kilometers. We ascribe this
deficit to uncertainties in the Tibet reconstruction, uncertainties in the timing
of Tibetan Himalaya (TH) shortening, and the fact that balanced cross-sections
provide minimum estimates. STD, South Tibetan Detachment; IYSZ, Indus–
Yarlung Suture Zone; MBT, main boundary thrust.
7660 ∣www.pnas.org/cgi/doi/10.1073/pnas.1117262109 van Hinsbergen et al.
India, separating a microcontinent from cratonic India (Fig. 3).
The paleogeography of extended Greater India may have con-
sisted of one or more deep basins alternating with (stretched)
continental fragments, similar to Mediterranean paleogeography
(35). The Cretaceous extension that would have opened the GIB
(2;675 700 km) encompasses the approximately 2,350 km of
plate convergence that remains undocumented in the geological
record of the Himalaya (Fig. 1). Numerical models demonstrate
that continental lithosphere can subduct if its buoyant upper crust
becomes decoupled from the subducting continent, but denser
oceanic or thinned continental lithosphere can subduct without
accretion (36). These modeling results are consistent with the ob-
servation that many convergent margins with oceanic subduction
are nonaccreting (37). If 2;675 700 km of extension created up
to approximately 2,350 km of oceanic crust and highly extended
continental margins in the GIB, subduction of that crust thus pro-
vides a straightforward explanation for the discrepancy between
convergence and shortening.
Subduction History and Mantle Tomography
Subduction of the GIB following the approximately 50 Ma
Tibetan Himalaya–Lhasa collision is consistent with seismic
tomographic images of the mantle beneath India and Tibet.
These images reveal three prominent velocity anomalies that
have been interpreted as subducted Indian plate lithosphere
(38–40) (Fig. 4). The deepest anomaly, at approximately 900 km
and greater depth, is generally considered subducted Neotethyan
oceanic lithosphere, detached sometime after the Tibetan
Himalaya–Lhasa collision (38–40). A shallower anomaly between
approximately 400- and 850-km depth would hence be Cenozoic
in age (38–40); it projects directly below the reconstructed posi-
tion of the Tibetan Himalaya between 50 and 25 Ma (Fig. 4). A
third conspicuous body imaged horizontally below Tibet over a
distance of approximately 500 km from the Himalayan front
represents Indian continental lithosphere that has underthrust
Asia since the last phase of slab break-off (39, 41). Taking Asian
shortening (14) into account, this horizontal body represents the
last 10–15 Ma of India–Asia convergence.
Seismic tomographic images also show anomalies in the lower
mantle at equatorial latitudes (Fig. 4A), generally interpreted as a
result of Cretaceous intraoceanic subduction (40, 42). Aitchison
et al. (28) suggested that this anomaly resulted from intraoceanic
subduction that terminated with ophiolite obduction onto the
leading edge of Greater India. We agree that this anomaly repre-
sents the relict of an intraoceanic subduction zone, but we note
that the much larger size of Greater India in late Cretaceous time
shown by paleomagnetic data (Fig. 2) positions the northern mar-
gin of the Tibetan Himalaya above this equatorial lower mantle
anomaly around 70 Ma, not 55 Ma (Figs. 3 and 4). This reconstruc-
tion is consistent with structural and stratigraphic evidence for
ophiolite obduction onto the Tibetan Himalaya at approximately
70 Ma (5, 6).
Fig. 2. Paleolatitude evolution of India, Greater India, Greater Asia, and Asia. (A) Paleomagnetic poles for the Tibetan Himalaya (magneta upright triangle)
and cratonic India (white downward triangle) at 118 Ma, compared to the Indian apparent polar wander path (APWP) (blue squares), indicating that the net N-
S drift of the Tibetan Himalaya relative to India since 118 Ma was negligible; Greater India was not larger than approximately 900 km (SI Text). Ninety-five
percent confidence intervals of the pole positions are also shown. (Band C) Same as A, but at 68 and 59 Ma, respectively, indicating 2;675 699 km of north-
ward drift of the Tibetan Himalaya relative to continental India compared to their 118 Ma position. (D) Paleolatitudes of a reference site (29°N, 88°E) located on
the present-day position of the Indus–Yarlung Suture Zone in Eurasian, Greater Asian, Tibetan-Himalayan, and Indian reference frames. Numbers correspond
to paleomagnetic poles described and listed in the SI Text. Pl, Pliocene; Mio, Miocene; Oligo, Oligocene; Paleoc., Paleocene; MCT, main central thrust; STD,
South Tibetan detachment. Age uncertainties are based on the age of the units determined from either radiometric dates or geologic stages (52, 53), and
latitude uncertainties are calculated from the corresponding poles at the 95% confidence level. Question mark next to estimate 7 indicates that this pole may
insufficiently average paleosecular variation. See SI Text for details.
van Hinsbergen et al. PNAS ∣May 15, 2012 ∣vol. 109 ∣no. 20 ∣7661
EARTH, ATMOSPHERIC,
AND PLANETARY SCIENCES
Location of a Paleogene Subduction Zone Within the
Himalaya
The collision between the Tibetan Himalayan microcontinent
and Asia at approximately 50 Ma, and the subsequent closure
of the Greater India Basin south of this microcontinent, requires
a Paleogene to Miocene subduction zone south of (i.e., structu-
rally below) the Tibetan Himalaya. The location of this subduc-
tion zone must be located somewhere within the modern
Himalayan thrust belt, and therefore has implications for the in-
terpretation of existing geological evidence on the contempora-
neous evolution of the Himalaya (Fig. 3).
Since the onset of thrusting of the Greater Himalaya over the
Lesser Himalaya along the Main Central thrust approximately
25–20 Ma ago, cumulative shortening in the Himalaya (19, 20)
and Asia (14) is close to contemporaneous India–Asia conver-
gence (4) (approximately 1,000 km). The continental clastic rocks
of the Lesser Himalaya demonstrate continuous subduction of
continental lithosphere and accretion of its upper crust to the
Himalaya in this time interval. Thus, the discrepancy between
predicted convergence and measured shortening is largely con-
centrated between 50 and approximately 25–20 Ma. Because
the Tibetan Himalaya collided with Asia at or before approxi-
mately 50 Ma, and the deformation of the Lesser Himalaya is
Miocene in age (43), 50–25 Ma convergence must have been ac-
commodated structurally below (i.e., south of) the Tibetan, but
above (i.e., north of) the Lesser Himalaya. As mentioned before,
the total displacement along the Main Central thrust, as well as
convergence accommodated within the Greater Himalaya, re-
mains unknown due to severe Miocene deformation; however,
well-dated prograde mineral growth and magmatism (21–23,
44) in the Greater Himalaya shows evidence for burial and heat-
ing between 45 and 25 Ma. If Cretaceous extension opened a sin-
gle Greater Indian Ocean separating a microcontinent from India
(Fig. 3), then these Paleogene metamorphic ages would suggest
that the Greater and Tibetan Himalaya both belonged to a single
microcontinent that collided with and accreted to Asia around
50 Ma. During the 50–25 Ma subduction of the Indian Plate
(i.e., the GIB), the Greater and Tibetan Himalaya would then
have thickened and metamorphosed as part of the compressed
overriding plate (i.e., together with the Tibetan plateau to the
north). Subduction—without accretion—would have been con-
centrated along a precursor of the Main Central thrust. If the
GIB had a more complex paleogeography, then the Greater
Himalaya could contain a ductile thrust stack (duplex) of deep
marine rocks underplated and metamorphosed below the
Tibetan Himalaya throughout the Paleogene. Pre-25-Ma pro-
grade metamorphic ages in the Greater Himalaya vary consider-
ably, from 45 Ma migmatitic rocks in gneiss domes in the Tibetan
Himalaya (21) to 25 Ma eclogites in Nepal (23); these ages may
represent the different times at which Greater Himalayan rocks
were structurally buried in the duplex. In any case, the GIB suture
lies either structurally within or below the Greater Himalaya.
Microcontinent–Continent Collision (50 Ma) and India–Asia
Collision (25–20 Ma)
In our model, the approximately 50 Ma India–Asia collision
epresents collision of an extended microcontinental fragment
(that contained the rocks now found in the Tibetan Himalaya)
and continental Asia. This collision led to orogeny (14, 45),
but was followed by ongoing oceanic subduction of the GIB.
The hard continent–continent collision followed approximately
25 Ma later with the arrival of much less extended, contiguous
continental Indian lithosphere to the collision zone, leading to
increased coupling at the plate contact (46). Such increased cou-
pling is consistent with (i) the similarity between shortening and
convergence since 25–20 Ma; (ii) the onset of Greater Himalayan
extrusion along the South Tibetan detachment and Main Central
thrust (2); and (iii) the simultaneous and sudden onset of far-field
Fig. 3. Plate reconstructi on of the India–Asia collision (4, 54). A small Greater
India (A) was extended in the Cretaceous (B), leading to a “soft”collision and
ongoing subduction around 50 Ma (C) and a “hard”collision with thick, con-
tinuous Indian lithosphere between 25 and 20 Ma (D). MCT, main central
thrust; Md, madagascar; STD, South Tibetan detachment; Sy, Seychelles.
7662 ∣www.pnas.org/cgi/doi/10.1073/pnas.1117262109 van Hinsbergen et al.
deformation into Central Asia (14) (e.g., deforming and uplifting
the Tien Shan, ref. 47) (Fig. 1). Comparison of our plate and de-
formation reconstructions with seismic tomography suggests that
this latter collision was followed by slab break-off at approxi-
mately 15–10 Ma and horizontal underthrusting of India below
Tibet thereafter. The onset of this last phase of India–Asia colli-
sion is contemporaneous with a period of outward growth and
extension of the Tibetan plateau (10, 12, 48).
