Implications of Shortening in the Himalayan Fold-Thrust Belt for Uplift of the Tibetan Plateau

TECTONICS, VOL. 21, NO. 6, 1062, doi:10.1029/2001TC001322, 2002

Implications of shortening in the Himalayan fold-thrust belt for uplift of the Tibetan Plateau

Peter G. DeCelles, Delores M. Robinson, and George Zandt Department of Geosciences, University of Arizona, Tucson, Arizona, USA Received 20 August 2001; revised 13 September 2002; accepted 18 September 2002; published 31 December 2002. [1] Recent research in the Himalayan fold-thrust belt Plateau. The need for Tibetan crust to stretch laterally provides two new sets of observations that are crucial as the Greater Indian lower crust was inserted may to understanding the evolution of the Himalayan- explain the widespread east-west extension in the Tibetan orogenic system. First, U-Pb zircon ages and southern half of the Plateau. Our reconstruction of the Sm-Nd isotopic studies demonstrate that (1) Greater Himalayan fold-thrust belt suggests that Indian cratonic Himalayan medium- to high-grade metasedimentary lower crust, of presumed mafic composition and high rocks are much younger than true Indian cratonic strength, should extend approximately halfway across basement; and (2) these rocks were tectonically the Tibetan Plateau, to the Banggong suture. From mobilized and consolidated with the northern margin there northward, we predict that the Tibetan Plateau is of Gondwana during early Paleozoic orogenic activity. underlain by more felsic, and therefore weaker, lower These observations require that Greater Himalayan crust of Greater Himalayan affinity. Two slab break-off rocks be treated as supracrustal material in restorations events are predicted by the model: the first involved of the Himalayan fold-thrust belt, rather than as Indian Neotethyan oceanic lithosphere that foundered 45– cratonic basement. In turn, this implies the existence 35 Ma, and the second consisted of Greater Indian of Greater Himalayan lower crust that is not exposed lithosphere (most likely composed of Greater anywhere in the fold-thrust belt. Second, a regional Himalayan material) that delaminated and foundered compilation of shortening estimates along the 20–10 Ma. Asthenospheric upwelling associated Himalayan arc from Pakistan to Sikkim reveals that with the break-off events may explain patterns of (1) total minimum shortening in the fold-thrust belt is Cenozoic volcanism on the Tibetan Plateau. Although up to 670 km; (2) total shortening is greatest in the model predicts a northward migrating topographic western Nepal and northern India, near the apex of the front due solely to insertion of Greater Indian lower Himalayan salient; and (3) the amount of Himalayan crust, the actual uplift history of the Plateau was shortening is equal to the present width of the Tibetan complicated by early-middle Tertiary shortening of Plateau measured in an arc-normal direction north of Tibetan crust. INDEX TERMS: 8102 Tectonophysics: the Indus-Yalu suture zone. This information suggests Continental contractional orogenic belts; 8110 Tectonophysics: that a slab of Greater Indian lower crust (composed of Continental tectonics—general (0905); 8120 Tectonophysics: both Indian cratonic lower crust and Greater Dynamics of lithosphere and mantle—general; 9320 Information Himalayan lower crust) with a north-south length of Related to Geographic Region: Asia; 9604 Information Related to 700 km may have been inserted beneath the Tibetan Geologic Time: Cenozoic; KEYWORDS: collisional orogenic belts, crust during the Cenozoic orogeny. We present a Tibetan Plateau, Himalaya, lithosphere dynamics, delamination, modified version of the crustal underthrusting model thrust belts. Citation: DeCelles, P. G., D. M. Robinson, and G. Zandt, Implications of shortening in the Himalayan fold-thrust belt for Himalayan-Tibetan orogenesis that integrates for uplift of the Tibetan Plateau, Tectonics, 21(6), 1062, surface geological data, recent results of mantle doi:10.1029/2001TC001322, 2002. tomographic studies, and broadband seismic studies of the crust and upper mantle beneath the Tibetan Plateau. Previous studies have shown that incremental Mesozoic and early Cenozoic shortening had probably 1. Introduction thickened Tibetan crust to 45–55 km before the [2] The origin of high-elevation orogenic plateaus is a onset of the main Cenozoic orogenic event. Thus, the topic of considerable interest, given the potential climato- insertion of a slab of Greater Indian lower crust with logical, geochemical, and environmental side effects of maximum thickness of 20 km (tapering northward) plateau growth and maintenance. Explanations of orogenic could explain the Cenozoic uplift of the Tibetan plateaus have been best refined for the case of the Tibetan Plateau, which has been studied intensively since the mid- 1970’s [e.g., Molnar and Tapponnier, 1975]. The Hima- Copyright 2002 by the American Geophysical Union. layan fold-thrust belt, which forms the southern rim of the 0278-7407/02/2001TC001322$12.00 Plateau (Figure 1), is obviously a result of shortening of

12 - 1 12 - 2 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY

Figure 1. Regional tectonic map of the Tibetan Plateau, showing major suture zones, terranes, fault systems, Cenozoic magmatic centers, and the Gangdese magmatic arc [after Yin and Harrison, 2000; Hacker et al., 2000; Tapponnier et al., 2001]. Elevations after Fielding et al. [1994].

Indian rocks to the south of the suture zone that marks the butions of Greater Indian crust to thickening of the Plateau. paleosubduction zone between the Indian and Eurasian Greater India is regarded as the landmass of the Indian plates [Gansser, 1964]. However, no consensus exists on subcontinent before the onset of the Indo-Eurasian collision the timing and mechanism(s) of formation of the Tibetan [Veevers et al., 1975]. The north-south length of Greater Plateau and the nature of its relationship to the Himalaya Indian lower crust available to thicken the Tibetan Plateau [e.g., Dewey et al., 1988; Harrison et al., 1992; Molnar et should be equivalent to the amount of supracrustal shortening al., 1993; Matte et al., 1997; Yin and Harrison, 2000; in the Himalayan fold-thrust belt [Klootwijk et al., 1985]. Tapponnier et al., 2001]. Moreover, the kinematic history of the fold-thrust belt should [3] Our corporate understanding of the development of the provide a gauge of the timing of addition of Greater Indian Tibetan Plateau has developed to its current state largely crustal material to the Tibetan Plateau. These simple concepts unaided by a thorough consideration of the potential contri- have been difficult to exploit because estimates of shortening DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 3 based on geological data from the Himalayan fold-thrust belt contributed to growth of the Plateau, are the main subjects of are sparse and vary by a factor of roughly two. Instead, the this paper. principal constraints on amounts of Himalayan shortening have come from paleomagnetic data, which are inherently 2. Models for Tibetan Plateau Uplift subject to large uncertainties and provide only the coarsest kinematic information [Molnar and Tapponnier, 1975; Ach- [5] Models for the origin of the Tibetan Plateau begin with ache et al., 1984; Patriat and Achache, 1984; Klootwijk et the following boundary conditions: (1) approximately 2500 al., 1985; Besse et al., 1984; Besse and Courtillot, 1988, km of post-collision convergence has taken place between 1991; Patzeltetal., 1996]. In addition, the high-grade India and Eurasia [Molnar and Tapponnier, 1975; Patriat metamorphic rocks of the Greater Himalayan sequence that and Achache, 1984; Besse and Courtillot, 1988, 1991; crop out in the medial part of the fold-thrust belt historically Patzelt et al., 1996]; (2) the Plateau is underlain by con- have presented a puzzle in terms of how to restore cross tinental crust that is approximately twice as thick as normal sections through the Himalaya: Should these rocks be treated crust [Hirn, 1988; Owens and Zandt, 1997; Zhu, 1998; Kind as Indian cratonic basement [e.g., Gansser, 1964; Mattauer, et al., 2002]; and (3) seismic phase velocities in the upper 1986; Srivastava and Mitra, 1994; Hauck et al., 1998], or mantle (at 100–300 km depth) are generally fast compared to should they be treated as an exotic tectonostratigraphic adjacent regions, and Pn wave velocity and Sn propagation terrane that was structurally elevated prior to the Cenozoic are relatively slow and inefficient, respectively, beneath the orogenic event [e.g., Parrish and Hodges, 1996; DeCelles et northern half of the plateau but relatively fast and efficient al., 2000]? Adding to the dilemma is the uncertainty of the beneath the southern half [Ni and Barazangi, 1984; Molnar, structure in the mantle and lithosphere beneath the Tibetan 1988; Holt and Wallace, 1990; McNamara et al., 1997; Plateau. However, recent progress on the Greater Himalayan Owens and Zandt, 1997; Tapponnier et al., 2001]. issue, as well as recent deep crustal seismic reflection profil- [6] Five general categories of models for the Tibetan ing [Nelson et al., 1996; Hauck et al., 1998] and broadband Plateau have been proposed: (1) crustal thickening by pure seismic experiments across the Tibetan Plateau [Owens and shear during the Mesozoic and Cenozoic [Murphy et al., Zandt, 1997; Kosarev et al., 1999; Kind et al., 2002], and 1997] or entirely during the Cenozoic [England and House- mantle tomographic studies [Grand et al., 1997; Van der Voo man, 1988; Molnar et al., 1993]; (2) crustal underthrusting, et al., 1999] provide an opportunity to integrate geological or injection, from the south [Argand,1924;Powell and and geophysical data sets at an unprecedented level. Conaghan, 1973; Ni and Barazangi, 1984; Coward and [4] In this paper we summarize the available geological Butler, 1985; Mattauer, 1986; Zhao and Morgan, 1987], and geophysical data on the timing and magnitude of crustal perhaps accompanied by phase transitions in the lower crust shortening in the Himalayan fold-thrust belt and the mantle [LePichon et al., 1997; Chemenda et al., 2000]; (3) crustal structure beneath the Tibetan Plateau, with the goal of underthrusting from the north [Willett and Beaumont, 1994; quantitatively assessing potential contributions of Greater Tapponnier et al., 2001]; (4) uplift owing to dynamic pro- Indian lower crust to thickening of the Tibetan Plateau. cesses in the mantle [Dewey et al., 1988; Molnar et al., 1993; Although our perspective is informed mainly by geological Turner et al., 1993; Platt and England, 1994]; and (5) investigations of the Himalayan fold-thrust belt in Nepal, the indentation of a weak Eurasian plate by a rigid, strong Indian work of many previous investigators throughout the Hima- plate, possibly accompanied by intraplate subduction and laya and Tibetan Plateau helps to regionalize some of the key lateral tectonic escape of large Eurasian crustal blocks observations. We propose an integrated model that connects [Tapponnier et al., 1982; Matte et al., 1996] and local thrust thickening and uplift of the Tibetan Plateau with the develop- faulting and sediment aggradation [Me´tivier et al., 1998; ment of the Himalayan fold-thrust belt and suggests timing Tapponnier et al., 2001]. A number of studies have suggested for at least a major increment of Plateau uplift. What is that the lower crust of Tibet is extremely hot, rheologically different about our model is the way we treat Indian lower weak, and decoupled from the mantle, which helps to explain crust. Previous studies of the Himalayan fold-thrust belt have the remarkably low-relief internal topography of much of the considered the high-grade metamorphic rocks of the Greater Tibetan Plateau (Figure 2) [Zhao and Morgan, 1987; Bird, Himalaya to be Indian cratonic basement that has been thrust 1991; Fielding et al., 1994; Royden et al., 1997; Kirby et al., upward and southward, removing the need to dispose of more 2000; Clark and Royden, 2000]. Some studies have proposed than 7.0 Â 106 km3 of Indian lower crust (roughly one- combinations of more than one of these mechanisms for tenth of the Plateau volume) beneath the Tibetan Plateau. thickening and uplift of the Plateau. Most of these models Several recent geochronological and geochemical studies, have difficulty accounting for all of the known features of the however, show that Greater Himalayan rocks should be Tibetan Plateau and the Himalaya (see, for example, discus- considered as supracrustal material, much younger than sions by Matte et al. [1997] and Yin and Harrison [2000]). cratonic India, that was stripped from its lower crustal base- [7] The crustal thickening by pure shear model implies ment during the Cenozoic orogeny. This new way of looking that the Tibetan Plateau has thickened either incrementally at the Greater Himalayan rocks revives the need to dispose of since early Mesozoic time through the accretion of Eurasian a large amount (>1.4 Â 107 km3) of lower continental crust crustal blocks or entirely during the Cenozoic collision and its associated mantle lithosphere beneath the Tibetan between the Indian and Eurasian plates. Recent studies Plateau [Ni and Barazangi, 1984]. Exactly how much of this indicate that incremental shortening and magmatism may crust must be accounted for, and how much this may have have played important roles in thickening the crust of the 12 - 4 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY

Figure 2. Digital elevation model topography draped onto a cartoon of the crustal and upper mantle structure in the Himalayan-Tibetan region along a north-south swath between 88°E and 93°E. Note the difference between the topographic and subsurface vertical scales. Structure in upper mantle is modified after Jin et al. [1996], Owens and Zandt [1997], Kosarev et al. [1999], and Kind et al. [2002]. Arrows represent possible west-east flow to account for east-west anisotropy in mantle [McNamara et al., 1994]. Abbreviations as follows: MFT, Main Frontal thrust; MBT, Main Boundary thrust; MCT, Main Central thrust; ISZ, Indus-Yalu suture zone; BSZ, Banggong suture zone; QA, Qiangtang anticlinorium; JS, Jinsha suture; KF, Kunlun fault; SGT, Songpan-Ganzi terrane; QB, Qaidam basin; QS, Qilian Shan.

Plateau during Mesozoic time [Murphy et al., 1997; Kapp et because it has been thrust (or ‘‘extruded’’) southward in the al., 2000, 2002; Yin et al., 1999; Robinson et al., 2002; form of the Greater Himalayan high-grade metamorphic Tapponnier et al., 2001]. However, the presence of mid- rocks that constitute much of the higher part of the Himalayan Cretaceous marine rocks at many localities in the southern fold-thrust belt [Gansser, 1964; Dewey et al., 1988; Hauck et part of the Tibetan Plateau indicates that much of the region al., 1998]. Models that require mantle dynamics include must have been at or below sea level until late Cretaceous or delamination-induced isostatic rebound in combination early Tertiary time, which implies that the crust was not thick. with crustal thickening by pure shear [Molnar et al., 1993] Models that invoke large amounts of pre-Cenozoic crustal and those that inflate the crust by significant amounts of thickening are not necessarily at odds with other mechanisms magmatic underplating [e.g., Powell and Conaghan, 1973; of Plateau development [e.g., Yin and Harrison, 2000]. The Molnar, 1988; Bird, 1991]. The rigid indentation model crustal underthrusting model, which in its most basic form [Tapponnier et al., 1982] requires long-distance eastward was first presented by Argand [1924], proposes an increase of tectonic escape of large crustal fragments accommodated by crustal thickness by tectonic insertion [Powell and Cona- strike-slip faulting. Elements of this model have been sup- ghan, 1973; Ni and Barazangi, 1984; Chemenda et al., 2000] ported by recent studies of the geology of the Tibetan Plateau or ductile injection [Zhao and Morgan, 1987] of Indian lower and by neotectonic studies of fault slip rates. However, the crust beneath the Tibetan crust, with or without Indian litho- actual amounts of lateral escape versus distributed shortening spheric mantle. Uplift of the Plateau has also been attributed and rotation are still debated [e.g., Yin and Harrison, 2000]. It to a phase transition from eclogite to granulite in an inter- is quite plausible that elements of several of the existing mediate composition lower crust that has been underthrust models for the origin of the Tibetan Plateau may have from the south [LePichon et al., 1997]. The underthrusting conspired to produce today’s Plateau. model has been challenged because geophysical studies reveal that Indian lithosphere is probably not present under the northern part of the Tibetan Plateau [e.g., Molnar, 1988; 3. Geology of the Himalayan Fold-Thrust Belt McNamara et al., 1997; Owens and Zandt, 1997], and and Tibetan Plateau insertion of Indian lithosphere beneath the Plateau would 3.1. Himalayan Fold-Thrust Belt have been impeded by the presence of Eurasian lithosphere. In addition, it is generally thought that no need exists to [8] The Himalayan fold-thrust belt stretches for an arc- account for Indian crustal basement beneath the Plateau length distance of 2400 km between the Hazara (in the DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 5

