Earthquake distribution in southern Tibet and its tectonic

implications

Xiaofeng Liang,1 Shiyong Zhou,1 Y. John Chen,1 Ge Jin,1 Liang Xiao,1 Pingjiang Liu,1

Yuanyuan Fu,1 Youcai Tang,1 Xiaoting Lou,1,2 and Jieyuan Ning1

Received 6 April 2008; revised 2 July 2008; accepted 31 July 2008; published 23 December 2008.

[1] A temporary 37-station seismic array was operated in southern Tibet from June 2004to August 2005 close to Xigaze along a traverse from Tangra Yum Co rift in the west toYadong-Gulu rift in the east. This lateral array of the international Hi-CLIMB projectrecorded a total of 885 local earthquakes during the 14-month deployment. Hypocenterlocations of these events were obtained using HYPOINVERSE2000, and about half ofthem were relocated by the joint hypocenter determination method. Lateral variations incrustal seismic P and S wave velocities beneath the array were obtained simultaneouslyas part of the station corrections during the relocation procedure. While earthquakeswere scattered in the region, more than 250 earthquakes were clustered within a small areaat about 50 km north of the Indus-Yalu Suture and west of the Pumqu-Xianza rift.Crustal and uppermost mantle earthquakes are concentrated beneath the Himalayan crest.Projection of all earthquakes to a north–south profile shows that the subducting Indianplate passes the Indus-Yalu Suture below the Lhasa Terrane. While earthquakes areconcentrated in the upper crust and upper mantle beneath the Himalayas, seismicity ismostly restricted to the upper crust to the north, which is consistent with a ductilemiddle/lower crust beneath southern Tibet. About 20 deep crustal and uppermost mantleevents (at depths of 50–70 km) were observed to be associated with the subductingIndian lithosphere. Our relocation results of these local earthquakes in southern Tibetsupport the popular ‘‘jelly-sandwich’’ model for the rheology of continental lithosphere inwhich the strong seismogenic layers of upper curst and uppermost mantle are separated bya weak lower crust.

Citation: Liang, X., S. Zhou, Y. J. Chen, G. Jin, L. Xiao, P. Liu, Y. Fu, Y. Tang, X. Lou, and J. Ning (2008), Earthquake distribution

in southern Tibet and its tectonic implications, J. Geophys. Res., 113, B12409, doi:10.1029/2007JB005101.

1. Introduction

[2] The Himalayan Range and the Tibetan Plateau, builtby the continental collision between the Indian and theEurasian continents, are the two of the most active tectonicregions in the world. While significant progress has beenobtained by previous international studies many key ques-tions about this mountain-building process have yet to beanswered, such as the age of the onset of the collision withestimates ranging from �65 Ma to �43 Ma [Yin andHarrison, 2000; Yin, 2006], and what is the fate of theIndian lithosphere beneath the Tibet.[3] One important aspect for understanding the mountain-

building process is the rheology structure of the Tibetanlithosphere. The conventional ‘‘sandwich model’’ proposedoriginally by Chen and Molnar [1983] divides the conti-nental lithosphere into two seismogenic layers: the upper

crust and the uppermost mantle, which are separated by aductile/weaker lower crust. This concept has been used toconstruct numerical models of Tibetan Plateau [e.g., Zhaoand Morgan, 1985, 1987] and to interpret seismic resultsfrom the INDEPTH projects at southern Tibet. For examplethe ‘‘bright spots’’ in seismic reflection profiles and a lowelectromagnetic resistance layer in the middle crust wereused as evidence for the existence of a ductile lower crust inTibet [Brown et al., 1996; Chen et al., 1996]. Recently, thissimple conceptual model was challenged by Jackson [2002]and Jackson et al. [2004], who argued that there is probablyjust one seismogenic layer (within the crust) for the conti-nental lithosphere. The key issue of this debate is whetherthere exists a ductile lower crust. Precise determination ofearthquake distribution in depth could provide importantevidence to discriminate between these two models.[4] Seismicity across the Himalaya Range has been

investigated by using several local permanent seismic net-works in Nepal [Pandey et al., 1999], in Sikkim [De andKayal, 2003], and in Bhutan [Drukpa et al., 2006]. Hypo-center depths of central and western Nepal are wellconstrained at depths shallower than 30 km and in particular,intense and continuous seismic activity along the Pumqu-

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 113, B12409, doi:10.1029/2007JB005101, 2008ClickHere

for

FullArticle

1Institute of Theoretical and Applied Geophysics, School of Earth andSpace Sciences, Peking University, Beijing, China.

2Now at Department of Earth and Planetary Sciences, NorthwesternUniversity, Evanston, Illinois, USA.

Copyright 2008 by the American Geophysical Union.0148-0227/08/2007JB005101$09.00

B12409 1 of 11

Xianza rift was reported by Pandey et al. [1999], whointerpreted that the deep events in the southern Tibet couldbe related to compressional bending stresses in the upper-most mantle of the Indian lithosphere, induced by thelithospheric flexure at the front of the mountain range. Incontrast, earthquakes in Sikkim are continuous from thesurface to the lower crust at 45 km depth [De and Kayal,2003] and events at midcrustal and deep crustal depths werealso reported by Drukpa et al. [2006] in Bhutan.[5] Several temporary seismic arrays were operated from

southern Himalaya in Nepal to Lhasa Terrane of southernTibet in the past two decades. As a Sino-U.S. collaborationa dozen of portable broadband seismic stations weredeployed along the Lhasa-Golmud highway in the early1990s [Owens et al., 1993]. Since then several major fieldprograms have been conducted in southern Tibet such as theINDEPTH I, II, and III [Zhao et al., 1993; Nelson et al.,1996; Huang et al., 2000], the Sino-French projects[Herquel et al., 1995; Hirn et al., 1995; Wittlinger et al.,2004], the HIMNT project [de la Torre and Sheehan, 2005],and the Namcha Barwa Tibet project [Sol et al., 2007] toimage the Indian lithosphere beneath the Tibet. Significantunderstanding about the structure beneath Himalayas andTibet has been obtained by extensive analyses of the datacollected by these field experiments including local seis-micity from Nepal to Lhasa Terrane. For example, a total of267 earthquakes were located by a 1-D array as a part of theINDEPTH III studies. All of them were located in the uppercrust from 0 to 25 km depth and several earthquake clusterswere found near the Pumqu-Xianza rift zone in the centralTibet [Langin et al., 2003]. Thousands of local earthquakeswere recorded by the HIMNT stations in Nepal and southernTibet [Monsalve et al., 2006]. The depth distribution ofthese earthquakes indicated the existence of double seismo-genic layers beneath the Himalayan Range: one layer fromsurface to 25 km depth, and the deeper layer from 50 km to100 km depth.[6] In this study we used the joint hypocenter determina-

