JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 113, F04018, doi:10.1029/2007JF000897, 2008

Quantifying landscape differences across the Tibetan plateau: Implications for topographic relief evolution J. Liu-Zeng,1,2 P. Tapponnier,2 Y. Gaudemer,2 and L. Ding1 Received 28 August 2007; revised 4 September 2008; accepted 14 October 2008; published 31 December 2008.

[1] We quantify the bulk topographic characteristics of the Tibet-Qinghai plateau with specific focus on three representative regions: northern, central, and southeastern Tibet. Quantitative landscape information is extracted from Shuttle Radar Topography Mission digital elevation models. We find that the morphology of the Tibetan plateau is nonuniform with systematic regional differences. The northern and central parts of the plateau are characterized by what we suggest to call ‘‘positive topography,’’ i.e., a topography in which elevation is positively correlated with relief and mean slope. A major change from the internally drained central part of Tibet to the externally drained part of eastern Tibet is accompanied by a transition from low to high relief and from positive to ‘‘negative topography,’’ i.e., a topography where there is an inverse or negative correlation between elevation and relief and between elevation and mean slope. Relief in eastern Tibet is largest along rivers as they cross an ancient, eroded plateau margin at high angle to the major strike-slip faults, the Yalong-Yulong thrust belt, implying strong structural control of regional topography. We propose that the evolution of river systems and drainage efficiency, the ability of rivers to transport sediments out of the orogen, coupled with tectonic uplift, is the simplest mechanism to explain systematic regional differences in Tibetan landscapes. Basin filling due to inefficient drainage played a major role in smoothing out the tectonically generated structural relief. This mode of smoothing started concurrently with tectonic construction of the relief, as most clearly illustrated today in the Qilian Shan-Qaidam region of the northeastern plateau. In the interior of Tibet, further ‘‘passive’’ filling, due to internal drainage only, continued to smooth the local relief millions of years after the cessation of major phases of surface uplift due to crustal shortening. Thus, diachronous beveling at high elevation produced the low-relief surface of the high plateau. In southeast Tibet, headward retreat of external drainages brought back ‘‘in’’ the global ocean base level, first disrupting then interrupting the relief-reduction process. It produced a transitional topography by dissecting the ‘‘old’’ remnant plateau surface, which introduced younger and steeper incision of this hitherto preserved high base level. This provides a unifying mechanism for the formation of the low-relief plateau interior, and for the origin of the high-elevation, low-relief relict surface in southeastern Tibet. Our analysis brings forth the importance of surface processes, in particular drainage efficiency, in shaping plateau morphology and landscape relief. Such key processes appear to have been mostly ignored in numerical models of plateau deformation. Our results also cast doubt on and provide a more realistic alternative to the fashionable contention that a continuous preuplift, low-relief surface first formed at low elevation, extending all the way to the South China Sea shore, before being warped upward in the late Miocene-Pliocene by lower crustal channel flow. Citation: Liu-Zeng, J., P. Tapponnier, Y. Gaudemer, and L. Ding (2008), Quantifying landscape differences across the Tibetan plateau: Implications for topographic relief evolution, J. Geophys. Res., 113, F04018, doi:10.1029/2007JF000897.

1. Introduction 1Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing, China. [2] The Tibetan plateau is the world’s highest and largest 2Laboratoire de Tectonique, Institut de Physique du Globe, Paris, orogenic plateau. It has a low-relief internally drained France. interior, flanked in the north and south by steep-sided edges [Fielding et al., 1994], the southern one including Earth’s Copyright 2008 by the American Geophysical Union. 0148-0227/08/2007JF000897 highest summits with elevations greater than 8 km. Toward

F04018 1of26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018 the southeast and northeast, the plateau shows more gently models addressing plateau uplift and growth. In this paper, tapered topography, providing escape routes for some of the we engage in a quantitative description of the bulk land- largest Asian rivers. scape characteristics of different regions of the Tibetan [3] Continuum models of Tibet plateau deformation have plateau. Our morphometric analysis of topography across been primarily motivated by these general topographic the plateau demonstrates the existence of systematic and features and their gravitational implications, and sometimes meaningful regional differences. The northern and central based solely upon them. For instance, the low-relief interior parts of the plateau can be characterized as regions of of the plateau has often been taken as evidence of viscous ‘‘positive’’ topography, i.e., where there is a positive corre- flow in the lower crust, a mechanism thought necessary to lation between relief and elevation, and between local slope smooth out irregularities in topography and crustal thick- and elevation. The change from the internally drained ness [Bird, 1991; Shen et al., 2001; McKenzie and Jackson, central plateau to the externally drained part of eastern 2002]. At the plateau scale, these continuum models, Tibet is accompanied by a transition from positive to whether the vertically averaged ‘‘thin viscous’’ sheet or ‘‘negative’’ topography, i.e., with inverse correlation be- the decoupled weaker lower crust flow classes of models, tween elevation, relief and slopes. Fluvial relief in eastern view the lateral growth of Tibet as being driven by stored Tibet is largest across a NE-trending topographic step that gravitational potential energy, through ‘‘concentric’’ out- may mark an ancient, eroded rim of the plateau. Regions ward spreading occurring more or less simultaneously along located southeast of this step do not show negative topog- all edges [e.g., England and Houseman, 1986, 1989; raphy. We propose that the evolution of river systems and Royden et al., 1997; Shen et al., 2001; Jimenez-Munt and drainage efficiency, the ability of rivers to transport sedi- Platt, 2006]. In such models, the southeastern plateau, ments out of the orogen, coupled with tectonic uplift, like the northeastern plateau, would have been rising fast provides a robust unifying mechanism to explain both the throughout the late Cenozoic to the present-day. In partic- way in which the plateau morphology and relief were ular, the more gently sloping topography in southeastern shaped and the systematic regional differences observed in Tibet has been taken as evidence for tilting of a low- the Tibetan landscape. elevation, low-relief surface driven by recent and ongoing [6] Similar techniques of large-scale topographic analysis lower crustal channel flow [Clark and Royden, 2000; have been applied to various regions [e.g., Ahnert, 1970, Clark et al., 2005, 2006]. In all these models, the current 1984; Koons, 1989; Ohmori, 1993; Brozovic´etal., 1997; topography and landscape morphology of the plateau are Brocklehurst and Whipple, 2004]. An overall study of the inferred to be a straight forward reflection of deep seated topography of the Tibet plateau was previously performed lower crust and mantle processes. By contrast, the impor- by Fielding et al. [1994] and Fielding [1996] using a 90 m tance of surficial erosion processes in shaping the landscape resolution digital elevation model (DEM). Their data (not is downplayed. publicly available at the time) was 100 times better in [4] Block models that include geological and tectonic resolution than ETOPO5 (5 min latitude/longitude gridded evidence make quite different predictions, in particular, that elevation data). These authors focused mainly on the the plateau has grown mostly toward the northeast, in flatness of Tibet, however, without characterizing regional stepwise fashion [Lacassin et al., 1997; Meyer et al., differences, even though the total area of Tibet is almost half 1998; Metivier et al., 1998; Tapponnier et al., 2001; Pares that of the contiguous United States. Yet inspection of et al., 2003]. On the basis of strike-slip driven mountain Tibet’s morphology shows that the plateau exhibits signif- growth and not just foreland thrust migration, Tapponnier et icant variations in interior relief. Besides, it is an assem- al. [2001] proposed a ‘‘three-step’’ model of sequential blage of tectonic terranes with different Phanerozoic uplifting (Figure 1a): south Tibet would have been the first histories that likely responded differently to the strain part of the plateau to rise during the Eocene. The region generated by the India/Eurasia collision, and that likely north and east of the Kunlun range would be the youngest rose diachronously since 55 Ma ago [e.g., Lacassin et part of the plateau, and would have risen fast due to crustal al., 1996; Meyer et al., 1998; Tapponnier et al., 2001]. With shortening, between the Pliocene and present. The region in the advent of freely accessible Shuttle Radar Topography between would have gone through its main phase of uplift Mission DEMs, which provide public access to a spatial between the Eocene and the Miocene. The southeastern part resolution similar to that used by Fielding et al. [1994], it is of the plateau would have been shortened and uplifted now possible to explore such questions in greater detail on coevally with south Tibet mostly before the Miocene, as it the basis of a uniform high-quality digital data set. was being extruded southeastward along the Red River and Xianshuihe faults. In the stepwise model, intermontane 2. Methods ‘‘bathtub sediment infilling’’ is an integral part of elevation gain and relief smoothing, a process essential to the building [7] Our regional scale landscape analysis of the Tibetan of plateau morphology [Meyer et al., 1998; Metivier et al., plateau is carried out using Shuttle Radar Topography 1998; Tapponnier et al., 2001]. However, while such Mission (SRTM) DEMs. The SRTM (http://www2.jpl.nasa. surface processes are clearly at work in shaping the topog- gov/srtm/index.html) collected elevation data over 80% of raphy of northern Tibet, it has been less straightforward to Earth’s land area by means of radar interferometry. The data assess what role they might have played, if any, in central set is at present the best quality, freely available, digital and southeastern Tibet. elevation model with 3 arc sec resolution and consistent [5] Understanding the coevolution of plateau topography accuracy outside the United States of America. We use the and relief structure, within a sound geological and tectonic unedited version, which contains holes or voids, especially in framework, is a starting point to assess the validity of

2of26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 1. (a) Regions chosen for bulk landscape characteristics analysis, superimposed on ‘‘three- Tibet’’ model of principal plateau building epochs. From Tapponnier et al. [2001]. Reprinted with permission from AAAS. Boxes N1–N4 are samples of the northern plateau, C1–C5 are samples of the plateau interior, and E1–E6 are samples of the SE Tibet. Major rivers draining from the plateau are shown in green. (b) Map of slopes in the plateau [from Fielding et al., 1994]. Red line delineates current internally drained area. Blue lines trace boundary between shallow and steep slopes within externally drained area. Analysis boxes are also shown. areas with very steep terrain, water bodies and sand dune 10 km window size scale captures the most common ridge- fields. valley spacing. Local relief is widely accepted as a basic [8] The geomorphic indices we discuss in some detail are parameter that reflects the interplay between tectonics and elevation histograms, local relief, and local slope. The erosion [e.g., Burbank, 1992; Montgomery, 1994; Whipple topographic hypsometry quantifies the distribution of ele- et al., 1999]. It is a measure of roughness in topography at vations within a region. We take the local relief to be the kilometer scale. The value of local relief has been shown to elevation difference between the highest and lowest points depend on the scale or window size, possibly as a self-affine in 10 10 km square sampling windows, assuming that a fractal [Ahnert, 1984; Weissel et al., 1994]. Our choice of a

