Geodesy and Geodynamics 2010 , 1 (1 ) : 1 http://www. jgg09. com Doi:10.3724/SP.J.1246.2010.00034

Effect of Kunlun Ms 8. 1 earthquake on crustal deformation in northeastern edge region of Qinghai-Tibet plateau

Duxin Cui , Qingliang Wang and Wenping Wang Second Crust Monitoring and Application Center, China Earthquake Administration ,Xi' an 710054

Abstract: Seismic fault parameters can he inversed with Okada model based on deformation data before and af­ ter earthquakes in focal region and its adjacent area. Co-seismic displacements can be simulated by using these parameters , and then regional velocity field obtained by deducting the co-seismic displacements from the ob­ served displacements by GPS method. We processed and analyzed the data in the northeastern edge region of the Qinghai-Tibet plateau observed during 2001 - 2003 in two steps : firstly, the displacements generated by Kunlun Ms8. 1 earthquake of 2001 in this region was simulated, and secondly, deducted the co-seismic dis­ placements from it and obtained the horizontal crustal velocity field. The results reveal: 1) the effect of Kunlun Ms8. 1 earthquake on crustal deformation in this region is significant; 2) the velocity field obtained with this method is better than the original GPS velocity field in reflecting the status of regional crustal movement and strain. Key wards: northeastern edge of the Qinghai-Tibet block; co-seiamic displacement; GPS velocity field; Kun­ lun Ms8. 1 earthquake; crustal movement

When using the GPS observation data that are not ac­ 1 Introduction quired during the same time period to calculate crustal­ movement velocity field, the result is probsbly affected by According to processed GPS data observed during 1999 co-seismic deformation, if the effect is ignored. Howev­ -2001 before the Ms8. 1 Kunlun earthquake on 14 er, to assess future earthquake trend in the region, peo­ November 2001 and data observed in 2003 after the ple are more interested in the status of regional stress earthquake , the direction of horizontal velocity field in and strain after earthquake. 1f we can simulate the co­ the western part of northeast edge region of the Qing­ seismic displacement field caused by the earthquake hai-Tibet plateau has changed a lot, from northeast to and deduct it from the observed velocity field, we can southwest. For several reasons, post-earthquake GPS obtain approximate velocity field after the earthquake. measurement was made only in the epicentml area for To remove the effect of co-seismic deformation on the the purpose of studying co-seismic deformation , where­ calculated velocity field , one may use the software 1 21 as in most surronnding region , including most part of GLOBK and QOCA in different ways[ • • But for a northeastern edge region of the Qinghai-Tibet plateau, large and complex earthquake , these methods can re­ no GPS measurement was made until 2003. duce the effect of the earthquake on the solution of ve­ locity field only to a certain extent, but can not truly e­ Received :2010-08-10; Accepted :2010-09-10 liminate it. The key issue in obtaining velocity field Corresponding author: Tel. + 86-29-85506514; E-mail: cuiduxin@ 163. com without earthquake effect is to obtain the co-seiamic This study was supported by the National Natural Science Fundati.on of displacement field and to remove its effect. China ( 40674057 ) and Earthquake Science Joint Fundation of China (A07132) In the present study, on the basis of geodetic data No. I Iluxin lli, et al.Effect of Kunhm Ms8.1 earthquake oo crustal dd"anmtioo in mrtlreastem ecJw:: "'19m of Qingbal-Tiliet plateau 35 observed in the epicentral area of the Kunlun Ms8. 1 the earthquake during 1999 and 2001 in the northeast­ earthquake and its adjacent region near the time of em edge region of the Qinghai-Tibet plateau under earthquake occurrence , we obtained the co-seismic slip "Project 973 " and " China Crustal Movement Net­ and geometric parameters of the earthquake fault by in­ work" , and in 2003 after the earthquake under "Project version. Then we used the resultant model parameters 973" . Thus we may use these data to calculate the ve­ to simulate the displacement field affected by the earth­ locity field in this area during 2001 -2003 after deduc­ quake on the northeastern edge region of the Qinghai­ ting the displacement related to the earthquake. Tibet plateau. Finally we deducted the velocity changes 3. 1 Calculation of the observed crustal movement due to co-seismic and post-seimic displacements from velocity the observed velocity field to obtain the current velocity field. Since in this procedure the geometry of the fault The GPS data are processed in following steps: 1 ) Cal­ 16 and the constraint of the observation data are fully con­ culating daily solution with GAMIT software l ; 2) sidered, the simulated displacement field is as close to combining global H file with regional H file from IGS the actual one as possible. By comparing the results be­ sites by GLOBK; 3 ) calculateing time series with fore and after deducting the earthquake effect from the GRED to check the repeatability of baseline and coor­ velocity and strain fields, we found this method to be dinate and recalculating when the standard is not satis­ quite satisfactory. fied; 4) deleting the data that are still substandard ; 5 ) calculating site velocity with GWBK. In processing 2 Calculation method GAMT data, about 10 to 15 sites in and near China ma­ inland were used. The constraint of IGS site coordinate The observed crustal displacement ( D,) in and near the is X =0. 10 m,Y =0. 10 m,Z =0. 10m; the constraint of regional site coordinate is = I. 0 m , = I. 0 m, epicentral area of an earthquake consists of normal tec­ X Y Z =I. 0 m; and the constraint of satellite obit is 10 -•. tonic displacement ( D t ) , co-seismic displacement (D,), and post-seismic displacement ( D,), as ex­ The steps of calculating velocity are as follows : pressed by: I) Combining H file of global IGS site with regional H file by GWBK, setting the constraint of site coordi­ (1) nate asX=lO m,Y=lO m,Z=IO m,and setting the