tion (JHD) method to relocate earthquakes in southern Tibetthat were recorded by 37 portable seismic stations whichwas the lateral 2-D array of the international Hi-CLIMBproject (Himalayan-Tibetan continental lithosphere duringmountain building [Nabelek et al., 2005]). These stationswere deployed for a period of 14 months between July 2004and August 2005. The main 1-D array of Hi-CLIMB wascomposed of closely spaced (4–5 km) broadband seismicstations from Lower Himalaya in Nepal extending intothe Qiangtang Terrane in Tibet (Figure 1a). The roughly700-km-long north – south profile crosses the MainBoundary Thrust (MBT), the Main Central Thrust (MCT),the Indus-Yalu Suture (IYS) and the Bangong-NujiangSuture (BNS) [Nabelek et al., 2005].[7] The 37 portable stations were deployed along a tra-

verse from Tingri to Xigaze in southern Tibet (Figure 1b).The main objectives of this 2-D array are to better locatethe local earthquakes and to image lateral lithosphericheterogeneity beneath southern Tibet. Mantle earthquakes(events below the Moho) had occurred in this area accord-ing to previous studies [e.g., Chen and Molnar, 1983;Chen and Kao, 1996; Zhu and Helmberger, 1996; Chenand Yang, 2004; Monsalve et al., 2006]. Therefore, anaccurate determination of the focal depth of intermediate-

depth earthquakes within our seismic array will provide aunique test for the occurrence of mantle earthquakes here,which could provide observational constraint to the ongoingdebate of the rheology of continental lithosphere [e.g.,Jackson, 2002; Jackson et al., 2004; Chen and Yang, 2004].[8] A Total of 885 local earthquakes were located by

HYPOINVERSE2000 and about half of them were relo-cated by the JHD method. Built on previous studies, ourresults will illuminate the earthquake (depth) distributionfrom the Himalayas to the southern Lhasa Terrane and willprovide observational evidence for the debate of seismo-genic layer(s) in continental lithosphere. It could also help usto understand the rheological structure of the underthrustingIndian lithosphere as well as the Tibetan lithosphere.

2. Tectonics Setting

[9] The main Cenozoic structures in the southern Tibetinclude both north trending rifts and NWW trending right-slip faults (Figure 1b) [Armijo et al., 1986, 1989] The activeMain Frontal Thrust (MFT) along the Himalaya marks thesouthern edge of the Himalayan Range. The north trendingrifts in the study area are the Yadong Gulu rift, the PumquXianza rift, and the Tangra Yum Co rift, from east to west.The origin of these rifts has been debated in the past threedecades. Several mechanisms such as a gravitational col-lapse model [Molnar and Tapponnier, 1978; Tapponnier etal., 1981], a lower crust flow model with east–westextension, a convective removal model of the upper mantle[England and Houseman, 1989] and an oblique conver-gence model of Indian subduction [McCaffrey and Nabelek,1998] have all been proposed as the origin for these rifts.Whether these rifts are restricted to the upper crust or areinvolving the mantle lithosphere has also been under debate[Masek et al., 1994; Yin, 2000]. The northern ends of theTangra Yum Co rift and the Pumqu Xianza rift areconnected by a NWW trending right-slip fault, the GyaringCo fault, which belongs to the central Tibet conjugate faultzone [Taylor et al., 2003]. This strike-slip fault zone hasbeen suggested as a transfer zone linking north trendingCenozoic extensional structures in the Lhasa and Qiangtangterranes [Yin and Harrison, 2000].[10] In this paper, we adopt Yin’s [2006] definition that

the Indus-Yalu Suture (IYS) divides the Himalayan orogenin the south and the Tibetan Plateau in the north. We furtherrefer to the crestal area of the Himalayan orogen as theHigher Himalaya, its south slope as the South Himalayaconsisting of the Lower Himalaya in the north and Sub-Himalaya in the south, and its north slope as the NorthHimalaya (Figure 3). Thus our array transverses two tec-tonic units: the Himalayan orogen and the Tibet Plateau. Forconvenience, southern Tibet is used in this paper to describethe location of our array, including the southern TibetanPlateau and the North Himalaya.

3. Data and Methods

[11] The lateral 2-D array of Hi-CLIMB includes 32broadband seismic stations from Peking University (PKU)and the Institute of Earth Sciences (IES), Academia Sinica,Taiwan and five short-period seismic stations from ChinaAcademy of Geological Sciences (CAGS). The 16 PKU

B12409 LIANG ET AL.: EARTHQUAKES IN SOUTHERN TIBET

2 of 11

B12409

seismographs were equipped with Guralp CMG-3ESPthree-component sensors, with a bandwidth from 0.033 to50 Hz, and a REF TEK 130–01 DAS recorder. The IESseismographs include 12 that equipped with StreckeisenSTS-2 three-component sensors, with a bandwidth from0.0086 to 50 Hz, and a REF TEK RT72-A DAS recorder,and 4 that equipped with Nanometrics Trillium 40 three-component sensors, with a bandwidth from 0.025 to 40 Hz,and a Quanterra Q330 recorder. All the five short-periodseismographs were equipped with CDJ-S1 1 Hz three-component sensors and a LEAS Hathor 3 recorder. These37 stations were operated with a sampling rate of 40 or50 sps from July 2004 to August 2005 for a period of14 months. Average station spacing was approximately

40 km (Figure 1). Because of instrument problems orvandalism, recordings from four broadband stations andall the five short-period stations are not used in this study.[12] Because of the absence of permanent seismic stations

in Tibet except the Lhasa station, global seismicity catalogscould not provide a reliable list of local events. Thus weused an event detection algorithm based on the ratio of theaverage short-term amplitude (STA) to the average long-term amplitude (LTA) to identify candidates for local events[Langin et al., 2003; Turkelli et al., 2003]. Then the Pphases and possible S phases of the selected events werepicked by SAC2000 [Goldstein et al., 2003]. In general, aband pass of 0.8- to 2-Hz filter was applied before thepicking to eliminate the background noise. If the data were

Figure 1. (a) Overview of Tibetan Plateau with the research area outlined by the box. All of the stationsof Hi-CLIMB were shown as light red circles on the map. The directions of plate movements are fromZhang et al. [2004]. (b) Active structures and station locations of the lateral seismic array of Hi-CLIMB.Stations used in this research are shown as red triangles (the stations with little data owning to equipmentproblems and vandalism are show as purple triangles). The rifts are Tangra Yum Co rift, Pumqu-Xianzarift, and Yadong-Gulu rift from left to right.