Figure 3. (a) Cross sectional restoration of the northern margin of Greater India prior to the Cenozoic Himalayan-Tibetan orogenic event. This architecture was disrupted by development of the Himalayan fold-thrust belt during Cenozoic time. Abbreviations as follows: LHT, Lesser Himalayan terrane; GHT, Greater Himalayan terrane; THT, Tibetan Himalayan terrane. (b) Generalized regional cross section of the Himalayan fold-thrust belt and Tibetan Plateau at the longitude of western Nepal-southwestern Tibet, after Yin and Harrison [2000], DeCelles et al. [2001], and Kapp et al. [2002]. Note that the hypothetical Greater Himalayan rocks that form the lower part of the Lhasa terrane in (b) are not structurally contiguous with rocks of the Greater Himalayan zone. Abbreviations as follows: MFT, Main Frontal thrust; LHZ, Lesser Himalayan zone; GHZ, Greater Himalayan zone; STDS, South Tibetan detachment system; THZ, Tibetan Himalayan zone; ISZ, Indus-Yalu suture zone; BSZ, Banggong suture zone; JSZ, Jinsha suture zone; AKMS, Anyimaqin-Kunlun-Muztagh suture; KF, Kunlun fault; QS, Qilian Shan thrust belt; ATF, Altyn Tagh fault. west) and Namche Barwa (in the east) syntaxes. Its northern Tibetan Himalayan rocks are commonly referred to as the boundary is defined by the Indus-Yalu suture zone and its Tethyan sequence [Gaetani and Garzanti, 1991; Brookfield, southern boundary by the Main Frontal thrust system and 1993]. largely undocumented blind thrusts beneath the northern part [9] The southern boundary of the Tibetan Himalayan of the Indo-Gangetic foreland basin system (Figures 1 and 3). zone is marked by the South Tibetan detachment system The Indus-Yalu suture zone contains Cretaceous forearc (Figure 3b), a family of top-to-the-north normal faults that basin deposits and ophiolites that formed above the north modified the pre-existing, probably depositional, contact dipping subduction zone between the Neotethyan ocean and between the Tethyan sequence and the underlying Greater the southern, Andean-style margin of Eurasia [Gansser, Himalayan sequence (GHS) [Burchfiel et al., 1992; Hodges 1964; Searle, 1986; Searle et al., 1997]. South of the suture et al., 1996; Godin et al., 1999; Hodges, 2000; Grujic et al., zone lies the Tibetan Himalayan zone, which comprises 2002]. The GHS consists of a 5–20 km thick succession of Cambrian-Eocene rocks in numerous south-vergent thrust metasedimentary rocks and orthogneiss that were metamor- sheets and related folds [Gansser, 1964; Burg and Chen, phosed to amphibolite grade during Eocene and Oligocene 1984; Ratschbacher et al., 1994; Searle et al., 1997]. The time [LeFort, 1986; Peˆcher, 1989; Vannay and Hodges, 12 - 6 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY

1996; Coleman, 1998; Godin et al., 1999; Catlos et al., Schelling, 1992; Srivastava and Mitra, 1994; DeCelles et al., 2001]. Early to middle Miocene leucogranites locally 1998b, 2001]. Where best documented in northern India and intruded the GHS and the overlying Tethyan sequence western Nepal, this duplex consists of a roof thrust sheet (the [LeFort, 1986; Harrison et al., 1999]. Eclogites have been Ramgarh sheet) and several internal horses of the older LHS documented locally in structural culminations in the Tibetan rocks. The faults within the duplex have fed slip southward Himalaya (the north Himalayan gneiss domes). The rocks into the frontal part of the fold-thrust belt. that contain the eclogites are probably of Greater Himalayan [13] The Main Boundary thrust marks the boundary affinity [Tonarini et al., 1993; de Sigoyer et al., 1997; between the LHS and the Neogene Siwalik Group foreland O’Brien et al., 2001]. These rocks were evidently meta- basin deposits [e.g., Burbank et al., 1996]. The portion of the morphosed during Eocene time and subsequently exhumed Himalayan fold-thrust belt between the Main Boundary and from depths of up to 100 km to mid-crustal levels during Main Frontal thrusts is known as the Subhimalayan zone, and late Eocene-Oligocene time. Any model to explain the includes several southward verging thrust sheets composed Himalayan orogenic belt must incorporate a mechanism to almost exclusively of the Siwalik Group [e.g., Schelling and exhume these rocks [Chemenda et al., 2000; O’Brien et al., Arita, 1991; Mugnier et al., 1993; DeCelles et al., 1998b]. 2001; Kohn and Parkinson, 2002]. The structural front of the fold-thrust belt is generally [10] The rocks of the GHS have traditionally been considered to be the Main Frontal thrust, although an regarded as Indian basement that was sliced off of the emerging consensus places the true front in the subsurface northern edge of the subcontinent and thrust southward on beneath the northern part of the Indo-Gangetic foreland basin top of the underlying Lesser Himalayan sequence (LHS) system [Lillie et al., 1987; Powers et al., 1998; Wesnousky et during the Cenozoic orogeny [Gansser, 1964; Dewey et al., al., 1999; Lave´ and Avouac, 2000]. 1988]. Recent U-Pb zircon geochronologic studies [Parrish and Hodges, 1996; DeCelles et al., 2000] and Sm-Nd 3.2. Pre-Cenozoic Structural Architecture of isotopic studies [Parrish and Hodges, 1996; Whittington et Greater India al., 1999; Ahmad et al., 2000; Robinson et al., 2001], [14] Prior to analyzing the contribution of Himalayan however, indicate that the GHS is nearly 1 Gyr younger than shortening to thickening of crust beneath the Tibetan Plateau, LHS rocks. Moreover, the GHS protoliths are upper crustal we must outline the pre-Cenozoic structural architecture of rocks (mainly sediments) that are possibly exotic to India and the northern part of the Indian subcontinent. As discussed experienced regional, medium- to high-grade metamorphism above, the new geochronologic and petrogenetic data from and igneous intrusion during Cambrian-Ordovician time the GHS and LHS suggest that the Greater Himalayan rocks along the tectonically active northern margin of Gondwana were tectonically mobilized during early Paleozoic orogenic [Manickavasagam et al., 1999; DeCelles et al., 2000; Mar- activity along the northern margin of Gondwana. In turn, this quer et al., 2000]. The southern boundary of the Greater requires that the present boundary between the GHS and LHS Himalayan zone is the Main Central thrust, which places the (i.e., the Main Central thrust) be a tectonic overprint of an high-grade metamorphic rocks of the GHS on top of low- older orogenic structure [DeCelles et al., 2000]. The bulk of grade metasedimentary rocks of the LHS. Recent studies the Tethyan sequence was deposited unconformably on top of have proposed that the GHS was extruded southward from this early Paleozoic orogenic terrane. Thus, the pre-Cenozoic beneath the southern Tibetan Plateau, contemporaneous with tectonic architecture of the northern margin of Greater India motion on the Main Central thrust (below) and the South included three tectonostratigraphic terranes (Figure 3a): the Tibetan detachment system (above) [Beaumont et al., 2001; Lesser Himalayan terrane, which rested depositionally upon Hodges et al., 2001; Grujic et al., 2002]. Indian cratonic lower crust; the Greater Himalayan terrane, [11] The LHS is composed mainly of a thick (10 km) which was separated from the Lesser Himalayan terrane by a succession of early to middle Proterozoic greenschist-grade paleostructural zone and consisted of supracrustal (mainly metasedimentary rocks, with local mafic and felsic intru- sedimentary) rocks and underlying Greater Himalayan lower sions. Detrital zircons from the lower part of the succession in crust; and the Tethyan Himalayan terrane, which was depos- NepalaredominatedbyU-Pbagesof1.86 Ga, and ited on top of the Greater Himalayan terrane (and probably mylonitic augen gneisses that intruded into these rocks have partly overlapped onto the Lesser Himalayan terrane) largely yielded U-Pb zircon ages of 1.83 Ga [DeCelles et al., 2000; after the latter had been consolidated with India. These three DePietro and Isachsen, 2001]. On top of the Proterozoic LHS terranes formed the northern part of Greater India as defined rocks in Nepal are the Permian and Paleocene Gondwanas, by Veevers et al. [1975]. Tectonic stripping of the Greater which consist of a thin succession of sedimentary and Indian cover rocks from their lower crust during the Cenozoic volcanic rocks [Sakai, 1989]. On top of the Gondwanas are orogenic event would have produced a slab of lower crustal Eocene and early Miocene foreland basin deposits [Najman rocks that we refer to as Greater Indian lower crust (GILC), et al., 1997; DeCelles et al., 1998a; Najman and Garzanti, which is a composite of Greater Himalayan and Indian 2000]. In northwestern India and Pakistan the LHS is over- cratonic lower crust (Figure 3b). lain directly by lower Paleozoic rocks [Pogue et al., 1999]. [12] The structure of the Lesser Himalayan zone through- 3.3. Tibetan Plateau out much of the central part of the Himalayan fold-thrust belt is dominated by a large antiformal duplex directly south of [15] In this section, we briefly highlight the key geological the Main Central thrust (Figure 3b) [Dhital and Kizaki, 1987; features of the Tibetan Plateau and adjacent areas that must be DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 7 explained (or at least accommodated) by models for the experienced pressures of >8 kbar and temperatures of origin of the Plateau. Our objective is not to provide an 350°–550°C[Kapp et al., 2000]. It is possible that the exhaustive account of Plateau geology; for this the reader is footwall rocks include Songpan-Ganzi flysch, oceanic mel- referred to reviews by Sengor and Natal’in [1996], Yin and ange materials, and Paleozoic passive margin deposits that Harrison [2000], and Tapponnier et al. [2001]. were subducted beneath the upper part of the Qiangtang [16] We adopt the Fielding et al. [1994] definition of the terrane during the Triassic and later unroofed by low-angle Tibetan Plateau as the contiguous region of high (>4.5 km) normal faulting [Kapp et al., 2000]. Hacker et al. [2000] elevation that stretches between the longitudes 72°E and reported xenoliths of metamorphosed sediments in the 99°E, and from the junction of the Altyn Tagh strike-slip central part of the Qiangtang terrane that were heated up fault and the Kunlun Mountains on the north to the Indus- to 1100°C at depths of 30–50 km and then rapidly Yalu suture zone on the south (Figure 1). The Plateau is conveyed (by volcanic eruption) to the surface at approx- characterized by low slopes, average relief of 1 km at long imately 3.5 Ma. This, together with the presence of Pana- (100 km) wavelengths, and internal drainage in its central frican zircons [Kapp et al., 2000] suggests that the lower part. In contrast, the edges of the Plateau have extremely crust of the Qiangtang terrane has affinities with Greater steep slopes and relief locally greater than 6 km [Fielding et Himalayan rocks as well as the Songpan-Ganzi flysch. al., 1994]. In detail, local topographic relief decreases Kapp et al. [2002] documented significant post-mid-Creta- systematically northward across the Plateau (Figure 2), from ceous shortening in the central Qiangtang terrane, and 1–2 km in the southern half, to <1 km in the northern half Horton et al. [2002] showed that 55 km of shortening [Fielding et al., 1994]. The northeastern margin of the occurred in its eastern part during Paleocene and Eocene Tibetan Plateau is bordered by a 500,000 km2 region in time. In addition, the Qiangtang anticlinorium (Figure 1) which average elevation is 2–4 km. Tapponnier et al. involves Tertiary rocks and therefore must have formed [2001] referred to this region as ‘‘Plio-Quaternary Tibet,’’ partly during the Cenozoic [Yin and Harrison, 2000]. and proposed that it consists of alternating north verging However, significant growth of the Qiangtang anticlinorium thrust systems and intervening flexural basins. Shortening in occurred prior to the Indo-Eurasian collision during Creta- Plio-Quaternary Tibet is on the order of 200 km [Me´tivier et ceous northward underthrusting of the Lhasa terrane al., 1998; Tapponnier et al., 2001]. beneath the southern Qiangtang terrane [Kapp et al., 2002]. [17] The Tibetan Plateau consists of three major crustal [19] The Lhasa terrane consists of medium- to high-grade blocks: the Lhasa, Qiangtang, and Songpan-Ganzi terranes metasedimentary rocks of late Precambrian-early Paleozoic [Matte et al., 1996; Yin and Harrison, 2000] (Figures 1 age overlain by Ordovician and Carboniferous to Creta- and 2). The Songpan-Ganzi terrane is bounded on the north ceous sedimentary rocks [Dewey et al., 1988]. Panafrican- by the Kunlun suture zone and on the south by the Jinsha age zircons in the footwall of the Nyainqentanghla detach- suture zone. It consists mainly of thick Triassic turbidites ment [D’Andrea et al., 1999] and the Amdo gneiss [Xu et (the ‘‘Songpan-Ganzi flysch’’) that were deposited in a al., 1985] suggest the presence of crust with Greater remnant ocean basin to the west of the Qinling-Dabie Himalayan affinities beneath the cover rocks. The Lhasa orogenic belt, which marked the collision zone between terrane probably rifted off of northern Gondwana during the North and South China cratons [Yin and Nie, 1996; Zhou Triassic time and collided with the southern margin of the and Graham, 1996; Hacker et al., 1996]. The Songpan- Qiangtang terrane in Late Jurassic-middle Cretaceous time Ganzi basin was tectonically collapsed during late Triassic- [Dewey et al, 1988, Gaetani et al., 1993; Matte et al., 1996; early Jurassic time (see summaries by Zhou and Graham Kapp et al., 2002]. The southern fringe of the Lhasa terrane [1996]; Sengor and Natal’in [1996]; Yin and Nie [1996]; is intruded by the extensive Cretaceous-Eocene Gangdese and Yin and Harrison [2000]). In its eastern part, the batholith belt, which formed an Andean-style continental Songpan-Ganzi terrane experienced thin-skinned thrusting margin arc above the subducting Neotethyan oceanic slab and shortening of several tens of kilometers during early [Alle´gre et al., 1984]. Murphy et al. [1999] reported that as Tertiary time [Yin and Harrison, 2000]. much as 187 km of shortening may have occurred in the [18] South of the Songpan-Ganzi terrane lies the Qiang- Lhasa terrane during Late Jurassic-Cretaceous time. If this tang terrane, which rifted off of northern Gondwana during shortening was accommodated entirely within the Lhasa late Paleozoic time and collided with Eurasia in late terrane, then the thickness of Lhasa terrane crust prior to the Triassic-early Jurassic time [Yin and Nie, 1996; Yin and Himalayan orogeny could have been 55 km [Murphy et Harrison, 2000]. The Qiangtang terrane consists of meta- al., 1997], enough to support regional elevations of 3 km, morphic rocks overlain in structural contact by upper assuming Airy isostasy. However, mid-Cretaceous shallow Paleozoic and Mesozoic strata. The metamorphic rocks marine limestones are widespread in the Lhasa terrane, consist of metasedimentary and mafic schists that enclose indicating that regions of low elevation persisted until the less deformed blocks of blueschist-bearing metabasites, Late Cretaceous [Zhang, 2000]. Kapp et al. [2002] pre- upper Paleozoic strata, and minor ultramafic rocks [Kapp, sented evidence for northward underthrusting of Lhasa 2001; Kapp et al., 2000, 2001]. Kapp et al. [2000] inter- terrane lower crust beneath the Qiangtang terrane; this preted these rocks as a metamorphosed melange, and the process would minimize the amount of Cretaceous crustal contacts between the melange and the overlying Paleozoic- thickening in the Lhasa terrane. Cenozoic shortening in the Mesozoic rocks as low-angle normal faults that were active Lhasa terrane is minimal [Dewey et al., 1988; Pan, 1993; during Late Triassic-Early Jurassic time. The footwall rocks Murphy et al., 1997; Yin and Harrison, 2000], being largely 12 - 8 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY confined to displacements on thrust systems that over- Tibetan detachment system and numerous north-south to printed the Indus-Yalu suture zone [Yin et al., 1999]. northeast-southwest striking graben. These basins are largest [20] From the above summary, we conclude that large and most numerous in the southern part of the Plateau [Yin, portions of the Tibetan Plateau are underlain by rocks that 2000], where they cut across the Tibetan Himalaya and the have upper crustal compositions and physical properties Lhasa terrane (Figure 1). Shallow earthquakes in the modern [e.g., Owens and Zandt, 1997; Hacker et al., 2000; Kapp et Tibetan Plateau are dominated by strike-slip and extensional al., 2000]. The isotopic and geochronologic characteristics of focal mechanisms [e.g., Armijo et al., 1986; Molnar and these rocks, where known, suggest affinities with Greater Lyon-Caen, 1989]. The direction of extension is approxi- Himalayan (Late Proterozoic-early Paleozoic) and Mesozoic mately east-west, at nearly a right angle to the direction of metamorphic rocks. The Mesozoic-Cenozoic orogenic compression in the Himalayan fold-thrust belt [Armijo et al., events of central Eurasia, therefore, may be viewed as a 1986; Molnar and Lyon-Caen, 1989]. Total east-west exten- piecemeal redistribution of exotic tectonostratigraphic sion amounts to less than 3%, or 40 km [Armijo et al., 1986; assemblages from Gondwana to Eurasia [e.g., Sengor and Molnar et al., 1993]. Natal’in, 1996; Yin and Nie, 1996]. Documented shortening [22] Because of a possible linkage between gravitational of upper crust in the central and northern Tibetan Plateau extension in the upper crust and attainment of a critical during Cenozoic time is <100 km [Horton et al., 2002; Kapp thickness of the crust [Molnar and Lyon-Caen, 1989; Stu¨we et al., 2002]. To the north of the Kunlun fault in Plio- and Barr, 2000], much interest has been focused on the Quaternary Tibet, an additional 200 km of shortening timing of extension in the Tibetan Plateau. North-south occurred during mid-late Cenozoic time [Me´tivier et al., extension along the South Tibetan detachment system 1998; Tapponnier et al., 2001]. In the southern part of the commenced 22 Ma and occurred episodically during early Plateau, Cenozoic shortening was <50 km [Pan, 1993; Yin and middle Miocene time along the length of the fault and Harrison, 2000]. Evidence for Mesozoic shortening in system [Hodges et al., 1996; Edwards and Harrison, 1997; the upper crust of the Plateau (and presumably crustal Coleman, 1998; Hodges, 2000; Yin and Harrison, 2000; thickening) is widespread. However, even where the esti- Grujic et al., 2002]. Neotectonic studies in central Nepal mates of pre-Cenozoic crustal shortening are greatest (in the suggest that faults in this system may still be active Lhasa terrane [Murphy et al., 1999]) the presence of mid- [Hurtado et al., 2001]. Most of the available evidence Cretaceous shallow-marine sedimentary rocks suggests that suggests that east-west extension in the Plateau began in parts of the region lay close to sea level until the late mid- to late Miocene time (14–8 Ma) [Harrison et al., Cretaceous. The oldest widespread nonmarine sedimentary 1992, 1995; Coleman and Hodges, 1995; Edwards and rocks with relevance for the timing of uplift in the Lhasa Harrison, 1997; Garzione et al., 2000a, 2000b; Williams et terrane are Paleocene-Eocene in age, and even these were al., 2001; Blisniuk et al., 2001]. With respect to the issue of probably deposited close to sea level [Willems et al., 1996] the timing of elevation gain, Garzione et al. [2000a, 2000b] (summary by Rowley [1996]). On the other hand, the pres- reported oxygen isotopic data that indicate the depositional ence of the Gangdese batholith along the southern margin of floor of Thakkhola graben (Figure 1) has been at elevations Eurasia prior to the Cenozoic collision, evidence for up to >4 km since at least 11 Ma. 60% pre-Cenozoic horizontal shortening in the central [23] Several major strike-slip faults are present on the Lhasa terrane, and evidence for Mesozoic crustal thickening Tibetan Plateau, including the Altyn Tagh, Kunlun, Xing- and extensional detachment faulting in other parts of the shuihe, Jiali, Red River, and Karakoram fault systems. In Tibetan Plateau imply that the crust in the southern and eastern Tibet left- and right-lateral faults translate Tibetan central parts of the Plateau may have been as thick as terranes eastward, although the amounts of translation versus 50–55 km before the Himalayan orogeny [Murphy et al., rotation and distributed motion are still debated [e.g., Tap- 1999; Yin et al., 1999; Robinson et al., 2002; Kapp et al., ponnier et al., 1982, 1990; Burchfiel et al., 1995; Yin and 2000, 2002]. Unless a dense eclogitic root or dynamic mantle Harrison, 2000; Wang et al., 2001]. The left-lateral Altyn processes held this thickened crust at low elevations, it is Tagh fault, which forms the northwestern boundary of the likely that local elevation may have been on the order of 2–4 Tibetan Plateau, has accommodated 280–550 km [Peltzer km. For the crust of the Tibetan Plateau (excluding Plio- and Tapponnier, 1988; Yin and Harrison, 2000] or 375 ± 25 Quaternary Tibet) to reach its present thickness and high km [Yue et al., 2001] of slip since Oligocene time. The right- regional elevation, it must have thickened by an additional lateral Karakoram fault in the western part of the Plateau may 10–20 km during the Himalayan orogenic event, without transfer slip between the Himalayan fold-thrust belt and the shortening internally by more than 15–20%. Pamir Range [Ratschbacher et al., 1994; Yin et al., 1999; Murphy et al., 1999]. [24] In addition to the widespread evidence for extensional 3.4. Extension and Strike-Slip Faulting in the and strike-slip faulting at shallow levels (<20 km) in Tibetan Tibetan Plateau crust [Molnar and Lyon-Caen, 1989], a small number of [21] Two sets of normal faults and several major strike-slip moderate-sized earthquakes have been documented at depths faults dominate the present structure of the Tibetan Plateau of 70–113 km beneath the Lhasa terrane and Himalayan [Armijo et al., 1986; Molnar and Lyon-Caen, 1989; Tappon- fold-thrust belt (Figure 2) [Ekstrom, 1987; Chen and Molnar, nier et al., 2001]. The normal fault systems include the 1983; Chen and Kao, 1996; Zhu, 1998]. These earthquakes generally northwest-southeast to east-west striking South occurred within the mantle lithosphere, and demonstrate that DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 9 the upper mantle is cold and strong enough in this region to store elastic strain energy. The handful of these earthquakes for which fault plane solutions are available exhibit horizon- tal, east-west extension, similar to the faults in the upper crust of Tibet [Molnar and Chen, 1983; Chen and Molnar, 1983; Chen and Kao, 1996].