The recognition of major Cretaceous extension in India and a
multistage India–Asia collision history no longer requires models
invoking major continental extrusion from Tibet predicting
>2;000 km of Cenozoic intra-Asian convergence (10–12), but
is consistent with kinematic and paleomagnetic reconstructions
showing <1;000 km of intra-Asian shortening (8, 14, 15, 49).
Our model does not require younger initial collision ages (13)
and is in line with geological data from the Indus–Yarlung suture
zone (3), suggesting an approximately 50 Ma Tibetan Himalaya–
Asia collision. It identifies the Greater Himalaya as the exhumed
mid- to lower-continental crust that resided directly above a sub-
duction zone from approximately 50 to 25 Ma. Finally, the hard
collision was followed by a substantial increase in erosion rates
within the Himalayan system, which may reflect a more virgorous
South Asian monsoon (50), highlighting another potential link
between geodynamic processes forming the Tibetan-Himalayan
mountain belt and climate evolution.
ACKNOWLEDGMENTS. Paul Kapp, Pete DeCelles, Douwe van der Meer, and
Rob van der Voo are thanked for discussion and comments on previous ver-
sions of this manuscript. D.J.J.v.H. and T.H.T. acknowledge financial support
from Statoil (SPlates Model Project). D.J.J.v.H., P.V.D., and T.H.T. were sup-
ported by the Center for Advanced Study of the Norwegian Academy of
Science and Letters. T.H.T., P.V.D., and W.S. acknowledge support through
the TOPO-Europe TOPO-4D program. Reasearch of W.S. was carried out in
context of the Netherlands Research School of Integrated Solid Earth
Sciences. G.D.-N. acknowledges support from the Dutch Netherlands Organi-
zation for Scientific Research, the French Centre National de la Recherche
Scientifique, and the Chinese National Science Foundation. N.M. acknowl-
edges National Science Foundation Grant EAR 0738522. P.C.L. acknowledges
National Science Foundation Grant EAR 1008527. D.J.J.v.H. thanks the Carne-
gie Institution for Science in Washington, DC for their hospitality.
1. Gansser A (1964) The Geology of the Himalayas (Wiley Interscience, New York).
2. Hodges KV (2000) Tectonics of the Himalaya and southern Tibet from two perspec-
tives. Geol Soc Am Bull 112:324–350.
3. Najman Y, et al. (2010) The timing of India–Asia collision: Geological, biostratigraphic
and palaeomagnetic constraints. J Geophys Res 115:B12416.
4. van Hinsbergen DJJ, Steinberger B, Doubrovine PV, Gassmöller R (2011) Acceleration
and deceleration of India–Asia convergence since the Cretaceous: Roles of mantle
plumes and continental collision. J Geophys Res 116:B06101.
5. Searle MP, Treloar PJ (2010) Was Late Cretaceous—Palaeocene obduction of ophiolite
complexes the primary cause of crustal thickening and regional metamorphism in
the Pakistan Himalaya? The Evolving Continents: Understanding Processes of Conti-
nental Growth, eds TM Kusky, MG Zhai, and W Xiao (Geological Society, London),
Special Publication, Vol 338, pp 345–359.
6. Cai F, Ding L, Yue Y (2011) Provenance analysis of upper Cretaceous strata in the
Tethys Himalaya, southern Tibet:Implications for timing of India–Asia collision. Earth
Planet Sci Lett 305:195–206.
7. Abrajevitch AV, et al. (2005) Neotethys and the India–Asia collision: Insights from a
palaeomagnetic study of the Dazhuqu ophiolite, southern Tibet. Earth Planet Sci Lett
233:87–102.
8. Dewey JF, Shackleton RM, Chang C, Sun Y (1988) The tectonic evolution of the
Tibetan Plateau. Philos Trans R Soc A Math Phys Eng Sci 327:379–413.
9. Guillot S, et al. (2003) Reconstructing the total shortening history of the NW
Himalaya. Geochem Geophys Geosyst 4:1064.
10. Royden LH, Burchfiel BC, Van der Hilst RD (2008) The geological evolution of the
Tibetan Plateau. Science 321:1054–1058.
11. Replumaz A, Tapponnier P (2003) Reconstruction of the deformed collision zone
Between India and Asia by backward motion of lithospheric blocks. J Geophys
Res 108:2285.
12. Tapponnier P, et al. (2001) Oblique stepwise rise and growth of the Tibet Plateau.
Science 294:1671–1677.
13. Aitchison JC, Ali JR, Davis AM (2007) When and where did India and Asia collide?
J Geophys Res 112:B05423.
14. van Hinsbergen DJJ, et al. (2011) Restoration of Cenozoic deformation in Asia, and
the size of Greater India. Tectonics 30:TC5003.
15. Dupont-Nivet G, van Hinsbergen DJJ, Torsvik TH (2010) Persistently shallow paleo-
magnetic inclinations in Asia: Tectonic implications for the Indo-Asia collision.
Tectonics 29:TC5016.
16. Myrow PM, et al. (2003) Integrated tectonostratigraphic analysis of the Himalaya and
implications for its tectonic reconstruction. Earth Planet Sci Lett 212:433–441.
17. Garzanti E (1999) Stratigraphy and sedimentary history of the Nepal Tethys Himalaya
passive margin. J Asian Earth Sci 17:805–827.
18. Gaetani M, Garzanti E (1991) Multicyclic history of the northern India continental
margin (northwestern Himalaya). AAPG Bull 75:1427–1446.
19. DeCelles PG, Robinson DM, Zandt G (2002) Implications of shortening in the
Himalayan fold-thrust belt for uplift of the Tibetan Plateau. Tectonics 21:TC1062.
20. Long SP, McQuarrie N, Tobgay T, Grujic D (2011) Geometry and crustal shortening of
the Himalayan fold-thrust belt, eastern and central Bhutan. Geol Soc Am Bull
123:1427–1447.
21. Pullen A, Kapp P, DeCelles PC, Gehrels GE, Ding L (2011) Cenozoic anatexis and
exhumation of Tethyan Sequence rocks in the Xiao Gurla Range, Southwest Tibet.
Tectonophysics 501:28–40.
22. Lee J, Whitehouse MJ (2007) Onset of mid-crustal extensional flow in southern Tibet:
Evidence from U/Pb zircon ages. Geology 35:45–48.
23. Corrie SL, Kohn MJ, Vervoort JD (2010) Young eclogite from the Greater Himalayan
Sequence, Arun Valley, eastern Nepal: P-T-t path and tectonic implications. Earth
Planet Sci Lett 289:406–416.
24. Long S, et al. (2011) Tectonostratigraphy of the Lesser Himalaya of Bhutan: Implica-
tions for the along-strike stratigraphic continuity of the northern Indian margin.
Geol Soc Am Bull 123:1258–1274.
Fig. 4. Seismic tomographic images of subducted slabs compared to India–Asia restorations. (Aand C) Sixty and 30 Ma plate reconstructions of the India–Asia
collision zone in a moving hotspot reference frame (54) projected above seismic tomographic images, derived from the UU-P07 model (55), at 1,175 (0.6%
contours) and 710 km (1% contours) depth, respectively, with previously identified intraoceanic subduction zones (40, 42). (B) Vertical cross-section (0.6%
contours) of the mantle below the India–Asia collision zone indicating the three major Indian plate mantle anomalies. TH, Tibetan Himalaya; LH, Lesser
Himalaya; GH, Greater Himalaya.
van Hinsbergen et al. PNAS ∣May 15, 2012 ∣vol. 109 ∣no. 20 ∣7663
EARTH, ATMOSPHERIC,
AND PLANETARY SCIENCES
25. Yin A (2006) Cenozoic tectonic evolution of the Himalayan orogen as constrained
by along-strike variation of structural geometry, exhumation history, and foreland
sedimentation. Earth Sci Rev 76:1–131.
26. Webb AAG, et al. (2011) Cenozoic tectonic history of the Himachal Himalaya (north-
western India) and its constraints on the formation mechanism of the Himalayan oro-
gen. Geosphere 7:1013–1061.
27. Ali JR, Aitchison JC (2005) Greater India. Earth Sci Rev 72:169–188.
28. Aitchison J, Xia X, Baxter AT, Ali JR (2011) Detrital zircon U-Pb ages along the
Yarlung-Tsangpo suture zone, Tibet: Implications for oblique convergence and colli-
sion between India and Asia. Gondwana Res 20:691–709.
29. Dupont-Nivet G, Lippert P, van Hinsbergen DJJ, Meijers MJM, Kapp P (2010) Paleo-
latitude and age of the Indo-Asia collision: Paleomagnetic constraints. Geophys J Int
182:1189–1198.
30. Patzelt A, Li H, Wang J, Appel E (1996) Palaeomagnetism of Cretaceous to Tertiary
sediments from southern Tibet: Evidence for the extent of the northern margin of
India prior to the collision with Eurasia. Tectonophysics 259:259–284.
31. Yi Z, Huang B, Chen J, Chen L, Wang H (2011) Paleomagnetism of early Paleogene
marine sediments in southern Tibet, China: Implications to onset of the India–Asia
collision and size of Greater India. Earth Planet Sci Lett 309:153–165.
32. Hu X, et al. (2010) Provenance of Lower Cretaceous Wölong volcaniclastics in the
Tibetan Tethyan Himalaya: Implications for the final breakup of Eastern Gondwana.