B12409 LIANG ET AL.: EARTHQUAKES IN SOUTHERN TIBET

3 of 11

B12409

too noisy different band-pass filters were tried such as 1 to 5Hz before picking more phases.[13] As a first step, we used the Hypoinverse2000

[Klein, 2002] (code recompiled by F. Aldersons, http://faldersons.net/Software/Hypoinverse/Hypoinverse.html) forthe analysis of all the events from STA/LTA ratio checkingto obtain a single event location (SEL) list. Here, Zhu andHelmberger’s [1996] model was selected as the reference1-D velocity model since it has been used in theirwaveform simulation analysis for the events close to ourstudy area [Zhu and Helmberger, 1996; Zhu et al., 2006].Data used for the SEL analysis were selected by thefollowing criteria to obtain as many events as possible:(1) the events have a clear P phase, (2) the number of Pwave phases exceed four, and (3) at least one S phase ispicked. As a consequence, a total of 885 local events werelocated by the SEL method.[14] Next, these 885 events were relocated by the joint

hypocenter determination (JHD) method, which was origi-nally developed by Douglas [1967] and later modified byPujol [1988]. Because the mathematical basis and descrip-tion of this algorithm had been discussed in great details inprevious reports [e.g., Pujol, 1988, 1992, 1995, 1996;Ratchkovsky et al., 1997], it will not be repeated here.The JHD method can, in general, simultaneously solve forboth the hypocenter parameters and station corrections.Here the P and S station corrections represent the accumu-lated deviations of the seismic velocity distribution from thesimple 1-D layered model along the raypaths. In otherwords, the positive and negative corrections correspond tolow- and high-velocity anomalies beneath the stationsrelative to the original 1-D velocity model, respectively.While S wave station corrections are calculated separatelyfrom the P wave station corrections they are not indepen-dent of each other because the S wave arrival times arecalculated using a constant Vp/Vs ratio during the JHDinversion.

4. Results

[15] A total of 437 local earthquakes were relocated bythe JHD inversion. The distribution of these events seems tobe related to the local geological structures (Figure 2). Oneobvious feature of all the JHD inversions is that more than250 events were clustered within a small area, at a distanceof 50 km north of IYS and west of the Pumqu-Xianza rift.About 40 earthquakes are also concentrated along theHigher Himalaya, which is parallel to the Main FrontalThrust fault (MFT) and could be related to the MainBoundary Thrust, similar to the events that have beenrelocated at Bhutan [Drukpa et al., 2006]. A small clusterof events was also noticeable between the Pumqu-Xianzarift and the Yadong-Gulu rift, north of IYS.[16] Most significantly, about 20 intermediate-depth

(deeper than 50 km) events were located at the southernends of both the Pumqu-Xianza rift and the Tangra Yum Corift (Figure 2), where intermediate depth events have beenreported [Monsalve et al., 2006]. There was also a deepearthquake close to the IYS near 89�E, where three deepearthquakes (shown with their focal mechanisms inFigure 2) have been studied in great details [Zhu andHelmberger, 1996]. We noticed that almost all of the

reported intermediate-depth earthquakes at southern Tibetare from these two areas, suggesting the undergoing of adistinct dynamic process here.[17] All relocated earthquakes are projected to a roughly

north–south cross section along the receiver function pro-file of Schulte-Pelkum et al. [2005] with a strike of N18�E(Figure 2). The cross section shows that the Indian plate,which is subducting beneath the Higher Himalaya, has pastthe IYS into the Lhasa Terrane (Figure 3). From the MFT tothe Higher Himalaya, seismicity seems to be continuouslydistributed within the whole crust of the subducting Indianlithosphere, consistent with previous studies [Maggi et al.,2000a, 2000b; Jackson 2002; Mitra et al., 2005], indicatinga strong/cold Indian lithosphere. However, earthquakes aredivided into two layers beneath southern part of the NorthHimalaya. There are no earthquakes between these twoseismogenic layers from our relocation results (Figure 3).Interestingly, most of the seismicity is limited only to theupper crust (less than 20 km) in the Lhasa Terrane, exceptthe one event close to the IYS. Our result is also consistentwith the location result of INDEPTH III studies [Langin etal., 2003].[18] An apparent gap in the upper crustal seismicity is

observed at the North Himalaya near the Southern TibetDetachment System (STDS) (Figure 3), indicating an inac-tive STDS. Interestingly, this seismic gap in the uppercrustal seismicity is located above the area with a clusterof the intermediate-depth earthquakes. These deeper eventsare near the concave upward section of the subducted Indianlithosphere and could be related to the bending stressesalthough this explanation requires further study of themechanisms of these intermediate-depth events.[19] As shown in Figure 3, seven earthquakes (from

Harvard CMT catalog) beneath the Lower and HigherHimalaya are clearly correlated with the subduction of theIndian lithosphere because of their downdip faulting mech-anisms [Baranowski et al., 1984; Molnar and Chen, 1983;Chen et al., 1981; Ni and Barazangi, 1984; Dziewonski etal., 1987; Chen and Kao, 1996; Zhu and Helmberger, 1996](see also the Harvard CMT catalog). On the other hand,focal mechanism solutions of both the intermediate depthevents beneath the North Himalaya and the shallow uppercrustal events beneath the Lhasa Terrane are consistent witha regional stress field of east–west extension [Chen andYang, 2004]. The maximum compression axes of thesemoderate-sized, intermediate-depth events are orientedalong the downdip direction of the underthrusting Indianplate. Finally, the location of these intermediate-depthevents indicates that they are associated with the subductionof the Indian lithosphere (Figure 3).[20] During the JHD inversion both P and S wave station

corrections were calculated simultaneously in addition tothe hypocenter location and the origin time of each event[e.g., Pujol, 1996; Ratchkovsky et al., 1997]. Station cor-rections in general can be attributed to crustal velocityvariations beneath these stations since positive and negativecorrections correspond to lower- and higher-velocityanomalies relative to the original 1-D velocity model,respectively. For comparison, six different 1-D velocitymodels were used for the JHD inversion in this study andthe results are shown in Figure 4. All the JHD inversionresults from the six different 1-D velocity models show a

B12409 LIANG ET AL.: EARTHQUAKES IN SOUTHERN TIBET

4 of 11

B12409

Figure 3. Cross section of seismicity across southern Tibet. Hypocenters are shown as circles, stationsare shown as squares, and focal mechanisms within 150 km width are projected onto this profile.Location of this profile is shown in Figure 2. Hypocenters are relocated by JHD using the velocity modelfrom Zhu and Helmberger [1996]. The focal mechanisms with dark compressional quadrants are fromChen and Yang [2004], and the small gray ones are from Harvard CMT catalog. The Moho and MHT,both are shown as dashed lines, are from receiver function results of Schulte-Pelkum et al. [2005] and Jinet al. [2006].

Figure 2. Results of JHD relocation plus focal mechanisms. Hypocenters and focal mechanisms arecolor coded according to the depth. The stations used in JHD inversions are shown as dark triangles. Thegreen line shows the profile position in Figures 3 and 5, which is from 26.873�N, 86.517�E to 30.939�N,88.059�E. The red line shows the receiver function profile of Jin et al. [2006] used in Figure 3, and thered ellipse shows the region of strong converted phases from Jin et al. [2006] mentioned in section 5.Velocity model from Zhu and Helmberger [1996] is used in JHD relocation. Focal mechanisms are fromChen and Yang [2004], which were collected from Baranowski et al. [1984], Molnar and Chen [1983],Chen et al. [1981], Ni and Barazangi [1984], Dziewonski et al. [1987], Chen and Kao [1996], and Zhuand Helmberger [1996], and Harvard CMT catalog. The compressional quadrants of focal mechanismsare color coded by depth. The extensive quadrants of focal mechanism from Chen and Yang [2004] arewhite and the ones from Harvard CMT catalog are gray.