This relation can be rewritten as : constraint of velocity in three directions as 0 m. 2) Calculating velocity field with GWRG, choosing D, =D,- (D, +D,) (2) ten sites of steady velocity located in Eurasian plate Speicfic calculation procedure is as follows : 1 ) The surrounding China, taking the velocity from ITRF2000 observed crustal movement velocity is calculated with as an approximate value, and taking the coordinate and the GLOBK software without considering the earth­ twice of the velocity errors from ITRF2000 as con­ quake effect1'l. 2) From the observed results, the co­ straint. The constraints of site coordinate and velocity seismic displacement is calculated with the Okada are 1. 0 m and 0. 1 m,respectively. model 1'l. 3) The post-earthquake displacement is esti­ 3 ) The velocity calculated by step 2 ( deducting the mated with Pollitz ' s stratified spheroid model. 4 ) By affect by Eurasian plate rotation) is taken as the veloci­ deducting the velocity caused by the co-seismic and ty value of regional crustal movement. The constraint of post-earthquake deformation from the observated veloc­ IGS site near the focal region is taken to be that of re­ ity , the velocity field unaffected by the co-seismic gional site in the above steps. Otherwise the calculated movement is obtained 1•.s]. In the present study, we did velocity would be unrealistic , for the earthquake caused not deduct the post-earthquake effect. displacement of the GPS site is very large. The effect of earthquake is ignored in the velocity field calculation 3 Data processing with GWBK. Fig. I and fig. 2 show, respectively, the obtained ve­ GPS observations were carried out in two stages before locity field of crustal movement in the northeastern edge 36 Geodesy and Geodynamics Vol. 1