Sediment Geol 223:193–205.
33. Jadoul F, Berra F, Garzanti E (1998) The Tethys Himalayan passive margin from Late
Triassic to Early Cretaceous (South Tibet). J Asian Earth Sci 16:173–194.
34. Torsvik TH, Müller RD, Van der Voo R, Steinberger B, Gaina C (2008) Global plate
motion frames: Toward a unified model. Rev Geophys 46:RG3004.
35. Stampfli GM, Hochard C (2009) Plate tectonics of the Alpine realm. Ancient Orogens
and Modern Analogues, eds JB Murphy, JD Keppie,and AJ Hynes (Geological Society,
London), Special Publication, Vol 327, pp 89–111.
36. Capitanio FA, Morra G, Goes S, Weinberg RF, Moresi L (2010) India–Asia convergence
driven by the subduction of the Greater Indian continent. Nat Geosci 3:136–139.
37. Clift P, Vannucchi P (2004) Controls on tectonic accretion versus erosion in subduction
zones: Implications for the origin and recycling of the continental crust. Rev Geophys
42:RG2001.
38. van der Voo R, Spakman W,Bijwaard H (1999) Tethyan slabs under India. Earth Planet
Sci Lett 171:7–20.
39. Replumaz A, Negredo AM, Villasenor A, Guillot S (2010) Indian continental subduc-
tion and slab break-off during Tertiary collision. Terra Nova 22:290–296.
40. Hafkenscheid E, Wortel MJR, Spakman W (2006) Subduction history of the Tethyan
region derived from seismic tomography and tectonic reconstructions. J Geophys Res
111:B08401.
41. Zhao J, et al. (2010) The boundary between the Indian and Asian tectonic plates
below Tibet. Proc Natl Acad Sci USA 107:11229–11233.
42. van der Meer DG, Spakman W, van Hinsbergen DJJ, Amaru ML, Torsvik TH (2010)
Toward absolute plate motions constrained by lower mantle slab remnants. Nat
Geosci 3:36–40.
43. Caddick MJ, et al. (2007) Burial and exhumation history of a Lesser Himalayan schist:
Recording the formation of an inverted metamorphic sequence in NW India. Earth
Planet Sci Lett 264:375–390.
44. Imayama T, et al. (2012) Two-stage partial melting and contrasting cooling history
within the Higher Himalayan Crystalline Sequence in the far-eastern Nepal Himalaya.
Lithos 134–135:1–22.
45. Wang C, et al. (2008) Constraints on the early uplift history of the Tibetan Plateau.
Proc Natl Acad Sci USA 105:4987–4992.
46. Molnar P, Lyon-Caen H (1988) Some simple physical aspects of the support, structure,
and evolution of mountain belts. Spec Pap Geol Soc Am 218:179–207.
47. Sobel ER, Chen J, Heermance RV (2006) Late Oligocene-Early Miocene initiation of
shortening in the Southwestern Chinese Tian Shan: Implications for Neogene short-
ening rate variations. Earth Planet Sci Lett 247:70–81.
48. Lease RO, et al. (2011) Middle Miocene reorganization of deformation along the
northeastern Tibetan Plateau. Geology 39:359–362.
49. Johnson MRW (2002) Shortening budgets and the role of continental subduction
during the India–Asia collision. Earth Sci Rev 59:101–123.
50. Clift PD, et al. (2008) Correlation of Himalayan exhumation rates and Asian monsoon
intensity. Nat Geosci 1:875–880.
51. Yin A, et al. (2010) Geologic correlation of the Himalayan orogen and Indian craton:
Part 2. Structural geology, geochronology, and tectonic evolution of the Eastern
Himalaya. Geol Soc Am Bull 122:360–395.
52. Gradstein FM, et al. (1994) A Mesozoic time scale. J Geophys Res 99:24051–24074.
53. Cande SC, Kent DV (1995) Revised calibration of the geomagnetic polarity timescale
for the Late Cretaceaous and Cenozoic. J Geophys Res 100:6093–6095.
54. O’Neill C, Muller D, Steinberger B (2005) On the uncertainties in hot spot reconstruc-
tions and the significance of moving hot spot reference frames. Geochem Geophys
Geosyst 6:Q04003.
55. Amaru ML (2007) Global travel time tomography with 3-D reference models. PhD
thesis Geol Ultraiectina, 274 (Utrecht Univ, Utrecht, The Netherlands).
7664 ∣www.pnas.org/cgi/doi/10.1073/pnas.1117262109 van Hinsbergen et al.
Supporting Information
van Hinsbergen et al. 10.1073/pnas.1117262109
SI Text
Paleomagnetic Constraints on the Age of the India–Asia Collision and
Size of Greater India. The relative positions of continents that are
directly connected by ocean basins are determined by plate cir-
cuits based on marine magnetic anomalies and fracture zones (1).
The paleolatitude of a continent is paleomagnetically determined
from global synthetic apparent polar wander paths (APWPs),
which are obtained by rotating all available high-quality paleo-
magnetic poles using Euler poles from the plate circuit to a cho-
sen reference location (2, 3). This paper uses the most recent
global synthetic APWP of Torsvik et al. (3), which covers the last
320 Ma. The global synthetic APWP (3) did not include the minor
changes in the Indo-Atlantic plate circuit of van Hinsbergen et al.
(4) used in our plate reconstruction. These changes, however, do
not significantly influence the analyses below. From 320 to
440 Ma, the analysis below uses the APWPs for Gondwana and
Laurasia of Torsvik and van der Voo (5), and before 440 Ma, uses
the Gondwana poles from Torsvik et al. (6). These APWPs are
determined for individual continents based on paleomagnetic
poles of that continent alone.
The global synthetic APWPs do not include poles from conti-
nental blocks that are not bounded (or not anymore) by oceans,
but are now instead separated by deformed belts, such as the
northern edge of Greater India (i.e., the Himalaya) and the
southern edge of Greater Asia (i.e., the Tibetan orogen). These
blocks are frequently internally deformed by shortening and may
have experienced regional and local vertical axis rotations during
tectonism. Although paleomagnetic data from these deformed
blocks cannot be used in APWP compilations, the inclination
component can be used to calculate the paleolatitude of these
blocks through time for comparison with plates in the plate
circuit.
We assess the relative paleolatitude of the Indian and Siberian
cratons that bound the India–Asia collision zone by using the
APWPs described above to calculate the predicted paleolatitude
of a reference location presently located on the Indus–Yarlung
Suture Zone (IYSZ) for each continent (29°N, 88°E) (Table S1).
We test the accuracy of these APWP predictions with paleolati-
tudes for the reference location calculated from robust lava-
based paleomagnetic poles from cratonic India. We then add
paleomagnetic information from rocks directly bordering the
IYSZ, that is, paleomagnetic poles determined from lavas and
sediments from the Lhasa and Qiangtang blocks and the Tibetan
Himalaya. The selection of poles from the Lhasa and Qiangtang
blocks, the Tibetan Himalaya, and cratonic India are described
below, with sampling localities shown in Fig. S1.
Paleomagnetic Data Procedures. A robust paleomagnetic pole is
calculated from primary, well-determined, independent readings
of the geomagnetic field over a time period that is long enough to
average the effect of paleosecular variation of the geomagnetic
field, but short enough so as to not integrate appreciable plate
motion. It is generally assumed that continuous sections of sedi-
mentary rocks average secular variation, such that a stratigraphic
series of site mean directions, or a magnetostratigraphic column,
will provide a good characterization of the magnetic field. This
assumption is only valid, however, if the preserved magnetic
directions (i) are primary and (ii) have not been biased by sedi-
mentary inclination shallowing caused by particle settling and
compaction (7, 8). When magnetic directions are collected at
the site level (in contrast to a continuous stratigraphic succession
in magnetostratigraphic studies), site mean directions must be of
comparable reliability before they can be transformed to virtual
geomagnetic poles (VGPs) to calculate the average paleomag-
netic pole at a study location. In the compilations below, we apply
the following filtering criteria to site mean directions determined
from sedimentary rocks, by excluding sites that (i) are not used by
the original authors, if reason for exclusion is provided; (ii) con-
tain directions of mixed polarity [i.e., a single sample of the geo-
magnetic field (site mean direction) should be short enough to
exclude a reversal event]; (iii ) are characterized by fewer than
five samples; (iv) have α95 values (the radius of the 95% confi-
dence interval on the mean direction) greater than 20°; (v) or de-
viate more than 45° from the locality mean direction.
Paleomagnetic results from lavas must be treated differently
than results from sediments, because the processes and time-
scales by and over which magnetic directions are recorded are
principally different. Several independent readings of the geo-
magnetic field are required to accurately characterize its time-
averaged behavior, which is fundamental to determining a robust
paleomagnetic pole. Cooling of a single, thin (1–3 m) lava unit
will record a geologically instantaneous measurement of the geo-
magnetic field. Furthermore, subsequent lavas can extrude within
a short time period (years), such that several stratigraphically
successive lavas often record the same magnetic field direction,
meaning that each lava unit may not provide an independent
reading of the geomagnetic field. Hence, paleomagnetic data
from volcanic units must first be filtered so that all site mean di-
rections are of comparably high quality, and then, if the sampling
strategy permits, stratigraphically successive site mean directions
that are statistically indistinguishable should be combined into
direction groups. In the compilations below, we apply the follow-
ing filtering criteria to site mean directions determined from lavas
by excluding sites that (i) are not used by the original authors, if
reason for exclusion is provided; (ii) contain directions of mixed
polarity; (iii) are characterized by fewer than five samples;
(iv) have k values (Fisher’s precision parameter, ref. 9) less than
50; (v) or deviate more than 45° from the mean site mean di-
rection.