B12409 LIANG ET AL.: EARTHQUAKES IN SOUTHERN TIBET

5 of 11

B12409

similar variation in both P and S wave station corrections.There is a general trend that the S wave velocity decreasestoward the east–northeast direction. More importantly, Pwave station corrections show a negative value (earlyarrival) along the Pumqu-Xianza rift (except models 1 and5), corresponding to a higher P velocity zone at the rift,

which will be further elaborated in section 5. However, thisfeature is not apparent in the S wave station corrections.

5. Discussion

[21] In order to verify the robustness of our JHD inver-sion results we have applied different 1-D velocity modelsand also added random noises to the arrival time picks forthe JHD inversion (Figure 5). Inversion with differentvelocity models could examine the dependence of theinversion to the reference 1-D velocity models and inver-sion using data with random noise provides us with anestimate of the sensitivity of inversion results to the phasepick errors. Several different 1-D velocity models of south-ern Tibet have been reported using different geophysicalmethods, including the receiver function inversion andsurface wave inversions [Cotte et al., 1999; Rapine et al.,2003; Zhu and Helmberger, 1996; Bassin et al., 2000;Monsalve et al., 2006]. It is noticed that the stations usedin the previous receiver function studies to obtain the 1-Dvelocity models were mostly deployed along the Yardong-Gulu rift. Therefore, velocity models derived from thesereceiver function studies might not be a good representativefor the entire southern Tibet because of the possible east–west variation in seismic velocities, particularly across therifts. On the other hand, significant difference does existamong the velocity models obtained from surface wavegroup velocity inversion for the Himalayas, the LhasaTerrane and the Qiangtang Terrane [Cotte et al., 1999;Rapine et al., 2003]. From the depth distribution of relo-cated hypocenters as shown in Figure 5 it is clear that thegeneral pattern of the event distribution from all inversionsof the 6 different velocity models is quite similar. We hadalso added a ±0.2 s random average distributed noise to thearrival time picks for a stability check and found that theadded noise had little effect on the inversion results. Thesetests show that our relocation results are quite stable to theuncertainties from the simple 1-D velocity model and phasepicking errors.[22] We have also relocated the events from the Prelim-

inary Determination of Epicenter (PDE, from NationalEarthquake Information Center of USGS) catalog duringthe operation of our temporary seismic array, which are

Figure 4. P and S wave station corrections from sixdifferent velocity models. (a) P wave station correction.(b) S wave station correction. From the P wave stationcorrection, especially models 2, 3, 4, and 6 in Figure 4a,there is a negative anomaly beneath Pumqu-Xianza rift.From the S wave station corrections there is no significantlypattern correlated with active structure, but northeast strikeincreasing trend is found from all of them except model 1 inFigure 4b. The different velocity model 1 in Figures 4a and4b is the velocity model of Zhu and Helmberger [1996];model 2 in Figures 4a and 4b is the Crust 2.0 model [Bassinet al., 2000]; model 3 in Figures 4a and 4b is the velocitymodel 1 of Cotte et al. [1999]; model 4 in Figures 4a and 4bis velocity model 2 of Cotte et al. [1999]; model 5 inFigures 4a and 4b is velocity model of Jin et al. [2006];and model 6 in Figures 4a and 4b is velocity model ofMonsalve et al. [2006].

B12409 LIANG ET AL.: EARTHQUAKES IN SOUTHERN TIBET

6 of 11

B12409

listed in Table 1. In general, our results from the JHDinversion using data from the local seismic network haveproduced a better location for these events compared to thePDE results. Combination of a good station azimuth cov-erage with a high-quality record at the nearest station couldprovide a good constraint on the depth determination. Whenthere are enough effective P and S picks, even inversion ofone particular event with an azimuthal gap and a largerdistance from the nearest station, such as the event20050208 and 20050530, could still produce a reasonablelocation result. The location errors depend on the stationdistribution for each event (Table 1). We noticed that theresiduals of JHD results are in general larger than the

residuals of HYPOINVERSE2000 results because of thetrade off among different events during the JHD inversions,which locate all events at the same time with the samestation corrections.[23] The most striking feature of our relocation results

(Figures 3 and 5) is that there are almost no earthquakes atdepths of 30 to 50 km in the Tibetan crust north of the IYS,a good indication for the presence of a midcrustal or lowercrustal ductile layer as suggested by many previous studies[e.g., Zhao and Morgan, 1985, 1987]. The aseismic middlecrust is also consistent with the crustal channel flow modelsfor the expansion of the Tibetan Plateau [Royden et al.,1997; Clark and Royden, 2000]. Our result is consistent

Figure 5. Relocation results from six different velocity models. The distribution pattern is as same asthe result using the velocity model from Zhu and Helmberger [1996] shown in Figure 3, so it is clear thatthe influence from different velocity models is not so significant. (a) Velocity model of Zhu andHelmberger [1996]. (b) Crust 2.0 model [Bassin et al., 2000]. (c) Velocity model 1 of Cotte et al. [1999].(d) Velocity model 2 of Cotte et al. [1999]. (e) Velocity model of Jin et al. [2006]. (f) Velocity model ofMonsalve et al. [2006].

B12409 LIANG ET AL.: EARTHQUAKES IN SOUTHERN TIBET

7 of 11

B12409

with previous studies of major earthquakes in this regionusing stations outside of Tibet, which had argued fortwo seismogenic layers beneath southern Tibet [Chen andMolnar, 1983; Chen and Kao, 1996; Chen and Yang, 2004].In addition, our relocated events show a strong cluster nearthe Moho, both above and beneath the Moho (Figures 3 and5), but whether some of these events are truly mantle eventswould have to require further study, such as the waveformfitting analysis [Zhu and Helmberger, 1996], given theuncertainties in both the JHD depths and the Moho depthderived from the receiver function study [Schulte-Pelkum etal., 2005; Jin et al., 2006]. Another feature of our inversionsis that there are no intermediate-depth earthquakes north ofIYS where all of our relocated events were within the uppercrust (Figures 3 and 5), indicating one seismogenic layer inthe Tibetan crust [e.g., Maggi et al., 2000a].[24] In summary the distribution of local seismicity from