3of26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

10 km window size is consistent with those made in trates into the easternmost part of the region. Tributaries previous studies, such as Montgomery and Brandon of the Yellow River are few, and do not extend far from [2002] and Clark et al. [2006], who chose a 10 km the trunk. diameter circular window, though 20 and 5 km square [12] In map view, the local relief, on length scales of sampling window sizes have also been used [Ahnert, 10 km, mimics closely the topography (Figure 2b). Areas 1970; Sobel et al., 2003]. There still lacks a thorough of relatively high local relief coincide with mountain ranges, assessment of the exact effect of sampling window size. and low local relief with intermontane basins. Furthermore, Despite potential biases related to scale dependence, how- since the mountain ranges of the region are bounded by ever, a comparison between regions can be performed and is active thrust faults, and continue to rise as large-scale ramp meaningful for the general, first-order exploration under- anticlines [Tapponnier et al., 1990b; Meyer et al., 1998; van taken in this paper, provided the same horizontal window der Woerd et al., 2001], local relief is clearly controlled by scale is kept for all regions. these faults. Hence, the local, actively growing relief of the [9] We capture the local slope from the average gradient Qilian Shan area can be directly linked to ‘‘structural relief.’’ in 3 3° cell grids, whose dimensions are converted into [13] Another important feature in the local relief distri- meters and adjusted for latitude dependence. The calculated bution is that the largest local relief is localized along the local slope is averaged over a 200 m scale, hence under- plateau’s mountain rim. Inside and away from this rim, estimates the actual slope. In addition, points with no data toward the Qaidam and Gonghe basins, the relief is smaller, (NaNs) in the grid are ignored and not taken in account in though locally high along the Yema Shan, Danghe Nan the averaging. While elevation histograms are a measure of Shan, and Qaidam Shan. Two NE–SW transects illustrate the overall relief at the regional scale, local slopes charac- the spatial correlation between elevation, local relief and terize relief at the smallest scale (200 m), and local relief slope (Figure 3). In both transects, the areas of highest at an intermediate scale. Together, these three parameters elevation are located 100–200 km inward from the plateau thus quantify relief at three different spatial scales. edge but correspond to more moderate, rather than to the [10] We target three different regions of the Tibet plateau, largest, local relief. This illustrates the fact that relief in which we further separate subregional areas as analysis development results from the competition between relief boxes (Figure 1): (1) northern Tibet, including the Qilian growth by tectonic uplift and relief reduction by erosional Shan region (N1 to N4); (2) Tibet’s interior, including both mass wasting. That local relief is generally smaller inside internally drained (C1 to C3) and externally drained regions the plateau despite higher elevation suggests that relief (C4 and C5); and (3) southeastern Tibet, i.e., the area reduction continues in these areas of increased regional bounded to the north and east by the Xianshuihe fault (E1 topography at a rate that has probably outpaced that of through E6) with one box south of the Red River-Ailao relative uplift by thrusting. Shan zone in the Shan Thai plateau. Each box is 2 by [14] Elongated intermontane basins, perched at elevations 3 degrees in size, except for N1 (3 by 3 degrees) and N2. of 3000–4000 m, are prominent in the Qilian Shan regional The selection of rectangular analysis boxes is somewhat landscape. They usually correspond to the local maxima in arbitrary and they span multiple drainage basins. Such large, the elevation histograms (Figure 4). Because the basins are arbitrary boxes may not capture fine local variations [e.g., flat or with only gently dipping slopes, the mean slope at the Brocklehurst and Whipple, 2004], but are appropriate for corresponding elevations is less than it would be if such regional to mesoscale quantification [Brozovic´etal., 1997; basins did not exist. This is clear in Figure 4 where the Montgomery et al., 2001], which is the spatial scale of elevation histograms and average slope distributions are on interest here. In all subregions, nondata points make up only average anticorrelated. The mirror images of the elevation 2% of the surface, or less, except for E3, where this per- histogram and mean slope distributions are a clear indica- centage is somewhat higher (5.6%). tion of the dominance of intermontane basins at 3000 and 4000 m. Above 4000 m, the average slope increases 3. Morphological Characteristics of Selected monotonically with elevation. The dominant areal extent Regions of basins is also clear on the slope histograms. These are exponential-type distributions, characteristic of depositional 3.1. Northern Tibetan Plateau topography [Montgomery, 2001]. About 50% of the local [11] Northern Tibet, north of the Kunlun range and slopes are gentler than 15°. Qaidam basin and south of the Hexi corridor and Gobi [15] That such intermontane flats, which are characterized desert, consists of rugged roughly parallel mountain ranges, by high elevation, low local relief and gentle slope, are the separated by subparallel intermontane basins (Figure 2a). main component, in areal extent, of the regional topography, The ranges are usually bounded by growing folds and active ahead of the rugged mountain ranges, suggests that their thrusts [e.g., Tapponnier et al., 1990b; Meyer et al., 1998; development is a key factor of the regional topographic van der Woerd et al., 2001]. The drainage network is not evolution. Not only do they provide a sedimentary record of well developed in this area, because of the arid to semiarid uplift and erosion in nearby mountain ranges but, as shown climate (annual rainfall is 50–400 mm/a, with a maximum in the next section, they offer a clue to the origin of the of 400–600 mm/a along the Qilian Shan range front [World larger and flatter areas that are ubiquitous in the interior of Meteorological Organization, 1981]). A few mountain rivers the highest part of Tibet. drain north into the Hexi corridor and Ala Shan platform. Other rivers flow parallel to the orientation of mountain 3.2. Tibet Interior ranges and thrusts, some to the northwest into the Tarim basin. [16] The most striking feature of Tibet’s interior is its The externally drained Yellow River catchment only pene- great flatness [e.g., Fielding et al., 1994]. Tibet is in fact

4of26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 2. (a) Topography and major faults of northern Tibetan plateau. This region is generally characterized by narrow, linear mountain ranges separated by intermontane basins. The Yellow River is the only externally draining river that flows across the eastern part of the region, cutting at high angle across several ranges. Active faults are modified after Meyer et al. [1998] and Tapponnier et al. [2001]. (b) Map of local relief using 10 10 km size moving window with 25% overlap. In northern Tibet, areas of high relief are mainly associated with active faults. To the northwest of Qinghai Lake is a high- elevation, low-relief region that is tectonically ‘‘quieter’’ and not yet invaded by external drainage. Inset shows the histograms of local relief in subregions.

5of26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 3. Local relief, elevation, and slope along two transects across Qilian Shan region. Locations of transects are shown in Figure 2a. Horizontal lines at 5000 m for elevation and 1000 m for local relief are arbitrarily chosen for references. Regions of highest relief are associated with faults. Along A-A0,in particular, the highest relief is near the edge of the plateau, which is also the locus of frontal thrusts and active uplift. Moving inward from the plateau edge on this section, regions of highest elevation are associated with more moderate local relief. best described as a mosaic of flat basins, separated by a relatively abrupt low-elevation cut-off at 4000–4500 m mountain ranges tens to hundreds of kilometers apart and a longer tail toward high elevations slightly above (Figure 5). At a more detailed level, close inspection reveals 6000 m (Figure 7). The 4000–4500 m elevations corre- geomorphic differences both between the northern and spond to low-lying, low-relief basins and piedmonts. They southern areas of the high plateau, and between internally represent base levels for local drainages, and include lakes drained and externally drained regions. and sediment-floored flats. The regional morphology is best [17] Figure 6 shows the north–south differences in to- described to result from sediment ‘‘flooding’’ in the closed pography and local relief inside the internally drained basins that nearly ‘‘submerged’’ former peaks (Figure 5). plateau interior. The primary factor for high relief here is These basins have average slopes of generally less than 7°, active normal faulting. Such faulting typically produces and large areal extents, making up 63–65 % of the total structural relief in the 1–2 km range [Armijo et al., 1986; surface area. Slope distributions show that the mean slope see also Taylor et al., 2003; Kapp et al., 2008]. Away from increases with elevation. Slopes are gentle to flat at low the faults, the relief is generally less than 900 m. A relative elevation, but generally become larger as elevation secondary factor is fluvial incision, as revealed by a patch increases, up to around 15° on average. The increasing trend of moderate local relief in the SE corner of C1: this small of slope toward high topography is similar to that observed area is within the uppermost reach of the externally drained in the Qilian Shan region of NE Tibet (Figure 4). The Yarlung Tsangbo River catchment. Toward the north, local difference, however, is that most topographic highs are less relief decreases, from 500 to 900 m in C1 to 300–500 m in steep, with slopes 10° gentler than in the Qilian Shan C2, past the Bangong suture, then to an even narrower 200– region. 300 m range in C3, north of the Tanggula Mountains. The [18] In summary, the internally drained Tibet interior is elevation histograms show an asymmetric distribution with characterized by positive topography, i.e., with elevation

6of26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 4. Thick dark lines with error bars show average slope as a function of elevation for regions N1–N4. The linear increasing trend of average slope with elevation is interrupted by local minima at midaltitudes, because of flat intermontane basins. Gray dashed lines show histograms of elevation distribution. Intermontane basins correspond to local maxima in elevation histogram. Also shown are slope histograms of subregions in the northern Tibetan plateau, showing dominance of slopes of less than 10°.

Figure 5. Photos showing general geomorphic characteristics in Tibetan plateau interior: (a and b) flat basins and far apart steep mountain ranges, as compared with (c) that in the Qaidam basin.

7of26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 6. (a) Topography and (b) local relief, with active faults, of selected regions in Tibet plateau interior (between 30–36°N and 85–88° E). The region is mostly internally drained, with a mean elevation of 4700 m and relief less than 1 km, except near active normal faults.

positively correlated with relief and slope (Figure 8). Low [19] Internally and externally drained zones differ mainly elevations are exclusively located in close basins, which in slope gradient. The most prominent difference is seen at have low or flat relief and gentle slope. At the other end of the low end of the topographic distribution. In the internally the spectrum, higher-elevation hills or ‘‘small’’ mountains drained zones, the 4000 m elevation areas consist of closed that stand above the flat basins show greater relief. basin flats where sediments accumulate, often in and around

8of26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 7. Average slopes (dark solid lines with error bars) increase nearly monotonically with elevation in typical internally drained regions (C3 and C4). Externally drained regions show increased slopes at 4000 m elevation and below. Also shown are histograms of elevation distribution in each region (gray dashed lines) with modal elevations indicated. Depending on location within the plateau interior, elevations of 4300–4900 m have maximal areal extent, with average slopes of 5–10°. Areas with elevations higher than 5500 m have steeper slopes, on order of 15–25° in general. As in northern Tibet, the slope histogram is dominated by shallow slopes, of less than 10°. lakes, with slopes that can be as small as 2°. But in the metrical distribution in internally drained zones, the eleva- externally drained zones, relative lowlands are associated tion histogram in externally drained zones shows more with slopes in excess of 5–10°(e.g., C5, Figure 7), indicating symmetry. This is because incision introduces a low topog- a shift from depositional to erosional regime [Montgomery, raphy end ‘‘tail’’ to the distribution. The difference between 2001]. Low-elevation surfaces are most affected by this the two externally drained regions (C4, C5) is that C4 has shift. The overall effect on topography is a slight change in higher mean elevation (by 300 m), with a significant the regional elevation distribution. Compared to the asym- fraction of lowland flats still unaffected by river down

Figure 8. Local relief, elevation, and local slope along 650 km long N-S profile at longitude 83.6°. Location of profile is shown in Figure 6. The plateau interior is still characterized by ‘‘positive’’ topography, with high elevation corresponding to high relief and steeper slope and vice versa.