This filtering procedure is commonly used by the geomagnetic
field community (10, 11). The assignment of direction groups
follows from the original authors of each of the studies described
below, and in general follows the procedures described by
Dupont-Nivet et al. (12) and Lippert et al. (13).
Paleomagnetic studies of the present and ancient magnetic
field indicate that VGPs fit a Fisherian distribution better than
the parent directions (11, 14). Therefore, we transform site
mean and direction group directions into VGPs and calculate
the Fisher mean VGP for each sampling locality. We evaluate
the averaging of secular variation of a mean lava-based VGP
by comparing the calculated dispersion (S) of VGPs to the dis-
persion observed in ancient lavas that yield good estimates of a
geocentric axial dipole field when averaged (11, 14), and to the
dispersion determined from paleosecular variation models (15).
If the dispersion of the VGPs is consistent with dispersion ob-
served in young lava successions, then we conclude that the sam-
ple set represents and has averaged secular variation.
We calculate the expected inclination and its uncertainty at
the reference location on the IYSZ (29°N, 88°E), as well as the
expected paleolatitude and corresponding uncertainty of the re-
ference locality from each tilt-corrected locality paleomagnetic
pole following Butler (16). Paleomagnetic poles, predicted incli-
nations, and paleolatitudes are listed in Table S1. All age assign-
ments follow the original authors unless otherwise noted and are
van Hinsbergen et al. www.pnas.org/cgi/doi/10.1073/pnas.1117262109 1of10
keyed to the same timescales as used in the global synthetic
APWP (3), that is, Gradstein et al. (17) and Cande and Kent
(18), unless the section is dated using radiometric techniques.
Paleomagnetic Poles Used in This Study.
Paleomagnetic poles from
cratonic India.
Two robust paleomagnetic poles from the Rajmahal
(118 1Ma) and Deccan (65 1Ma) Traps on cratonic India
are recalculated based on the most recent paleomagnetic data.
These poles are used to test the accuracy of the global synthetic
APWP (3) for predicting Indian paleolatitudes.
The Deccan Traps.
The end-Cretaceous Deccan Traps in western
India have been the focus of many paleomagnetic (see reviews
by Vandamme et al., ref. 19, and Chenet et al. ref. 20) and geo-
chronologic (see reviews by Hofmann et al., ref. 21, and Chenet et
al., ref. 22) studies. High-precision 40 Ar∕39 Ar geochronology
consistently yield eruption ages of 65 1Ma for the Deccan
Traps, indicating that the >3-km thick unit erupted rapidly,
but still spans several Ma. Paleomagnetic data from the Deccan
Traps (19, 20, 23) (Tables S2–S4) should therefore sample suffi-
cient time to accurately characterize secular variation of the geo-
magnetic field but not so much time that the rapid northward
motion of India is of concern.
For the exceptionally large and high-quality dataset of Chenet
et al. (20), we recalculated the single eruptive event directions
assigned by the original authors by (i) selecting only site mean
directions that met our quality criteria (Table S2), then (ii) recal-
culating direction group directions as assigned by the original
authors using these filtered data (Table S3), and (iii) finally by
recalculating the single eruptive events as assigned by the original
authors in pole space. These new single eruptive event directions,
as well as site mean directions from Vandamme et al. (19, 23)
(Table S4) that met our quality criteria, are shown in Table S5.
Paleomagnetic directions from these latter two studies have
not been binned into direction groups due to lack of detailed stra-
tigraphic information and may contain duplicate spot readings of
the geomagnetic field. The 65 VGPs resulting from the site mean
directions that pass our data quality filter provide a paleomag-
netic pole for the Deccan Traps located at 37.1°N, 285.3°E
(A95 ¼3.3°, K¼29.7,S¼14.9°). Although the position of this
paleomagnetic pole is not statistically different from previous
Deccan poles (19, 20), the VGP scatter is more consistent with
secular variation studies of young lavas at the same latitude (11).
The 65 1Ma paleolatitude of the reference location on the
IYSZ calculated from this pole is −21.93.3°N, in contrast to
the more northerly −15.12.4°N latitude calculated from the
coeval Indian APWP pole. We prefer to use the Deccan pole
as the cratonic Indian reference pole over the coeval Indian
APWP pole because it is based on direct measurements of well-
dated lava units on the Indian craton and is one of the most ro-
bust paleomagnetic poles in the geologic record. The Deccan
pole therefore overcomes potential uncertainties in the global
synthetic APWP caused by (i) small errors in the relative plate
motions between continents; (ii) the inclusion of paleomagnetic
poles in the compilation with unrecognized errors, for instance
due to sedimentary inclination shallowing biases or insufficient
averaging of secular variation in volcanic-based datasets; and
(iii) the relatively low number of individual poles used in each
averaging window. Further discussion on uncertainties in syn-
thetic APWPs can be found elsewhere (3, 12, 13, 24, 25).
The Rajmahal Traps.
The Aptian Rajmahal Traps of eastern India
have been accurately dated to 118 1Ma using single-crystal ar-
gon geochronology (26). Paleomagnetic results from the Rajma-
hal Traps are published by Klootwijk (27) and Sherwood and
Malik (28) and are presented in Ta b l e S 6 . Thirty-three of the
48 site mean directions met our quality criteria, resulting in a pa-
leomagnetic pole located at 8.1°N, 297.9°E (A95 ¼4.1°,
K¼37.4,S¼13.3°). The calculated dispersion of the VGPs
can be straightforwardly explained by secular variation typical
for the Cretaceous Superchron (14). The 118 1Ma paleolati-
tude of the reference location on the IYSZ calculated from
this pole for cratonic India is −43.04.1°N. This paleolatitude
is statistically indistinguishable from the −43.42.6°N latitude
predicted by the global synthetic APWP (3).
Paleomagnetic Poles from Southern Tibet (Greater Asia).
Eocene-
Oligocene lavas.
Paleomagnetic data from well-dated upper
Eocene-lower Oligocene (35 3Ma) basaltic lavas from the
Qiangtang block are described by Lippert et al. (13). Thirty-three
site mean directions from three sampling localities have been
filtered and binned into 20 independent direction groups with a
dispersion that is consistent with modern secular variation stu-
dies. These lavas yield a maximum likelihood estimate of paleo-
latitude of 28.73.7°N at a present reference site at 33°N, 88°E
on the Lhasa-Qiangtang suture. Because there has been negligi-
ble north-south-directed upper crustal shortening or rotation
within the Lhasa block throughout the Neogene (29–31), we
can transfer this paleolatitude directly to our reference position
on the IYSZ, 4° of latitude to the south, resulting in a paleola-
titude estimate for the reference location of 24.73.8°N at ap-
proximately 35 3Ma.
Paleogene lavas.
Basaltic to andesitic lavas, tuffs, and volcanoclas-
tic units of the upper Paleocene to lower Eocene (56–47 Ma) Lin-
zizong formation (LZF) blanket a large area of the southern
Lhasa block (32, 33). These volcanic units have been the focus
of several paleomagnetic studies because of their potential to re-
veal the paleolatitude of the southern margin of Asia at approxi-
mately 56–47 Ma (Fig. S1) (34–39). Paleolatitude estimates for
the reference location on the IYSZ from these individual studies
range from approximately 12°N to as high as approximately 32°N,
mainly as a result of the low number of independent site mean
directions in each study.
We compiled all reliable data from the six studies of the upper
LZF (i.e., Pana and upper Nianbo formations) listed above. After
applying the data quality filter described above, excluding 37 of
89 reported lava site mean directions, the mean paleomagnetic
pole in tilt-corrected coordinates is located at 80.1°N, 248.6°E
(K¼26.8,A95 ¼3.9°, S¼15.7°, n¼52). The 52 site mean di-
rections used to calculate this filtered paleomagnetic pole include
data from each of the six independent studies, pass a reversal test
at 95% confidence (class C: Ycalculated ¼7.6°, Ycritical ¼10.9°;
ref. 40), and pass a regional fold test at 95% and 99% confidence
(X2test of McFadden, ref. 41) (Table S7). Importantly, the cal-
culated dispersion of this filtered VGP is consistent with known
dispersion caused by paleosecular variation (11). We conclude
that this paleopole more accurately characterizes the time-
averaged behavior of the geomagnetic field than any previous in-
dividual study of the LZF. The 56–47 Ma paleolatitude of the re-
ference site on the IYSZ predicted from this pole is 19.63.9°N.
Some recent studies (36) have reported reconnaissance paleo-
magnetic data from the lower LZF (approximately 64–60 Ma),
suggesting paleolatitudes as low as 6.6°N 8.5°N (reference
locality 30.0°N, 91.2°E). These anomalously low latitudes would
imply unlikely large and fast northward motion (>1;500 km) of
southern Tibet from 60 to 50 Ma, the evidence for which is either
lacking or remains unrecognized. These results are based on a
limited number of sites from which our data quality filter removes
too many site mean directions to provide a statistically robust pa-
leomagnetic pole. We therefore choose to not incorporate these
data into our compilation.
Upper Cretaceous redbeds, Maxiang.