this study seems to suggest that two seismogeneic layersexist at the collision front, associated with the subduction ofthe Indian lithosphere. There are different explanations forthese double seismogenic layers at the collision front. Oneis that the base of the southern Tibetan crust comprises amafic layer generated during basaltic ponding at the base ofthe Gangdese arc [Yin and Harrison, 2000]. The problemwith this explanation is that the basement of the Gangdesearc may not be likely to be present below the NorthHimalaya where earthquakes were present at both the lowercrust and uppermost mantle depths (Figure 3). The secondexplanation is due to the initiation of the transformationfrom granulite to eclogite when hydrous fluid is present[Jackson et al., 2004]. This appears to be the most satis-factory model as it is most consistent with our observations.Further north of IYS there is only one seismogenic layer atthe upper crust, that is, local earthquakes are limited to theupper crust (0–20 km) similar to observations at othercontinental active regions such as the southern Californiain the North America [Wong and Chapman, 1990] or theIran Plateau of Middle East [Maggi et al., 2000b].[25] Dip-slip fault plane solutions of those intermediate-

depth earthquakes shown in Figures 2 and 3 have been usedto suggest a regional stress field of east–west extensionwithin the subducting Indian lithosphere [e.g., Chen andKao, 1996], which supports a model for rifts within thesubducting Indian lithosphere [Yin, 2000]. Drukpa et al.[2006] had also reported the midcrustal to deep crustalstrike-slip events in Bhutan, interpreted as the result ofoblique convergence. If the subducting Indian lithospherewas torn apart laterally in the east–west direction, then thelithosphere on either side of the rift could be subductingindependently from each other. As a consequence strike-slipdisplacement could take place along the conjugation faults,as suggested by the strike-slip character of the fault planesolutions of these intermediate-depth earthquakes that areclustered at the southern end of the rifts (Figure 2). On theother hand, the water content required to facilitate the phasechange from granulite to eclogite phase change [Jackson etal., 2004] could be introduced by rifting of the subductingIndian lithosphere. Stresses associated with the eclogitephase transformation could be at least partially responsiblefor these intermediate-depth earthquakes.[26] Over 60% of our relocated events (250 of 400 from

June 2004 to August 2005) are shallow and clustered nearTable

1.ComparisonoftheDifferentLocations

Date

PDEOrigin

Tim

e(U

T)

Longitude

Latitude

Depth

(km)

Residual

(s)

HYPOIN

VERSE2000

JHD

Longitude

Latitude

Depth

(km)

Residual

(s)

Horizontal

Error

(km)

Vertical

Error

(km)

Maxim

um

Azimuthal

Gap

(deg)

Nearest

Station

(km)

Effective

PandS

Picksa

Longitude

Latitude

Depth

(km)

Residual

(s)

8Jul2004

215000.52

87.997

29.844

32

0.94

88.1658

30.0273

6.91

0.64

1.73

21.27

181

20

15

88.173

30.030

4.1

0.49

20Jul2004

033551.08

85.793

27.938

13

1.07

85.9105

27.9413

6.98

0.63

2.87

42.21

292

114

16

85.906

27.900

32.5

0.92

23Jul2004

012532.60

88.118

30.175

111.01

88.1992

30.0277

4.41

0.15

0.42

12.04

187

17

10

88.174

30.010

3.9

0.68

21Aug2004

090711.11

88.086

29.944

26

1.07

88.1622

30.0310

6.44

0.71

4.47

40.78

182

20

11

88.166

30.018

3.8

0.84

10Nov2004

042111.12

87.778

27.929

45

0.83

87.8107

28.0165

59.58

0.20

1.14

0.90

201

13

987.838

28.023

62.7

0.68

8Feb

2005

041356.39

86.072

27.760

14

1.20

86.1132

27.7967

34.26

0.35

2.78

1.54

294

177

15

86.159

27.756

29.5

0.77

30May

2005

105729.93

88.180

31.568

10

1.25

88.2087

31.5708

6.36

0.25

1.15

1.47

279

175

16

88.241

31.564

8.0

0.50

aEffectivePandSpicksisthenumber

ofPandStimes

withfinal

weightsgreater

than

0.1

duringHYPOIN

VERSE2000inversion.

B12409 LIANG ET AL.: EARTHQUAKES IN SOUTHERN TIBET

8 of 11

B12409

the Pumqu-Xianza rift at 30�N (Figure 2). Interestingly,more than 40% (219 of 500 from 1995 to 2004) of theevents reported in the ISC catalog within our local networkwere also from this small area. The active seismicity at thisarea was also noticed by Langin et al. [2003] and Yang etal. [2002], who had suggested that the tidal stresses couldbe the cause for these microearthquakes. Guo and Chen[2008] found thousands of tiny events in this area from thedata recorded by nearby stations of our array and hadspeculated that these little events could have been dynam-ically triggered by the great 2004 Sumatra earthquake,which was at several thousands kilometers to the southof our stations. These events could reflect the activities ofgeofluids such as magma or hot springs in response to thestrong seismic waves from such a large event. Thisphenomenon has been reported at other continents suchas the activities recorded at northern California after the1992 Landers earthquakes [Hill et al., 1993] and at theYellow Stone National Park after the 2002 Alaska Denaliearthquake [Husen et al., 2004]. Boiling springs and othergeothermal activities have been reported at the Pumqu-Xianza rift by Hou et al. [2004]. On the other hand, Zhouand Murphy’s [2005] tomography results showed a wedge-shaped slow velocity anomaly within the lower crust from85�E to 93�E and north of IYS, which is consistent withthe presence of partial melting in the middle crust as hadbeen suggested by INDEPTH II deep seismic reflectionexperiment along the Yadong-Gulu rift [Nelson et al.,1996].[27] Jin et al. [2006] also reported a northeast trending

strong converted phases in their east–west profiles (shownas the red ellipse in Figure 2), which seems to be correlatedwith the northeast increasing gradient of our S wavecorrections. The negative P wave station correction alongthe Pumqu-Xianza rift, corresponding to a high P velocityzone, derived from our JHD inversion is a little surprisesince low P velocity is usually expected at a rift zone. Sincemost of the intermediate-depth earthquakes were concen-trated at the southern end of the Pumqu-Xianza rift onepossibility could be a shallow Moho beneath the Pumqu-Xianza rift, as a consequence of a thinner crust from theeast–west rifting, which could induce a localized mantleupwelling. Receiver function study of the same seismicarray does indeed reveal a shallower Moho (at least 5 kmshallower) at the rift [Jin et al., 2006]. On the other hand,there is a low seismic velocity anomaly extending to about300 km depth beneath the Coma rift to the east of theYadong Gulu rift (Figure 2) from both finite frequency Pand S wave tomography inversions, which has been inter-preted as localized mantle upwelling due to continentaldelamination [Ren and Shen, 2008]. All of the abovediscussion points to the case that mantle process plays animportant role during the formation of these north–southtrending rifts in southern Tibet. Whether the subductingIndian lithosphere was torn apart and lithospheric rupturezone was established that had induced localized mantleupwelling or the process of continental delamination hadtriggered the mantle upwelling requires further study. Bothmechanisms would have produced a shallower Moho be-neath rifts and the localized mantle upwelling could providethe heat/melts to the rifts and neighbor region, which could

explain the existence of geothermal activities and hotsprings in the southern Tibet.