3.5. Magmatism on the Tibetan Plateau

[25] The southern margin of Eurasia was occupied by the >2000 km long, calc-alkaline Gangdese magmatic arc from mid-Cretaceous through mid-Eocene time (Figure 1) [Chang and Zheng, 1973; Tapponnier et al., 1981; Alle´gre et al., 1984]. Silicic volcanic rocks, mainly ash flow tuffs of the Upper Cretaceous-Paleocene Linzizong Formation, are widespread in the Lhasa terrane and the southern part of the Qiangtang terrane [Pan, 1993; Kapp et al., 2002; Ding et al., 2002]. High-potassium magmas mixed with a mantle component have erupted on the Tibetan Plateau since 45 Ma, with major pulses of activity during late Eocene- Oligocene time in the Qiangtang terrane and during the Miocene-Quaternary in the Songpan-Ganzi and Lhasa ter- ranes [Deng, 1978; Coulon et al., 1986; Harris et al., 1988; Turner et al., 1993; Chung et al., 1998; Miller et al., 1999, 2000; Ding et al., 2002]. Widespread geothermal activity and several independent geophysical and petrological data sets indicate that the crust of the Tibetan Plateau is abnor- mally hot, although the debate persists whether partial melt is pervasive throughout the plateau or restricted to the extensional grabens [Nelson et al., 1996; Owens and Zandt, 1997; Alsdorf and Nelson, 1999; Hacker et al., 2000; Wei et al., 2001; Kind et al., 2002].

4. Geophysical Constraints on Sub-Tibetan Lithospheric and Crustal Structure Figure 4. (a) Tectonic map of India and central Eurasia, showing major terrane boundaries. North-south line X-Y [26] Figures 2, 3, and 4 depict an interpretation of the structure of the lithosphere and upper mantle beneath India indicates location of tomographic profile shown in (b). (b) and central Eurasia based on recent geophysical studies. Simplified version of tomographic model of Van der Voo et Results of the 1991–92 Sino-American PASSCAL broad- al. [1999] along profile X-Y. Vertical ruled zones are band experiment [McNamara et al., 1994, 1997; Owens and characterized by slower than average seismic wave Zandt, 1997], the international and multidisciplinary propagation; stippled areas are zones of faster than average INDEPTH experiments [Nelson et al., 1996; Kosarev et al., wave propagation. (c) Geological interpretation of tomo- 1999; Zhao et al., 2001], and the series of Sino-French graphic profile, showing the Greater Indian lithospheric slab seismic studies [Hirn, 1988; Hirn et al., 1995; Wittlinger et (anomaly I) and the Neotethyan oceanic slab (anomaly II). al., 1996] have begun to delineate the lithospheric-scale Details of the lithosphere under Tibet are from Jin et al. structure beneath the Tibetan Plateau, principally in the [1996], Owens and Zandt [1997] and Kosarev et al. [1999]. central and eastern Plateau. A comprehensive review of even Sloping lines labeled 45° and 55° represent plausible these selected studies is beyond the scope of this paper; projections to the surface of the Neotethyan oceanic slab instead a summary is given of the results most relevant to the (anomaly II) before it broke off and foundered into the thesis of this paper. Most studies agree that the Tibetan crust mantle. Abbreviations: MFT, Main Frontal thrust; ISZ, is up to 75 km thick in the southern Lhasa terrane and thins Indus-Yalu suture zone; BSZ, Banggong suture zone; ATF, to the north [Owens and Zandt, 1997; Zhao et al., 2001]. The Altyn Tagh fault. crust is 65 km thick in the Qiangtang terrane, and thins to 60 km in the Songpan-Ganzi terrane [Zhao et al., 2001; abrupt step in the Moho beneath the boundary between the Kind et al., 2002]. Some studies have suggested that the North Kunlun Mountains and Qaidam Basin [Zhu and crustal thickness variations occur as abrupt steps correspond- Helmberger, 1998], but the data for the Jinsha suture are ing to suture boundaries such as the Jinsha suture and major more equivocal owing to larger station spacing. The Moho strike-slip faults such as the Kunlun fault [Hirn et al., 1995; step at the Kunlun-Qaidam border is associated with a change Wittlinger et al., 1996]. The seismic evidence is strong for an in topography, so it is not particularly surprising. The 12 - 10 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY presence or absence of a step beneath the Jinsha suture zone, functions to image a shallow north dipping interface (their where there is no major change in elevation, is potentially Zangbo Conversion Boundary) at a depth of 80 km just north more significant but still uncertain. The 10–15 km gradual of the Indus-Yalu suture zone to a depth of 200 km beneath variation in crustal thickness across the Plateau may owe to the Banggong suture. They suggested that the Zangbo Con- the Moho adjusting independently of the surface to provide version Boundary is evidence for the presence of subducting, isostatic compensation for lateral density contrasts in the cold Indian mantle lithosphere beneath the Plateau, support- upper mantle [Molnar, 1988; Bird, 1991; Owens and Zandt, ing the model of Jin et al. [1996]. Zandt et al. [2000] further 1997; Zhao et al., 2001]. supported this interpretation with geoid modeling that sug- [27] Although the details of the plateau crustal structure gests a high density, north dipping Indian lithospheric slab are still sketchy, a few characteristic traits have been (stripped of most of its original crust) underlying much of the documented. The bulk crustal seismic velocities are gen- Plateau. In the combined data analysis, the Zangbo Conver- erally low, consistent with an average composition more sion Boundary was not enhanced, and the weight of the felsic than average continental crust [Owens and Zandt, seismic evidence has swung back to a horizontal Indian 1997; Rodgers and Schwartz, 1998; Zhao et al., 2001]. lithosphere beneath the southern plateaus [Kind et al., Seismic reflection profiling in the Lhasa terrane observed 2002]. The alternative interpretations of the distribution of mid-crustal ‘‘bright spots’’ interpreted as evidence for Greater Indian lithosphere beneath Tibet are depicted in pervasive partial melting in the crust [Nelson et al., 1996] Figure 3b. For the purposes of this paper, the alternatives or a combination of fluids and melt [Makovsky and Klem- are not significantly different from each other; the amount of perer, 1999]. Magnetotelluric studies have found the middle shortening predicted in the Himalaya is essentially the same and lower crust of the plateau are anomalously conductive, regardless of which interpretation is adopted. most likely because of the presence of aqueous fluids and [29] An important additional piece of evidence on the partial melt [Weietal., 2001]. Combining seismic conver- nature of the plateau mantle is the presence of east-west sion data from several international experiments, Kind et al. oriented seismic anisotropy in the upper mantle under the [2002] concluded that over most of the central plateau, the plateau that is especially strong beneath the northern plateau average Vp/Vs is near normal, indicating that despite the [McNamara et al., 1994]. Two competing interpretations apparent high conductivity, the volume of fluid and melt is have been offered for these observations. One interpretation not greater than a few percent. Available data indicate that is that the upper mantle beneath northern Tibet is composed much of the Lhasa terrane is underlain by a 15 km thick of a thick keel of north-south shortened Eurasian lithosphere lower crustal layer where the seismic velocity increases up and the seismic anisotropy is reflecting the resulting east- to 7.2 km/s or higher, causing the ‘‘doublet’’ arrival in the west deformation fabric in the mantle [McNamara et al., seismic conversion data [Owens and Zandt, 1997; Kind et 1994]. An alternative interpretation is that the horizontal al., 2002]. The correlation of the high velocity lower crustal motion of rigid Tibetan lithospheric blocks shears the layer and high velocity upper mantle in southern Tibet asthenosphere below and produces the east-west anisotropy suggests the presence of seismically fast lower Indian in the asthenospheric mantle [Hirn et al., 1995; Lave´etal., cratonic crust and lithosphere beneath southern Tibet (Fig- 1996]. Holt [2000] rejected this latter mechanism based on ures 2 and 3). The northern termination of the high velocity the apparent absence of strong anisotropy beneath the much layer may represent the end of cratonic Indian lower crust, more rapidly moving Indian subcontinent [Chen and O¨ za- but the Greater Himalayan lower crust, of differing compo- laybey, 1998]. However, recent studies show that astheno- sition, may continue farther north (Figure 3b). spheric-flow induced anisotropy appears less related to [28] The interpretation that the Indian mantle lithosphere absolute plate motions and more related to forced flow (including Indian lower crust) underthrusts the Tibetan Pla- around lithospheric slabs and keels [Russo and Silver, 1994; teau nearly horizontally as far north as the Banggong suture Fouch et al., 2000]. This supports an alternative interpreta- zone (32°N) is based on numerous seismological studies that tion that north of the Banggong suture, weak asthenosphere indicate a seismically fast (cold) upper mantle beneath the flows eastward in response to compression between con- southern part of the Plateau and a seismically slow (warm) verging strong and thick Indian and Eurasian lithospheres mantle beneath its northern part [e.g., Ni and Barazangi, (Figures 2 and 3b). In this model, the anisotropy is attrib- 1984; McNamara et al., 1997; Owens and Zandt, 1997; Kind uted to a combination of sublithospheric mantle flow et al., 2002]. Even among those who agree on the existence of [Owens and Zandt, 1997] and lithospheric fabric associated the Indian lithosphere beneath southern Tibet, the dip of the with a south dipping Eurasian lithosphere [Furlong and underthrust Indian lithosphere has been the subject of debate. Owens, 1997]. This idea is consistent with Holt’s [2000] Owens and Zandt [1997] and others have suggested that the conclusion that the vertical coherence of deformation indi- Indian slab is horizontal to the latitude of the Banggong cators in Tibet ‘‘may be influenced more by the velocity suture. Based on gravity data, Jin et al. [1996] suggested that boundary conditions imposed on both crust and mantle than the Indian plate is subducting under the Lhasa terrane at a by coupling between crust and upper mantle.’’ Recently moderate angle instead of horizontally. In their model, the produced seismic images showing a southward dipping Indian-Eurasian mantle suture is located beneath the Indus- converter boundary beneath northern Tibet [Kosarev et Yalu suture zone and the Greater Indian lithosphere continues al., 1999; Tapponnier et al., 2001; Kind et al., 2002] have to subduct into the upper mantle for an indeterminate bearing on this interpretation. A logical interpretation of this distance. Kosarev et al. [1999] modeled migrated receiver seismic feature is that it marks the top of southward DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 11 subducting Eurasian lithosphere. Magnetotelluric studies plate, this separation would have developed since 10 Ma, have documented a broad conductivity anomaly centered more than 40 Myr after the initial collision. It seems unlikely beneath the Qiangtang terrane extending into the upper that Neotethyan lithosphere would have continued to subduct mantle. The anomaly is interpreted as partial melting due beneath central Eurasia until late Miocene time. Moreover, to localized upwelling of asthenosphere, and is bounded on this reconstruction would predict only a few hundred kilo- the north and south by relatively resistive zones that may meters of underthrusting of Greater India beneath the Hima- indicate where Eurasian and Indian lithosphere descends laya, which conflicts with the conservative estimates of [Wei et al., 2001]. Based on this interpretation of multiple shortening in the Himalaya discussed in the following data sets there is direct evidence that the Eurasian litho- section. We therefore concur with Van der Voo et al. [1999] sphere, beneath the crust, is behaving like a plate and not that tomographic anomaly I is composed of Greater Indian deforming by pure shear. lithosphere. In a later section, we further postulate that this [30] Recent work on the seismic anisotropy in the Plateau fragment of lithosphere is composed of Greater Himalayan crust using data from the 1991–92 PASSCAL experiment is material. also relevant to some of these questions (H. Folsom et al., [32] Neither the seismic data nor the geoid observations manuscript in preparation, 2002). Middle to lower crustal can resolve in detail the presence or absence of GILC beneath anisotropy is present at most stations, with a unique fast axis the Tibetan Plateau or on the detached slab of inferred Greater trending north-south to northwest-southeast in the south, Indian lithosphere (anomaly I). However, the broadband nearly east-west in the central Plateau, and north-south to seismic results dictate that a slab of GILC must extend at northeast-southwest in the northern Plateau. This pattern least as far north as 32°N directly beneath the Plateau. From appears consistent with recent ductile deformation due to this latitude northward, there is no lithospheric barrier that both topographically induced flow and to boundary forces would have blocked further northward insertion of GILC from subducting lithosphere at the northern and southern beneath the Plateau. We therefore suggest that a northward margins of the Plateau. The orientations of crustal anisotropy tapering slab of GILC continues northward beneath the are not entirely consistent with shear wave splitting fast Tibetan crust, perhaps as far north as the Kunlun fault polarization directions, potentially implying distinct motions (Figure 3b). in the crust and mantle (Folsom et al., manuscript in prepa- ration, 2002). [31] The structure of the deeper mantle beneath India and 5. Shortening in the Himalayan central Eurasia is revealed by recent tomographic studies. Fold-Thrust Belt Tomographic models of Grand et al. [1997] and Van der Voo et al. [1999] suggest that relatively cold, seismically fast [33] The timing of initial contact between Greater India lithosphere of the Indian plate plunges northward into the and Eurasia remains a subject of debate. The usual argument asthenosphere to a depth of several hundred kilometers employed for initial contact is the timing of disappearance of beneath the central part of the Himalaya (Figure 4, anomaly marine waters between the two continents. However, many I). The location of this anomaly 500 km south of the modern foreland basin systems are filled partly or wholly by northern end of Greater Indian lithosphere beneath the marine waters [DeCelles and Giles, 1996; Sinclair, 1997], Tibetan Plateau (Figure 3b) suggests that the anomaly repre- and so this is not a meaningful indicator of contact between sents a slab of Greater Indian lithosphere that is no longer Greater India and Eurasia. We prefer to use the timing of attached to the Indian plate. A second, deeper (1000–2000 initial foreland basin development. In Nepal and northern km) region of cold, fast lithosphere that was imaged to the India, it has been shown that shallow marine sediments of south of the inferred Greater Indian slab was interpreted by early to middle Eocene age were derived from the nascent Van der Voo et al. [1999] as Neotethyan oceanic lithosphere Himalayan fold-thrust belt and the Indus-Yalu suture zone that broke off of the Indian plate at the onset of collision and deposited in a southward migrating foreland basin (Figure 4, anomaly II). This deeper slab of (presumably) system [Pivnik and Wells, 1996; Najman et al., 1997; oceanic lithosphere can be traced continuously in tomo- DeCelles et al., 1998a; Najman and Garzanti, 2000]. We graphic images from northern Indonesia to the eastern therefore place the time of initial thrusting in the Himalaya as Mediterranean region [Spakman, 1991; Van der Voo et al., 55 Ma, which is consistent with several previous independ- 1999]. Continued northward motion of the Indian plate ent estimates [Klootwijk et al., 1985; Besse and Courtillot, during early-middle Tertiary time carried the lower end of 1988; LePichon et al., 1992; Patzelt et al., 1996]. the inferred Indian slab (anomaly I) northward away from the [34] Plate tectonic reconstructions based on paleomag- detached Neotethyan slab (Figure 4). Alternatively, anomaly netic data from the Himalayan orogenic belt suggest that I (Figure 4) may be interpreted as a slab of Neotethyan 2600 ± 900 km of post-collision convergence has taken oceanic lithosphere. This interpretation, however, conflicts place between Eurasia and Greater India, with 1700 ± 610 with known rates of Indian plate migration and the offsets km of this total accommodated by north-south shortening between anomalies I and II and the northern end of Greater in the Tibetan Plateau and lateral tectonic escape [Patriat Indian lithosphere beneath the Tibetan Plateau. The lateral and Achache, 1984; Achache et al., 1984; Besse and offset between anomaly I and the northern end of Greater Courtillot, 1988, 1991]. The 900 km difference between Indian lithosphere is only 500 km (Figure 5, point 6). At a these average values is available for shortening in the rate of 50–55 mm/yr of northward migration of the Indian Himalayan fold-thrust belt [LePichon et al., 1992]. Pub- 12 - 12 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY

Figure 5. Simplified version of the cross section depicted in Figure 3, annotated with the key geometric constraints that must be explained in a kinematic model of the India-Eurasia collision. Sources of information on Tibetan shortening: Yin and Harrison [2000], Tapponnier et al. [2001], Horton et al. [2002]. See text for discussion. lished estimates of shortening in the Himalaya based on derived from geologic data from the Himalayan fold-thrust paleomagnetic data range between 700 km and 1500 km belt. [Patzelt et al., 1996]. [36] Only a handful of regional balanced cross-sections [35] An alternative approach to predicting the amount of through the Himalayan fold-thrust belt in Nepal, India, and shortening in the Himalayan fold-thrust belt is given in Pakistan have been published (Figure 1). Table 1 lists the Figure 6. The present length of Indian crust at the surface is available balanced cross sections and estimates of shortening. 2304 km (measured along 83°E). Paleomagnetic data Coward and Butler [1985] produced the only complete indicate a long-term convergence rate of 55 mm/yr between structural transect of the Himalayan fold-thrust belt, in north- Greater India and Eurasia since early Tertiary time [e.g., ern Pakistan, which yielded a total shortening estimate of 470 Patriat and Achache, 1984; Dewey et al., 1989; Molnar et al., km. Complete sections in the remainder of the Himalaya must 1993]. This rate can be used to hindcast the position of the be stitched together by combining cross sections by different Indian continent at 55 Ma, placing its southern tip at investigators. The available studies are generally divided into latitude 25°S at the onset of the collision. Current estimates those that cover the Indus-Yalu suture zone, the Tibetan of the paleolatitude of the Indus-Yalu suture zone place it at Himalaya, or the portion of the fold-thrust belt south of the 6.5°N±2.5° at the onset of collision [Klootwijk et al., 1985; South Tibetan detachment system. Van der Voo, 1993]. The distance between this paleolatitude [37] Searle [1986] estimated that 126 km of shortening and the paleolatitude of the southern tip of the continent occurred in the Tibetan Himalaya of northern India (near provides the length of Indian continental crust at the onset of longitude 76°E), and Searle et al. [1997] increased this collision, 3075 ± 192 km. The 771 ± 192 km difference amount to 150–170 km based on more recent work. These between this length and the present length of Indian con- authors argued that it is not feasible to balance cross sections tinental crust at the surface is the predicted total shortening in this part of the Himalaya because the deformation is so (Figure 6), which is within error of the estimate by LePichon intense. Based on detailed mapping, Steck et al. [1998] et al. [1992]. If we assume that the onset of collision was at 50 suggested that hundreds of kilometers of shortening have Ma, then the shortening estimate reduces to 446 ± 192 km. occurred in the Tibetan Himalayan zone of northern India but Let us see how these estimates compare with estimates they presented no balanced cross sections. Ratschbacher et DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 13

Figure 6. Plot showing the northward trajectory of Greater India, the Indus-Yalu suture zone, and the southern tip of India since 55 Ma, based on discussion in text. The shaded regions reflect uncertainties in paleomagnetic data. al. [1994] produced shortening estimates of 133 km and 139 ening between the Main Central and Main Frontal thrusts of km for balanced cross sections in the Tibetan Himalaya near 228 km. DeCelles et al. [2001] constructed a second cross longitudes 88°E and 90°E. section, 50 km to the east of the previous section, that [38] In the Indus-Yalu suture zone, post-collisional short- suggests 418–493 km of shortening between the South ening has been documented by Yin et al. [1999]. Several tens Tibetan detachment system and the Main Frontal thrust. of kilometers of shortening occurred during Oligocene and [40] By cobbling together the shortening estimates derived middle Miocene thrusting. These thrusts reactivated and from nearby cross sections in the Indus-Yalu suture zone, crosscut suture-related features, suggesting that they are Tibetan Himalaya, and the portion of the Himalaya south of related to internal shortening in the Himalayan hinterland. the South Tibetan detachment system, estimates for the entire [39] Aside from Coward and Butler [1985], seven regional fold-thrust belt may be obtained in four districts (Figure 7). balanced cross sections of the Himalaya south of the South Hauck et al. [1998] combined the cross sections of Schelling Tibetan detachment system have been published. Schelling [1992], Schelling and Arita [1991], and Ratschbacher et al. [1992] constructed two regional sections in eastern Nepal [1994] with the INDEPTH seismic reflection profile to yielding 210–280 km of shortening between the South construct a crustal-scale cross section with shortening of Tibetan detachment system and the Main Frontal thrust. 323 km from the Indus-Yalu suture zone to the Main Schelling and Arita [1991] published a cross section in Frontal thrust in the eastern Himalaya. Greater shortening Sikkim (northeastern India) that yielded shortening estimates estimates are obtained in northern India (480 to 547 km)[Sri- of 185–245 km. Srivastava and Mitra [1994] published a vastava and Mitra, 1994] and western Nepal (556–623 km pair of cross sections separated by 100 km in Kumaon, [DeCelles et al., 1998b]; and 643–669 km [DeCelles et al. northern India, which yielded a range of shortening between [2001]) (Table 1). All of these estimates are minima because 353–421 km. DeCelles et al. [1998b] published a balanced they do not include penetrative strain or small-scale folds and cross section from far western Nepal that indicated short- faults, which could significantly increase the total shortening. 12 - 14 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY

Table 1. Shortening Estimates in the Himalayan Fold-Thrust Belt

Reference in Figures 1 and 7 Reference Location Structural Boundariesa Tectonostratigraphic Zonesb Shortening, km

1 Coward and Butler [1985] Pakistan MMT-MFT GHZ, LHZ, SHZ 470 km 2 Searle [1986] India, Zanskar and Ladakh ISZ-MMT THZ 126 km 3 Searle et al. [1997] India, Zanskar and Ladakh ISZ-MCT THZ 150–170 4 Srivastava and Mitra [1994] India, Kumaon and Garhwal STDS-MCT GHZ + Almora thrust sheet 193–260 km 5 Srivastava and Mitra [1994] India, Kumaon and Garhwal MCT-MFT LHZ, SHZ 161 km 6 DeCelles et al. [1998a] Western Nepal MCT-MFT LHZ, SHZ 228 km 7 Murphy and Yin [2003] Tibet, Mt. Kailas region ISZ-STDS THZ + Indus Suture 176 km 8 DeCelles et al. [2001] Western Nepal, Seti River STDS-MCT GHZ + Dadeldhura thrust sheet 131–206 km 9 DeCelles et al. [2001] Western Nepal, Seti River MCT-MFT LHZ, SHZ 287 km 10 Ratschbacher et al. [1994] Tibet, north of Arun River ISZ-MCT THZ 133–139 km 11 Schelling [1992] Central Nepal STDS-MCT GHZ 140–210 km 12 Schelling [1992] Central Nepal MCT-MFT LHZ, SHZ 70 km 13 Schelling and Arita [1991] Far Eastern Nepal STDS-MCT GHZ 140–175 km 14 Schelling and Arita [1991] Far Eastern Nepal MCT-MFT LHZ, SHZ 45–70 km

aMMT, Main Mantle thrust; ISZ, Indus-Yalu suture zone; MFT, Main Frontal thrust; MCT, Main Central thrust; STDS, South Tibetan detachment system. bGHZ, Greater Himalayan zone; LHZ, Lesser Himalayan zone; SHZ, Subhimalayan zone; THZ, Tibetan Himalayan zone.

[41] Figure 7 provides an along-strike comparison of the consistent with paleomagnetic data that indicate rotations in available estimates of shortening in the Himalayan fold- a counterclockwise sense in the eastern half of the orogenic thrust belt, and the corresponding width of the Tibetan belt and a clockwise sense in the western half of the belt Plateau. Several features of this diagram stand out. First, [Klootwijk et al., 1985]. Third, the width of the Tibetan the eastern third of the fold-thrust belt is terra incognita from Plateau north of the Indus-Yalu suture zone and the amount of the standpoint of crustal shortening estimates. The vastness shortening in the Himalayan fold-thrust belt are remarkably of this region and its obvious importance for the origin of the similar for the western half of the orogenic belt. The mis- Tibetan Plateau beckon geological and geophysical studies. match, by a factor of almost two, in the eastern part of the Second, the diagram suggests that the greatest amount of fold-thrust belt and the adjacent Tibetan Plateau we attribute shortening may be accommodated in the central part of the to the following. The estimates of shortening in the Himalaya Himalayan arc, in northern India and western Nepal. The of eastern Nepal are based on cross sections [Schelling and increase in displacement toward the apex of the arc is Arita, 1991; Schelling, 1992] that do not include two major

Figure 7. Compilation of shortening estimates along the Himalayan fold-thrust belt from west to east between the Hazara and Namche Barwa syntaxes. Dashed line indicates corresponding width of the Tibetan Plateau as measured in an arc-normal direction. Numbers indicate sources of data as listed in Table 1. DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 15

al. [1997], it seems quite likely that the existing values of shortening for the Tibetan Himalayan zone are substantial underestimates. If the actual rate of shortening in the Tibetan Himalayan zone were equal to the Neogene short- ening rates in the remainder of the Himalaya, then the total shortening would be on the order of 900–1000 km. More work in the Tibetan Himalaya is needed to clarify this issue.

6. An Integrated Model for Uplift of the Tibetan Plateau

[43] In the following, we present a hypothetical, modified version of the crustal underthrusting model (Figure 3) that can accommodate the kinematic constraints and shortening estimates from balanced cross sections in the Himalayan fold-thrust belt (Figures 5–8), recent revisions in our under- standing of the origin of Greater Himalayan rocks, and the available geophysical data from the Tibetan Plateau (Figures 3 and 4). Figure 5 illustrates the geometrical con- straints on possible kinematic models for the Tibetan Plateau. Analog modeling provides additional insight into the machi- nery of the Indo-Eurasian collisional process [Chemenda et al., 2000]. [44] The Greater Indian lithosphere subducts at a low angle beneath the Lhasa terrane and continues subhorizon- Figure 8. Plot of cumulative shortening in the Himalayan tally at least as far north as the Banggong suture zone fold-thrust belt of western Nepal, based on kinematic (Figure 3b). Eurasian lithosphere subducts at a moderate evidence summarized by Ratschbacher et al. [1994] and angle to a depth of 300 km beneath the northern part of DeCelles et al. [1998b, 2001]. Shown along left-hand  the Plateau. The intervening region in the upper mantle vertical axis are locations (present coordinates) of major between approximately 33°N and 35°N is occupied by hot tectonic features in the Tibetan Plateau. asthenosphere (Figures 2 and 3b) [Owens and Zandt, 1997; Kosarev et al., 1999; Weietal., 2001; Kind et al., 2002]. The tomographic images of Van der Voo et al. [1999] thrust systems that have been mapped in western Nepal and combined with the broadband seismology results suggest northern India: the Ramgarh thrust and the Almora-Dadeld- that the total length of subducted Greater Indian lithosphere hura thrust [Valdiya,1980;Srivastava and Mitra,1994; (beneath the Plateau and detached in the mantle) is on the DeCelles et al., 1998b, 2001]. Where mapped, these faults order of 600–900 km (Figures 4 and 5). The older, Neo- accommodated up to 250 km of shortening. Reconnais- tethyan oceanic slab is depicted as a vertical planar domain sance mapping in central and eastern Nepal [Pearson, 2002], at depths of 1000–2000 km in the mantle at latitude 20°N U-Pb detrital zircon ages [DeCelles et al., 2000], and Sm-Nd (Figures 4 and 9f). isotopic studies [Robinson et al., 2001] indicate that struc- [45] The length of the GILC beneath the Plateau is tural and stratigraphic equivalents of both of these thrust predicted to be at least 650–700 km, consistent with sheets are indeed present in eastern Nepal. The state of the the hypothesis that its length should approximately equal mapping is not sufficient to produce balanced cross sections, the shortening in the Himalayan fold-thrust belt (Figure 5, but if displacements on these two thrusts in eastern Nepal are point 8). This slab of crust is probably a composite of roughly equivalent to their displacements in western Nepal, Archean Indian cratonic lower crust (Figure 5, point 9) and then the total minimum shortening in eastern Nepal should late Proterozoic-Cambrian lower crust of Greater Hima- increase to 650 km (Figure 7). layan affinity (Figure 5, point 10). The length of Greater [42] The long-term average rate of shortening in the Himalayan lower crust is expected to be on the order of Himalayan fold-thrust was 20 mm/yr throughout most several hundred kilometers (Figure 5, point 10), which is of Neogene time and decreased to about 13–20 mm/yr consistent with the length of the detached Greater Indian during Pliocene to Recent time [Powers et al., 1998; lithosphere slab (anomaly 1). We therefore suggest that DeCelles et al., 1998b, 2001; Lave´ and Avouac, 2000] this detached slab mainly consists of Greater Himalayan (Figure 8). The present rate of shortening in the central part lithosphere. of the orogenic belt is 17 mm/yr [Bilham et al., 1997; [46] Combining some reasonable assumptions about the Larson et al., 1999]. Prior to the Neogene, however, the motion of the Neotethyan slab and kinematic information available shortening estimates from the Tibetan Himalaya from the Himalayan fold-thrust belt, we can constrain the suggest that shortening was more than three times slower, tempo of Neotethyan slab break-off, Greater Indian sub- only 6 mm/yr. Based on arguments presented by Searle et duction, break-off of the Greater Indian lithosphere slab, 12 - 16 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY

Figure 9. Kinematic history of the crust, lithosphere, and upper mantle in the India-Eurasia collision zone since 55 Ma, beginning with present-day interpretation shown in Figure 3b. Upward pointing arrow tracks the northern edge of the hypothesized Greater Indian lower crust. Abbreviations not explained in legend are as follows: FTF, frontal thrust fault (equivalent to today’s Main Frontal thrust, MFT); ISZ, Indus-Yalu suture zone; BSZ, Banggong suture zone; KF, Kunlun fault; PQT, Plio-Quaternary Tibet [Tapponnier et al., 2001]. and continued insertion of the GILC beneath the Tibetan shown plummeting vertically from the upper mantle, with Plateau (Figure 9). Following Van der Voo et al. [1999], we its lower tip fixed since 40 Ma at its present latitude assume that the mantle has not imparted any significant (Figure 9). Recalling that the location of the Indus-Yalu horizontal velocity to the Neotethyan slab. Thus, the slab is suture zone at the onset of collision was 6.5°N±2.5° DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 17

Figure 9. (continued)

[Klootwijk et al., 1985; Van der Voo, 1993], we position the tion, perhaps in latest Eocene time (Figures 9b and 9c). A Neotethyan slab at 55 Ma in the upper mantle between minimum age for the break-off of the Neotethyan slab can be depths of 350 km and 1200 km with its lower end at 12°N calculated by dividing the present offset between the Neo- and its upper end projecting to the surface at the Indus-Yalu tethyan slab and the northern limit of inferred GILC (1650 suture zone. This configuration imparts a 55° northward dip km) by the rate of northward convergence between India and to the slab (Figure 9a). Eurasia (50–55 mm/yr) (Figure 5, point 11), which yields a [47] Subsequent time steps in Figure 9 are calibrated minimum age range of 33–30 Ma for the break-off event. according to the rates of northward migration of the Indus- There is little constraint on the maximum age of Neotethyan Yalu suture zone, the southern tip of the Indian continent, and slab break-off. It is plausible that break-off occurred imme- the rate of subduction of Greater Indian lithosphere as shown diately after the initial impingement of Greater India against in Figure 6. These diagrams show the Neotethyan slab Eurasia, and that the slab rotated freely into a vertical breaking off after it had been rotated into a vertical orienta- orientation. 12 - 18 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY

[48] Following Davies and von Blanckenburg [1995], buoyancy with respect to the asthenosphere. Beneath the several workers have suggested that initial exhumation of northern Plateau, younger, denser Eurasian lithosphere (and Himalayan eclogites was associated with isostatic rebound eventually Greater Himalayan lithosphere as well) may have of partially subducted Indian crust during and immediately been more susceptible to gravitational foundering. after the break-off event [Chemenda et al., 2000; O’Brien [50] Another factor that may influence whether delamina- et al., 2001; Kohn and Parkinson, 2002]. This would tion takes place under Tibet is the well-documented differ- imply a late Eocene-Oligocene maximum age for this ence in mantle temperatures between the northern and event, based on the ages of the eclogites [Tonarini et al., southern parts of the Plateau and its effects on phase 1993; de Sigoyer et al., 1997]. The minimum age of break- transitions. Petrologic and thermal modeling [LePichon et off of the Greater Indian lithosphere slab can be calculated al., 1997] suggests that underthrust GILC would have by dividing the offset between it and the northern end of become eclogitized at depths of 60–75 km, but thermal inferred GILC (500 km) by the convergence rate between relaxation within 20 Myr would have converted GILC India and Eurasia, yielding an age of 9–10 Ma (Figure 5, back into granulite that was too light to sink into the mantle. point 6). Greater Indian slab break-off probably com- On the other hand, thermal modeling by Henry et al. [1997] menced sometime after 20 Ma and was complete by suggests that GILC has never been in the eclogite stability 10 Ma (Figures 9d and 9e). Modeling by Chemenda et field. At depths of 80–100 km the gabbro-eclogite phase al. [2000] suggests that break-off of the Greater Indian slab transition is nearly isothermal at a temperature of 500°C. involved progressive delamination followed by complete The intermediate-depth earthquakes beneath the Lhasa ter- detachment. Alternatively, Greater Indian lithosphere could rane indicate upper mantle temperatures <400°C, which have collided with Eurasian lithosphere beneath the central should inhibit eclogitization. Under northern Tibet, higher part of the Plateau (near the Banggong suture zone) at 20 mantle temperatures would trigger the phase change at Ma, causing both to thicken and become gravitationally equivalent depths. Any mafic lower crust remaining on top unstable. Greater Indian lithosphere began to delaminate of the southward underthrusting Eurasian lithosphere would and eventually broke off, while Eurasian lithosphere began eclogitize and increase the negative buoyancy of the slab. We to subduct southward. In any case, ongoing northward suggest that the petrologic differences between the northern migration of the remainder of Greater India has overrun and southern parts of the GILC slab can explain the contrast- the upper end of the detached Greater Indian lithospheric ing behavior of the lithosphere beneath Tibet. The southern slab at the longitude of western Nepal (Figures 9e and 9f) part of the GILC slab is underlain by cold cratonic Indian [Van der Voo et al., 1999]. lithosphere that has not been eclogitized, whereas the north- [49] Insertion of Greater Indian mantle lithosphere and ern part of the slab was, until mid-Miocene time, underlain by lower crust beneath the Tibetan Plateau could not have taken younger, Greater Himalayan, eclogite-prone lithosphere. The place unless this region had been evacuated of Eurasian documented presence of eclogites in Greater Himalayan lithosphere. Removal of Eurasian lithosphere may have been upper crustal rocks supports this view. Thus, conversion of facilitated by Late Cretaceous-Paleocene low-angle north- Greater Himalayan lithosphere to eclogite could have driven ward subduction of the Neotethyan slab prior to initial the Miocene delamination event. impingement of Greater India [Ding et al., 2002]. In this model, the widespread Paleocene ignimbrites (Linzizong 7. Implications of the Model Formation) in the Lhasa terrane and similar rocks in the southern part of the Qiangtang terrane were produced during [51] The proposed model is similar in various respects to a regional ‘‘flare-up’’ in response to southward rollback and previously published models involving underthrusting of steepening of a formerly flat Neotethyan oceanic slab, India beneath Tibet [e.g., Argand, 1924; Powell and Con- analogous to models for mid-Cenozoic ignimbrites in the aghan, 1973; Seeber et al., 1981; Ni and Barazangi, 1984; North American Cordillera [e.g., Coney and Reynolds, 1977; Zhao and Morgan, 1987; Matte et al., 1997; Chemenda et al., Constenius, 1996]. The Ding et al. [2002] model suggests 2000]. The model makes several testable predictions about that Eurasian lithosphere was either not present or greatly the history of deformation and magmatism in the Tibetan attenuated beneath the southern part of the Tibetan Plateau at Plateau. Before discussing these, however, we emphasize the onset of the Himalayan orogenic event (Figure 9a). One that we are not suggesting crustal underthrusting operated or more segments of Eurasian lithosphere, too small to be alone to thicken the Tibetan Plateau. The pre-Cenozoic detected by present tomographic models, also could have history of shortening and stresses transmitted ahead of the been removed by delamination during early-middle Tertiary underthrusting GILC undoubtedly complicate the tectonic shortening of Tibetan upper crust [Tapponnier et al., 2001]. history of Tibetan upper crust. Nevertheless, the fact that The contrasting behavior of the relatively young (Paleozoic Himalayan shortening has continued unabated since 55 Ma and Mesozoic) Eurasian lithosphere and the subhorizontally implies that underthrusting of the GILC is one of the few subducting Greater Indian lithosphere may have been pro- processes in the Himalayan-Tibetan orogenic system that moted by density differences. Archean cratonic lithosphere, may have operated predictably. In the following, we briefly like that hypothesized to lie beneath the southern part of the outline seven predictions that derive from the crustal under- Plateau, is thought to be less dense than younger lithosphere thrusting model. owing to long-term melt extraction [Jordan, 1978]. Stripped 1. The similarity of the width of the Tibetan Plateau to the of its lighter crust, such lithosphere may have near neutral amount of shortening in the corresponding portion of the DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 19

Himalayan fold-thrust belt is circumstantial evidence in favor have isostatically increased regional elevation in the Plateau of the crustal underthrusting model (Figure 5, point 8). One during mid- to late-Miocene time, consistent with evidence obvious explanation is that the width of the Plateau faithfully for late Miocene regional environmental and climatic reflects the presence of the underthrust slab of GILC. If this is changes in central Eurasia and changes in the regional force the case, then the cumulative shortening in the Himalaya at distribution surrounding the Plateau [Harrison et al., 1992; any given time should provide an estimate of the location of Molnar et al., 1993]. the northern edge of the underthrusting slab. Thus, the tip of 3. The crustal underthrusting model has implications for the slab reached the Banggong suture by 20 Ma, the the distribution of mafic Indian cratonic lower crust beneath Qiangtang anticlinorium by 14 Ma, and the Jinsha suture the Tibetan Plateau. The amount of shortening of the Lesser within the last million years or so (Figures 8 and 9). This Himalayan and Subhimalayan zones (300 km), which were aspect of the model predicts that a front of upper crustal strain deposited on cratonic India, should approximately equal the should have advanced across the Tibetan Plateau as the GILC length of Indian cratonic lower crust that has been inserted slab propagated northward. The nature of the strain beneath the Plateau. This suggests that strong, mafic, associated with this front might vary according to the Archean lower crust of cratonic India extends 300 km to strength of the underthrusting GILC and the location relative the north of the Indus-Yalu suture zone (Figure 5, point 9), to the leading edge of the slab. Directly above the leading coincident with the distribution of high-velocity lower crust edge, we would expect a crustal-scale monocline in the beneath the Lhasa terrane [Owens and Zandt, 1997; Kind et overlying Tibetan crust, keeping in mind the possibility that al., 2002]. The remainder of the Plateau is probably underlain pre-existing structures could render the recognition of such a by Greater Himalayan lower crust (Figure 3b) composed of monocline difficult. A candidate for a structure affected by weak, felsic metasedimentary rocks that may be close to the proposed underthrusting GILC is the Qiangtang anticli- solidus conditions. This is consistent with the change in the norium (Figure 1), which experienced broadening and velocity structure in the lower crust north of the Banggong topographic rejuvenation during the mid-Tertiary [Kapp et suture zone [Owens and Zandt, 1997; Tapponnier et al., al., 2002]. This is not to say that all deformation in the 2001] and crustal anisotropy data that suggest a component of Tibetan Plateau followed an orderly northward march. It is eastward lower crustal flow (Folsom et al., manuscript in plausible that far-field stresses created contractional strain preparation, 2002). The presence of strong lower crust and strike-slip faulting in advance of the underthrusting beneath the Lhasa terrane may have increased its ability to GILC slab tip [e.g., Yin et al., 1999; Horton et al., 2002; transmit compressive stresses northward, accounting for the Robinson et al., 2002; Yin et al., 2002]. A key challenge in lack of major Cenozoic horizontal shortening in the Lhasa proving the validity of the crustal underthrusting model terrane compared to the significant Cenozoic shortening in would be to distinguish between strain that resulted directly the Qiangtang and Songpan-Ganzi terranes [Kapp et al., from crustal underthrusting and strain that resulted from the 2000; Yin and Harrison, 2000; Horton et al., 2002]. propagation of stresses far in front of the GILC. However, this explanation is only relevant for post-20 Ma 2. The model predicts two slab break-off events. The first deformation in the Plateau, because the insertion of the strong involved Neotethyan lithosphere and occurred during late mafic portion of the GILC did not commence until about that Eocene-Oligocene time. This prediction is consistent with the time. Restriction of stronger cratonic GILC to the southern timing of Neotethyan slab break-off proposed by previous part of the Plateau is also consistent with the modern workers based on petrologic considerations [O’Brien et al., topography of the Plateau, with highest local relief in the 2001; Kohn and Parkinson, 2002], but relies on an south and very low relief in the northern part [Fielding et al., independent, geometrically and kinematically reasonable 1994]. reconstruction. The second predicted break-off event in- 4. The crustal underthrusting model predicts that the volved mainly Greater Himalayan lithosphere and took place elevation of the Tibetan Plateau should have increased in during mid- to Late Miocene time. A similar break-off event isostatic proportion with the thickness of the GILC slab. If we was predicted by Chemenda et al. [2000] during late assume that the crust of the Lhasa terrane was 50–55 km Oligocene-early Miocene time. Crustal tectonic events that thick prior to the Cenozoic orogeny [Murphy et al., 1997], might correlate with these break-off events include Eocene- then underthrusting of the GILC slab would have added Oligocene eclogite exhumation by crustal-scale duplexing in 15–20 km of crust to the southern portion of the Plateau. the North Himalayan domes of the Tibetan Himalayan zone Addition of this crustal mass without the need for surficial [Steck et al., 1998; Chemenda et al., 2000; Kohn and thrusting above the leading edge of the slab would result in Parkinson, 2002], and regional extension in the southern the northward propagation of a topographic front without Tibetan Plateau and high Himalaya and activation of the attendant upper crustal shortening and flexural subsidence in Main Central thrust during early and mid-Miocene time a migrating foreland basin system. Noteworthy in this context [Hodges, 2000]. Insofar as break-off events should cause is the lack of evidence for active, large-scale shortening and isostatic rebound of the orogenic belt [Davies and von flexural subsidence along the northern [Li et al., 1996; Jiang Blanckenburg, 1995], we would expect that the Himalayan et al., 1999] and eastern [Royden et al., 1997; Kirby et al., fold-thrust belt was driven into a supercritical state promoting 2000] margins of the modern Plateau. Moreover, GPS forward propagation [Davis et al., 1983] during the proposed studies, earthquake seismology, and neotectonic studies break-off events. Because it must have occurred beneath the along the eastern margin of the Plateau, in the Longmen Tibetan Plateau, the second break-off event is expected to Shan and Min Shan, indicate that the upper crust is not 12 - 20 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY shortening despite the presence of an immense topographic break along the edge of the Plateau [Chen et al., 2000; Kirby et al., 2000]. Royden et al. [1997] and Clark and Royden [2000] proposed that the 3.0 Â 106 km2 region of eastward sloping, moderate to high elevation terrain to the east of longitude 96°E is being uplifted from below by eastward flowing lower crust (Figure 10). Similar models involving lower crustal flow have been proposed for other large mountain belts comprising mid- to lower crustal rocks [e.g., Hodges and Walker, 1992; Wernicke and Getty, 1997]. We suggest that lower crustal flow beneath the region to the east of the Plateau is driven by the insertion beneath Tibet of the GILC (Figure 10), which is essentially a geographic modification of the ‘‘hydraulic piston’’ model of Zhao and Morgan [1987]. Our model is consistent with results of recent GPS studies that show eastward and southeastward rotation of surficial velocity vectors [Wang et al., 2001], suggesting that GILC underthrusting and lower crustal lateral extrusion are not completely coupled with deformation in the Tibetan upper crust [Holt, 2000]. Insofar as the present volume of crust beneath the Tibetan Plateau west of 96°E can be accounted for by our model, the volume of excess crustal material that has been extruded eastward and southeastward [Clark and Royden, 2000] implies that our estimates of Greater Indian additions to Tibet, and/or the pre-Cenozoic crustal thickness of Tibet, may be bare minima. 5. The underthrusting model suggests that east-west extension in the crust of the Tibetan Plateau may be a result of the need for Tibetan crust to stretch in this direction in order to accommodate the insertion of the GILC slab. If the slab is thickest and widest toward the south, then we would expect that east-west extension should be greatest in the southern part of the Plateau. A careful inspection of the digital elevation model of Fielding et al. [1994] strongly suggests that this is the case (Figure 2). The most prominent north-south striking grabens in the Plateau are those that rim its southern edge, just north of the South Tibetan detachment system (e.g., the Thakkhola and Yadong-Gulu grabens; Figure 1). North of the Banggong suture, evidence for east- west extension, though still present, is much more subdued [Yin, 2000]. This mechanism for east-west extension is also consistent with the radial shortening directions in the Himalayan arc [Molnar and Lyon-Caen, 1989].