40"N ~ 1 ,_} t I~yu~ " 1 \ ~ ~ aotoa / ~~~~ -:J{q~- . . ' / ~- -- ~ i:( 3 ~ ...._''(2 _ / / ~ Wuhai Dongsheng • • , Al:Xl _ .I s~ ~h~gye~ . 1 • asnanzuoqt 38"N f .., ~~ Jinchav-g 1 Y." / lt... 6( "'- - ' ..f; 'YYinchuan Yulin ~ue;m_r / ~---::,.. '-.:; , ,} Wuwei 1Wuzhong 1 // ~ ' ~ \ \ { __;-- -~.l- ~ ....: -- ; - " -"-( 1~~· ~ \ 9 36"N 1 ~--4 ~_{Xini~~~1:~, ~ \\ -Gon~ _; __.. • -Bai{~ Yan·an . ---....._ --- ...:::_--.L~o~ " Guyuan ~~~ ~ ' Ton~ -~~a .D~gxi \ ~\ \Pinglian~feng ~~ ~ "- - -~ "\\ " Tongch 34 • N \."'-..._'"\_ , 1 '- ~ 1 """"- - -.... '-.. \.~iansh ~ .. . We_,-- ~ (!)Altyn fault '\ Sh ~ ~B a?1_1 ~y~ man... eet (2)Norlh side fault ofHexi Corridor ./"' '- \. ' ~~ City (3)Norlhern edge region fault ofQilian Mountain _).:__ Shangzho Xi" an (4)Qilian Mountain fault \ Chen!;;JUilll __ main fault (5)Thole Mountain-Lnglong Ridge fault '\ \ - _ velocityvector (6)NorlhsidefaultofQilianMountain "- ~ (7)Huangcheog-Ta'er zhuan fault Hanzho~ 10 mm/a (8)Riyue Mountain-Laji Mountain e Earthquake (9)E'la Mountain fault Ankang (IO)Xiangshan Mountain-Rianjing Mountain fault Guangyuan ---..._ --- motive edge zone (II )Haiyuan fault ""- 96"E 98"E lOO"E 102"E 104"E 106"E 108"E llO"E Figure 1 Horizontal movement rate field in northeastern margin of Qinghai-Tibet block( 199 -10 -2001 - 09) region of the Qinghai-Tibet plateau from 1999 to 2001 time period between two sets of GPS observation. and 2001 to 2003. 3. 3 Strain calculation 3. 2 Calculation of co-seismic displacement and We first fitted the GPS horizontal velocity field with the velocity field after deducting co-seismic dis­ 14 15 cubic-spline model[ ' l, then interpolated velocities at placement. 16 grid points[ l ,and finally based on the strain-displace­ Before calculating the co-seismic displacement, the pa­ ment relation in elastic mechanics calculated the plane rameters of the earthquake fault model must be deter­ apparent strain during 2001 - 2003 in the northeast mined. Many researchers had obtained the fault param­ edge region of Qinghai-Tibet plateau by using the origi­ eters of this earthquake by inversion[? -IJ] • nal GPS horizontal velocity field and the velocity field Here we adopted the model parameters of refer­ after deducting co-seismic deformation. ence [!3] to calculate the co-seismic displacement , which is shown in Fig. 3. Although the co-seismic displace­ 4 Analysis of calculation result ment occurred instantaneously, its effect on the velocity calculated from the GPS data observed before and after 4.1 Regional horizontal velocity field before the the earthquake is related to the observation time. Kunlun earthquake The following formula is used to calculate the veloci­ ty of a GPS station after deducting co-seismic displace­ In order to understand the effect of the Kunlun earthquake ment: on regional velocity field, we need first to know the char­ (3) acteristics of velocity field before the earthquake. Fig. 1 shows the velocity field at and near the northeastern edge Where denotes the observed velocity ( calculated with P. region of the Qinghai-Tibet block relative to the Euro-Asi­ GLOBK) at j(j = 1 ,2, 3 ·· ·, m) site; V{, the velocity of an plate during the period of 1999 - 2001. This horizontal tectonic movement; D~ , the co-seismic displacement at movement has the following features : j (j =1 '2 '3 ... ' m) site from model calculation ; !':::,_ r ' 1 ) West of Wuwei, Menyuan, and Qinghai Lake, in- No. 1 Duxin Cui, et al. Effect of Kunlun Ms8. 1 earthquake on crustal deformation in northeastern edge region of Qinghai-Tibet plateau 37 eluding Qaidamu Basin , Qilian Mountain and Hexi Cor­ tion, and the eastern part moves towards SE. This GPS ridor, the crust basically moves in the northeast direc­ observation result is basically consistency with the as­ tion, and is accompanied by crustal shortening. The sumption that the Gansu -Qinghai block is in clockwise southern edge of Qaidamu Basin moves to the NE at 8 rotation. But the rotation involves only the eastern re­ mm/a. In the Tuole Mountain-Lenglong Ling fault gion of Gansu-Qinghai block, not the western region of zone , the north -eastern crustal movement velocity de­ Qilian Mountain and Qaidamu Basin. creases to 3 - 6 mm/ a. In Hexi Corridor region and 4. 2 Co-seismic displacement Alashan block, the crustal velocity is reduced signifi­ cantly. The velocity at the observation points in the Fig. 2 shows the co-seismic displacement of GPS obser­ block is very small ( mostly < 3 mm/ a) . There is no vation points caused by the Kunlun earthquake. From block rotation , and the movement direction is irregulari­ the figure it may be seen that the displacement is the ty and unsystematic. The crustal velocity gradually largest in Delinha area , and becomes smaller toward changes from northeast to southwest , showing that the east, west and north. The displacement is largest ( 20 - northern area of Qilian Mountain is probably in a 50 mm) in Qaidamu Basin south of the Qilian Moun­ process of strong compression, overthrust, and nappe. tainand , and reduces to 3 - 5 mm in Alashan block of 2) North of the main fault of Altyn fault, Duhuang north of Hexi Corridor. The displacement is small (less and Anxi , the observation points moved in the NNW di­ than 10 mm ) north of the Altyn fault and gets even rection. The velocity is generally about 6 mm/a, without smaller( generally less than 3 mm) southeast of Xining­ obvious differential movement on different sides of the Wuwei and Ningxia area. Therefore, the main areas af­ fault. fected by the co-seismic movement are located in the 3 ) The eastern region of Gansu -Qinghai block , from east of Altyn fault , south of Alashan block and west of east of Qaidamu Basin to Liupan Mountain, mainly ro­ Wuwei -Gonghe. tates clockwise. The western part moves towards NE, The degree of influence of the earthquake on velocity the center part moves approximately in the EW direc- field is inversely proportional to the time interval of the