9of26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018 cutting, likely due to its position closer to the internally by the perched Gongjue and Changdu Paleocene-Eocene drained interior. This difference likely illustrates the transi- basins (Figure 11), which remained relatively uneroded by tional landscape in the upper reach of external drainages either river [Bureau of Geology and Mineral Resources of that penetrate headward into the plateau from the east and Xizang Autonomous Region, 1993]. The low-relief Dao- southeast. cheng surface between the second-order tributaries of the Jinsha around 29°N caps large bodies of slightly more 3.3. Eastern and Southeastern Tibetan Plateau dissected granite and granodiorite that were intruded into [20] The southeastern part of Tibet, between the eastern the Triassic flysch during the early Jurassic [Bureau of Himalayan syntaxis and the Sichuan basin, is a region of Geology and Mineral Resources of Sichuan Province, rugged topography with well-developed drainage and valley 1991]. Part of this surface above 4000 m was occupied by networks (Figure 9). Three of the largest rivers of Asia, the a small ice cap during the Last Glacial Maximum, and Jinsha Jiang (Yangtze), the Lancang Jiang (Mekong), and previous maxima [Luo and Yang, 1963; Derbyshire et al., the Nu Jiang (Salween), flow out of the high plateau across 1991; Li et al., 1991]. Along the southeast side of the this region, following nearly parallel courses, and carving Yalong thrust belt, the Yanyuan basin occupies a prominent deep gorges into the terrain. Between these large rivers, the relatively flat surface area of 450 km2. It is filled with 1– mean elevation decreases gradually southeastward from 4– 2 km of Paleogene red molasse and up to 650 m of 5 km to 2–3 km, across a NE-trending approximately 200– Neogene coal-bearing fluviolacustrine deposits. [Bureau of 250 km wide zone that roughly coincides with the Yalong Geology and Mineral Resources of Sichuan Province, 1991; thrust belt. Southeast of this belt, the average elevation is Xu et al., 1992]. Southeast of the Yalong thrust belt, low- mainly between 1 and 3 km, with 2–3 km high surface relief terrain caps parts of the folded Cretaceous-lower- patches that loose continuity, shrink and finally vanish Tertiary red beds of the Chuxiong basin. altogether southeastward. The relatively gentler topographic [24] The composite NW- to SSE-trending swath profile of gradient across SE Tibet has been interpreted to result from Figure 12 clearly shows the coincidence between the the uplift, due to lower crustal channel flow, of a low- Yalong-Yulong thrust belt and a 200–250 km wide topo- elevation and low-relief ‘‘relict’’ surface that once extended graphic ‘‘step.’’ The steepest part of this profile, with a continuously, without major disruption, over thousands of mean elevation drop of 2400 m from 4200 m to 1800 m, kilometers from the interior of Tibet to the South China Sea across a series of 3–4 parallel Cenozoic thrust faults (e.g., [e.g., Clark and Royden, 2000; Clark et al., 2006]. Lijiang-Muli and Jin He-Jian He thrust faults), unambigu- [21] Clark et al. [2006] used topographic indices, among ously defines the southeastern margin of the high Tibet other criteria, to characterize such a regional ‘‘preuplift’’ plateau. The Lijiang-Muli fault shows evidence of Quater- surface. But these indices, often single values for the entire nary activity, with a still prominent thrust component [Xiang region, were mainly based on the coarser 1 km resolution et al., 2002; Shen et al., 2005]. In addition to this relatively GTOPO30 DEM [Clark, 2003]. Below we assess the spatial long wavelength slope break, a minor gentle decrease in variation of the morphological characteristics of the regional mean elevation is observed toward the northwest, possibly topography. Such variations define a topographic step and due to regressive fluvial erosion by the Mekong and ‘‘slope break’’ (Figure 9), which seems to have been over- Yangtze drainages. Along the swath profile, the relief looked in previous studies and appears to mark an eroded increases toward the plateau edge. plateau rim. The region northwest of this rim has a different [25] The low-relief terrain northwest of the Yalong thrust topographic signature from that to the southeast. belt is quite different from that to the southeast. In sub- [22] In the map of local relief (Figure 10), linear high- regions on the plateau (E1–E3), there is a marked local relief swaths generally follow the courses of the large rivers. minimum in slope hypsometry around 4200 m elevation Along any given river course, the relief is largest in the (Figure 13). This minimum’s mean slope is 10° less than middle reaches, roughly between 25 and 31°N latitudes, and that at other elevations. This altitude coincides with the decreases both up and downstream. In the upper reaches, modal elevation of these subregions (Figure 13). A similar from latitude 31°N northward into Tibet, relief along the pattern with a local slope minimum close to the modal rivers starts decreasing to vanish into the background relief. elevation is also found in the northwest Himalaya, at 4000– A similar, more rapid decrease is observed along the 5000 m elevation [Brozovic´etal., 1997]. It has been downstream reaches in the relative low-lying lands of interpreted to reflect extensive glacial scouring of the Yunnan south of 26°N. In particular, the Jinsha river and landscape coupled with vigorous freeze-thaw action be- the major tributaries of Yangtze, such as the Yalong Jiang tween 4000 and 6000 m [Brozovic´etal., 1997], rather than and Dadu He cut deepest across a NE-trending zone the signature of a relict plateau surface. But it is clear that coinciding with the Yalong-Yulong thrust belt. Interestingly, here the surface at 4200 m elevation with minimum slopes both the Jinsha and Yalong Jiang make sharp north directed of 20° is simply the continuation of the plateau interior hairpin bends as they cut across this belt. If, as long inferred high base level. It may have been locally modified and [e.g., Clark et al., 2004], such bends formed as a result of further smoothed by glacial action since, as mentioned river capture, then deformation along the thrust belt likely above, ice caps of significant size appear to have occupied played a role. areas northwest of Daocheng and Yajiang. But while the [23] Between these major river valleys are areas of relative flatness and high elevation between 4000 and relatively subdued relief [Clark et al., 2006]. These areas, 5000 m, close to the ELA (equilibrium line altitude) during however, correspond to low-relief terrains of different origin the LGM or previous maxima [Derbyshire et al., 1991; Li et and age. For example, the relatively flat surface between the al., 1991], may have facilitated the accumulation of stagnant Lancang and Jinsha rivers north of 30°N is partly occupied

10 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 9. Smoothed topography (1 km resolution) of southeastern Tibetan plateau. Active faults are shown by dark lines; thick lines are major faults [after Lacassin et al., 1998; Wang et al., 1998; Tapponnier et al., 2001; Deng et al., 2003; Xu et al., 2003]. Red lines are main tectonic boundaries [Chengdu Institute of Geology and Mineral Resources, 2004; Pan et al., 1990]. ALS-RRF, Ailaoshan- Red River fault; JF, Jiali fault; JJF, Jinhe-Jianhe fault; KLF, Kunlun fault; LF, Longquan Shan fault; LMF, Lijiang Muli fault; LTF, Litang fault; NTF, Nanting fault; SF, Sagaing fault; WDF, Wanding fault; XSH- XJF, Xianshuihe-Xiaojiang fault; YJF, Yunijin fault.

11 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 10. Map of local relief in southeast Tibetan plateau. Thick white lines are major river courses. Dark lines indicate major active faults. Regional relief is mainly controlled by fluvial incision with highest elevation differences chiefly along the largest rivers. High relief is also prominent along major active faults, for example, the Xianshuihe or Red River faults. The moving window size (10 10 km square) we used for local relief calculation is slightly larger than the 65 km2 circular window in Clark et al.’s [2006] analysis. Fault names as in Figure 9.

12 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 11. Locations of main Mesozoic and Cenozoic basins and intrusives in eastern and southeastern Tibet. Modified from Chengdu Institute of Geology and Mineral Resources [2004], Bureau of Geology and Mineral Resources of Yunnan Province [1990], and Bureau of Geology and Mineral Resources of Sichuan Province [1991]. Fault names as in Figure 9.

ice, the primary origin of the broad plateau surface cannot [27] Two E–W profiles illustrate further the difference in be solely due to local ice abrasion. topography between regions northwest and southeast of the 0 [26] Southeast of the Yalong-Yulong belt, down below Yalong-Yulong thrust belt. Profile A-A , just north of 30 N, the plateau edge, the variation of slopes as a function of shows clearly that the elevation is inversely correlated with elevation is quite different from those in the highlands to the local relief and with slopes (Figure 14). High elevation northwest. Slopes tend to be more uniform at most eleva- corresponds to relatively small relief and gentler slope, tions. The relatively flattest areas in subregion E5 are at whereas low elevation corresponds to large relief and steep elevations around 1850 m. This corresponds to the Chuxiong slope. This pattern, which we call negative topography, is sedimentary basin. But a continuation of this surface at this not observed in other parts of the plateau. It thus provides elevation does not show up in the adjacent subregion E6. strong evidence for a remnant high and flat surface that is in

13 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 12. Topographic swath profile (100 km width) across plateau margin to the east and southeast of Sichuan basin. Location of swath is shown on Figures 9 and 10. (top) Curves represent maximum, minimum, and mean elevation profiles projected onto axis of swath. Minimum elevation profile samples parts of long profiles of the Lancang-Mekong, Yangtze, and Red rivers. (bottom) Relief difference between maximum and minimum elevation profiles along swath. The maximum relief (3000 m) occurs where the mean elevation starts to drop. Toward the plateau interior, the relief decreases linearly to a uniform low value (800 m) in the upper reaches of the external drainage catchments. Southeastward from the plateau margin relief decreases to 1000–1500 m in the Chuxiong basin, then increases again farther south, likely because of an incision by the Red River catchment and to differential tectonic exhumation along the Red River fault and Ailao Shan shear zone [Leloup et al., 1995, 2001]. the process of being actively dissected by fluvial incision, present-day clue to relief smoothing is observed in the but is still preserved in patches. By contrast, as illustrated on Qilian Shan area of the northeastern plateau (‘‘bathtub’’ profile B-B0, negative topography, characteristic of a dis- basin-filling hypothesis of Meyer et al. [1998], Metivier et sected remnant surface, is not observed in regions south of al. [1998], and Tapponnier et al. [2001; see also Sobel et al., the plateau edge. South of the Yalong-Yulong belt at 2003]). Intermontane catchments are either disconnected 24.4°N, there is no distinctive inverse correlation between from or only partially connected with external drainage elevation and local relief or between elevation and slope. toward the foreland. Small catchments and short rivers perform only short-distance mass transport from topographic 4. Interpretation and Model of Topographic highs to adjacent basins. While vigorous fluvial down Development of Tibetan Plateau cutting by external drainage systems may enhance or maintain relief [Molnar and England, 1990; Gilchrist et [28] Our morphometric analysis and comparison of the al., 1994; Montgomery, 1994; Whipple et al., 1999], the characteristics of topography in three typical regions of the restricted NE Tibet mass transfer process ‘‘erode high and Tibet plateau shows clear variations in local relief and slope/ fill low’’ reduces regional relief instead. elevation patterns. In SE Tibet, average slopes at all [30] In the Qilian Shan region, relief reduction through elevations are generally steep, mostly greater than 10°. This basin filling starts concurrently and in competition with, the region is different from both the internally drained, rela- growth of tectonic relief along thrust-bounded mountain tively stable plateau interior and the Qilian Shan zone of ranges, as indicated by regional basin development [Meyer active thrusting and mountain growth in the northeastern et al., 1998]. Such growth, as well as bathtub infilling, plateau. There, slopes are dominantly gentler than 10°, raises the average elevation, such that the combination of indicating that depositional processes prevail. both processes results in high elevation and low relief. Inward from the plateau rim, relief reduction will in general 4.1. Formation of the Low-Relief Surface in the prevail, because of slower tectonic uplift than along the Plateau Interior plateau-bounding ranges and more effective intermontane [29] The low-relief interior morphology of Tibet and its basin filling, by rivers that are more likely to become close association with internal drainage indicate that drain- completely cut off from their foreland outlet. Such an age efficiency plays a key role in relief smoothing. A interplay accounts well for the landscape evolution of

14 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 13. Average slope and elevation histograms, plotted as a function of elevation, for regions E1– E6. Average slope is steeper than 20° at almost all elevations in E1–E3. Local minimum at 4200 m corresponds to remnant plateau surface. At or south of former plateau edge (E4, E5, and E6), slope and elevation curves are markedly different from those in E1–E3. Also shown are histograms of slope distribution in each region. Elevation histograms of regions C4–C5 and E1–E3 show a gradual change with a decreasing percentage of elevations at 4000–4500 m concurrent with an increasing percentage of lower elevation. northeastern Tibet, where relief becomes smaller as eleva- [32] Low-relief interior morphology associated with in- tion increases from the Gobi rim toward the plateau interior ternal drainage appears to be typical of other continental (Figure 2). In the highest internally drained part of central orogenic plateaus. Particularly clear examples are the Ira- Tibet, amalgamated closed basins form a coalescent and nian plateau and the Puna-Altiplano plateau in South relatively uniform high-elevation base level. The restricted America [Meyer et al., 1998; Sobel et al., 2003; Garcia- drainage catchments are effective in further smoothing relief Castellanos, 2007]. Though smaller in area and lower in at even higher elevation. This is in a way perhaps similar to mean elevation than Tibet, the latter is a high (3800 m) ‘‘high-altitude peneplanation,’’ which is thought to explain low-relief orogenic plateau with internal drainage, large smaller low-relief surfaces within orogenic belts [e.g., lakes or swamps (e.g., Titicaca, Salar de Ulyuni, Salar de Babault et al., 2005, 2007]. Coipasa) and aridity [e.g., Jordan and Alonso, 1987; Isacks, [31] Our topographic analysis thus supports the inference 1988; Alonso et al., 1991; Masek et al., 1994; Vandervoort that, instead of having been raised after formation at low et al., 1995]. An erode-high, fill-low type of surface process, elevation, the low-relief surface of the Tibetan plateau i.e., intermontane bathtub-filling and piedmont growth due formed by diachronous beveling at high elevation. to inefficient drainage evacuation of sediments, is thought to

15 of 26 F04018 I-EGE L:LNSAEDFEECSI TIBET IN DIFFERENCES LANDSCAPE AL.: ET LIU-ZENG 6o 26 of 16

Figure 14. Local relief, elevation, and local slope along two E–W profiles at latitude 30.2° (A-A0) and latitude 24.4° (B- 0 0 B ); locations given on Figures 9 and 10. Profile A-A shows the characteristic signature of ‘‘negative’’ topography, high F04018 elevations correspond to low relief and shallow slopes while low elevations are associated with high relief and steep slopes. The low-relief, high-elevation remnant plateau surface is clear. Profile B-B0, south of the plateau edge, shows no such negative topography. F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 15. Comparison of mean slope: as a function of elevation (a) between northeastern plateau and plateau interior and (b) between plateau interior and southeastern plateau. HPBL is the high plateau base level.