6. Conclusions

[28] A total of 885 local earthquakes were located byHYPOINVERSE2000 and 437 of them were relocated byjoint hypocenter determination (JHD) method. While mostof these earthquakes are scattered within the seismic array,more than 250 earthquakes were clustered within a smallarea near the Pumqu-Xianza rift, which could be related toactivities of geofluids there. Earthquakes are also concen-trated along the Higher Himalaya, which is in parallel to thenorth of the MFT fault.[29] Projection of all earthquakes to a north–south profile

shows that the subducted India plate has extended into theLhasa Terrane. While earthquakes are divided into twolayers beneath the Himalayas most of the seismicity islimited only to the upper crust north of IYS. About 20intermediate depth events (50–70 km) were observed andthey could be associated with the combination of separatedsubduction of the Indian lithosphere and petrological phasetransformation. There are no earthquakes at depths between20�50 km beneath the array, which is consistent with aductile middle/lower crust beneath southern Tibet. The lackof deeper events north of the IYS is probably related to theductile and partially molten middle crust of the southernTibet where water could play a major role. Relocationresults of these local earthquakes in southern Tibet supportthe popular ‘‘jelly-sandwich’’ model for the rheology ofcontinental lithosphere in which the strong seismogeniclayers of upper curst and upper mantle are separated by aweak lower crust.[30] Finally, P wave station corrections show a high-

velocity zone along the Pumqu-Xianza rift zone, whichcould be interpreted as a thinner crust from an east–westextension at the rift. Whether east–west variations in theMoho depth exist across these rifts in southern Tibet is animportant question in understanding the N–S continentalcollision and the E–W extension seen at the surfacetectonics and deserves serious attention from the seismotec-tonics community.

[31] Acknowledgments. We would like first to thank John Nabelekfor his leadership in the fieldwork of Hi-CLIMB and all the members of theHi-CLIMB team who collected the data used in this study. We would alsolike to thank the colleagues from CAGS and IES for their instruments anddata sharing. An Yin, James Ni, Rodolfo Console, and an anonymousreviewer provided constructive comments that improved the manuscript.Zhuqi Zhang at Peking University helped in making the figures and theGeneric Mapping Tools (GMT) is used in producing all the figures. Thefieldwork is supported by a NSFC grant for international collaboration(40520120222) and data processing and analyses were also supported byadditional NSFC grants of (40125011, 40521002).

ReferencesArmijo, R., P. Tapponnier, J. P. Mercier, and T. Han (1986), Quaternaryextension in southern Tibet, J. Geophys. Res., 91, 13,803 –13,872,doi:10.1029/JB091iB14p13803.

Armijo, R., P. Tapponnier, and T. Han (1989), Late Cenozoic right-lateralstrike-slip faulting across southern Tibet, J. Geophys. Res., 94,2787–2838, doi:10.1029/JB094iB03p02787.

Baranowski, J., J. Armbruster, L. Seeber, and P. Molnar (1984), Focaldepths and fault plane solutions of earthquakes and active tectonicsof the Himalaya, J. Geophys. Res., 89, 6918 – 6928, doi:10.1029/JB089iB08p06918.

B12409 LIANG ET AL.: EARTHQUAKES IN SOUTHERN TIBET

9 of 11

B12409

Bassin, C., G. Laske, and G. Masters (2000), The current limits of resolu-tion for surface wave tomography in North America, Eos Trans. AGU,81(48), Fall Meet. Suppl., Abstract S12A-03.

Brown, L. D., W. Zhao, K. D. Nelson, M. Hauck, D. Alsdorf, A. Ross, M.Cogan, M. Clark, X. Liu, and J. Che (1996), Bright spots, structure, andmagmatism in southern Tibet from INDEPTH seismic reflection profil-ing, Science, 274, 1688–1690, doi:10.1126/science.274.5293.1688.

Chen, W., and H. Kao (1996), Seismotectonics of Asia: Some recentprogress, in The Tectonic Evolution of Asia, edited by A. Yin and T. M.Harrison, pp. 37–62, Cambridge Univ. Press, Cambridge, U.K.

Chen, W., and P. Molnar (1983), Focal depths of intracontinental and in-traplate earthquakes and their implications for the thermal and mechanicalproperties of the lithosphere, J. Geophys. Res., 88, 4183 – 4214,doi:10.1029/JB088iB05p04183.

Chen, W., and Z. Yang (2004), Earthquakes beneath the Himalayas andTibet: Evidence for strong lithospheric mantle, Science, 304, 1949–1952,doi:10.1126/science.1097324.

Chen, W., J. Nabelek, T. Fitch, and P. Molnar (1981), An intermediatedepth earthquake beneath Tibet: Source characteristics of the event ofSeptember 14, 1976, J. Geophys. Res., 86, 2863–2876, doi:10.1029/JB086iB04p02863.

Chen, L., J. R. Booker, A. G. Jones, N. Wu, M. J. Unsworth, W. Wei, andH. Tan (1996), Electrically conductive crust in southern Tibet fromINDEPTH magnetotelluric surveying, Science, 274(5293), 1694–1696,doi:10.1126/science.274.5293.1694.

Clark, M. K., and L. H. Royden (2000), Topographic ooze: Building theeastern margin of Tibet by lower crustal flow, Geology, 28(8), 703–706,doi:10.1130/0091-7613(2000)28<703:TOBTEM>2.0.CO;2.

Cotte, N., H. Pedersen, M. Campillo, J. Mars, J. Ni, R. Kind, E. Sandvol,and W. Zhao (1999), Determination of the crustal structure in southernTibet by dispersion and amplitude analysis of Rayleigh waves, Geophys.J. Int., 138, 809–819, doi:10.1046/j.1365-246x.1999.00927.x.

De, R., and J. R. Kayal (2003), Seismotectonic model of the Sikkim Hi-malaya: Constraint from microearthquake surveys, Bull. Seismol. Soc.Am., 93(3), 1395–1400, doi:10.1785/0120020211.

de la Torre, T., and A. Sheehan (2005), Broadband seismic noise analysis ofthe Himalayan Nepal Tibet seismic experiment, Bull. Seismol. Soc. Am.,95(3), 1202–1208, doi:10.1785/0120040098.

Douglas, A. (1967), Joint epicenter determination, Nature, 215, 47–48,doi:10.1038/215047a0.

Drukpa, D., A. A. Velasco, and D. I. Doser (2006), Seismicity in theKingdom of Bhutan region (1937–2003), Evidence for crustal trans-current deformation, J. Geophys. Res., 111, B06301, doi:10.1029/2004JB003087.