Figure 10. (opposite) Schematic map showing the hy- pothetical eastward and southeastward flow (arrows) of lower crustal material into the region east of the Tibetan Plateau, with the 1.5 km elevation contour shown for present conditions, after Clark and Royden [2000]. The Tarim and Sichuan basins are underlain by strong rigid lower crust (black), whereas the lower crust north and south of the Sichuan basin is relatively weak. Crustal flow is driven by insertion of the slab of GILC (light gray). Portion of the GILC that is predicted to be composed of strong, cratonic Indian lower crust is shown by diagonal ruling. Present location of the Indus-Yalu suture zone is shown by heavy line labeled ISZ. Present coastline is kept fixed. Approximately 150 km of shortening in the upper crust of Tibet is included. DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 21

6. The reconstruction shown in Figure 9 has bearing on terrane. The tempo of Himalayan thrusting provides (1) an interpretations of the Tertiary magmatism in the Tibetan estimate of the position of the northern edge of the slab Plateau. The model satisfies existing geochronologic con- through time, and (2) a chronometer of events in the upper straints on the Gangdese arc by predicting that calc-alkaline mantle beneath Tibet. Two key predictions of our model are magmatism should have persisted in the southern Plateau break-off of the Neotethyan oceanic slab at 45–35 Ma, and until late Eocene time, when the Neotethyan slab broke off delamination and break-off of a several hundred kilometer (Figures 9b and 9c). High-potassium magmatism associated long slab of Greater Indian lithosphere between 20 and 10 with rapid convective flow in the mantle and melting of Ma. metasomatized upper mantle is expected in the wake of slab 3. The crustal underthrusting model helps to explain the break-off events [Davies and von Blanckenburg, 1995]. Such distribution and cause of east-west extension in the Tibetan a process could have pooled magmas at the base of the crust Plateau. Insertion of a tapering (both vertically and and produced the outbursts of Eocene-Oligocene high- horizontally) slab of Greater Indian lower crust beneath the potassium volcanic rocks in the central part of the Plateau Plateau forced the Tibetan upper crust to stretch laterally to (Figure 1). It has also been suggested that the Banggong accommodate the excess crustal mass. This suggests that suture zone was partially reactivated during the mid-Tertiary extension should be greatest in the southern part of the [Yin and Harrison, 2000], possibly generating Eocene- Plateau, and that the age of east-west extension should Oligocene melts in central Tibet [Hacker et al., 2000; Kapp et decrease northward across the Plateau. The region of al., 2000; Tapponnier et al., 2001; Ding et al., 2002]. A maximum east-west extension should also correlate with second phase of high-potassium magmatism during Mio- the predicted distribution of strong, relatively thick, cratonic cene-Quaternary time might be explained by the break-off Indian lower crust, which should extend as far north as the and foundering of Greater Himalayan lithosphere and the Banggong suture zone. onset of southward subduction of Eurasian lithosphere 4. Cenozoic magmatism in the Tibetan Plateau spans the (Figures 9d–9f) [Tapponnier et al., 2001]. last 45 Myr. We tentatively note that the ages and spatial 7. Tapponnier et al. [2001] highlighted the potential distributions of these rocks crudely correlate with mantle significance for growth of the Tibetan Plateau of late upwelling events associated with predicted Neotethyan and Miocene-Quaternary crustal shortening in the region to the Greater Indian lithospheric slab break-off events. northeast of the Kunlun fault and associated southward 5. Several processes have collaborated to produce the high subduction of Eurasian lithosphere beneath the Plateau. elevation of the Tibetan Plateau, including Mesozoic and These processes are analogous to the crustal shortening and Cenozoic upper crustal shortening in Tibetan terranes, underthrusting that have taken place along the Himalayan Cenozoic upper crustal shortening in the Himalaya, crustal margin of the Plateau since Eocene time. If Eurasian thickening by northward underthrusting of Greater Indian lithosphere were to peel back northward, additional space lower crust, and generally eastward flow of lower crustal would be created for further northward insertion of GILC. material that thickened during the Mesozoic and early In this sense, perhaps the tectonic processes operating Tertiary and was subsequently displaced by the under- between the Kunlun fault and the northern margin of the thrusting Greater Indian lower crust. In addition, since late Nan Shan thrust belt are analogous to those which operated Miocene time, southward underthrusting of Eurasian litho- in the upper mantle and crust of the Tibetan Plateau prior to sphere and associated upper crustal shortening has begun to insertion of the GILC slab. add a half-million square kilometer region to the northeastern part of the Plateau [Tapponnier et al., 2001]. Although the regional pattern of deformation in the Tibetan Plateau seems 8. Conclusions chaotic, incorporation of the temporal predictions made by the crustal underthrusting model may help to identify order [52] Our main conclusions are as follows: and process as more kinematic information is obtained from 1. Reconstruction of pre-Himalayan structural architec- the Tibetan Plateau. ture of Greater India indicates that the northern margin of 6. Based in part on the great breadth of continental Gondwana constituted the Lesser, Greater, and Tibetan orogenic systems (relative to orogenic systems developed Himalayan terranes. These terranes were underlain by a between converging oceanic plates), a number of models composite lower crust that included Archean and early have suggested that continental crust and mantle lithosphere Proterozoic cratonic Indian lower crust and late Proterozoic- are weak and incapable of maintaining sharp plate early Paleozoic Greater Himalayan lower crust. This Greater boundaries [e.g., England and Houseman, 1988]. If the Indian lower crust has been stripped of its supracrustal general aspects of our model are correct, then the sedimentary and metasedimentary cover during the Cenozoic Himalayan-Tibetan orogenic system resembles a flat-slab orogenic event, and underthrust beneath the Tibetan Plateau. subduction system, rather than the classic, pure-shear- 2. Himalayan shortening matches the width of the Tibetan dominated concept of a collisional orogen [e.g., Dewey and Plateau north of the Indus-Yalu suture zone. The similarity in Bird, 1970]. Moreover, the Greater Indian and Eurasian these two quantities suggests that a northward tapering slab lithospheres beneath the Himalaya and Tibet seem to behave of Greater Indian lower crust has been underthrust beneath like coherent plates. In this sense, the widespread expanse the Tibetan crust. Insertion of the slab probably increased the of intracontinental deformation in central Eurasia owes less thickness of Tibetan crust by up to 20 km beneath the Lhasa to the ‘‘softness’’ of Tibetan crust and mantle lithosphere 12 - 22 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY than it does to the bulk of continental material situated support was provided by ExxonMobil and Conoco. We are grateful to above the subduction system. Gautam Mitra, Kelin Whipple, and Greg Houseman for thoughtful reviews of the original manuscript. Our understanding of the Tibetan-Himalayan system has been improved by discussions with and the generous provision of preprints by Paul Kapp, Clem Chase, Ofori Pearson, Carmala Garzione, [53] Acknowledgments. Funding for research that contributed to this Rainer Kind, Doug Nelson, An Yin, Brian Horton, Heather Folsom, Kevin paper was provided by the U.S. National Science Foundation (grants EAR- Furlong, Tom Owens, Matt Spurlin, Ding Lin, Mark Harrison, Jessica 9814060, EAR-0105339, EAR-0125121, and EAR-0207179). Additional D’Andrea, Brad Ritts, and Brad Hacker.

References

Achache, J., V. Courtillot, and Y. Xiu, Paleogeographic Burg, J. P., and G. M. Chen, Tectonics and structural Coward, M. P., and R. W. H. Butler, Thrust tectonics and tectonic evolution of southern Tibet since mid- zonation of southern Tibet, Nature, 311, 219 – 223, and the deep structure of the Pakistan Himalaya, dle Cretaceous time: New paleomagnetic data and 1984. Geology, 13, 417 – 420, 1985. synthesis, J. Geophys. Res., 89, 10,311– 10,339, Catlos, E. J., T. M. Harrison, M. J. Kohn, M. Grove, D’Andrea, J., T. M. Harrison, and M. Grove, The ther- 1984. R. J. Ryerson, C. Manning, and B. N. Upreti, Geo- mal and physical state of the South Tibetan middle Ahmad, T., N. Harris, M. Bickle, H. Chapman, J. Bun- chronologic and thermobarometric constraints on crust between 20-8 Ma: U-Th-Pb and Nd isotopic bury, and C. Prince, Isotopic constraints on the the evolution of the Main Central thrust, central evidence from the Nyainqentanghla Massif, in 14th structural relationships between the Lesser Himala- Nepal Himalaya, J. Geophys. Res., 106, 16,177 – Himalaya-Karakoram-Tibet Workshop, edited by E. yan Series and the High Himalayan Crystalline Ser- 16,204, 2001. Sobel, et al., Kloster Ettal, Germany, 1999. ies, Garhwal Himalaya, Geol. Soc. Am. Bull., 112, Catlos, E. J., T. M. Harrison, C. Manning, M. Grove, Davies, J. H., and F. von Blanckenburg, Slab breakoff: 467 – 477, 2000. S. M. Rai, M. S. Hubbard, and B. N. Upreti, Re- A model of lithosphere detachment and its test in Alle`gre, C. J., et al., Structure and evolution of the cords of the evolution of the Himalayan orogen the magmatism and deformation of collisional oro- Himalaya-Tibet orogenic belt, Nature, 307,17– from in situ Th-Pb ion microprobe dating of mon- gens, Earth Planet. Sci. Lett., 129, 85 – 102, 1995. 22, 1984. azite: Eastern Nepal and western Garhwal, J. Asian Davis, D., J. Suppe, and F. A. Dahlen, Mechanics of Alsdorf, D., and D. Nelson, Tibetan satellite magnetic Earth Sci., 20, 459 – 479, 2002. fold-and-thrust belts and accretionary wedges, low: Evidence for widespread melt in the Tibetan Chang, C.-F., and S.-L. Zheng, Tectonic features of the J. Geophys. Res., 88, 1153 – 1172, 1983. crust?, Geology, 27, 943 – 946, 1999. Mount Jolmu Lungma region in southern Tibet, DeCelles, P. G., and K. N. Giles, Foreland basin sys- Argand, E., La Tectonique de l’Asie, Proc. Int. Geol. China (in Chinese), Sci. Geol. Sinica, 1,1–12, tems, Basin Res., 8, 105 – 123, 1996. Cong., 7, 171 – 372, 1924. 1973. DeCelles, P. G., G. E. Gehrels, J. Quade, and T. P. Ojha, Armijo, R., P. Tapponnier, J. L. Mercier, and H. Han Chemenda, A. I., J.-P. Burg, and M. Mattauer, Evolu- Eocene-early Miocene foreland basin development Tonglin, Quaternary extension in southern Tibet: tionary model of the Himalaya-Tibet system: Geo- and the history of Himalayan thrusting, western and Field observations and tectonic implications, poem based on new modelling, geological and central Nepal, Tectonics, 17, 741 – 765, 1998a. J. Geophys. Res., 91, 13,803 – 13,872, 1986. geophysical data, Earth Planet. Sci. Lett., 174, DeCelles, P. G., G. E. Gehrels, J. Quade, P. A. Kapp, T. P. Beaumont, C., R. A. Jamieson, M. H. Nguyen, and B. 397 – 409, 2000. Ojha, and B. N. Upreti, Neogene foreland basin de- Lee, Himalayan tectonics explained by extrusion of Chen, W.-P., and H. Kao, Seismotectonics of Asia: posits, erosional unroofing, and the kinematic his- a low-viscosity crustal channel coupled to focused Some recent progress, in The Tectonic Evolution tory of the Himalayan fold-thrust belt, western surface denudation, Nature, 414, 738 – 742, 2001. of Asia, edited by A. Yin and T. M. Harrison, pp. Nepal, Geol. Soc. Am. Bull., 110, 2 – 21, 1998b. Besse, J., and V. Courtillot, Paleogeographic maps of 37 – 62, Cambridge Univ. Press, New York, 1996. DeCelles, P. G., G. E. Gehrels, J. Quade, B. LaReau, the continents bordering the Indian Ocean since the Chen, W.-P., and P. Molnar, Focal depths of intraconti- and M. Spurlin, Tectonic implications of U-Pb zir- Early Jurassic, J. Geophys. Res., 93, 11,791 – nental and intraplate earthquakes and their implica- con ages of the Himalayan orogenic belt in Nepal, 11,808, 1988. tions for the thermal and mechanical properties of Science, 288, 497 – 499, 2000. Besse, J., and V. Courtillot, Revised and synthetic ap- the lithosphere, J. Geophys. Res., 88, 4183 – 4214, DeCelles, P. G., D. M. Robinson, J. Quade, T. P. Ojha, parent polar wander paths of the African, Eurasian, 1983. C. N. Garzione, P. Copeland, and B. N. Upreti, North American and Indian Plates, and true polar Chen, W.-P., and S. O¨ zalaybey, Correlation between Stratigraphy, structure, and tectonic evolution of wander since 200 Ma, J. Geophys. Res., 96, 4029 – seismic anisotropy and Bouguer anomalies in Tibet the Himalayan fold-thrust belt in western Nepal, 4050, 1991. and its implications for lithosphereic structures, Tectonics, 20, 487 – 509, 2001. Besse, J., V. Courtillot, J. P. Pozzi, M. Westphal, and Geophys. J. Int., 135, 93 – 1010, 1998. Deng, W., A preliminary study on the petrology and Y. X. Zhou, Palaeomagnetic estimates of crustal Chen, Z., B. C. Burchfiel, Y. Liu, R. King, L. H. Roy- petrochemistry of the Quaternary volcanic rocks shortening in the Himalayan thrusts and Zangbo den, W. Tang, J. Zhao, and X. Zhang, GPS mea- of northern Tibet autonomous region (in Chinese), Suture, Nature, 311, 621 – 626, 1984. surements from eastern Tibet and their implications Acta Geol. Sinica, 52, 148 – 162, 1978. Bilham, R., K. Larson, and J. Freymuller, Indo-Asian for India/Eurasia intracontinental deformation, DePietro, J. A., and C. E. Isachsen, U-Pb zircon ages convergence rates in the Nepal Himalaya, Nature, J. Geophys. Res., 105, 16,215 – 16,227, 2000. from the Indian plate in northwest Pakistan and 386, 61 – 64, 1997. Chung, S.-L., C.-H. Lo, T.-Y. Lee, Y. Zhang, Y. Xie, X. their significance to Himalayan and pre-Himalayan Bird, P., Lateral extrusion of lower crust from under Li, K.-L. Wang, and P.-L. Wang, Diachronous uplift geologic history, Tectonics, 20, 510 – 525, 2001. high topography in the isostatic limit, J. Geophys. of the Tibetan plateau starting 40 Myr ago, Nature, de Sigoyer, J., S. Guillot, J.-M. Lardeaux, and G. Mascle, Res., 96, 10,275 – 10,286, 1991. 394, 769 – 773, 1998. Glaucophane-bearing eclogites in the Tso Morari Blisniuk, P. M., B. R. Hacker, J. Glodny, L. Ratschba- Clark, M. K., and L. H. Royden, Topographic ooze: dome (eastern Ladakh, NW Himalaya), Eur. cher,S.Bi,Z.Wu,M.O.McWilliams,andA. Building the eastern margin of Tibet by lower crus- J. Mineral., 9, 1073 – 1083, 1997. Calvert, Normal faulting in central Tibet since at tal flow, Geology, 28, 703 – 706, 2000. Dewey, J. F., and J. M. Bird, Mountain belts and the least 13.5 Myr ago, Nature, 412, 628 – 632, 2001. Coleman, M. E., U-Pb constraints on Oligocene-Mio- new global tectonics, J. Geophys. Res., 75, 2625 – Brookfield, M. E., The Himalayan passive margin from cene deformation and anatexis within the central 2647, 1970. Precambrian to Cretaceous times, Sediment. Geol., Himalaya, Marsyandi valley, Nepal, Am. J. Sci., Dewey, J. F., R. M. Shackleton, C. Chang, and Y. Sun, 84, 1 – 35, 1993. 298, 553 – 571, 1998. The tectonic development of the Tibetan Plateau, in Burbank, D. W., R. A. Beck, and T. Mulder, The Hi- Coleman, M. E., and K. V. Hodges, Evidence for Tibe- The Geological Evolution of Tibet,editedbyC. malayan foreland basin, in The Tectonics of Asia, tan Plateau uplift before 14 Myr ago from a new Chang, et al., Philos. Trans. R. Soc. London, Ser. edited by A. Yin and T. M. Harrison, pp. 205 – 226, minimum estimate for east-west extension, Nature, A, 327, 379 – 413, 1988. Cambridge Univ. Press, New York, 1996. 474, 49 – 52, 1995. Dewey, J. F., S. Cande, and W. C. Pitman, Tectonic Burchfiel, B. C., C. Zhiliang, K. V. Hodges, L. Yuping, Coney, P. J., and S. J. Reynolds, Cordilleran Benioff evolution of the India/Eurasia collision zone, Eclo- L. H. Royden, D. Changrong, and X. Jiene, The zones, Nature, 270, 403 – 406, 1977. gae Geol. Helv., 82, 717 – 734, 1989. South Tibetan detachment system, Himalyan oro- Constenius, K. N., Late Paleogene extensional collapse Dhital, M. R., and K. Kizaki, Structural aspect of the gen: Extension contemporaneous with and parallel of the Cordilleran foreland fold and thrust belt, northern Dang, Lesser Himalaya, Bull. Coll. Sci., to shortening in a collisional mountain belt, Spec. Geol. Soc. Am. Bull., 108, 20 – 39, 1996. Univ. Ryukyus, 45, 159 – 182, 1987. Pap. Geol. Soc. Am., 269, 51 pp., 1992. Coulon, C., H. Maluski, C. Bollinger, and S. Wang, Ding, L., P. Kapp, A. Yin, D. Wangming, and Z. Dalai, Burchfiel, B. C., Z. Chen, Y. Liu, and L. H. Royden, Mesozoic and Cenozoid volcanic rocks from central Early Tertiary volcanism in the Qiantang terrane of Tectonics of the Longmen Shan and adjacent re- and southern Tibet: 40Ar/39Ar dating, petrological central Tibet: Evidences for a transition from ocea- gions, central China, Int. Geol. Rev., 37, 661 – characteristics and geodynamical significance, nic to continental subduction, J. Petrol., in press, 735, 1995. Earth Planet. Sci. Lett., 79, 281 – 302, 1986. 2002. DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 23