have played a critical role in shaping the Altiplano subdued [35] Figure 16 illustrates a simple interpretation of the landscape [e.g., Riller and Oncken, 2003; Sobel et al., 2003; gradual changes in the slope-elevation functions between Hilley and Strecker, 2005]. Therefore, there is growing the three regions. In cross-sectional view, northern Tibet has consensus that surface processes play a major role in the a steep rim and hanging intermontane basins separated by geomorphic evolution of orogenic plateaux, with high base rugged, towering mountain ranges. Due to the inefficient levels connected to short local drainage systems being the drainage network that progressively becomes internal, sedi- key condition to maintain high-elevation, low-relief plateau ments eroded from the high mountains fill the basins and morphology. flatten the low elevation end of the spectrum. Depending on the period during which internal drainage is maintained, the 4.2. A Model of Topographic Evolution Linking mountain top slopes should become smoother with time. the Three Regions of Tibetan Plateau Later on, as tectonic uplift at the nm slows down or stops [33] The lateral growth of the Tibet plateau since the altogether or possibly due to wetter climate, external drain- onset of collision between India and Eurasia implies that its age grows back vigorously headward into the plateau in- topography is not in steady state flux. Clearly therefore, the terior [Hilley and Strecker, 2005]. The subdued ‘‘protected’’ landscape morphology reflects not only the current climatic topography of the isolated, high base level succumbs to and tectonic setting and surficial processes, but also their incision, which starts at the relatively lowest elevations history [e.g., Howard, 1997]. An evolutionary sequence of where steep slopes reappear. In this interpretation, the plateau development emerges from the quantitative com- drainage evolution, from external to restricted then internal parison and characterization of geomorphic differences due to mountain rim growth, then back to external through between the three parts of Tibet: northeastern plateau, headward conquest, is the key factor linking the three plateau interior and southeastern plateau. regions of the Tibetan plateau. Each region may in fact [34] A connection between these three regions may be be taken to reflect a time step in plateau building, from birth inferred from the variation of mean slope as a function of to demise. elevation. Figure 15a shows that the topography of central [36] Seen from this timescale point-of-view, the Qilian Tibet is characterized by a monotonic increase of mean Shan and Qaidam regions, where active crustal shortening slope with elevation above 4000 m. This is similar to what and thickening occur prominently over a surface area of is observed in northeastern Tibet at elevations above that of about 5.4 105 km2 [e.g., Meyer et al., 1998], provide a intermontane basins. Both northeastern and central Tibet typical example of plateau landscape at an early stage of show positive topography. In Figure 15b, a comparison of topographic development under relatively dry climate con- the slope-elevation functions for southeastern Tibet and ditions (annual rainfall 400 mm/a and less [World Meteo- central Tibet shows the gradual transition from the C rological Organization, 1981]). Regional topography is regions of smooth high topography to the E regions with characterized by relatively large structural relief, still rela- degraded topography. Despite steeper slopes at all eleva- tively low mean elevation (2700m, 1500 m lower than tions and the change to negative topography induced by the Tibet proper), and a steep outer rim. Subparallel narrow external drainage ‘‘reconquest’’ of formerly isolated basins, mountain ranges bounded by active thrusts and growing the remnant plateau surface is still preserved, as demon- folds isolate long and narrow, often rhomb-shaped inter- strated by the relative flatness at 4000–4500 m. montane basins.

17 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

of the interior, and of the southeastern marginal part of the plateau. Southeastern Tibet would in fact represent the third elevation stage of plateau topography: the ultimate, demise period. Intensive fluvial down cutting by external drainages attacks and dismembers the elevated and smoothed plateau surface thus far protected by isolation. The Sichuan and northern Yunnan regions of SE Tibet are characterized by mean elevation lowering and renewed relief buildup. Such fluvially driven increase of relief, or ‘‘negative topogra- phy,’’ is radically different from the tectonically controlled structural relief of the NE plateau. The fluvial incision dominance is clear from the slope-elevation functions (Figure 15b). Down cutting associated with fluvial incision starts at low elevation, and progresses into the highlands through headward retreat and capture. This eventually increases the slopes at all elevations, though remnants of the plateau surface are preserved for a time. The ‘‘flat’’ highland limit that subsists 50–400 km east of the internal drainage divide in east-central Tibet attests to such recent rapid headward conquest of the plateau by the upper catch- ments of the Yellow, Jisha, Lancang and Nu rivers (Figure 1). 4.3. Origin of the High-Elevation, Low-Relief Remnant Surface in Southeastern Tibet [39] The relict high-elevation, low-relief, and low-slope plateau surface of SE Tibet clearly extends out from the plateau interior (Figures 9 and 10). In contrast with previous inferences of a remnant, formerly continuous low-elevation surface having reached as far south as Vietnam (20°N) without major disruption prior to the uplift of Tibet [Clark et al., 2006], our quantitative analysis suggests that the Tibet plateau surface terminates along a 200 km wide, NE- Figure 16. Cartoons illustrating sequential buildup of trending topographic edge, which coincides with the plateau relief, reflecting interplay between sedimentation, Yalong-Yulong or Jinhe-Jianhe (in Chinese literature) belt. drainage efficiency, and crustal shortening, as well as Though eroded, this fold and thrust belt, which comprises possible evolutionary links between northern plateau, plateau several parallel ridges, likely represents the Eocene-Miocene interior, and southeastern plateau, as a function of age. structural barrier that once separated the SE plateau region from the Chuxiong Cretaceous-Eocene foreland basin [Leloup et al., 1995; Lacassin et al., 1996, 1997]. The [37] The plateau interior represent the next evolution Yalong range formed during the Cenozoic [Pan et al., 1990; stage of plateau morphology, essentially driven by internal Xu et al., 1992] as a result of crustal shortening on south- drainage with vanishingly small fault uplift or even tectonic eastward younging thrusts (Z. Q. Xu, personal communica- quiescence, as the plateau grows and the deformation front tion, 2005). Synmetamorphic deformation on the Yulong moves outward. The region undergoes further relief smooth- Shan de´collement was dated at 35 Ma [Lacassin et al., ing while the mean elevation keeps increasing. Highlands 1996], while39Ar/40Ar cooling ages indicate later exhuma- become more isolated and separated by wider basins. tion by folding at 17 Ma. The uplift of the range is also Highland disintegration brings their summits down while recorded in the sediments of the Yanyuan piggyback basin, ponding of the resulting debris and infilling of adjacent on the hanging wall of the Jinhe-Jianhe thrust. Still preserved lowlands raises the high-elevation base level to even greater are 1–2 km of Paleogene red molasses unconformably height. As in the Qaidam and Qilian Shan, diffusive slope capped by a several hundred meters thick fault-bounded and short fluvial transport are the dominant modes of mass wedge of fluvial, coal-bearing Neogene clastics [Bureau of wasting. The characteristic ‘‘positive topography’’ remains Geology and Mineral Resources of Sichuan Province, 1991; (Figure 15a). The internal drainage smoothing process can Xu et al., 1992; Li et al., 2001]. continue for protracted periods, as long as no outbound river [40] That the high and relatively flat SE plateau surface penetrates into the high, isolated base level and tectonic did not continue south of the Yalong thrust belt is supported quiescence prevails. by several lines of evidence. Topographic parameters NW [38] Internal relief smoothing can later be disrupted by and SE of the thrust belt are markedly different. The slope- resumed tectonic deformation as is observed with normal elevation relationships in Figure 15b clearly show the faulting in southern Tibet, but more efficiently and com- gradual transition from the internally drained, to the shal- monly by the invasion of external drainage. Our study lowly incised, then highly dissected regions, as one suggests that such vigorous headward invasion provides approaches the Yalong belt from the northwest. Despite the most plausible link between the geomorphic signatures increasing erosion, remnants of the plateau surface are still

18 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018 recognizable from the 4000–4500 m HPBL (high plateau modern large river catchments into a plateau interior, from base level) trough in the slope-elevation functions, up to which they had long been excluded. 28°N southward. But there is no trace of them SE of the [44] This interpretation requires that the modern external Yalong belt (E5 and E6). The topography there lacks the drainages of the Jinsha, Lancang, and Nu Jiang did not contrast between a relatively flat highland and steeply reach into the plateau while its surface formed, a view incised river valleys seen northwest of the belt (Figure 14). radically different from previous interpretations surmising In fact, Clark et al.’s [2006] best examples of low-relief antecedence of these rivers [e.g., Hallet and Molnar, 2001]. surfaces are from this NW region. To the southeast, there is Studies by Horton et al. [2002] indeed suggest that the no quantitative evidence; such a surface is questionable at present headwaters of the Lancang Jiang south of Yushu best (e.g., Figures 14 and 15b). were established during the Neogene, well after the devel- [41] Although Tertiary sedimentary records in SE Tibet opment of the red-bed basins. In the Hoh Xil basin (90– are few, perhaps due to extensive fluvial erosion, the largest 94°E and 34–36°N), part of which is currently drained by Tertiary basins are localized along and southeast of the the Jinsha Jiang, widely distributed lacustrine carbonates of Longmenshan-Yalong-Yulong thrust belt (Figure 11). The early Miocene age (23.0–16 Ma) and sandstones includ- depositional environments were clearly different northwest ing organic rich shales [Yi et al., 2000; Liu and Wang, 2001; and southeast of the plateau edge. To the northwest, the Liu et al., 2001; Zhenhan et al., 2008] imply restricted Cretaceous-Eocene basins share some common features. internal drainage until at least 15 Ma. Roughly coeval lake They are narrow and elongated, intermontane basins where deposits were found in regions to the east (Z. Liu, personal internal drainage led to diverse paleocurrent flow directions, communication, 2004), but the evidence becomes patchier, apparently under depositional and climatic conditions sim- possibly because of later removal by erosion. This would be ilar to those currently acting in the interior of Tibet [Yi et al., broadly consistent with the mid- to late Miocene (12–7Ma) 2000; Horton et al., 2002; Liu and Wang, 2001; Liu et al., termination of planation long inferred by Armijo et al. 2001; Zhou et al., 2003]. In contrast, southeast of the [1986] and Shackleton and Chang [1988]. plateau edge (i.e., Lanping-Simao, Chuxiong, and Jinggu basins further to the southeast), sedimentary records of 5. Discussion similar ages indicate open, broad alluvial fan and fan delta, 5.1. Processes Responsible for the Low-Relief or large lake environments [Zhou et al., 2003], as befits a Topography of Tibet’s Interior foreland piedmont. Besides, the appearance of coarse con- glomerates, indicative of proximal deposition, occurs later [45] The interior flatness of the highest part of the Tibet in the southeast than in the northwest [Zhou et al., 2003]. plateau has often been taken as evidence of viscous flow in Such differences in depositional environments across a the lower crust, a mechanism thought necessary to smooth regional topographic rim are analogous to those now out irregularities in topography and crustal thickness [Bird, observed across the Qilian Shan, south of the Hexi corridor 1991; Fielding, 1996; Shen et al., 2001; McKenzie and foreland. Jackson, 2002]. The possibility of a combined effect of [42] Rocks exposed on either side of the Yalong-Yulong lower crustal flow and passive sediment infilling cannot be thrust belt imply different amounts of denudation. To the ruled out, as has been proposed to account for the subdued northwest and on the plateau, large granite plutons that likely relief in the Altiplano plateau [Lamb and Hoke, 1997]. crystallized below 10 km depth [Mattauer et al., 1992] are Importantly however, several recent tomographic studies exposed at the surface over hundreds of square kilometers, [Kind et al., 2002; Vergne et al., 2002; Wittlinger et al., whereas to the SE, the Jurassic-Cretaceous sedimentary 2004] confirm the existence of steps and relief of the Moho cover is extensive. Large early Tertiary basins are preserved, interface beneath the plateau. and there are only small leucrogranite plugs along major faults, [46] That surface processes play an important role in such as the Xianshuihe and Red River faults (Figure 11). smoothing the surface of the Tibetan plateau has been Therefore, denudation plateauward from the Yalong-Yulong dismissed by some [e.g., Fielding et al., 1994; Fielding, thrust belt must have been at least 10 km greater than 1996] in part because of low erosion rates associated with to the southeast, requiring a substantial amount of tectonic the aridity of the current climate. The rather small amounts structural relief across this belt, probably since at least the of denudation, as indicated by widely preserved outcrops of early Miocene (20 Ma [Lacassin et al., 1996]). In turn, the Triassic to middle Cretaceous sediments in most places 2 km high topographic step observed today across it is [e.g., Tapponnier et al., 1981, 1986; Kidd and Molnar, likely a residual of a formerly higher plateau rim. 1988; Dewey et al., 1988, Matte et al., 1996] have been [43] Thus, contrary to previous suggestions of a now taken to support this in inference [e.g., Fielding et al., uplifted surface that formed at low elevation [Clark et al., 1994]. However, Sobel et al. [2003] argued exactly the 2006], both geological and quantitative geomorphic evi- opposite in the Puna plateau, where there is a spatial dence suggest that the surface of the southeastern part of the correlation of low-relief landscape with dryness in the Tibet plateau NW of the Yalong-Yulong belt formed at high plateau interior and high relief with high precipitation near elevation. It was part of the elevated base level of the the windward plateau edges. Second, extrapolation of current southern plateau interior, beveled at 4000 m or more present-day erosion rates back in time is uncertain, because in the Eocene-Oligocene, at a time when it was largely erosion rates generally decrease as relief decreases for unconnected with external drainage systems. Dissection of tectonically inactive areas [e.g., Ahnert, 1970; Montgomery this Sichuan high plateau surface is now in an advanced and Brandon, 2002; Schaller et al., 2001]. Clearly, the stage, due to particularly aggressive headward retreat of the current dry climate in Tibet is not typical of that of the past. For instance, during the Miocene, a humid to semihumid