Sun et al. (42) report paleo-
magnetic data from the 100–72 Ma Shexing formation of redbeds
that outcrop along the Lhasa-Golmud highway near Maxiang.
van Hinsbergen et al. www.pnas.org/cgi/doi/10.1073/pnas.1117262109 2of10
Site mean directions calculated from the high-temperature char-
acteristic remanent magnetization (ChRM) component pass a
fold test at high confidence. We evaluated the distribution of
high-temperature specimen directions for inclination shallowing
following the elongation/inclination (E∕I) method of Tauxe and
Kent (15). Samples with mean angular deviations of ChRM di-
rections >15° and with reversed polarity ChRM directions were
removed from the population to exclude poorly resolved or tran-
sitional directions. The Vandamme cutoff procedure for identify-
ing outliers in the direction population (43) removed another
seven samples, resulting in a mean direction of the Shexing
formation with declination ¼343.4°, inclination ¼31.2°(α95 ¼
2.8°, k¼26.7,n¼100)(Fig.S3A). We calculated a flattening
factor of 0.66 (0.53 and 0.91 95% confidence interval), resulting
in an E∕I corrected inclination of 40.5° (33.3–47.7° 95% confi-
dence interval) (Fig. S3 Cand D). These results indicate approxi-
mately 9° of inclination shallowing in the fluvial-lacustrine redbeds
of the Shexing formation, which is typical of these facies in Asia
(12, 13). The formation mean direction calculated from the
E∕I-corrected directions is declination ¼343.4°, inclination ¼
41.6°, α95 ¼2.8°, k¼26.0,n¼100 (Fig. S3B), with a correspond-
ing pole at 74.1°N, 342.7°E (A95 ¼2.3). The calculated paleolati-
tude for the reference location on the IYSZ is 23.82.3°N at
100–72 Ma, which is consistent with E∕I-corrected data from
upper Cretaceous redbeds from the Penbo regions of the Lhasa
terrane (see below).
Upper Cretaceous redbeds, Penbo.
We use paleomagnetic data from
the nonmarine Takena formation and calculations of a late Cre-
taceous (105–90 Ma, ref. 44) paleolatitude for the Lhasa Block of
Tan et al. (34). The distribution of the sample directions from this
dataset are consistent with inclination shallowing that is readily
attributed to sedimentary depositional processes. These direc-
tions have been corrected by Tan et al. (34) using the E∕I method
(15). The paleomagnetic pole calculated from these tilt and in-
clination shallowing corrected sediments is located at 79.6°N,
329.9°E (A95 ¼2.2). The paleolatitude of the reference location
on the IYSZ in Lhasa-terrane coordinates calculated from this
pole is 23.72.2°N at approximately 105–90 Ma.
Lower Cretaceous Lavas.
Paleomagnetic results from volcanic rocks
of Cretaceous age have been reported in two recent studies. Sun
et al. (45) describe 15 site mean directions from the 114.2
1.1Ma Woronggou formation rhyolites near Deqing, on the west
shore of Nam Co (Table S8). Ten site mean directions have either
too few samples defining the site mean direction or precision
parameters <50, leaving only five directions for further analysis.
Chen et al. (46) describe 19 site mean directions from 130–
110 Ma tuffs and lavas of the Zenong Group near the town of
Cuoqin in the central Lhasa terrane (Table S8). Nine of these site
mean directions have precision parameters <50°, and another di-
rection is considered an outlier by the original authors, leaving
only nine site mean directions to describe the geomagnetic field
at this locality. We combine the filtered results of Sun et al. (45)
and Chen et al. (46) to calculate a lava-based paleomagnetic pole
for the Lower Cretaceous of the Lhasa terrane. Although both
localities show similar inclinations, the declinations vary by ap-
proximately 30–40°, indicating significant rotations between the
localities (Fig. S4). We follow Doubrovine and Tarduno (24)
and Lippert et al. (13), and calculate a maximum likelihood vir-
tual geomagnetic latitude (VGL) using the Arason and Levi
method (47). First, we calculate the VGPs for the filtered site
mean directions and then calculate the VGL of the reference site
on the IYSZ. These VGLs are then used as input to calculate the
maximum likelihood latitude for the reference site from these
data: 16.23.6°N at approximately 130–110 Ma. We note, how-
ever, that the dispersion of these VGLs is only 7.0°, indicating
that paleosecular variation has been undersampled by these
igneous units, and that this paleolatitude estimate may not be ro-
bust. Additional paleomagnetic studies focusing on Cretaceous
volcanic rocks from the Lhasa terrane are clearly warranted.
General comment on paleomagnetism of Cretaceous rocks of the Lha-
sa terrane.
We note that many historical paleomagnetic studies of
Cretaceous rocks of the Lhasa terrane have been excluded from
our compilation; for example, Achache et al. (35), Pozzi et al.
(48), Westphal and Pozzi (49), Lin and Watts (50), and Chen
et al. (51). The reasons for these exclusions are because (i) most
of these studies produced very small sample sets with few site
mean directions such that paleosecular variation and inclination
shallowing could not be evaluated; and (ii) the more recent stu-
dies that we summarize above supersede all of these previous stu-
dies in terms of paleomagnetic field tests, rock magnetic tests,
and overall data quality.
Paleomagnetic Poles from the Tibetan Himalaya (Greater India).
Below, we summarize all available paleomagnetic poles for the
Tibetan Himalaya (Fig. S1). Inclination shallowing corrections
are only available for the late Cretaceous to Paleocene poles.
Poles of early Cretaceous and older age have not been corrected
for inclination shallowing, and the effects of sedimentary inclina-
tion shallowing on the paleolatitude estimations require future
work. We note, however, that the effect of possible inclination
shallowing can only decrease the paleomagnetically estimated
size of Greater India, as its Paleozoic to Cretaceous position was
exclusively on the southern hemisphere. Moreover, the majority
of paleomagnetic data from the Triassic Tibetan Himalaya are
from marine limestones, which typically show only small (<5°)
amounts of inclination shallowing (12, 52).
Several coeval Tibetan Himalaya poles plot east or west of
the Indian APWP. This relationship indicates significant, local-
ity-specific vertical axis rotation of Tibetan Himalayan rocks with
respect to the Indian craton. These tectonic rotations must post-
date the age of the rocks in which they are found and probably
accrued during the intense deformation of the Tibetan Himalaya
that began with the obduction of ophiolites during the latest Cre-
taceous to early Paleogene (53, 54) and continued during the col-
lision of the Tibetan Himalaya with the Lhasa block (12, 55–58).
We are primarily interested in the latitudinal separation between
the Tibetan Himalaya and cratonic India, so here we correct for
these vertical axis rotations by moving the Tibetan Himalayan
pole by an angular distance equal to the rotation of the Tibetan
Himalayan pole with respect to the coeval pole from cratonic
India (e.g., Indian APWP or Deccan or Rajmahal poles) about
a small circle that is centered on the sampling site for the Tibetan
Himalayan pole. The rotation-corrected Tibetan Himalayan
poles are used to calculate the paleolatitudes shown in Fig. 2
of the main paper, Fig. S2, and Table S1, and are shown in Fig. S5;
the inclination and paleolatitude estimates from the uncorrected
poles are shown for comparison in Table S1. We note that ac-
counting for these vertical axis rotations typically changes the pa-
leolatitude estimates by less than one degree.
Upper Cretaceous–upper Paleocene marine sediments.
Paleomag-
netic data from the Maastrichtian to Paleocene Zongshan and
Selandian-Thanetian Zongpu formations near Gamba and Duela
in Southern Tibet are described by Patzelt et al. (59) and Yi et al.
(52). Both datasets pass paleomagnetic fold tests and reversal
tests at high confidence as described in the original studies, con-
sistent with a primary magnetization. Dupont-Nivet et al. (12)
tested the sediment-based data of Patzelt et al. (59) using the
E∕I procedure and found only minor (<5°) amounts of inclina-
tion shallowing, as common for marine limestones. Yi et al. (52)
also evaluated their data using the E∕I procedure and also found
minor (4.4°) inclination shallowing. Here we combine the ChRM
specimen directions of Patzelt et al. (59) and Yi et al. (52) from
van Hinsbergen et al. www.pnas.org/cgi/doi/10.1073/pnas.1117262109 3of10
the Zongpu formation to recalculate the magnitude of inclination
shallowing. Tilt-corrected outlier directions in this combined
dataset were removed using the Vandamme cutoff criteria
(43), resulting in 245 individual directions that define a slightly
elongated distribution (Fig. S6A). This filtered dataset passes a
reversals test at high confidence (Y¼2.4°, Yc¼12.0°, class
C) (40). We calculated a flattening factor of 0.69 (0.61% and
0.80 95% confidence interval), resulting in an E∕I-corrected in-
clination of 15.6° (12.6–18.5° 95% confidence interval) (Fig. S6 C
and D). These results indicate approximately 4° of inclination
shallowing in the marine limestones of the Zongpu formation.
Two outlier directions were removed from the population of
E∕I-corrected directions, resulting in a formation mean direction
of declination ¼358.6°, inclination ¼15.5°, a95 ¼2.3°, k¼16.6,
n¼243 (Fig. S6B). The corresponding poles for the Zongpu for-
mation calculated from this direction, as well as the for the Zong-
shan formation previously calculated by Dupont-Nivet et al. (12)
are shown in Table S1. We calculate a paleolatitude for the re-
ference location on the IYSZ of 8.71.7°N with an age range
of 62–56 Ma, and −5.02.8°N for 71–65 Ma (12). The age of the
Zongpu formation latitude is constrained by recent magnetostra-
tigraphy (52). The upper Cretaceous pole is rotated east of the
Indian APWP, however, and the rotation-corrected paleolatitude
at 71–65 Ma is −4.92.8°N. Compared to the Indian APWP, the
Zongpu pole (59 3Ma) shows 16.62.9° of latitudinal dis-
tance between India and the Tibetan Himalaya, whereas the
Zongshan pole (68 3Ma) indicates 13.64.4°oflatitudinal
separation. (Fig. 2 and Figs. S2 and S5) (12).