Dziewonski, A. M., G. Ekstrom, J. E. Franzen, and J. H. Woodhouse(1987), Global seismicity of 1979: Centroid-moment tensor solution for524 earthquakes, Phys. Earth Planet. Inter., 48(1 – 2), 18 – 46,doi:10.1016/0031-9201(87)90109-9.

England, P., and G. Houseman (1989), Extension during continental con-vergence, with application to the Tibetan Plateau, J. Geophys. Res., 94,17,561–17,579, doi:10.1029/JB094iB12p17561.

Goldstein, P., D. Dodge, M. Firpo, and L. Minner (2003), SAC2000: Signalprocessing and analysis tools for seismologists and engineers, in TheIASPEI International Handbook of Earthquake and Engineering Seismol-ogy, edited by W. H. K. Lee et al., pp. 1613–1614, Academic, London.

Guo, Y., and Y. J. Chen (2008), Local seismicity in south Tibet remotelytriggered by the Great 2004 Mw 9.0 Sumatra earthquake, Acta Seismol.Sin., in press.

Herquel, G., G. Wittlinger, and J. Guilbert (1995), Anisotropy and crustalthickness of northern-Tibet: New constraints for tectonic modelling,Geophys. Res. Lett., 22, 1925–1928.

Hill, D. P., et al. (1993), Seismicity remotely triggered by the magnitude 7.3Landers, California earthquake, Science, 260, 1617–1623, doi:10.1126/science.260.5114.1617.

Hirn, A., et al. (1995), Seismic anisotropy as an indicator of mantle flowbeneath the Himalayas and Tibet, Nature, 375, 571–574, doi:10.1038/375571a0.

Hou, Z., Z. Li, X. Qu, Y. Gao, L. Hua, M. Zheng, S. Li, and W. Yuan(2004), The uplifting process of the Tibetan Plateau since 0.5 Ma B.P,Sci. China, Ser. D, 44, Suppl., 35–44.

Huang, W.-C., et al. (2000), Seismic polarization anisotropy beneaththe central Tibetan Plateau, J. Geophys. Res., 105, 27,979–27,989,doi:10.1029/2000JB900339.

Husen, S., R. Taylor, R. B. Smith, and H. Healser (2004), Changes ingeyser eruption behavior and remotely triggered seismicity in Yellow-stone National Park produced by the 2002 M 7.9 Denali fault earthquake,Alaska, Geology, 32(6), 537–540, doi:10.1130/G20381.1.

Jackson, J. (2002), Strength of the continental lithosphere: Time toabandon the jelly sandwich?, GSA Today, 12, 4 –10, doi:10.1130/1052-5173(2002)012<0004:SOTCLT>2.0.CO;2.

Jackson, J. A., H. Austrheim, D. McKenzie, and K. Priestley (2004),Metastability, mechanical strength, and the support of mountain belts,Geology, 32(7), 625–628, doi:10.1130/G20397.1.

Jin, G., Y. J. Chen, E. Sandvol, S. Zhou, Y. Tang, X. Liang, G. Ye, D.Wilson, and J. Nabelek (2006), Receiver Function Images of the Crustand Upper Mantle Structure of Southern Tibet, Eos Trans. AGU, 87(52),Fall Meet. Suppl., Abstract T53F–04.

Klein, F. W. (2002), User’s guide to HYPOINVERSE-2000, a FORTRANprogram to solve for earthquake locations and magnitudes, U.S. Geol.Surv. Open File Rep., 02-171.

Langin, W., L. Brown, and E. Sandvol (2003), Seismicity in central Tibetfrom project INDEPTH III seismic recordings, Bull. Seismol. Soc. Am.,93(5), 2146–2159, doi:10.1785/0120030004.

Maggi, A., J. Jackson, D. McKenzie, and K. Priestley (2000a), Earth-quake focal depths, Effective elastic thickness and the strength of thecontinental lithosphere, Geology, 28(6), 495–498, doi:10.1130/0091-7613(2000)28<495:EFDEET>2.0.CO;2.

Maggi, A., J. Jackson, K. Priestley, and C. Baker (2000b), A reassessmentof focal depth distributions in southern Iran, the Tien Shan and northernIndia: Do earthquakes really occur in the continental mantle?, Geophys. J.Int., 143, 629–661, doi:10.1046/j.1365-246X.2000.00254.x.

Masek, J. G., B. L. Isacks, and E. J. Fielding (1994), Rift flank uplift inTibet: Evidence for a viscous lower crust, Tectonics, 13(3), 659–667,doi:10.1029/94TC00452.

McCaffrey, R., and J. Nabelek (1998), Role of oblique convergence in theactive deformation of the Himalayas and southern Tibet plateau, Geology,26(8), 691 –694, doi:10.1130/0091-7613(1998)026<0691:ROOCIT>2.3.CO;2.

Mitra, S., K. Priestley, A. Bhattacharyya, and V. Gaur (2005), Crustalstructure and earthquake focal depths beneath northeastern India andsouthern Tibet, Geophys. J. Int., 160, 227–248, doi:10.1111/j.1365-246X.2004.02470.x.

Molnar, P., and W. Chen (1983), Focal depths and fault plane solutions ofearthquakes under the Tibetan Plateau, J. Geophys. Res., 88, 1180–1196,doi:10.1029/JB088iB02p01180.

Molnar, P., and P. Tapponnier (1978), Active tectonics of Tibet, J. Geophys.Res., 83, 5361–5375, doi:10.1029/JB083iB11p05361.

Monsalve, G., A. Sheehan, V. Schulte-Pelkum, S. Rajaure, M. R. Pandey,and F. Wu (2006), Seismicity and one-dimensional velocity structure ofthe Himalayan collision zone: Earthquakes in the crust and upper mantle,J. Geophys. Res., 111, B10301, doi:10.1029/2005JB004062.

Nabelek, J., J. Vergne, G. Hetenyi, and the Hi-CLIMB Team (2005), Asynoptic view of the Himalayan collision zone and southern Tibet, EosTrans. AGU, 86(52), Fall Meet. Suppl., Abstract T52A–02.

Nelson, K., et al. (1996), Partially molten middle crust beneath southernTibet: Synthesis of project INDEPTH results, Science, 274, 1684–1688,doi:10.1126/science.274.5293.1684.

Ni, J., and M. Barazangi (1984), Seismotectonics of the Himalayan collisionzone D Geometry of the underthrusting Indian Plate beneath the Hima-laya, J. Geophys. Res., 89, 1147–1163, doi:10.1029/JB089iB02p01147.

Owens, T. J., G. E. Randall, F. T. Wu, and R. Zeng (1993), PASSCALinstrument performance during the Tibetan Plateau passive seismic ex-periment, Bull. Seismol. Soc. Am., 83(6), 1959–1970.

Pandey,M. R., R. P. Randukar, J. P. Avouac, J. Vergne, and T. Heritier (1999),Seismotectonics of the Nepal Himalaya from a local seismic network,J. Asian Earth Sci., 17, 703–712, doi:10.1016/S1367-9120(99)00034-6.