19 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018 climate prevailed on the forested plateau, suggesting higher assumption that the low-relief surface visible in SE Tibet precipitation rates than that at present [Bureau of Geology must have been at low elevation prior to regional uplift can and Mineral Resources of Xizang Autonomous Region, hardly be justified. 1993; Quade et al., 1989]. In any event, with shortening [49] Second, offshore integrated sedimentation flux and and relief buildup having occurred in the early Cenozoic geochemistry in southeast Asia [Metivier et al., 1999; Clift, [Liu et al., 2003; Wang et al., 2004], there would have been 2006; Clift and Sun, 2006; Clift et al., 2006] does not enough time to smooth the topography, even with slow support the existence of a protracted period of slow erosion erosion rates. The existence of the residual Tangula moun- before the mid-Miocene as would be in keeping with the tain range in central Tibet, with its Eocene granitic intru- hypothesis of a regional low-elevation, low-relief paleo- sives [Roger et al., 2000], and of the 5 km deep Hohxil landscape at that time. These authors find that the Paleogene foreland sedimentary basin [Wang et al., 2004] to corresponding clastic sediment flux first peaked in the early the north suggest that the now subdued topography of this to middle Miocene (24–11 Ma), and decreased consider- range once marked the former high-relief rim of the south- ably during the late Miocene (11–5Ma). Throughout the ern Tibet plateau in the early Tertiary [Roger et al., 2000; South China Sea, offshore of both the Red river and Zhu Tapponnier et al., 2001; Liu et al., 2003; Wang et al., 2004]. river deltas, peak sedimentation rates were inferred to date In the interior of Tibet, the thickness of Neogene deposits, back to a similar but slightly older period (29–16 Ma) [e.g., though quite variable, does reach up to 4 km [Bureau of Gong and Li, 1997; Zhang and Hao, 1997; Zhong et al., Geology and Mineral Resources of Xizang Autonomous 2004]. This Oligocene-Early Middle Miocene sedimenta- Region, 1993]. The existence of locally thick low-density tion flux peak, which appears to be a robust feature, likely sediment accumulations in the basins of Tibet is also indicates that the mean elevation in the headwaters regions supported by regional seismic waveforms, teleseismic re- of east and southeast Asian rivers increased before, rather ceiver function and teleseismic tomographic experiments than after the mid-Miocene, as inferred by Clark et al. [e.g., Zhu et al., 1995, 2006; Wittlinger et al., 1996, 2004]. [2006]. [50] Third, the diachronous ages of the proposed ‘‘low- 5.2. Style and Age of Surface Uplift in Southeastern elevation low-relief’’ (LELR) surface of Clark et al. [2006] Tibet undermine the inference that this surface is unique. Corre- [47] Clark and Royden [2000] interpreted the topography lation of patchy low-relief areas, up to hundreds of kilo- of SE Tibet as a response to regional long-wavelength tilt meters apart in highlands of different ages, into a single linked to thickening by lower crustal channel flow. The surface without discrimination between depositional and onset of such regional warping was inferred to be late erosional origin is highly uncertain, even more so where Miocene or younger, coeval with the onset of rapid incision such areas are on opposite sides of major tectonic divides of modern rivers [Clark et al., 2005; Kirby et al., 2002; such as the Yalong-Yulong or Ailao Shan ranges. Conversely, Schoenbohm et al., 2004]. This interpretation is under- it is hard to justify why the plateauward extent of the pinned by the inference that a low-relief preuplift surface, proposed preuplift low-relief surface would coincide with which formed at low elevation, on the order of a few the current outer limit of the internally drained part of Tibet, hundred meters only, did exist. Moreover that surface would which is in fact flatter and more continuous than the LELR have been continuous for thousands of kilometers from a remnants. Our results show that there is no sharp boundary former drainage divide somewhere west of eastern Tibet all between the ‘‘remnant surface’’ and the internally drained the way to the South China Sea [Clark et al., 2006]. In area, whether of tectonic or geomorphic origin. In Fielding actual fact, the evidence that ‘‘one’’ regionally extensive et al.’s [1994] map of slopes, the boundary between steep low-relief surface existed at low elevation across southeast- and gentle lies east of the current internal drainage divide ern Tibet and adjacent regions prior to the mid-Miocene is (Figure 1b). This suggests that the divide is ephemeral, its very tenuous. position being controlled by the degree of headward retreat [48] First, Clark et al.’s [2006] identification of such a of river catchments. relict low-relief landscape is based on the Davisian concept [51] In short, topographic data, morphometric measure- of ‘‘peneplain,’’ depicted to be the final result of landscape ments, and other evidence suggest that the high low-relief evolution as it is beveled ‘‘to sea level’’ during long tectonic surface in Sichuan, which is clearly the continuation of that quiescence [Davis, 1899]. As pointed out by Ritter [1988], in the plateau interior all the way to the Yalong-Yulong however, ‘‘because of this loose definition of peneplain, topographic step, is better explained as a surface that formed every geologist was free to establish his own meaning for at high elevation, rather than close to sea level prior to the the term,’’ and ‘‘almost every major proposed peneplain has Late Miocene. There is still no solid age constraint for been explained in different ways.’’ The revised definition by surface uplift in SE Tibet. Thermochronological data are Clark et al. [2006] and Widdowson [1997] of ‘‘a remnant of quite diverse [e.g., Arne et al., 1997; Lacassin et al., 1997; a regionally significant paleolandscape with a low-relief Xu and Kamp, 2000; Leloup et al., 2001; Kirby et al., 2002; topographic surface, of initially erosional and/or depositional Gilley et al., 2003; Wallis et al., 2003; Clark et al., 2005; origin, that is associated with a protracted period of ero- Lai et al., 2007; Godard, 2007; Richardson et al., 2008]. sion’’ is just as loosely defined, and leaves even more But a Late Miocene initiation of regional surface uplift is freedom in arguing for the existence of such a kind of clearly not in keeping with the integrated sedimentation surface. Several studies now show that low-relief surfaces budget representing the net flux of eroded material from can form at high elevations, provided a local base level can Southeast Asia into the surrounding marginal seas [Metivier be maintained at such elevations [e.g., Armijo et al., 1986; et al., 1999; Clift et al., 2006]. Late Cenozoic increased Meyer et al., 1998; Babault et al., 2005, 2007]. Thus, the river incision might just be one consequence of the estab-

20 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018 lishment of the monsoon climate system in East Asia since Tibetan orogen have been arbitrarily assumed to be uni- the Miocene [Sun and Wang, 2005]. formly low. [52] Studies of the structure and tectonics of the region [58] Three or four main competing mechanisms likely imply that the topography of SE Tibet attained its present controlled the drainage evolution at all times. The first, height or even greater height well before mid-Miocene. The headwater advection toward the mountain rim’s foreland evidence includes: [e.g., Lave and Avouac, 2001; Godard, 2007] played a [53] 1. Geochronological and thermochronological stud- major role in isolating the mountain hinterlands from ies north of the Red River that document a 36 Ma south- outbound erosion, thus allowing for the formation of the verging decollement exhumed in the core of the Yulong high internally drainage and high base levels of the plateau. Shan [Lacassin et al., 1996] along the topographic step The second, headward catchment retreat is responsible for singled out in Figure 12. Uplift and growth of the Yulong the present-day negative topography of SE Tibet and for the Shan anticline continued up to 17 Ma [Lacassin et al., relatively recent inward shift of the eastern internal drainage 1996]. divide from the flat highland boundary in Figure 1b. Today, [54] 2. Preserved sedimentary basins inside the SE pla- this mechanism leads to significant isostatic rock uplift in teau that record Paleogene shortening [e.g., Tapponnier et the border region between Sichuan and Yunnan. It is a al., 2001, and references therein; Liu et al., 2001; Horton et process that was probably kept at bay by the first one during al., 2002; Zhou et al., 2003]. most of the initial stages of growth of the high plateau [55] 3. Widespread shortening and transpression in the (35–40 Ma). The third mechanism is drainage capture Chuxiong and Lanping-Simao basins during the Paleogene which probably occurred at different times between the (Figure 11 [Pan et al., 1990; Tapponnier et al., 1986, 1990a; onset of collision and the present, particularly along the Leloup et al., 1995]). mid to lower catchments of the Jinsha, Lancang Jiang [e.g., [56] As long argued by the above authors, there can be no Brookfield, 1998; Clark et al., 2004] and Huang He, but doubt that the major deformation episodes involving large- also recently in the upper catchments of these and other scale crustal shortening in SE Tibet took place during the large rivers. Capture has probably been largely orchestrated Eocene-Oligocene and earliest Miocene. It is highly unlikely by headward catchment retreat (see van der Woerd et al. that the existence of south-directed shortening and of a steep [2001], for instance, for small-scale, present-day examples south-facing topographic break along the Yulong Yalung of this interaction). Finally, antecedence, across high moun- belt is a coincidence. Rather it has to reflect a causal link. tain rims of large trunk rivers that had developed broad Consistently, Moho depth underneath the Yulong belt has a upstream catchments is also observed. The most striking larger gradient than that to the northwest and southeast examples appear to be those of the Indus and Yarlung [Teng et al., 2002]. Finally, it must be noted that the Zangbo, across the western and eastern terminations of mechanism proposed for explaining broad surface uplift, the Himalayan range at Nanga Parbat and Namche Barwa, ‘‘lower crustal channel flow’’ [Clark and Royden, 2000], is respectively. We infer that both rivers had developed pale- not unambiguously supported by crustal anisotropy studies ocatchments along the south slope of the Gandese rim of the [Bhaskar et al., 2005, Lev et al., 2006; Sol et al., 2007]. Eocene plateau (stage 35–40 Ma, Figure 17) that resembled the present-day catchments of the Ganges and Punjab Indus 5.3. Evolution of Tibet Drainages and Plateau Growth south of the Himalayas, and were large enough for the trunk [57] The results of our morphometric analysis of the rivers to keep crossing the latter range as it grew to towering different areas of Tibet that can be discerned on Fielding heights since the Early Miocene (stage 15 Ma and present, et al.’s [1994] gradient map (Figure 1b) support a growth Figure 17). Note that this interpretation is different from model of the Tibet plateau similar to that idealized in previous inferences of river capture, particularly across the Figure 17. The basis for this model is the successive eastern syntaxis [e.g., Clark et al., 2004]. isolation, by the outward shifting of mountain rims of the [59] The recent work of Richardson et al. [2008], which orogen, of distinct internal drainage areas, hence local base implies unusual exhumation of the surface of the Sichuan levels disconnected from the global sea level. The model is basin, suggests that this basin may have once been inter- consistent with the geological and geophysical evidence nally drained (stage 15 Ma, Figure 17) and filled with discussed by Meyer et al. [1998], Metivier et al. [1998], and significant thicknesses of sediments of Oligo-Miocene Tapponnier et al. [2001, and references therein] and with the age, in much the same way as the Qaidam basin is today. results of recent paleoelevation or paleoenvironmental stud- Sediment infill removal, hence exhumation, would then ies [e.g., Spicer et al., 2003; Rowley and Currie, 2006; have occurred by drainage ‘‘flushing,’’ as the powerful DeCelles et al., 2007]. The primary emphasis of the Yangtze catchment retreated headward across the fold belts schematic, three-stage evolution depicted Figure 17 is on that bound the basin to the east and northeast, then into the age of high plateau topography in areas positioned in the southeastern Tibet (present, Figure 17). This would account present-day geographical reference frame. No attempt is for the inference of drainage ‘‘reversal’’ along the middle thus made to reconstruct the deformation of the region, reaches of the Jinsha-Yangtze River, discussed by Clark et whether on the large thrusts or strike-slip faults that have al. [2004] and Richardson et al. [2008], among others. As generated such deformation throughout the collision zone importantly, such a sequence of events might explain the since 55 Ma ago. For the same reason, the faults are not present-day morphology of the basin floor, most of which represented. As food for thought, we speculate on possible looks erosional rather than depositional, and the prevailing locations of paleodrainage networks. Clearly, the elevations evidence for drainage superimposition rather than anteced- depicted are only indicative and areas away from the ence throughout much of the basin.