'-=------

40°N

·v· · · ~ ·~Jiayu~ " /'. A ~ ' "· ~~~i~ Ia 38oN ------,' ,,.!'::~:~~~~~''\;;.~ . """" block / .-< • ..., ~ ~ Jmcb:ang / 6( )..,: ~ ~ ,/' y<'v / • Delingha -~ · Wuwei ,,/ /' .__,._;;."' ,,, .-<- . ~ ~ " . ) 1) ; · Ms8.1 ·-""' ~-- ..,~~v _.. / / ,< · ,, · ~/' en~ ---- -(10\ 36oN ,.V ,('> ~ ., , , 8) • .,, /-(· - ~ ~ / / {9~· ,.. Xining ~ 1) / .-<· \ Goni!:be Baiyhl <.., ------"' ' Sub- ock orelistem part -...Lanzhou

o ~ ~~ ...___~ \ Tongren Linxi~ Dingxi 34 N (!)Altyn fault " Sheet (2)North side fault ofHexi Corridor Xi'an City (3)NorthernedgercgionfaultofQilianMountain ~ (4)Qilian Mountain fault ~ --main fault (S)Tuole Mountain-Lnglung Ridge fault 6lNorth side fault of Qilian Mountain • • A' velocity vector 7 Hnonacheng-Ta'er zhuan fault ~ 20mm/a ~ ~ ~~~~':!l,~~jiMountain ~ • Earthquake IO)Xiangshan Mountain-Rianjing Mountain fault \.\" --- motive edge zone 11 )Haiyuan fault . ~ J 96°E 98°E lOOoE 102°E l04°E 106°E 108°E llOoE

Figure 2 Simulated co-seismic horizontal displacement field caused by Kunlun Ms8. 1 earthquake in northeastern margin of Qinghai-Tibet block 38 Geodesy and Geodynamics Vol. 1