Pujol, J. (1988), Comments of the joint determination of hypocenters andstation corrections, Bull. Seismol. Soc. Am., 78(3), 1179–1189.

Pujol, J. (1992), Joint hypocentral location in media with lateral velocityvariations and interpretation of the station corrections, Phys. EarthPlanet. Inter., 75, 7–24, doi:10.1016/0031-9201(92)90114-B.

Pujol, J. (1995), Application of the JHD technique to the Loma Prieta,California, mainshock-aftershock sequence and implications for earth-quake location, Bull. Seismol. Soc. Am., 85(1), 129–150.

Pujol, J. (1996), An integrated 3D velocity inversion– joint hypocentraldetermination relocation analysis of events in the Northridge area, Bull.Seismol. Soc. Am., 86(1B), S138–S155.

Rapine, R., F. Tilmann, M. West, J. Ni, and A. Rodgers (2003), Crustalstructure of northern and southern Tibet from surface wave dispersionanalysis, J. Geophys. Res., 108(B2), 2120, doi:10.1029/2001JB000445.

Ratchkovsky, N. A., J. Pujol, and N. N. Biswas (1997), Relocation of earth-quakes in the Cook Inlet area, south central Alaska, using the joint hypo-center determination method, Bull. Seismol. Soc. Am., 87(3), 620–636.

Ren, Y., and Y. Shen (2008), Finite frequency tomography in southeasternTibet: Evidence for the causal relationship between mantle lithospheredelamination and the north–south trending rifts, J. Geophys. Res., 113,B10316, doi:10.1029/2008JB005615.

Royden, L. H., B. C. Burchfiel, R. W. King, E. Wang, Z. Chen, F. Shen, andY. Liu (1997), Surface deformation and lower crustal flow in easternTibet, Science, 276, 788–790, doi:10.1126/science.276.5313.788.

B12409 LIANG ET AL.: EARTHQUAKES IN SOUTHERN TIBET

10 of 11

B12409

Schulte-Pelkum, V., G. Monsalve, A. Sheehan, M. Pandey, S. Sapkota, andR. Bilham (2005), Imaging the Indian subcontinent beneath the Hima-laya, Nature, 435, 1222–1225, doi:10.1038/nature03678.

Sol, S., et al. (2007), Geodynamics of the southeastern Tibetan Plateau fromseismic anisotropy and geodesy, Geology, 35(6), 563–566, doi:10.1130/G23408A.1.

Tapponnier, P., J. L. Mercier, R. Armijo, T. Han, and J. Zhao (1981), Fieldevidence for active normal faulting in Tibet, Nature, 294, 410–414,doi:10.1038/294410a0.

Taylor, M., A. Yin, F. J. Ryerson, P. Kapp, and L. Ding (2003), Conjugatestrike-slip faulting along the Bangong-Nujiang suture zone accommo-dates coeval east –west extension and north– south shortening in theinterior of the Tibetan Plateau, Tectonics, 22(4), 1044, doi:10.1029/2002TC001361.

Turkelli, N., et al. (2003), Seismogenic zones in Eastern Turkey, Geophys.Res. Lett., 30(24), 8039, doi:10.1029/2003GL018023.

Wittlinger, G., J. Vergne, P. Tapponnier, V. Farra, G. Poupinet, M. Jiang, H.Su, G. Herquel, and A. Paul (2004), Teleseismic imaging of subductinglithosphere and Moho offsets beneath western Tibet, Earth Planet. Sci.Lett., 221, 117–130, doi:10.1016/S0012-821X(03)00723-4.

Wong, I. G., and D. S. Chapman (1990), Deep intraplate earthquakes in thewestern United States and their relationship to lithospheric temperatures,Bull. Seismol. Soc. Am., 80(3), 589–599.

Yang, M., P. Cai, and W. Hua (2002), Research on current seismic activityof Xietongmen-Shenzha area of Tibet (in Chinese), South ChinaJ. Seismol., 22(2), 24–31.

Yin, A. (2000), Mode of Cenozoic east–west extension in Tibet suggestinga common origin of rifts in Asia during the Indo-Asian collision,J. Geophys. Res., 105, 21,745–21,759, doi:10.1029/2000JB900168.

Yin, A. (2006), Cenozoic tectonic evolution of the Himalayan orogen asconstrained by along-strike variation of structural geometry, exhumationhistory, and foreland sedimentation, Earth Sci. Rev., 76, 1 – 131,doi:10.1016/j.earscirev.2005.05.004.

Yin, A., and T. M. Harrison (2000), Geologic evolution of the Himalayan-Tibetan Orogen, Annu. Rev. Earth Planet. Sci., 28, 211 – 280,doi:10.1146/annurev.earth.28.1.211.

Zhang, P. Z., et al. (2004), Continuous deformation of the Tibetan Plateaufrom Global Positioning System data, Geology, 32(9), 809 – 812,doi:10.1130/G20554.1.

Zhao, W., and W. J. Morgan (1985), Uplift of Tibetan Plateau, Tectonics, 4,359–369, doi:10.1029/TC004i004p00359.

Zhao, W., and W. J. Morgan (1987), Injection of Indian crust into Tibetanlower crust: A two-dimensional finite element model study, Tectonics, 6,489–504, doi:10.1029/TC006i004p00489.

Zhao, W., K. D. Nelson, and Project INDEPTH Team (1993), Deep seismicreflection evidence for continental underthrusting beneath southern Tibet,Nature, 366, 557–559, doi:10.1038/366557a0.

Zhou, H., and M. A. Murphy (2005), Tomographic evidence for wholesaleunderthrusting of India beneath the entire Tibetan Plateau, J. Asian EarthSci., 25, 445–457, doi:10.1016/j.jseaes.2004.04.007.

Zhu, L., and D. V. Helmberger (1996), Intermediate depth earth-quakes beneath the India-Tibet collision zone, Geophys. Res. Lett., 23,435–438, doi:10.1029/96GL00385.

Zhu, L., Y. Tan, D. V. Helmbereger, and C. K. Saikia (2006), Calibrationof the Tibetan Plateau using regional seismic waveforms, Pure Appl.Geophys., 163, 1193–1213, doi:10.1007/s00024-006-0073-7.

�����������������������Y. J. Chen, Y. Fu, G. Jin, X. Liang, P. Liu, J. Ning, Y. Tang, L. Xiao, and

S. Zhou, Institute of Theoretical and Applied Geophysics, School of Earthand Space Sciences, Peking University, Beijing 100871, China. (johnyc@pku.edu.c; pkucugfyy@gmail.com; ge_jin@pku.edu.cn; liangxf@pku.edu.cn)X. Lou, Department of Earth and Planetary Sciences, Northwestern

University, 1850 Campus Drive, Evanston, IL 60208-0000, USA.

B12409 LIANG ET AL.: EARTHQUAKES IN SOUTHERN TIBET

11 of 11

B12409