Edwards, M. A., and T. M. Harrison, When did the roof Constraints related to metamorphism within the un- duction, oroclinal bending and back-arc spreading, collapse?: Late Miocene N-S extension in the High derthrust Indian crust and topographic elevation, Earth Planet. Sci. Lett., 75, 167 – 183, 1985. Himalaya revealed by Th-Pb monazite dating of the Tectonophysics, 273, 31 – 56, 1997. Kohn, M. J., and C. D. Parkinson, Petrologic case for Khula Kangri granite, Geology, 25, 543 – 546, 1997. Hirn, A., Features of the crust-mantle structure of Hi- Eocene slab breakoff during the Indo-Asian colli- Ekstrom, G., A broadband method of earthquake ana- malayas-Tibet: A comparison with seismic traverses sion, Geology, 30, 591 – 594, 2002. lysis, Ph.D. dissertation, 215 pp., Harvard Univ., of Alpine, Pyrenean and Variscan orogenic belts, Kosarev, G., R. Kind, S. V. Sobolev, X. Yuan, W. Han- Cambridge, Mass., 1987. Philos Trans. R. Soc. London, Ser. A, 326,17– ka, and S. Oreshin, Seismic evidence for a detached England, P., and G. Houseman, The mechanics of the 32, 1988. Indian lithospheric mantle beneath Tibet, Science, Tibetan Plateau, Philos. Trans. R. Soc. London, Ser. Hirn, A., et al., Seismic anisotropy as an indicator of 283, 1306 – 1309, 1999. A, 326, 301 – 320, 1988. mantle flow beneath the Himalayas and Tibet, Nat- Larson,K.M.,R.Burgmann,R.Bilham,andJ.T. Fielding, E., B. Isacks, M. Barazangi, and C. Duncan, ure, 375, 571 – 574, 1995. Freymueller, Kinematics of the India-Eurasia colli- How flat is Tibet?, Geology, 22, 163 – 167, 1994. Hodges, K. V., Tectonics of the Himalaya and southern sion zone from GPS measurements, J. Geophys. Fouch, M. J., K. M. Fischer, M. E. Wysession, and T. J. Tibet from two perspectives, Geol. Soc. Am. Bull., Res., 104, 1077 – 1093, 1999. Clarke, Shear wave splitting, continental keels, and 112, 324 – 350, 2000. Lave´, J., and J. P. Avouac, Active folding of fluvial patterns of mantle flow, J. Geophys. Res., 105, Hodges, K. V., and J. D. Walker, Extension in the Cre- terraces across the Siwaliks Hills, Himalayas of 6255 – 6276, 2000. taceous Sevier orogen, North American cordillera, central Nepal, J. Geophys. Res., 105, 5735 – 5770, Furlong, K. P., and T. J. Owens, Tibetan mantle fabric Geol. Soc. Am. Bull., 104, 560 – 569, 1992. 2000. and crust-mantle kinematics, Eos Trans. AGU, Hodges, K. V., R. R. Parrish, and M. P. Searle, Tectonic Lave´, J., J. P. Avouac, R. Lacassin, P. Tapponnier, and 78(46), F173, Fall Meet. Suppl., 1997. evolution of the central Annapurna Range, Nepa- J. P. Montagner, Seismic anisotropy beneath Tibet: Gaetani, M., and E. Garzanti, Multicyclic history of the lese Himalayas, Tectonics, 15, 1264 – 1291, 1996. Evidence for eastward extrusion of the Tibetan northern India continental margin (northwestern Hi- Hodges, K. V., J. M. Hurtado, and K. X. Whipple, lithosphere?, Earth Planet. Sci. Lett., 140, 83 – 96, malaya), AAPG Bull., 75, 1427 – 1446, 1991. Southward extrusion of Tibetan crust and its effect 1996. Gaetani, M., F. Jadoul, E. Erba, and E. Garzanti, Jurassic on Himalayan tectonics, Tectonics, 20, 799 – 809, LeFort, P., Metamorphism and magmatism during the and Cretaceous orogenic events in the North Kara- 2001. Himalayan collision, in Collision Tectonics, edited korum: Age constraints from sedimentary rocks, in Holt, W. E., Correlated crust and mantle strain fields in by M. P. Coward and A. C. Ries, Geol. Soc. Spec. Himalayan Tectonics, edited by P. J. Treloar and Tibet, Geology, 28, 67 – 70, 2000. Pub., 19, 159 – 172, 1986. M. P. Searle, Geol. Soc. Spec. Publ., 73,39–52, Holt, W. E., and T. C. Wallace, Crustal thickness and Le Pichon, X., M. Fournier, and L. Jolivet, Kinematics, 1993. upper mantle velocities in the Tibetan plateau re- topography, shortening, and extrusion in the India- Gansser, A., Geology of the Himalayas, 289 pp., Wiley- gion from inversion of regional Pn waveforms: Evi- Eurasia collision, Tectonics, 11, 1085 – 1098, 1992. Interscience, New York, 1964. dence for a thick upper mantle lid beneath southern Le Pichon, X., P. Henrey, and B. Goffe, Uplift of Tibet: Garzione, C. N., J. Quade, P. G. DeCelles, and N. B. Tibet, J. Geophys. Res., 95, 12,499 – 12,525, 1990. From eclogites to granulites—Implications for the English, Predicting paleoelevation of Tibet and the Horton, B. K., A. Yin, M. S. Spurlin, J. Zhou, and J. Andean Plateau and the Variscan belt, Tectonophy- Nepal Himalaya from @18O vs. altitude gradients in Wang, Paleocene-Eocene syncontractional sedimen- sics, 273, 57 – 76, 1997. meteoric water across the Nepal Himalaya, Earth tation in narrow, lacustrine-dominated basins of east- Li, D., D. Liang, C. Jia, G. Wang, Q. Wu, and D. He, Planet. Sci. Lett., 183, 215 – 230, 2000a. central Tibet, Geol. Soc. Am. Bull., 114, 771 – 786, Hydrocarbon accumulations in the Tarim basin, Garzione, C. N., D. L. Dettman, J. Quade, P. G. DeCel- 2002. China, AAPG Bull., 80, 1587 – 1603, 1996. les, and R. F. Butler, High times on the Tibetan Hurtado, J. M.Jr., K. V. Hodges, and K. X. Whipple, Lillie, R. J., G. D. Johnson, M. Yousuf, A. S. H. Zamin, Plateau: Paleoelevation of the Thakkhola graben, Neotectonics of the Thakkhola graben and implica- and R. S. Yeats, Structural development within the Nepal, Geology, 28, 339 – 342, 2000b. tions for recent activity on the South Tibetan fault Himalayan foreland fold-and-thrust belt of Pakistan, Godin, L., R. L. Brown, and S. Hanmer, High strain system in the central Nepal Himalaya, Geol. Soc. in Sedimentary Basins and Basin Forming Mechan- zone in the hanging wall of the Annapurna detach- Am. Bull., 113, 222 – 240, 2001. isms, edited by C. Beaumont and A. J. Tankard, ment, central Nepal Himalaya, in Himalaya and Jiang, X., M. K. McNutt, and Y. Jin, Models of litho- Mem. Can. Soc. Pet. Geol., 12, 379 – 392, 1987. Tibet: Mountain roots to mountain tops, edited by spheric deformation beneath the Altyn Tagh and Makovsky, Y., and S. L. Klemperer, Measuring the A. Macfarlane, R. B. Sorkhabi, and J. Quade, Spec. West Kunlun faults from recent gravity surveys, seismic properties of Tibetan bright spots: Evidence Pap. Geol. Soc. Am., 328, 199 – 210, 1999. Eos, Trans. AGU, 80(46), F1008, Fall Meet. Suppl., for free aqueous fluids in the Tibetan middle crust, Grand, S. P., R. D. van der Hilst, and S. Widiyantoro, 1999. J. Geophys. Res., 104, 10,795 – 10,825, 1999. Global seismic tomography: A snapshot of convec- Jin, Y., M. K. McNutt, and Y.-S. Zhu, Mapping the Manickavasagam, R. M., A. K. Jain, S. Singh, and A. tion in the Earth, GSA Today, 7, 1 – 7, 1997. descent of Indian and Eurasian plates beneath the Asokan, Metamorphic evolution of the northwest Grujic, D., L. S. Hollister, and R. R. Parrish, Himalayan Tibetan Plateau from gravity anomalies, J. Geo- Himalaya, India: Pressure-temperature data, in- metamorphic sequence as an orogenic channel: In- phys. Res., 101, 11,275 – 11,290, 1996. verted metamorphism, and exhumation in the Kash- sight from Bhutan, Earth Planet. Sci. Lett., 198, Jordan, T. H., Composition and develoment of the con- mir, Himachal, and Garhwal Himalayas, in 177 – 191, 2002. tinental tectosphere, Nature, 274, 544 – 548, 1978. Himalaya and Tibet: Mountain Roots to Mountain Hacker, B. R., X. Wang, E. A. Eide, and L. Ratschba- Kapp, P., Blueschist-bearing metamorphic core com- Tops, edited by A. Macfarlane, R. B. Sorkhabi, and cher, The Qinling-Dabie ultra-high pressure colli- plexes in the Qiangtang block reveal deep crustal J. Quade, Spec. Pap. Geol. Soc. Am., 328, 179 – sional orogen, in The Tectonic Evolution of Asia, structure of northern Tibet: Reply, Geology, 29, 91, 198, 1999. edited by A. Yin and T. M. Harrison, pp. 345 – 370, 2001. Marquer, D., H. S. Chawla, and N. Challandes, Pre- Cambridge Univ. Press, New York, 1996. Kapp, P., A. Yin, C. E. Manning, M. Murphy, T. M. alpine high-grade metamorphism in High Hima- Hacker, B. R., E. Gnos, L. Ratschbacher, M. Grove, M. Harrison, M. Spurlin, L. Ding, X.-G. Deng, and laya crystalline sequences: Evidence from Lower McWilliams, S. V. Sobolev, J. Wan, and W. Zhen- C.-M. Wu, Blueschist-bearing metamorphic core Palaeozoic Kinnaur Kailas granite and surround- han, Hot and dry deep crustal xenoliths from Tibet, complexes in the Qiangtang block reveal deep crus- ing rocks in the Sutlej Valley (Himachal Pra- Science, 287, 2463 – 2466, 2000. tal structure of northern Tibet, Geology, 28, 19 – 22, desch, India), Eclogae Geol. Helv., 93, 207 – Harris, N. B. W., R. Xu, C. L. Lewis, C. J. Hawkes- 2000. 220, 2000. worth, and Y. Zhang, Isotope geochemistry of the Kapp, P., A. Yin, and C. E. Manning, Blueschist-bearing Mattauer, M., Intracontinental subduction, crust-mantle 1985 Tibet Geotraverse, Lhasa to Golmud, Philos. metamorphic core complexes in the Qiangtang block de´collement and crustal-stacking wedge in the Hi- Trans. R. Soc. London, Ser. A, 327, 263 – 285, 1988. reveal deep crustal structure of northern Tibet: Reply, malayas and other collision belts, in Collision Tec- Harrison, T. M., P. Copeland, W. S. F. Kidd, and A. Yin, Geology, 29, 664, 2001. tonics, edited by M. P. Coward and A. C. Ries, Raising Tibet, Science, 255, 1663 – 1970, 1992. Kapp, P., M. A. Murphy, A. Yin, T. M. Harrison, D. Spec. Pub. Geol. Soc., 19, 37 – 50, 1986. Harrison, T. M., P. Copeland, W. S. F. Kidd, and O. M. Lin, and G. Jinghu, Mesozoic and Cenozoic tec- Matte, P., P. Tapponnier, N. Arnaud, L. Bourjot, J. P. Lovera, Activation of the Nyainqentanghla shear tonic evolution of the Shiquanhe area of western Avouac, P. Vidal, L. Qing, P. Yusheng, and W. Yi, zone: Implications for uplift of the southern Tibetan Tibet, Tectonics, 21, doi:10.1029/2001TC001332, Tectonics of western Tibet, between the Tarim and Plateau, Tectonics, 14, 658 – 676, 1995. in press, 2002. the Indus, Earth Planet. Sci. Lett., 142, 311 – 330, Harrison, T. M., M. Grove, K. D. McKeegan, C. D. Kind, R., et al., Comprehensive seismic images of the 1996. Lovera,O.M.Coath,andP.LeFort,Originand crust and upper mantle beneath Tibet, Science, 298, Matte, P., M. Mattauer, J. M. Olivet, and D. A. Griot, episodic emplacement of the Manaslu intrusive com- 1219 – 1222, 2002. Continental subduction beneath Tibet and the Hima- plex, central Himalaya, J. Petrol., 40, 3 – 19, 1999. Kirby, E., K. X. Whipple, B. C. Burchfiel, W. Tang, G. layan orogen: A review, Terra Nova, 9, 264 – 270, Hauck, M. L., K. D. Nelson, L. D. Brown, W. Zhao, and Berger, Z. Sun, and Z. Chen, Neotectonics of the 1997. A. R. Ross, Crustal structure of the Himalayan oro- Min Shan, China: Implications for mechanisms McNamara, D. E., T. J. Owens, P. G. Silver, and F. T. gen at 90° east longitude from Project INDEPTH driving Quaternary deformation along the eastern Wu, Shear wave anisotropy beneath the Tibetan deep reflection profiles, Tectonics, 17, 481–500, margin of the Tibetan Plateau, Geol. Soc. Am. Bull., Plateau, J. Geophys. Res., 99, 13,655 – 13,665, 1998. 112, 375 – 393, 2000. 1994. Henry, P., X. Le Pichon, and B. Goffe, Kinematic, Klootwijk, C. T., P. J. Conaghan, and C. McA. Powell, McNamara, D. E., W. R. Walter, T. J. Owens, and C. J. thermal and petrological model of the H-imalayas: The Himalayan arc: Large-scale continental sub- Ammon, Upper mantle velocity structure beneath 12 - 24 DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY

the Tibetan Plateau from Pn travel time tomography, Parrish, R. R., and K. V. Hodges, Isotopic constraints Schelling, D., and K. Arita, Thrust tectonics, crustal J. Geophys. Res., 102, 493 – 505, 1997. on the age and provenance of the Lesser and Great- shortening, and the structure of the far-eastern Nepal Me´tivier, F., Y. Gaudemer, P. Tapponnier, and B. er Himalayan sequences, Nepalese Himalaya, Geol. Himalaya, Tectonics, 10, 851 – 862, 1991. Meyer, Northeastward growth of the Tibet plateau Soc. Am. Bull., 108, 904 – 911, 1996. Searle, M. P., Structural evolution and sequence of deduced from balanced reconstruction of two de- Patriat, P., and J. Achache, Indian-Asia collision chron- thrusting in the High Himalayan, Tibetan Tethys positional areas: The Qaidam and Hexi Corrdor ba- ology has implications for crustal shortening and and Indus suture zones of Zanskar and Ladakh, wes- sins, China, Tectonics, 17, 823 – 842, 1998. driving mechanisms of plates, Nature, 311, 615 – tern Himalaya, J. Struct. Geol., 8, 923 – 936, 1986. Miller, C., R. Schuster, U. Klotzli, W. Frank, and R. 621, 1984. Searle, M. P., R. L. Corfield, B. Stephenson, and J. Purtscheller, Post-collisional potassic and ultrapo- Patzelt, A., H. Li, J. Wang, and E. Appel, Palaeomag- McCarron, Structure of the north Indian continental tassic magmatism in SW Tibet: Geochemical and netism of Cretaceous to Tertiary sediments from margin in the Ladakh-Zanskar Himalayas: Implica- Sr-Nd-Pb-O isotopic constraints for mantle source southern Tibet: Evidence for the extent of the north- tions for the timing of obduction of the Spontang characteristics and petrogenesis, J. Petrol., 40, ern margin of India prior to the collision with Eur- ophiolite, India-Asia collision and deformational 1399 – 1424, 1999. asia, Tectonophysics, 259, 259 – 284, 1996. events in the Himalaya, Geol. Mag., 134, 297 – Miller,C.,R.Schuster,U.Klotzli,W.Frank,andB. Pearson, O. N., Structural evolution of the central Nepal 316, 1997. Grasemann, Late Cretaceous-Tertiary magmatic fold-thrust belt and regional tectonic and structural Seeber, L., J. G. Armbruster, and R. C. Quittmeyer, and tectonic events in the Transhimalaya batholith sifnificance of the Ramgarh thrust, Ph.D. disserta- Seismicity and continental subduction in the Hima- (Kailas area, SW Tibet), Schweiz. Mineral. Petrogr. tion, 231 pp., Univ. of Ariz., Tucson, 2002. layan arc, in Zagros, Hindu Kush, Himalaya: Geo- Mitt., 80, 1 – 20, 2000. Peˆcher, A., The metamorphism in the central Himalaya, dynamic Evolution, Geodyn Ser., vol. 3, edited by Molnar, P., A review of geophysical constraints on the J. Metamorph. Geol., 7, 31 – 41, 1989. H. K. Gupta, and F. M. Delany, pp. 215 – 242, deep structure of the Tibetan Plateau, the Himalaya Peltzer, G., and P. Tapponnier, Formation and evolution AGU, Washington, D. C., 1981. and the Karakoram, and their tectonic implications, of strike-slip faults, rifts, and basins during the In- Sengor, A. M. C., and B. A. Natal’in, Paleotectonics of Philos. Trans. R. Soc. London, Ser. A, 326, 33 – 88, dia-Asia collision: An experimental approach, Asia: Fragments of a synthesis, in The Tectonic 1988. J. Geophys. Res., 93, 15,085 – 15,117, 1988. Evolution of Asia, edited by A. Yin and T. M. Har- Molnar, P., and W.-P. Chen, Focal depths and fault Pivnik, D. A., and N. A. Wells, The transition from rison, pp. 486 – 640, Cambridge Univ. Press, New plane solutions of earthquakes under the Tibetan Tethys to the Himalaya as recorded in northwest York, 1996. Plateau, J. Geophys. Res., 88, 1180 – 1196, 1983. Pakistan, Geol. Soc. Am. Bull., 108, 1295 – 1313, Sinclair, H. D., Tectonostratigraphic model for under- Molnar, P., and H. Lyon-Caen, Fault plane solutions of 1996. filled peripheral foreland basins: An Alpine per- earthquakes and active tectonics of the Tibetan Pla- Platt, J. P., and P. C. England, Convective removal of spective, Geol. Soc. Am. Bull., 109, 324 – 346, teau and its margins, Geophys. J. Int., 99, 123 – 153, lithosphere beneath mountain belts: Thermal and 1997. 1989. mechanical consequences, Am. J. Sci., 294, 307 – Spakman, W., Delay-time tomography of the upper Molnar, P., and P. Tapponnier, Cenozoic tectonics of 336, 1994. mantle below Europe, the Mediterranean, and Asia: Effects of a continental collision, Science, Pogue, K. R., M. D. Hylland, R. S. Yeats, W. U. Khattak, Asia Minor, Geophys. J. Int., 107, 309 – 332, 189, 419 – 426, 1975. and A. Hussain, Stratigraphic and structural frame- 1991. Molnar, P., P. England, and J. Martinod, Mantle dy- work of Himalayan foothills, northern Pakistan, in Srivastava, P., and G. Mitra, Thrust geometries and namics, uplift of the Tibetan Plateau, and the Indian Himalaya and Tibet: Mountain Roots to Mountain deep structure of the outer and lesser Himalaya, Monsoon, Rev. Geophys., 31, 357 – 396, 1993. Tops, edited by A. Macfarlane, R. B. Sorkhabi, and J. Kumaon and Garhwal (India): Implications for evo- Mugnier, J.-L., G. Mascle, and T. Faucher, Structure of Quade, Spec. Pap. Geol. Soc. Am., 328, 257 – 274, lution of the Himalayan fold-and-thrust belt, Tec- the Siwaliks of western Nepal: An intracontinental 1999. tonics, 13, 89 – 109, 1994. accretionary prism, Int. Geol. Rev., 35, 32–47, Powell, C. McA., and P. J. Conaghan, Plate tectonics and Steck, A., J.-L. Epard, J.-C. Vannay, J. Hunziker, M. 1993. the Himalayas, Earth Planet. Sci. Lett., 20, 1 – 12, Girard, A. Morard, and M. Robyr, Geological trans- Murphy, M. A., and A. Yin, Structural evolution and 1973. ect across the Tso Morari and Spiti areas: The nappe sequence of thrusting in the Tethyan fold-thrust belt Powers, P. M., R. J. Lillie, and R. S. Yeates, Structure structures of the Tethys Himalaya, Eclogae Geol. and Indus-Yalu suture zone, southwest Tibet, Geol. and shortening of the Kangra and Dehra Dun reen- Helv., 91, 103 – 121, 1998. Soc. Am. Bull., 115, 21 – 34, 2003. trants, Sub-Himalaya, India, Geol. Soc. Am. Bull., Stu¨we, K., and T. D. Barr, On the relationship between Murphy, M. A., A. Yin, T. M. Garrison, S. B. Durr, Z. 110, 1010 – 1027, 1998. surface uplift and gravitational extension, Tectonics, Chen, et al., Significant crustal shortening in south- Ratschbacher, L., W. Frisch, and L. Guanghua, Distrib- 19, 1056 – 1064, 2000. central Tibet prior to the Indo-Asian collision, uted deformation in southern and western Tibet dur- Tapponnier, P., et al., The Tibetan side of the India- Geology, 25, 719 – 722, 1997. ing and after the India-Asia collision, J. Geophys. Eurasia collision, Nature, 294, 405 – 410, 1981. Murphy, M. A., A. Yin, P. Kapp, T. M. Harrison, L. Din, Res., 99, 19,917 – 19,945, 1994. Tapponnier, P., G. Peltzer, A. Y. Le Dain, and R. Armi- and J. Guo, Southward propagation of the Karakor- Robinson, D. M., P. G. DeCelles, P. J. Patchett, and C. N. jo, Propagating extrusion tectonics in Asia: New am fault system into southwest Tibet: Timing and Garzione, The kinematic history of the Nepalese Hi- insights from simple experiments with plasticine, magnitude of slip, Geology, 27, 719 – 722, 1999. malaya interpreted from Nd isotopes, Earth Planet. Geology, 10, 611 – 616, 1982. Najman, Y., and E. Garzanti, Reconstructing early Hi- Sci. Lett., 192, 507 – 521, 2001. Tapponnier, P., B. Meyer, J. P. Avouac, G. Peltzer, Y. S. malayan tectonic evolution and paleogeography Robinson, D. M., G. Dupont-Nivet, G. E. Gehrels, and Gaudemer, H. X. Guo, K. Yin, Z. Chen, S. Cai, and from Tertiary foreland basin sedimentary rocks, Y. Zhang, The Tula uplift, northwestern China: Evi- H. Dai, Active thrusting and folding in the Qilian northern India, Geol. Soc. Am. Bull., 112, 435 – dence for regional tectonism of the northern Tibetan Shan, and decoupling between upper crust and 449, 2000. Plateau during late Mesozoic-early Cenozoic time, mantle in northeastern Tibet, Earth Planet. Sci. Najman, Y. M. R., M. S. Pringle, M. R. W. Johnson, Geol. Soc. Am. Bull., in press, 2002. Lett., 97, 382 – 403, 1990. A. H. F. Robertson, and J. R. Wijbrans, Laser Rodgers, A. J., and S. Y. Schwartz, Lithospheric struc- Tapponnier, P., X. Zhiqin, F. Roger, B. Meyer, N. Ar- 40Ar/39Ar dating of single detrital muscovite grains ture of the Qiangtang Terrane, northern Tibetan Pla- naud, G. Wittlinger, and Y. Jingsui, Oblique step- from early foreland basin sediments in India: Im- teau, from complete regional waveform modeling: wise rise and growth of the Tibet Plateau, Science, plications for early Himalayan evolution, Geology, Evidence for partial melt, J. Geophys. Res., 103, 294, 1671 – 1677, 2001. 25, 535 – 538, 1997. 7137 – 7152, 1998. Tonarini, S., I. Villa, M. Oberli, F. Meier, D. A. Spen- Nelson, K. D., et al., Partially molten middle crust be- Rowley, D. B., Age of initiation of collision between cer, U. Pognante, and J. G. Ramsey, Eocene age of neath southern Tibet: A synthesis of Project IN- India and Asia: A review of stratigraphic data, eclogite metamorphism in Pakistan Himalaya: Im- DEPTH results, Science, 274, 1684 – 1688, 1996. Earth Planet. Sci. Lett., 145, 1 – 13, 1996. plications for India-Eurasia collision, Terra Nova, Ni, J., and M. Barazangi, Seismotectonics of the Hima- Royden, L. H., B. C. Burchfiel, R. W. King, E. Wang, 5, 13 – 20, 1993. layan collision zone: Geometry of the underthrust- Z. Chen, F. Shen, and Y. Liu, Surface deformation Turner, S., C. J. Hawkesworth, J. Liu, N. Rogers, S. ing Indian plate beneath the Himalaya, J. Geophys. and lower crustal flow in eastern Tibet, Science, Kelley, and P. van Calsteren, Timing of Tibetan Res., 89, 1147 – 1163, 1984. 276, 788 – 790, 1997. uplift constrained by analysis of volcanic rocks, O’Brien, P. J., N. Zotov, R. Law, M. A. Khan, and M. Q. Russo, R. M., and P. G. Silver, Trench-parallel flow Nature, 364, 50 – 53, 1993. Jan, Coesite in Himalayan eclogite and implications beneath the Nazca plate from seismic anisotropy, Valdiya, K. S., Geology of the Kumaon Lesser for models of India-Asia collision, Geology, 29, Science, 263, 1105 – 1111, 1994. Himalaya, 291 pp., Wadia Inst. of Himalayan 435 – 438, 2001. Sakai, H., Rifting of the Gondwanaland and uplifting of Geol., Dehra Dun, 1980. Owens, T. J., and G. Zandt, Implications of crustal the Himalayas recorded in Mesozoic and Tertiary Van der Voo, R., Paleomagnetism of the Atlantic, Tethys property variations for models of Tibetan Plateau fluvial sediments in the Nepal Himalayas, in Sedi- and Iapetus Ocean, 411 pp., Cambridge Univ. evolution, Nature, 387, 37 – 43, 1997. mentary Facies in the Active Plate Margin, edited Press, New York, 1993. Pan, Y., Unroofing history and structural evolution of by A. Taira and F. Masuda, pp. 723 – 732, Terra Van der Voo, R., W. Spakman, and H. Bijwaard, Tethyan the southern Lhasa terrane, Tibetan Plateau: Impli- Sci., Tokyo, Japan, 1989. subducted slabs under India, Earth Planet. Sci. Lett., cations for the continental collision between India Schelling, D., The tectonostratigraphy and structure of 171, 7 – 20, 1999. and Asia, Ph.D. dissertation, 287 pp., State Univ. the eastern Nepal Himalaya, Tectonics, 11,925– Vannay, J.-C., and K. V. Hodges, Tectonometamorphic of New York, Albany, 1993. 943, 1992. evolution of the Himalayan metamorphic core be- DECELLES ET AL.: HIMALAYAN-TIBETAN OROGENY 12 - 25

tween Annapurna and Dhaulagiri, central Nepal, New constraints on the timing of east-west exten- Zandt, G., S. L. Beck, and C. Chase, Similarity of Deep J. Meteorol. Geol., 14, 635 – 656, 1996. sion and its relationship to postcollisional volcan- Structures of the High Plateaus in Tibet and the Veevers, J. J., C. McA. Powell, and B. D. Johnson, ism, Geology, 29, 339 – 342, 2001. Andes: Implications for Raising Plateaus, Eos Greater India’s place in Gondwanaland and in Asia, Wittlinger, G., et al., Seismic tomography of Northern- Trans. AGU, 81(48), Fall Meet. Suppl., Abstract Earth Planet. Sci. Lett., 27, 383 – 387, 1975. Tibet and Kunlun: Evidence for crustal blocks and T61D-07, 2000. Wang, Q., et al., Present-day crustal deformation in mantle velocity contrast, Earth Planet. Sci. Lett., Zhang, K. J., Cretaceous palaeogeography of Tibet and China constrained by global positioning system 139, 263 – 279, 1996. adjacent areas (China): Tectonic implications, Cre- measurements, Science, 294, 574 – 577, 2001. Xu, R. H., U. Scha¨rer, and C. J. Allegre, Magmatism taceous Res., 21, 23 – 33, 2000. Wei, W., et al., Detection of widespread fluids in the and metamorphism in the Lhasa block (Tibet): A Zhao, W., and W. J. Morgan, Injection of Indian crust Tibetan crust by magnetotelluric studies, Science, geochronological study, J. Geol., 93, 41 – 57, 1985. into Tibetan lower crust: A two-dimensional finite 292, 716 – 718, 2001. Yin, A., Mode of Cenozoic east-west extension in Tibet element model study, Tectonics, 6, 489 – 504, Wernicke, B., and S. R. Getty, Intracrustal subduction suggesting a common origin of rifts in Asia during 1987. and gravity currents in the deep crust: Sm-Nd, Ar- the Indo-Asia collision, J. Geophys. Res., 105, Zhao, W., et al., Crustal structure of central Tibet as Ar, and thermobarometric constraints from the Ska- 21,745 – 21,759, 2000. derived from project INDEPTH wide-angle seismic git Gneiss complex, Washington, Geol. Soc. Am. Yin, A., and T. M. Harrison, Geologic evolution of the data, Geophys. J. Int., 145, 486 – 498, 2001. Bull., 109, 1149 – 1166, 1997. Himalayan-Tibetan orogen, Annu. Rev. Earth Pla- Zhou, D., and S. A. Graham, the Songpan-Ganzi com- Wesnousky, S. G., S. Kumar, R. Mohindra, and V. C. net. Sci., 28, 211 – 280, 2000. plex of the western Qinling Shan as a Triassic Thakur, Uplift and convergence along the Himala- Yin, A., and S. Nie, A Phanerozoic palinspastic recon- remnant ocean basin, in The Tectonic Evolution yan Frontal thrust of India, Tectonics, 18, 967 – 976, struction of China and its neighboring regions, in of Asia,editedbyA.YinandT.M.Harrison, 1999. The Tectonic Evolution of Asia, edited by A. Yin pp. 281 – 299, Cambridge Univ. Press, New York, Whittington, A., G. Foster, N. Harris, D. Vance, and M. and T. M. Harrison, pp. 442 – 485, Cambridge Univ. 1996. Ayres, Lithostratigraphic correlations in the western Press, New York, 1996. Zhu, L., Broadband waveform modeling and its appli- Himalaya-An isotopic approach, Geology, 27, Yin, A., T. M. Harrison, M. A. Murphy, M. Grove, S. cation to the lithospheric structure of the Tibetan 585 – 588, 1999. Nie, F. J. Ryerson, W. X. Feng, and C. Z. Le, Ter- Plateau, Ph.D. dissertation, 141 pp., Calif. Inst. of Willems, H., Z. Zhou, B. Zhang, and K.-U. Gra¨fe, Stra- tiary deformation history of southeastern and south- Technol., Pasadena, Calif., 1998. tigraphy of the Upper Cretaceous and Lower Tertiary western Tibet during the Indo-Asian collision, Geol. Zhu, L., and D. V. Helmberger, Moho offset across the strata in the Tethyan Himalayas of Tibet (Tingri area, Soc. Am. Bull., 111, 1644 – 1664, 1999. northern margin of the Tibetan plateau, Science, China), Geol. Rundsch., 85, 723 – 754, 1996. Yin, A., et al., Tectonic history of the Altyn Tagh fault 281, 1170 – 1172, 1998. Willett, S. D., and C. Beaumont, Subduction of Asian system in northern Tibet inferred from Cenozoic lithosphere mantle beneath Tibet inferred from sedimentation, Geol. Soc. Am. Bull., 114, 1257 – models of continental collision, Nature, 369, 1295, 2002. ÀÀÀÀÀÀÀÀ 642 – 645, 1994. Yue, Y., B. D. Ritts, and S. A. Graham, Initiation and P. G. DeCelles, D. M. Robinson, and G. Zandt, Williams, H. S., S. Turner, S. Kelley, and N. Harris, long-term slip history of the Altyn Tagh fault, Int. Department of Geosciences, University of Arizona, Age and composition of dikes in southern Tibet: Geol. Rev., 43, 1087 – 1093, 2001. Tucson, AZ 85721, USA. ([email protected])