Cretaceous volcanoclastic sediments.
Klootwijk and Bingham (58)
sampled the Barremian to late Albian (127–112 Ma) Dzong for-
mation in the Thakkhola graben. Ninety-five paleomagnetic sam-
ples from the lower glauconitic sandstone part of the formation,
which has been assigned an Aptian age (121–112 Ma) (60), pass a
paleomagnetic fold test at high confidence (58). These data were
used to calculate a paleomagnetic pole located at −12.0°N,
108.6°E (A95 ¼3.7°). This pole is located east of the coeval In-
dian APWP pole, indicating significant rotation of the Tibetan
Himalaya with respect to cratonic India sometime after the de-
position of the Dzong formation. The rotation-corrected paleo-
latitude is −45.13.7°N; for comparison, the uncorrected
paleolatitude is −44.43.7°N. Both of these estimates overlap
within error with the coeval Indian APWP pole and the cratonic
India pole from the Rajmahal Traps. Notably, the paleolatitude
difference between rotation-corrected Dzong pole calculation
and the Rajmahal pole prediction is −2.15.5° of latitude.
Triassic marine sediments.
Paleomagnetic data from Triassic
strata throughout the Tibetan Himalaya were collected from several
published sources (Fig. S1 and Ta b l e S 9 ). Stratigraphic nomencla-
ture follows the original authors, whereas stage assignments follow
Garzanti (60), and stage ages follow Gradstein et al. (17).
Klootwijk et al. (61) describe paleomagnetic results from six
sampling localities from Kashmir in the western Tibetan Hima-
laya. Sampled lithologies include marine limestone, shales, and
sandstones of lowest Induan to Norian age (i.e., 248.2–209.6 Ma).
Most samples revealed high-temperature ChRM interpreted as
primary in tilt-corrected coordinates. The mean VGP calculated
from these six localities gives a paleolatitude of −55.28.7°N
for the reference location on the IYSZ. The Kashmir pole shows
a large rotation from the Indian APWP. The rotation-corrected
paleolatitude for the reference location is −49.68.7°N.
Induan to early Norian (248.2–215 Ma) greenschist-facies
meta-shales and limestones of the Tamba and Mukut formations
from the western Dolpo region of the Tibetan Himalaya record a
high-temperature magnetization carried by magnetite and inter-
preted to be prefolding in age (62). Ten of the 26 site mean direc-
tions meet our quality criteria. A new paleomagnetic pole
calculated from the 10 corresponding VGPs is located at −2.9°N,
281.4°E (A95 ¼18.5°, K¼7.8), but is located east of the Indian
APWP. The rotation-corrected paleolatitude for the reference
location is −55.918.5°N. This paleolatitude is the most south-
erly in this compilation, lying more than 25° of latitude south
of the APWP prediction. According to this pole, the Tibetan
Himalaya (TH) was derived from the center of the Indian craton;
this is geologically implausible. We choose to exclude these
Dolpo results from our compilation given the large number of
sites rejected, the low number of samples, the greenschist-facies
metamorphism of the limestones, and the weak fold test used to
suggest a primary magnetization. Future paleomagnetic sampling
of Triassic and younger rocks in the Dolpo region are needed to
verify these results.
Paleomagnetic results from two sampling localities in the
Ladinian to Norian (234.3–209.6 Ma) Thinigaon Limestone of
the Thakkhola region are described by Klootwijk and Bingham
(58). A high-temperature ChRM is resolved in most samples
and is interpreted to be primary in tilt-corrected coordinates.
The mean VGPs calculated from these two localities show
only small rotation with respect to India. The rotation-corrected
paleolatitude estimates for the reference location are −30.1
5.7°N and −26.910.0°N.
Lower Anisian to lower Norian (241.7–215 Ma) carbonates
from a single sampling locality near the Manang area of the
Tibetan Himalaya record a ChRM carried by magnetite (63).
A positive fold test of the ChRMs suggests that these directions
are primary and suitable for calculating a paleomagnetic pole.
This pole shows only minor rotation of the Manang region with
respect to cratonic India since the Triassic. The rotation-cor-
rected paleolatitude of the reference location is −36.24.2°N.
Thirty-four paleomagnetic sampling sites collected from
the lowest Induan to Norian (248.2–209.6 Ma) limestones of
the Tomba Kurkur and Mukut formations of Central Nepal are
described by Schill et al. (64). A high-temperature component
carried by magnetite was interpreted as primary by these authors.
Thirteen of these site mean directions meet our quality criteria,
resulting in a mean VGP located at 14.1°N, 256.3°E (A95 ¼10.8°,
K¼15.7). This pole is rotated east of the coeval Indian APWP.
The rotation-corrected paleolatitude of the reference location
is −47.510.8°N.
In summary, paleomagnetic data from several different packages
of Triassic rocks distributed throughout the Tibetan Himalaya re-
cord stable, primary remanences. The paleomagnetic poles from
each of the sampling locations indicate differential amounts of ver-
tical axis rotation with respect to cratonic India. We use the rota-
tion-corrected paleolatitudes from each of the studies (excluding
the results from Dolpo) described above to calculate the maximum
likelihood paleolatitude (47) of the reference site in a Tibetan
Himalaya reference frame of −38.69.9°N. In contrast, the max-
imum likelihood estimate of the reference location calculated from
the 250–200 Ma Indian APWP poles is −30.16.9°N (the paleo-
latitude calculated from the Fisher mean of the 250–200 Ma Indian
APWP poles is −29.97.3°N). Thus, we find that the mean
latitudinal separation between the reference location in a Tibetan
Himalaya and cratonic India reference frame during the Triassic is
−9.010.6° (Fig. S5). We note that several individual poles of high
quality lie closer to the Triassic APWP poles than indicated by this
comparison of the maximum likelihood estimates.
Ordovician redbeds.
Paleomagnetic data from the Shian formation
of the Tibetan Himalaya near Spiti, Northwest India are
described by Torsvik et al. (6), who assigned an Ordovician
(500–450 Ma) age to these rocks. High-temperature ChRM
van Hinsbergen et al. www.pnas.org/cgi/doi/10.1073/pnas.1117262109 4of10
directions carried by hematite pass a fold test at high confidence,
and the six paleomagnetic sampling sites yield a mean
VGP located at −20.4°N, 210.0°E (A95 ¼13.4°, K¼26.0)
(Tables S1 and S10), which is rotated away from the coeval part
of the Indian APWP. The paleolatitude of the reference location
calculated from the rotation-corrected pole is −37.713.4°N.
Implications of Paleomagnetic Data for the India–Asia Collision His-
tory.
Age of the Tibetan Himalaya–Lhasa collision.
The paleolatitude
separation between the late Cretaceous and Paleocene poles for
the TH and the coeval Indian APWP poles is remarkably similar
(23.63.7° at 59 Ma; 20.44.8° at 68 Ma, using the trend of the
Indian APWP adjusted to the Deccan Traps paleopole), suggest-
ing that over this time span, the Tibetan Himalaya drifted north-
ward at a similar rate as cratonic India (Fig. 2 and Figs. S2 and
S5). We use these paleolatitude separation values to calculate a
mean latitudinal separation between the Tibetan Himalaya and
cratonic India of 22.03.0° from 65 to 59 Ma (Fig. S5). In other
words, paleomagnetism shows that sometime after 59 Ma, 22.0
3.0°ofNorth–South (N-S) convergence occurred between the
northernmost exposed tip of the Tibetan Himalaya, located at
the modern suture zone, and continental India.
If we assume that there was no convergence accommodated
within Greater India prior to collision with Eurasia, we can cal-
culate a collision age by constructing a Tibetan Himalayan
“APWP”(or better, paleolatitude curve, Fig. S2) that is parallel
to the Indian APWP and is projected forward in time until it in-
tersects with the Asian paleolatitude curve (specifically, the pa-
leolatitude calculated from the Linzizong formation); the 95%
uncertainty in latitude assigned to the Tibetan Himalayan APWP
is 3°. The latitude of the northern Tibetan Himalaya overlaps
with the latitude of the southern margin of the Lhasa block at
48.66.2Ma at 19.63.9°N (95% confidence intervals, see
Fig. 2 and Fig. S2).
Paleolatitudinal drift of the Tibetan Himalaya relative to continental
India.
Paleolatitudes calculated from paleomagnetic data from
Ordovician, Triassic, and lower Cretaceous sediments of the
Tibetan Himalayan sequence all overlap within error with paleo-
latitudes predicted from coeval cratonic Indian reference poles
(Figs. S2 and S5). This overlap means that the N-S area of 22.0
3.0° that disappeared due to convergence within Greater India
sometime after 59 Ma (probably after 48.66.2Ma, see above)
must have been created by extension after approximately 120 Ma.
When making a direct comparison between the paleolatitude
of the Rajmahal Traps (118 1Ma) and the Dzong formation of
the Tibetan Himalaya (121–112 Ma) a statistically indistinguish-
able post-118 Ma net separation of −2.15.5° of latitude
(−233 877 km N-S separation) is obtained. This value shows
that the Greater India 118 Ma ago was not more than 644 km
(i.e., 877–233 km) larger than today. Today’s width of Greater
India is equal to the width of the Himalaya—i.e., approximately
250 km. Hence, paleomagnetism constrains the maximum width
(within 95% error bars) of Greater India 118 Ma ago at approxi-
mately 900 km. Using the 22.03.0° of post-59 Ma convergence
calculated above, the net increase in paleolatitudinal separation
(i.e., the amount of N-S extension) between the Tibetan Hima-
laya and cratonic India implied by these direct comparisons is
24.06.3°(2;675 699 km N-S) between 118 and 68 Ma.