21 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Figure 17. Schematic three-stage growth of Tibetan topography, based on stepwise tectonic model of Tapponnier et al. [2001]. Topographic evolution is represented in present-day geographic reference frame. No attempt is made at reconstructing deformation. Locations of advected or retreating drainage systems are inferred at each stage. (top) Older (35–40 Ma) internally drained Eocene plateau, bounded by Gangdese, Tanggula, and Yulong-Yalung ranges. (middle) Miocene (15 Ma) plateau, also internally drained, rimmed by Kunlun, Longmen Shan, and Himalayan ranges. At that time, possibly internally drained Sichuan basin may have been analogous to present-day Qaidam basin. (bottom) Present-day plateau has grown further toward NE, while headward catchment retreat has flushed out Sichuan sedimentary infill and is at work lowering average topography (small black arrows) and increasing relief in ancient, southeastern part of plateau.

[60] Research and fieldwork targeted at testing inferences But we suggest that, as previously proposed by Meyer et al. such as those depicted in Figure 17 has only begun. It [1998], Tapponnier et al. [2001], and Sobel et al. [2003], should be boosted by the development of reliable, direct among others, surface processes associated with long-last- paleoaltimeters [e.g., Ghosh et al., 2006; Blard et al., 2006]. ing internal drainage may well turn out to be the main

22 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018 mechanisms that lead to the ‘‘uplift’’ and areal growth of propagation of regressive erosion acting on an inherited other flat and horizontal plateaus around the world, be they topographic step. moderately high (e.g., Iran, Kalahari) or very high (e.g., Altiplano). [63] Acknowledgments. This work was funded by a grant from the Ministery of Education Nationale de Enseignement Superieur et de la Recherche and by the National Natural Science Foundation of China 6. Summary (NSFC 40625008) and Chinese Academy of Sciences Hundred Talent program. We benefited from help with coding from Paul Pinzuti and [61] Our morphometric analysis of the topography in Mathieu Daeron. We thank Rolando Armijo for an unpublished map of three typical regions of the Tibet plateau demonstrates the internal drained area in Tibet; Vincent Godard for comments on an earlier version; Mark Allen, Ed Sobel, Phillippe Davy, and Associate Editor Alex importance of drainage efficiency in controlling the plateau Densmore for reviews that improved the content and clarity of the morphology. Inefficient or internal drainage, short-distance manuscript; and Anne Claire Morillon for drafting Figure 17. The photo sediment transport and deposition can effectively smooth in Figure 5b was taken by Qinghai Zhang. out tectonically generated local relief, in areas where drainage headwaters are not connected to adjacent low- References lying forelands. In the internally drained plateau interior, Ahnert, F. (1970), Functional relationships between denudation, relief and uplift in large mid-latitute drainage basins, Am. J. Sci., 268, 243–263. local relief is further reduced, except in the vicinity of active Ahnert, F. (1984), Local relief and the height limits of mountain ranges, faults. Both regions display positive topography. High Am. J. Sci., 284, 1035–1055. elevations correspond to large local relief and steep slope, Alonso, R. A., T. E. Jordan, K. T. Tabbutt, and D. S. Vandervoort (1991), Giant evaporites of the Neogene central Andes, Geology, 19, 401–404, and low elevations to small relief and gentle slope. Because doi:10.1130/0091-7613(1991)019<0401:GEBOTN>2.3.CO;2. of the persistence of such positive topography, ‘‘planation’’ Armijo, R., P. Tapponnier, J. L. Mercier, and T.-L. Han (1986), Quaternary processes must have continued and are still going on in the extension in southern Tiber; field observations and tectonic implications, plateau interior, from a starting point that was likely similar J. Geophys. Res., 91, 13,803–13,872, doi:10.1029/JB091iB14p13803. Arne, D., B. Worley, C. Wilson, S. Chen, D. Foster, Z. Luo, S. Liu, and to that in the Qilian Shan. Clearly such planation can occur P. Dirks (1997), Differential exhumation in response to episodic thrusting at any elevation, as long as the corresponding regional along the eastern margin of the Tibetan plateau, Tectonophysics, 280, drainage base level is maintained at that altitude for sus- 239–256, doi:10.1016/S0040-1951(97)00040-1. Babault, J., J. van den Driessche, S. Bonnet, and S. Castelltort (2005), tained periods of time. Planation surfaces have no reason to Origin of the highly elevated Pyrenean peneplain, Tectonics, 24, be near-sea-level, low-elevation objects. TC2010, doi:10.1029/2004TC001697. [62] Rejuvenation of drainages, i.e., the capture of inter- Babault, J., et al. (2007), High-elevation of low-relief surfaces in mountain belts: Does it equate to post-orogenic surface uplift?, Terra Nova, 19, nal drainage systems by external drainage can disrupt the 272–277, doi:10.1111/j.1365-3121.2007.00746.x. planation process and produce a transitional topography by Bhaskar, A., B. Savage, and P. Silver (2005), Crustal structure beneath the dissecting an ‘‘old’’ remnant plateau surface, introducing southeastern Tibetan plateau and Yunnan province using teleseismic data, younger and steeper incision at that surface’s base level. In Eos Trans. AGU, 86(52), Fall. Meet. Suppl., Abstract T41A-1282. Bird, P. (1991), Lateral extrusion of lower crust from under high topogra- the southeastern plateau, where large rivers have carved phy in the isostatic limit, J. Geophys. Res., 96, 10,275–10,286, deep gorges, the regional topography is negative, with high- doi:10.1029/91JB00370. elevation terrain corresponding to smaller relief and gentler Blard, P. H., R. Pik, J. Lave, D. Bourles, P. G. Burnard, R. Yokochi, B. Marty, and F. Trusdell (2006), Cosmogenic He-3 production rates revisited from slope than at low elevation. This remnant SE Tibet plateau evidences of grain size dependent release of matrix-sited helium, Earth surface appears to terminate along a degraded plateau rim Planet. Sci. Lett., 247, 222–234, doi:10.1016/j.epsl.2006.05.012. that coincides with a major Oligo-Miocene thrust belt, the Brocklehurst, S. H., and K. X. Whipple (2004), Hypsometry of glaciated landscapes, Earth Surf. Processes Landforms, 29, 907–926, doi:10.1002/ Yalong-Yulong range. The development of the great drain- esp.1083. ages, Jinsha, Lancang, and Nu Jiang, appears to have been Brookfield, M. E. (1998), The evolution of the great river systems of south- due to headward retreat, probably controlled primarily by ern Asia during the Cenozoic India-Asia collision: Rivers draining south- the slowing or termination of transpressional tectonics wards, Geomorphology, 22, 285–312, doi:10.1016/S0169-555X(97) 00082-2. [Leloup et al., 1995] and correlative rim uplift, but possibly Brozovic´, N., D. W. Burbank, and A. J. Meigs (1997), Climatic limits on also in part by climate, as inferred in other areas [i.e., Sobel landscape development in the northwestern Himalaya, Science, 276, et al., 2003; Hilley and Strecker, 2005]. Renewed erosion of 571–574, doi:10.1126/science.276.5312.571. Burbank, D. (1992), Characteristic size of relief, Nature, 359, 483–484, the plateau inward from this rim could thus have lagged doi:10.1038/359483a0. considerably behind the construction of topography and Bureau of Geology and Mineral Resources of Sichuan Province (1991), uplift of the plateau, the preservation of high internal Regional geology of Sichuan province, pp. 730, Geological, Beijing. Bureau of Geology and Mineral Resources of Xizang Autonomous Region drainage base levels leading in effect to a decoupling be- (1993), Regional Geology of Xizang Autonomous Region, pp. 707, Geo- tween erosion and tectonics. Our analysis suggests that it logical, Beijing. should be difficult, in general, to infer deep-seated crustal or Bureau of Geology and Mineral Resources of Yunnan Province (1990), mantle processes from current surface topography without Regional geology of Yunnan province, pp. 728, Geological, Beijing. Chengdu Institute of Geology and Mineral Resources (2004), Geological first taking into account the time-integrated effect of upper map of the Qingha-Xizang (Tibet) plateau and adjacent areas, map, Geo- crustal deformation and surficial geomorphic evolution on logical, Beijing. topography. In fact, the local incision of rivers should not be Clark, M. K. (2003), Late Cenozoic uplift of southeastern Tibet, Ph. D. thesis, 226 pp., Mass. Inst. of Technol., Boston. readily used as a proxy of coeval uplift due to tectonic Clark, M. K., and L. H. Royden (2000), Topographic ooze: Building the shortening or other deep-seated process. In support of our eastern margin of Tibet by lower crustal flow, Geology, 28, 703–706, conclusions, and at a more local scale, recent thermochro- doi:10.1130/0091-7613(2000)28<703:TOBTEM>2.0.CO;2. Clark, M. K., L. M. Schoenbohm, L. H. Royden, K. X. Whipple, B. C. nogical studies in the Longmen Shan [Godard, 2007; Burchfiel, X. Zhang, W. Tang, E. Wang, and L. Chen (2004), Surface Richardson et al., 2008] suggest that the recent evolution uplift, tectonics, and erosion of eastern Tibet from large-scale drainage of this rim of Tibet may also be controlled by the westward patterns, Tectonics, 23, TC1006, doi:10.1029/2002TC001402.