The estimated size of the Greater India Basin will be refined by
future studies that provide robust paleomagnetic poles from Cre-
taceous rocks of the Tibetan Himalaya. Nevertheless, reviewing
the current state-of-the-art of published, robust paleomagnetic
data from India, the Tibetan Himalaya, and the Lhasa terrane
indicate that the Tibetan Himalaya must have drifted away from
cratonic India during the Cretaceous, leaving in its wake a basin
or several basins that accommodated a cumulative N-S extension
of 2;675 699 km.
1. Cox A, Hart RB (1986) Plate tectonics: How it works (Blackwell Science, Oxford).
2. Besse J, Courtillot V (2002) Apparent and true polar wander and the geometry of the
geomagnetic field over the last 200 Mya. J Geophys Res 107:2300.
3. Torsvik TH, Müller RD, Van der Voo R, Steinberger B, Gaina C (2008) Global plate
motion frames: Toward a unified model. Rev Geophys 41:RG3004.
4. van Hinsbergen DJJ, Steinberger B, Doubrovine PV, Gassmöller R (2011) Acceleration
and deceleration of India–Asia convergence since the Cretaceous: Roles of mantle
plumes and continental collision. J Geophys Res 116:B06101.
5. Torsvik TH, van der Voo R (2002) Refining Gondwana and Pangea palaeogeography:
Estimates of Phanerozoic non-dipole (octupole) fields Geophys J Int 151:771–794.
6. Torsvik TH, Paulsen TS, Hughes NC, Myrow PM, Ganerød M (2009) The Tethyan
Himalaya: palaeogeographical and tectonic constraints from Ordovician palaeomag-
netic data. J Geol Soc 166:679–687.
7. King RF (1955) The remanent magnetism of artificially deposited sediments. Mon Not
R Astron Soc 7:115–134.
8. Tauxe L (2005) Inclination flattening and the geocentric axial dipole hypothesis.
Earth Planet Sci Lett 233:247–261.
9. Fisher RA (1953) Dispersion on a sphere. Proc R Soc London A 217:295–305.
10. Deenen MHL, Langereis CG, van Hinsbergen DJJ, Biggin AJ (2011) Geomagnetic
secular variation and the statistics of palaeomagnetic directions. Geophys J Int
186:509–520.
11. Johnson CL, et al. (2008) Recent investigations of the 0–5 Ma geomagnetic field
recorded by lava flows. Geochem Geophys Geosyst 9:Q04032.
12. Dupont-Nivet G, van Hinsbergen DJJ, Torsvik TH (2010) Persistently shallow paleo-
magnetic inclinations in Asia: Tectonic implications for the Indo-Asia collision.
Tectonics 29:TC5016.
13. Lippert PC, Zhao X, Coe RS, Lo C-H (2011) Palaeomagnetism and 40 Ar∕39 Ar geochro-
nology of upper Palaeogene volcanic rocks from Central Tibet: Implications for the
Central Asia inclination anomaly, the palaeolatitude of Tibet and post-50 Ma short-
ening within Asia. Geophys J Int 184:131–161.
14. Biggin A, van Hinsbergen DJJ, Langereis CG, Straathof GB, Deenen MH (2008) Geo-
magnetic secular variation in the Cretaceous normal superchron and in the Jurassic.
Phys Earth Planet Int 169:3–19.
15. Tauxe L, Kent DV (2004) A simplified statistical model for the geomagnetic field and
the detection of shallow bias in paleomagnetic inclinations: Was the ancient mag-
netic field dipolar? Timescales of the Paleomagnetic Field, Geophysical Monograph,
eds Channell JET, Kent DV, Lowrie W, Meert JG (American Geophysical Union,
Washington, DC), Vol 145, pp 101–115.
16. Butler RF (1992) Paleomagnetism: Magnetic Domains to Geologic Terranes (Blackwell
Scientific, Boston), pp 195.
17. Gradstein FM, et al. (1994) A Mesozoic time scale. J Geophys Res 99:24051–24074.
18. Cande SC, Kent DV (1995) Revised calibration of the geomagnetic polarity timescale
for the Late Cretaceaous and Cenozoic. J Geophys Res 100:6093–6095.
19. Vandamme D, Courtillot V, Besse J, Montigny R (1991) Paleomagnetism and age de-
terminations of the Deccan Traps (India): Results of a Nagpur-Bombay traverse and
review of earlier work. Rev Geophys 29:159–190.
20. Chenet AL, et al. (2009) Determination of rapid Deccan eruptions across the Cretac-
eous-Tertiary boundary using paleomagnetic secular variation: 2. Constraints from
analysis of eight new sections and synthesis for a 3500-m-thick composite section.
J Geophys Res 114:B06103.
21. Hofmann C, Feraud G, Courtillot V (2000) 40 Ar∕39 Ar dating of mineral separates and
whole rocks from the Western Ghats lava pile: Further constraints on duration and
age of the Deccan traps. Earth Planet Sci Lett 180:13–27.
22. Chenet AL, Quidelleur X, Fluteau F, Courtillot V, Bajpai S (2007) K-40-Ar-40 dating of
the Main Deccan large igneous province: Further evidence of KTB age and short
duration. Earth Planet Sci Lett 263:1–15.
23. Vandamme D, Courtillot V (1992) Paleomagnetic constraints on the structure of the
Deccan traps. Phys Earth Planetary Int 74:241–261.
24. Doubrovine PV, Tarduno JA (2008) Linking the Late Cretaceous to Paleogene Pacific
plate and the Atlantic bordering continents using plate circuits and paleomagnetic
data. J Geophys Res 113:B07104.
25. Ganerød M, Smethurst MA, Rousse S, Torsvik TH, Prestvik T (2008) Reassembling the
Paleogene-Eocene North Atlantic igneous province: New paleomagnetic constraints
from the Isle of Mull, Scotland. Earth Planet Sci Lett 272:464–475.
26. Coffin MF, et al. (2002) Kerguelen hotspot magma output since 130 Ma. J Petrol
43:1121–1139.
27. Klootwijk CT (1971) Palaeomagnetism of upper Gondwana-Rajmahal Traps,
Northeast India. Tectonophysics 12:449–461.
28. Sherwood GJ, Malik SB (1996) A palaeomagnetic and rock magnetic study of the
northern Rajmahal volcanics, Bihar, India. J Southeast Asian Earth Sci 13:123–131.
29. Kapp P, Yin A, Harrison TM, Ding L (2005) Cretaceous-Tertiary shortening, basin
development and volcanism in central Tibet. Geol Soc Am Bull 117:865–878.
30. Kapp P, et al. (2007) The Gangdese retroarc thrust belt revealed. GSA Today 17:4–9.
31. Kapp P, DeCelles PG, Gehrels GE, Heizler M, Ding L (2007) Geological records of the
Cretaceous Lhasa-Qiangtang and Indo-Asian collisions in the Nima basin area, central
Tibet. Geol Soc Am Bull 119:917–932.
32. He S, Kapp P, DeCelles PG, Gehrels GE, Heizler MT (2007) Cretaceous–Tertiary geology
of the Gangdese Arc in the Linzhou area, southern Tibet. Tectonophysics 433:15–37.
33. Lee H-Y, et al. (2009) Eocene Neotethyan slab breakoff in southern Tibet inferred
from the Linzizong volcanic record. Tectonophysics 477:20–35.
van Hinsbergen et al. www.pnas.org/cgi/doi/10.1073/pnas.1117262109 5of10
34. Tan X, et al. (2010) New paleomagnetic results from the Lhasa block: Revised estima-
tion of latitudinal shortening across Tibet and implications for dating the India–Asia
collision. Earth Planet Sci Lett 293:396–404.
35. Achache J, Courtillot V, Zhou YX (1984) Paleogeographic and tectonic evolution of
southern Tibet since middle Cretaceous time: new paleomagnetic data and synthesis.
J Geophys Res 89:10311–10339.
36. Chen J, Huang B, Sun L (2010) New constraints to the onset of the India–Asia collision:
Paleomagnetic reconnaissance on the Linzizong Group in the Lhasa Block, China.
Tectonophysics 489:189–209.
37. Sun Z, Jiang W, Li H, Pei JL, Zhu Z (2010) New paleomagnetic results of Paleocene
volcanic rocks from the Lhasa block: Tectonic implications for the collision of India
and Asia. Tectonophysics 490:257–266.
38. Dupont-Nivet G, Lippert P, van Hinsbergen DJJ, Meijers MJM, Kapp P (2010) Paleo-
latitude and age of the Indo-Asia collision: Paleomagnetic constraints. Geophys J Int
182:1189–1198.
39. Liebke U, et al. (2010) Position of the Lhasa terrane prior to India–Asia collision de-
rived from palaeomagnetic inclinations of 53 Ma old dykes of the Linzhou Basin:
Constraints on the age of collision and post-collisional shortening within the Tibetan
Plateau. Geophys J Int 182:1199–1215.
40. McFadden PC, McElhinny MW (1990) Classification of the reversal test in palaeomag-
netism. Geophys J Int 103:725–729.
41. McFadden PL (1990) A new fold-test for palaeomagnetic studies. Geophys J Int
103:163–169.
42. Sun Z, et al. (2012) Palaeomagnetism of late Cretaceous sediments from southern
Tibet: Evidence for the consistent palaeolatitudes of the southern margin of Eurasia
prior to the collision with India. Gondwana Res 21:53–63.
43. Vandamme D (1994) A new method to determine paleosecular variation. Phys Earth
Planet Int 85:131–142.
44. Leier AL, DeCelles PG, Kapp P, Ding L (2007) The Takena formation of the Lhasa ter-
rane, southern Tibet: The record of a Late Cretaceous retroarc foreland basin. Geol
Soc Am Bull 119:31–48.
45. Sun Z, Jiang W, Pei JL, Li H (2008) New early Cretaceous paleomagnetic data from
volcanic of the eastern Lhasa Block and its tectonic implications. Acta Petrol Sin
24:1621–1626.
46. Chen W, et al. (2012) Paleomagnetic results from the Early Cretaceous Zenong Group
volcanic rocks, Cuoqin, Tibet, and their paleogeographic implications. Gondwana Res
(in press).
47. Arason P, Levi S (2010) Maximum likelihood solution for inclination-only data in
paleomagnetism. Geophys J Int 182:753–771.
48. Pozzi JP, Westphal M, Zhou YX, Xing SL, Chen XY (1982) Position of the Lhasa block,
South Tibet, during the late Cretaceous. Nature 297:319–321.
49. Westphal M, Pozzi JP (1983) Paleomagnetic and plate tectonic constraints on the
movement of Tibet. Tectonophysics 98:1–10.
50. Lin J, Watts DR (1988) Palaeomagnetic constraints on Himalayan-Tibetan tectonic
evolution. Philos Trans R Soc London 326:177–188.
51. Chen Y, Cogné JP, Courtillot V, Tapponnier P, Zhu XY (1993) Cretaceous paleomag-
netic results from western Tibet and tectonic implications. J Geophys Res
98:17981–17999.
52. Yi Z, Huang B, Chen J, Chen L, Wang H (2011) Paleomagnetism of early Paleogene
marine sediments in southern Tibet, China: Implications to onset of the India–Asia
collision and size of Greater India. Earth Planet Sci Lett 309:153–165.
53. Ding L, Kapp P, Wan X (2005) Paleocene-Eocene record of ophiolite obduction and
initial India–Asia collision, south central Tibet. Tectonics 24:TC3001.
54. Searle MP, Treloar PJ (2010) Was Late Cretaceous—Palaeocene obduction of ophiolite
complexes the primary cause of crustal thickening and regional metamorphism in
the Pakistan Himalaya? The Evolving Continents: Understanding Processes of Conti-
nental Growth, eds Kusky TM, Zhai MG, Xiao W (Geological Society Special Publica-
tions, London), Vol 338, pp 345–359.
55. Murphy MA, Yin A (2003) Structural evolution and sequence of thrusting in the Teth-
yan fold-thrust belt and Indus-Yalu suture zone, southwest Tibet. Geol Soc Am Bull
115:21–34.
56. Ratschbacher L, Frisch W, Liu G, Chen CC (1994) Distributed deformation in southern
and western Tibet during and after the India–Asia collision. J Geophys Res
99:19917–19945.
57. Schill E, Appel E, Zeh O, Singh VK, Gautam P (2001) Coupling of late-orogenic tec-
tonics and secondary pyrrhotite remanences: towards a separation of different rota-
tion processes and quantification of rotational underthrusting in the western
Himalaya (northern India). Tectonophysics 337:1–21.
58. Klootwijk CT, Bingham DK (1980) The extent of greater India, III. Palaeomagnetic
data from the Tibetan Sedimentary series, Thakkhola region, Nepal Himalaya. Earth
Planet Sci Lett 51:381–405.
59. Patzelt A, Li H, Wang J, Appel E (1996) Palaeomagnetism of Cretaceous to Tertiary
sediments from southern Tibet: Evidence for the extent of the northern margin of
India prior to the collision with Eurasia. Tectonophysics 259:259–284.
60. Garzanti E (1999) Stratigraphy and sedimentary history of the Nepal Tethys Himalaya
passive margin J Asian Earth Sci 17:805–827.
61. Klootwijk CT, et al. (1983) A palaeomagnetic reconnaissance of Kashmir, northwes-
tern Himalaya, India. Earth Planet Sci Lett 63:305–324.
62. Crouzet C, Gautam P, Schill E, Appel E (2003) Multicomponent magnetization in
western Dolpo (Tethyan Himalaya, Nepal): Tectonic implications. Tectonophysics
377:179–196.
63. Appel E, Müller R, Widder RW (1991) Palaeomagnetic results from the Tibetan sedi-
mentary series of the Manang area (north central Nepal). Palaeomagnetic results
from the Tibetan sedimentary series of the Manang area (north central Nepal). Geo-
phys J Int 104:255–266.
64. Schill E, Appel E, Gautam P, Dietrich D (2002) Thermo-tectonic history of the Tethyan
Himalayas deduced from the palaeomagnetic record of metacarbonates from Shiar
Khola (Central Nepal). J Asian Earth Sci 20:203–210.
van Hinsbergen et al. www.pnas.org/cgi/doi/10.1073/pnas.1117262109 6of10
Fig. S2. Paleolatitudes of a reference site (29°N, 88°E) located on the present-day position of the Indus–Yarlung Suture Zone in Eurasian, Greater Asian,
Tibetan Himalayan, and Indian reference frames. Numbers correspond to paleomagnetic poles described and listed in Table S1. Pl, Pliocene; Mio, Miocene;
Oligo, Oligocene; Paleo., Paleocene; Jr, Jurassic; Tr, Triassic; P, Permian; Carb., Carboniferous; Dev., Devonian; Sil., Silurian; Ord, Ordovician; C, Cambrian; MCT,
main central thrust; STD, South Tibetan detachment. Uncertainties in age are assigned based on the age of the units determined from either radiometric dates
or geologic stages (17, 18), whereas latitude uncertainties are at the 95% confidence level. Question mark next to estimate 7 indicates that this pole may not
fully average paleosecular variation.
AB
CD
Fig. S3. E∕I correction (15) of high-temperature, primary magnetic component of the Shexing formation redbeds at Maxiang, south, central Lhasa terrane
(42). (A) Equal area plot of original paleomagnetic directions in tilt-corrected coordinates, with a mean inclination of 31.2° (magenta star, 95% confidence
interval indicated in pink). Slight horizontal elongation of the direction distribution is consistent with sedimentary inclination shallowing. (B)E∕I-corrected
paleomagnetic directions after accounting for a flattening factor of 0.66. The E∕I-corrected inclination is 41.6°. (C) Plot of elongation versus inclination for the
TK03.geocentric axial dipole (GAD) paleosecular variation model (solid green line, ref. 15) and for the Shexing formation directions (barbed red line) for
different flattening factors, f. Yellow lines are the results from 20 boot-strapped datasets. The crossing points of the Shexing formation and boot-strapped
dataset with the TK03.GAD results represents the elongation/inclination pairs that are most consistent with the TK03. GAD field model. (D) Histogram of
crossing points from 5,000 boot-strapped datasets. The most frequent flattening factor is f¼0.66, resulting in the distribution of directions shown in B.
van Hinsbergen et al. www.pnas.org/cgi/doi/10.1073/pnas.1117262109 8of10
Fig. S4. Equal area plot of site mean directions from lower Cretaceous volcanic rocks from the central and western Lhasa terrane that pass our data quality
criteria (45, 46). Downward (upward) pointing directions are shown as filled (open) symbols. Star indicates the present dipole magnetic field direction. Sig-
nificant improvement in the clustering of inclinations from in situ coordinates (A) to tilt-corrected coordinates (B) is consistent with a prefoldin g magnetization.
The noticeably different declination values observed in the two localities are consistent with differential vertical axis rotation because the emplacement of
these lavas and justifies the use of latitude-only paleolatitude methods (SI Text).
Fig. S5. Rotation-corrected paleomagnetic poles for the Tibetan Himalaya showing latitudinal separation between the Tibetan Himalaya and cratonic India
for the (A) Paleocene, (B) latest Cretaceous, (C) early Cretaceous, and (D) Triassic. Ninety-five percent confidence limits of the pole positions are also shown.
van Hinsbergen et al. www.pnas.org/cgi/doi/10.1073/pnas.1117262109 9of10
AB
CD
Fig. S6. E∕I correction (15) of high-temperature, primary magnetic component of the Zongpu formation marine limestones at Gamba and Duela, southern
Tibetan Himalaya (52, 59). Format of figures follows Figs. S3 and S4. (A) Distribution of 245 original paleomagnetic directions in tilt-corrected coordinates, with
a mean inclination of 11.2°. Horizontal elongation of the direction distribution is consistent with sedimentary inclination shallowing. (B)E∕I-corrected pa-
leomagnetic directions after accounting for a flattening factor of 0.69 (Cand D). The E∕I-corrected inclination is 15.5°, indicating that inclination shallowing in
these marine limestones is <5°.
Other Supporting Information Files
Table S1 (DOC)
Table S2 (DOCX)
Table S3 (DOCX)
Table S4 (DOCX)
Table S5 (DOCX)
Table S6 (DOCX)
Table S7 (DOC)
Table S8 (DOCX)
Table S9 (DOCX)
Table S10 (DOCX)
van Hinsbergen et al. www.pnas.org/cgi/doi/10.1073/pnas.1117262109 10 of 10