23 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Clark, M. K., M. A. House, L. H. Royden, K. X. Whipple, B. C. Burchfiel, Jimenez-Munt, I., and J. P. Platt (2006), Influence of mantle dynamics on X. Zhang, W. Tang, E. Wang, and L. Chen (2005), Late Cenozoic uplift the topographic evolution of the Tibetan plateau: Results from numerical of southeastern Tibet, Geology, 33, 525–528, doi:10.1130/G21265.1. modeling, Tectonics, 25, TC6002, doi:10.1029/2006TC001963. Clark, M. K., L. H. Royden, K. X. Whipple, B. C. Burchfiel, X. Zhang, and Jordan, T. E., and R. N. Alonso (1987), Cenozioc stratigraphy and basin W. Tang (2006), Use of a regional, relict landscape to measure vertical tectonics of the Andes Mountains, 20–28 south latitude, AAPG Bull., 71, deformation of the eastern Tibetan plateau, J. Geophys. Res., 111, 49–64. F03002, doi:10.1029/2005JF000294. Kapp, P., M. Taylor, D. Stockli, and L. Ding (2008), Development of active Clift, P. D. (2006), Controls on the erosion of Cenozoic Asia and the flux of low-angle normal fault systems during orogenic collapse: Insight from clastic sediment to the ocean, Earth Planet. Sci. Lett., 241, 571–580, Tibet, Geology, 36, 7–10, doi:10.1130/G24054A.1. doi:10.1016/j.epsl.2005.11.028. Kidd, W. S. F., and P. Molnar (1988), Quaternary and active faulting ob- Clift, P. D., and Z. Sun (2006), The sedimentary and tectonic evolution of served on the 1985 Academia Sinica-Royal Society Geotraverse of Tibet, the Yinggehai-Song Hong basin and the southern Hainan margin, South Philos. Trans. R. Soc. London Ser. A, 327, 337–363, doi:10.1098/ China Sea: Implications for Tibetan uplift and monsoon intensification, rsta.1988.0133. J. Geophys. Res., 111, B06405, doi:10.1029/2005JB004048. Kind, R., et al. (2002), Seismic images of crust and upper mantle beneath Clift, P. D., J. Blusztajn, and N. A. Duc (2006), Large-scale drainage Tibet: Evidence for Eurasian Plate subduction, Science, 298(5596), capture and surface uplift in eastern Tibet-SW China before 24 Ma in- 1219–1221, doi:10.1126/science.1078115. ferred from sediments of the Hanoi basin, Vietnam, Geophys. Res. Lett., Kirby, E., P. W. Reiners, M. A. Krol, K. X. Whipple, K. V. Hodges, K. A. 33, L19403, doi:10.1029/2006GL027772. Farley, W. Tang, and Z. Chen (2002), Late Cenozoic evolution of the Davis, W. M. (1899), The geographical cycle, Geogr. J., 14, 481–504, eastern margin of the Tibetan plateau: Inferences from 40Ar/39Ar and (U/ doi:10.2307/1774538. Th)/He thermochronology, Tectonics, 21(1), 1001, doi:10.1029/ DeCelles, P. G., J. Quade, P. Kapp, M. Fan, D. L. Dettman, and L. Ding 2000TC001246. (2007), High and dry in central Tibet during the Late Oligocene, Earth Koons, P. O. (1989), The topographic evolution of collisional moutain Planet. Sci. Lett., 253, 389–401, doi:10.1016/j.epsl.2006.11.001. belts: A numerical look at the Southern Alps, New Zealand, Am. J. Deng, Q., P. Zhang, and Y. Ran (2003), Basic characteristics of active Sci., 289, 1041–1069. tectonics of China, Sci. China Ser. D, 46, 356–372. Lacassin, R., U. Schaerer, P. H. Leloup, N. Arnaud, P. Tapponnier, X. Liu, Derbyshire, E., Y. F. Shi, J. J. Li, B. X. Zheng, S. Li, and J. Wang (1991), and L. Zhang (1996), Tertiary deformation and metamorphism SE of Quaternary glaciation of Tibet: The geological evidence, Quat. Sci. Rev., Tibet: The folded Tiger-leap decollement of NW Yunnan, China, Tec- 10, 485–510, doi:10.1016/0277-3791(91)90042-S. tonics, 15, 605–622, doi:10.1029/95TC03749. Dewey, J. F., R. M. Shackleton, C. Chang, Y. Sun, and J. Yin (1988), The Lacassin, R., H. Maluski, P. H. Leloup, P. Tapponnier, C. Hinthong, tectonic evolution of the Tibetan plateau, Philos. Trans. R. Soc. London K. Siribhakdi, S. Chuaviroj, and A. Charoenravat (1997), Tertiary dia- Ser. A, 327, 379–413, doi:10.1098/rsta.1988.0135. chronic extrusion and deformation of western Indochina; structural and England, P. C., and G. A. Houseman (1986), Finite strain calculations of 40Ar/39Ar evidence from NW Thailand, J. Geophys. Res., 102, 10,013– continental deformation: 2. Comparison with the India-Asia collision 10,037, doi:10.1029/96JB03831. zone, J. Geophys. Res., 91, 3664–3676, doi:10.1029/JB091iB03p03664. Lacassin, R., A. Replumaz, and P. H. Leloup (1998), Hairpin river loops England, P. C., and G. A. Houseman (1989), Extension during continental and slip-sense inversion on southeast Asian strike-slip faults, Geology, convergence, with application to the Tibetan plateau, J. Geophys. Res., 26, 703–706, doi:10.1130/0091-7613(1998)026<0703:HRLASS>2.3. 94, 17,561–17,579. CO;2. Fielding, E. (1996), Tibet uplift and erosion, Tectonophysics, 260, 55–84, Lai, Q. Z., L. Ding, H. W. Wang, Y. H. Yue, and F. L. Cai (2007), Con- doi:10.1016/0040-1951(96)00076-5. straining the stepwise migration of the eastern Tibetan plateau margin by Fielding, E., B. Isacks, M. Barazangi, and C. Duncan (1994), How fl at is apatite fission track thermochronology, Sci. China Ser. D, 50, 172–183. Tibet?, Geology, 22, 163–167, doi:10.1130/0091-7613(1994)022<0163: Lamb, S., and L. Hoke (1997), Origin of the high plateau in the central HFIT>2.3.CO;2. Andes, Bolivia, South America, Tectonics, 16, 623–649, doi:10.1029/ Garcia-Castellanos, D. (2007), The role of climate during high plateau 97TC00495. formations Insights from numerical experiments, Earth Planet. Sci. Lett., Lave, J., and J.-P. Avouac (2001), Fluvial incision and tectonic uplift across 257, 372–390, doi:10.1016/j.epsl.2007.02.039. the Himalayas of central Nepal, J. Geophys. Res., 106, 26,561–26,591, Ghosh, P., C. N. Garzione, and J. M. Eiler (2006), Rapid uplift of the doi:10.1029/2001JB000359. Altiplano revealed through C-13-O-18 bonds in paleosol carbonates, Leloup, P. H., R. Lacassin, P. Tapponnier, U. Schaerer, D. Zhong, X. Liu, Science, 311, 511–515, doi:10.1126/science.1119365. L. Zhang, S. Ji, and P. T. Trinh (1995), The Ailao Shan-Red River shear Gilchrist, A. R., M. A. Summerfield, and H. A. P. Cockburn (1994), Land- zone (Yunnan, China), Tertiary transform boundary of Indochina, Tecto- scape dissection, isostatic uplift, and the morphologic development of nophysics, 251, 3–10, doi:10.1016/0040-1951(95)00070-4. orogens, Geology, 22, 963–966, doi:10.1130/0091-7613(1994)022< Leloup, P. H., N. Arnaud, R. Lacassin, J. R. Kienast, T. M. Harrison, T. T. 0963:LDIUAT>2.3.CO;2. Phan Trong, A. Replumaz, and P. Tapponnier (2001), New constraints on Gilley, L. D., T. M. Harrison, P. H. Leloup, F. J. Ryerson, O. M. Lovera, and the structure, thermochronology, and timing of the Ailao Shan-Red River J.-H. Wang (2003), Direct dating of left-lateral deformation along the shear zone, SE Asia, J. Geophys. Res., 106, 6683–6732, doi:10.1029/ RedRiver shear zone, China and Vietnam, J. Geophys. Res., 108(B2), 2000JB900322. 2127, doi:10.1029/2001JB001726. Lev, E., M. D. Long, and R. D. van der Hilst (2006), Seismic anisotropy Godard, V. (2007), Couplage e´rosion-tectonique en contexte de convergence from shear-wave splitting in Eastern Tibet reveals changes in lithospheric intracontinentale E´ tude compare´e de la chaıˆne Himalayenne et des Long- deformation, Earth Planet. Sci. Lett., 251, 293–304, doi:10.1016/ men Shan (est-Tibet), Ph. D. thesis, 247 pp., Univ. Paris XI, Paris. j.epsl.2006.09.018. Gong, Z., and S. Li (1997), Basin Analysis and Accumulation of Oil and Li, J. J., S. Zhou, and B. Pan (1991), The problems of Quaternary glaciation Gas of the Northern Margin of the South China Sea (in Chinese), 510 pp., in the eastern part of Qinghai-Xizang plateau, Quat. Sci., 3, 193–203. Science, Beijing. Li, Y., A. L. Densmore, P. A. Allen, and Q. Xu (2001), Sedimentary Hallet, B., and P. Molnar (2001), Distorted drainage basins as markers of responses to thrusting and strike-slipping of Longmen Shan along the crustal strain east of the Himalaya, J. Geophys. Res., 106, 13,697– eastern margin of Tibetan plateau and implications for accretion of Cim- 13,709, doi:10.1029/2000JB900335. merian continents and India/Eurasia collision, Dizhi Kexue, 10, 223–243. Hilley, G. E., and M. R. Strecker (2005), Processes of oscillatory basin Liu, Z., and C. Wang (2001), Facies analysis and depositional systems of filling and excavation in a tectonically active orogen: Quebrada del Toro Cenozoic sediments in the Hoh Xil basin, northern Tibet, Sediment. basin, NW Argentina, Geol. Soc. Am. Bull., 117, 887–901, doi:10.1130/ Geol., 140, 251–270, doi:10.1016/S0037-0738(00)00188-3. B25602.1. Liu, Z., C. Wang, and H. Yi (2001), Evolution and mass accumulation of Horton, B. K., A. Yin, M. S. Spurlin, J. Zhou, and J. Wang (2002), Paleo- the Cenozoic Hoh Xil basin, northern Tibet, J. Sediment. Res., 71, 971– cene-Eocene syncontractional sedimentation in narrow, lacustrine-domi- 984, doi:10.1306/030901710971. nated basins of east-central Tibet, Geol. Soc. Am. Bull., 114, 771–786, Liu, Z., X. Zhao, C. Wang, S. Liu, and H. Yi (2003), Magnetostratigraphy doi:10.1130/0016-7606(2002)114<0771:PESSIN>2.0.CO;2. of Tertiary sediments from the Hoh Xil basin: Implications for the Cen- Howard, A. D. (1997), Badland morphology and evolution: Interpretation ozoic tectonic history of the Tibetan plateau, Geophys. J. Int., 154, 233– using a simulation model, Earth Surf. Processes Landforms, 22,211– 252, doi:10.1046/j.1365-246X.2003.01986.x. 227, doi:10.1002/(SICI)1096-9837(199703)22:3<211::AID-ESP749> Luo, L., and J. Yang (1963), An approach of geomorphological formation 3.0.CO;2-E. in the region of West Suchan and North Yunnan, Geogr. Bull., 5, 34–46. Isacks, B. L. (1988), Uplift of the central Andean plateau and bending of Masek, J. G., B. L. Isacks, T. L. Gubbels, and E. J. Fielding (1994), Erosion the Bolivian orocline, J. Geophys. Res., 93, 3211–3231, doi:10.1029/ and tectonics at the margins of continental plateaus, J. Geophys. Res., 99, JB093iB04p03211. 13,941–13,956, doi:10.1029/94JB00461.

24 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

Mattauer, M., J. Malavieille, S. Calassou, J. Lancelot, F. Roger, Z. Hao, Z.-Q. Shen, F., L. H. Royden, and B. C. Burchfiel (2001), Large-scale crustal Xu, and L. Hou (1992), La chaine triasique de Songpan Garze (ouest deformation of the Tibetan plateau, J. Geophys. Res., 106, 6793–6816, Sechuan et est Tibet): Une chaine de plissement-decollement sur marge doi:10.1029/2000JB900389. passive, C. R. Acad. Sci. Paris, 314, 619–626. Shen, Z.-K., J. Lu¨, M. Wang, and R. Bu¨rgmann (2005), Contemporary Matte, P., P. Tapponnier, N. Arnaud, L. Bourjot, J. P. Avouac, P. Vidal, crustal deformation around the southeast borderland of the Tibetan pla- L. Qing, Y. S. Pan, and W. Yi (1996), Tectonics of western Tibet, between teau, J. Geophys. Res., 110, B11409, doi:10.1029/2004JB003421. the Tarim and the Indus, Earth Planet. Sci. Lett., 142, 311–330, Sobel, E. R., G. E. Hilley, and M. R. Strecker (2003), Formation of intern- doi:10.1016/0012-821X(96)00086-6. ally drained contractional basins by aridity-limited bedrock incision, McKenzie, D., and J. Jackson (2002), Conditions for flow in the continental J. Geophys. Res., 108(B7), 2344, doi:10.1029/2002JB001883. crust, Tectonics, 21(6), 1055, doi:10.1029/2002TC001394. Sol, S., et al. (2007), Geodynamics of the southeastern Tibetan plateau from Metivier, F., Y. Gaudemer, P. Tapponnier, and B. Meyer (1998), North- seismic anisotropy and geodesy, Geology, 35(6), 563–566, doi:10.1130/ eastward growth of the Tibet plateau deduced from balanced reconstruc- G23408A.1. tion of two depositional areas: The Qaidam and Hexi Corridor basins, Spicer, R. A., et al. (2003), Constant elevation of southern Tibet over the China, Tectonics, 17, 823–842, doi:10.1029/98TC02764. past 15 million years, Nature, 421, 622–624, doi:10.1038/nature01356. Metivier, F., Y. Gaudemer, P. Tapponnier, and M. Klein (1999), Mass Sun, X., and P. Wang (2005), How old is the Asian monsoon system?— accumulation rates in Asia during the Cenozoic, Geophys. J. Int., 137, Palaeobotanical records from China, Palaeogeogr. Palaeoclimatol. 280–318, doi:10.1046/j.1365-246X.1999.00802.x. Palaeoecol., 222, 181–222, doi:10.1016/j.palaeo.2005.03.005. Meyer, B., et al. (1998), Crustal thickening in Gansu-Qinghai, lithospheric Tapponnier, P., et al. (1981), The Tibetan side of the India-Eurasia collision, mantle subduction, and oblique, strike-slip controlled growth of the Tibet Nature, 294, 405–410, doi:10.1038/294405a0. plateau, Geophys. J. Int., 135, 1 – 47, doi:10.1046/j.1365-246X.1998. Tapponnier, P., G. Peltzer, and R. Armijo (1986), On the mechanics of the 00567.x. collision between India and Asia, Geol. Soc. Spec. Publ., 19, 115–157. Molnar, P., and P. C. England (1990), Late Cenozoic uplift of mountain Tapponnier, P., et al. (1990a), The Ailao Shan/Red River metamorphic belt: ranges and global climate change: Chicken or egg?, Nature, 346, 29–34, Tertiary left-lateral shear between Indochina and South China, Nature, doi:10.1038/346029a0. 343, 431–437, doi:10.1038/343431a0. Montgomery, D. R. (1994), Valley incision and the uplift of mountain Tapponnier, P., et al. (1990b), Active thrusting and folding in the Qilian peaks, J. Geophys. Res., 99, 13,913–13,921, doi:10.1029/94JB00122. Shan, and decoupling between upper crust and mantle in northeastern Montgomery, D. R. (2001), Slope distributions, threshold hillslopes, and Tibet, Earth Planet. Sci. Lett., 97, 382–403, doi:10.1016/0012- steady-state topography, Am.J.Sci., 301, 432–454, doi:10.2475/ 821X(90)90053-Z. ajs.301.4-5.432. Tapponnier, P., et al. (2001), Oblique stepwise rise and growth of the Tibet Montgomery, D. R., and M. T. Brandon (2002), Topographic controls on plateau, Science, 294, 1671–1677, doi:10.1126/science.105978. erosion rates in tectonically active mountain ranges, Earth Planet. Sci. Taylor, M., A. Yin, F. J. Ryerson, P. Kapp, and L. Ding (2003), Conjugate Lett., 201, 481–489, doi:10.1016/S0012-821X(02)00725-2. strike-slip faulting along the Bangong-Nujiang suture zone accommo- Montgomery, D. R., G. Balco, and S. D. Willett (2001), Climate, tectonics dates coeval east-west extension and north-south shortening in the inter- and the morphology of the Andes, Geology, 29, 579–582, doi:10.1130/ ior of the Tibetan plateau, Tectonics, 22(4), 1044, doi:10.1029/ 0091-7613(2001)029<0579:CTATMO>2.0.CO;2. 2002TC001361. Ohmori, H. (1993), Changes in the hypsometric curve through mountain Teng, J., R. Zeng, Y. Yan, and H. Zhang (2002), Depth distribution of Moho building resulting from concurrent tectonics and denudation, Geomor- and tectonic framework in eastern Asian continent and its adjacent ocean phology, 8, 263–277. areas, Sci. China Ser. D, 32(2), 89–100. Pan, G. T., P. S. Wang, Y. R. Xu, S. P. Jiao, and T. X. Xiang (1990), Vandervoort, D. S., T. E. Jordan, P. K. Zeitler, and R. N. Alonso (1995), Cenozoic Tectonic Evolution of Qinghai-Xizang Plateau, 190 pp., Geo- Chronology of internal drainage development and uplift, southern Puna logical, Beijing. plateau, Argentine central Andes, Geology, 23, 145–148, doi:10.1130/ Pares, J. M., R. VanderVoo, W. R. Downs, M. Yan, and X. Fang (2003), 0091-7613(1995)023<0145:COIDDA>2.3.CO;2. Northeastward growth and uplift of the Tibetan plateau: Magnetostrati- van der Woerd, J., X. Xu, H. Li, P. Tapponnier, B. Meyer, F. J. Ryerson, A.-S. graphic insights from the Guide basin, J. Geophys. Res., 108(B1), 2017, Meriaux, and Z. Xu (2001), Rapid active thrusting along the northwestern doi:10.1029/2001JB001349. range front of the Tanghe Nan Shan (western Gansu, China), J. Geophys. Quade, J., T. E. Cerling, and J. R. Bowman (1989), Development of Asian Res., 106, 30,475–30,504, doi:10.1029/2001JB000583. monsoon revealed by marked ecological shift during the latest Miocene Vergne, J., G. Wittlinger, Q. A. Hui, P. Tapponnier, G. Poupinet, J. Mei, in northern Pakistan, Nature, 342, 163–166, doi:10.1038/342163a0. G. Herquel, and A. Paul (2002), Seismic evidence for stepwise thickening Richardson, N., A. Densmore, D. Seward, A. C. Fowler, M. Wipf, M. Ellis, of the crust across the NE Tibetan plateau, Earth Planet. Sci. Lett., 203, L. Yong, and Z. Yi (2008), Extraordinary denudation in the Sichuan 25–33, doi:10.1016/S0012-821X(02)00853-1. basin: Insights from low-temperature thermochronology adjacent to the Wallis, S., T. Tsujimori, M. Aoya, T. Kawakami, K. Terada, K. Suzuki, and eastern margin of the Tibetan plateau, J. Geophys. Res., 113, B04409, H. Hyodo (2003), Cenozoic and Mesozoic metamorphism in the Long- doi:10.1029/2006JB004739. menshan Orogen: Implications for geodynamic models of eastern Tibet, Riller, U., and O. Oncken (2003), Growth of the central Andean plateau by Geology, 31, 745–748, doi:10.1130/G19562.1. tectonic segmentation is controlled by the gradient in crustal shortening, Wang, C., L. Zhu, and Z. Liu (2004), Tectonic and sedimentary evolution of J. Geol., 111, 367–384, doi:10.1086/373974. basins in the north of Qinghai-Tibet plateau and northward growing Ritter, D. F. (1988), Landscape analysis and the search for geomorphic process of the plateau (in Chinese with English abstract), Adv. Earth unity, Geol. Soc. Am. Bull., 100, 160–171, doi:10.1130/0016-7606 Sci., 19, 373–381. (1988)100<0160:LAATSF>2.3.CO;2. Wang, E., B. C. Burchfiel, L. H. Royden, L. Chen, J. Chen, W. Li, and Roger, F., P. Tapponnier, N. Arnaud, U. Schaerer, M. Brunel, Z. Q. Xu, and Z. Chen (1998), Late Cenozoic Xianshuihe-Xiaojiang, Red River and J. S. Yang (2000), An Eocene magmatic belt across central Tibet: Mantle Dali fault systems of southwestern Sichuan and central Yunnan, China, subduction triggered by the Indian collision?, Terra Nova, 12, 102–108, special paper, pp. 108, Geol. Soc. of Am., Boulder, Colo. doi:10.1046/j.1365-3121.2000.00282.x. Weissel, J. K., L. F. Pratson, and A. Malinverno (1994), The length-scaling Rowley, D. B., and B. S. Currie (2006), Palaeo-altimetry of the late Eocene properties of topography, J. Geophys. Res., 99, 13,997–14,012, to Miocene Lunpola basin, central Tibet, Nature, 439, 677–681, doi:10.1029/94JB00130. doi:10.1038/nature04506. Whipple, K. X., E. Kirby, and S. H. Brocklehurst (1999), Geomorphic Royden, L., B. C. Burchfiel, R. W. King, E. Wang, Z. L. Chen, F. Shen, and limits to climate-induced increases in topographic relief, Nature, 401, Y. P. Liu (1997), Surface deformation and lower crustal flow in eastern 39–43, doi:10.1038/43375. Tibet, Science, 276, 788–790, doi:10.1126/science.276.5313.788. Widdowson, M. (1997), The geomorphological and geological importance Schaller, M., F. von Blanckenburg, N. Hovius, and P. W. Kubik (2001), of paleosurfaces, in Paleosurfaces: Recognition, Reconstruction and Large-scale erosion rates from in situ-produced cosmogenic nuclides in Paleoenvironmental Interpretation, Spec.Publ.Geol.Soc.Am.,vol.120, European river sediments, Earth Planet. Sci. Lett., 188, 441–458, pp. 1–12, Geol. Soc. of Am., Boulder, Colo. doi:10.1016/S0012-821X(01)00320-X. Wittlinger, G., F. Masson, G. Poupinet, P. Tapponnier, J. Mei, G. Herquel, Schoenbohm, L. M., K. X. Whipple, B. C. Burchfiel, and L. Chen (2004), J. Guilbert, U. Achauer, X. Guanqi, and S. Danian (1996), Seismic to- Geomorphic constraints on surface uplift, exhumation, and plateau mography of northern Tibet and Kunlun: Evidence for crustal blocks and growth in the Red River region, Yunnan province, China, Geol. Soc. mantle velocity contrasts, Earth Planet. Sci. Lett., 139, 263–279, Am. Bull., 116, 895–909, doi:10.1130/B25364.1. doi:10.1016/0012-821X(95)00235-5. Shackleton, R. M., and C. Chang (1988), Cenozoic uplift and deformation Wittlinger, G., et al. (2004), Teleseismic imaging of subducting lithosphere of the Tibetan plateau: The geomorphological evidence, Philos. Trans. R. and Moho offsets beneath western Tibet, Earth Planet. Sci. Lett., 221, Soc. London Ser. A, 327, 365–377, doi:10.1098/rsta.1988.0134. 117–130, doi:10.1016/S0012-821X(03)00723-4.

25 of 26 F04018 LIU-ZENG ET AL.: LANDSCAPE DIFFERENCES IN TIBET F04018

World Meteorological Organization (1981), Climatic atlas of Asia, 28 pp., Zhong, Z., L. Wang, B. Xia, W. Dong, Z. Sun, and Y. Shi (2004), The Geneva, Switzerland. dynamics of Yinggehai basin formation and its tectonic significance, Xiang, H., X. Xu, S. Guo, H. Li, and G. Yu (2002), Sinistral thrusting along Dizhi Xuebao, 78,1–7. the Lijiang-Xiaojinhe fault since Quaternay and its geologic-tectonic sig- Zhou, J. Y., J. H. Wang, Y. An, B. K. Horton, and M. S. Spurlin (2003), nificance (in Chinese with English abstract), Seismol. Geol., 24, 188– Sedimentology and tectonic significance of Paleogene coarse clastic 198. rocks in eastern Tibet, Dizhi Xuebao, 77, 262–271. Xu, G., and P. J. J. Kamp (2000), Tectonics and denudation adjacent to the Zhu, L., T. J. Owens, and G. E. Randall (1995), Lateral variation in crustal Xianshuihe fault, eastern Tibetan plateau: Constraints from fission track structure of the northern Tibetan plateau inferred from teleseismic recei- thermochronology, J. Geophys. Res., 105, 19,231–19,251, doi:10.1029/ ver functions, Bull. Seismol. Soc. Am., 85, 1531–1540. 2000JB900159. Zhu, L., Y. Tan, D. V. Helmberger, and C. K. Saikia (2006), Calibration of Xu, X. W., X. Z. Wen, R. Z. Zheng, W. T. Ma, F. M. Song, and G. H. Yu the Tibetan plateau using regional seismic waveforms, Pure Appl. Geo- (2003), Pattern of latest tectonic motion and its dynamics for active phys., 163, 1193–1213, doi:10.1007/s00024-006-0073-7. blocks in Sichuan-Yunnan region, China, Sci. China Ser. D, 46, 210– 226. Xu, Z. Q., L. W. Hou, and Z. X. Wang (1992), Orogenic Processes of the L. Ding and J. Liu-Zeng, Institute of Tibetan Plateau Research, Chinese Songpan-Garze Orogenic Belt of China, 190 pp., Geological, Beijing. Academy of Sciences, 18 Shuang Qing Road, P.O. Box 2871, Beijing Yi, H., S. Liu, Z. Liu, C. Wang, and S. Wang (2000), Sedimentary record of 100085, China. ([email protected]) the planation surface in the Hoh Xil region of the northern Tibet plateau, Y. Gaudemer and P. Tapponnier, Laboratoire de Tectonique, Institut de Dizhi Xuebao, 74, 827–835. Physique du Globe, 4 Place Jussieu, F-75005 Paris, France. Zhang, Q., and F. Hao (1997), Evolution history of Yinggehai-Qiongdong- nan basin and oil storage, Sci. China Ser.D, 27, 149–154. Zhenhan, W., P. Barosh, Z. Wu, D. Hu, Z. Xun, and P. Ye (2008), Vast early Miocene lakes of the central Tibetan plateau, Geol. Soc. Am. Bull., 120, 1326–1337, doi:10.1130/B26043.1.

26 of 26