8 data used in calculation. For the velocity field of 2001 surface extension rate reaches 40 X 10 - I a. The rate is 8 - 2003 , the most influenced area is Delinha , reaching 5 X 10 - I a from southern area of Qilian Mountain to 20 mm/ a; the least influenced area ( only about 1 mm/ Delinha. A small area around Delinha shows compres­ 8 a) is in southeastern Gansu and Ningxia area. The de­ sion, with a rate of - 6 X 10 - I a. Another small area gree of influence is spatially proportional to the dis­ showing compression is in the vicinity of Qinghai Lake. placement. The areas north of Qilian Mountain and Hexi Corridor show weak compression. According to this deformation 4. 3 Post-earthquake velocity field pattern , the effect of the Kunlun earthquake on the The horizontal velocity field in the northeastern edge whole region is mainly extensional. region of Qinghai-Tibet block during 2001 - 2003 is 4. 4 Velocity field after deducting co-seismic shown in Fig. 3. It may be seen that Qaidamu Basin, deformation Qilian Mountain and Hexi Corridor in western area of The velocity field after dedecting the co-se1sm1c de­ the region and southern edge of Alashan block move to­ formation is shown in Fig. 5 , which shows that most are­ wards southwest. The crustal velocity is very small in as south of Qilian Mountain move towards northeast. the range of Delinha-Gonghe-Zhangye of Gansu-Qing­ The movement velocity is small in northern area of Qil­ hai block, and gradually increases in the range of Qil­ ian Mountain and Hexi Corridor. The northeastern crus­ ian Mountain and Qaidamu Basin from north to south, tal movement velocity gradually decreases from Qaid­ showing a feature of extension. The southeastern Gansu amu Basin to the south edge of Alashan block. The and Ningxia area mainly rotates clockwise. By compa­ southeastern Gansu and Ningxia area mainly rotates ring the co-seismic displacement ( Fig. 2 ) with the ve­ clockwise , similar to the horizontal movement before the locity field ( Fig. 1 ) before the earthquake , it becomes Kunlun earthquake ( Fig. 1 ) . As shown in Fig. 6 , most of clear that the velocity field of the western area is signif­ the western area of the northeastern edge region of Qing­ icantly affected by the earthquake. Fig. 4 shows the hai-Tibet block has negative values, especially the area horizontal surface extension rate during 2001 - 2003. near Delinha , where the surface extension rate reaches Most of the western area has positive values, especially 8 - 15 X 1 0 - I a. The compression zone includes Qinghai the area from south of Delinha to Geermu , where the

40oN ~ "" ~--Baotou- __-{ l ~~ayuojiai / ~ '"' ' ~~~2h , // Ww,.; DoosMoo 5 38°N // Jinchang \ ~ (6.)_~~ "- L~ ~ . A//~Yinchuan kzuoqi .Yulin Jf' Del~a "' - ~~"- "v Wuwei \ \vuzhong ' -- --\ s)~ ~~ ------( la~

36oN \ -.,Xinin: ~~ ~ ~ 11) ~ . Yan·an ~ ~ "'Gongh~ " ~ L=~: '\ G yuan ~ \ Tongren Linxi~ 0 . . ~· u. . Xifeng ~ ~ · Ingxi .Pmghang ~~ Tongchu 34oN ~~~""'-""'- ~ ~ ~ Tiansh ~ . ..r-- Sheet;2~Wh~\'i,tfaultofHexiCorridor ,./'/ ~ltao'i 'i(i;ny~n)}'em~m-- . C'ty 3 Northern edge region fault of Qilian Mountain ~ Xran 1 4 Qilian Mountain fault . Chengxian Shangzho main fault 5 Tuole Mountain-Lnlll~ng Ridge fault --=:::::- 6 North side fault ofQilian Mountain ------velocity vector Huangcheng· Ta'er zhuan fault ~ 10 mm/a 8 Riyue Mountain-Laji Mountain Hanzhong e Earth uak 9 E'la Mountain fault ~- Ankang . qed e IO)Xiangshan Mountain-Rianjiug Mountain fault ~ / Guangyuan -...... ---motive ge zone 11 )Haiyuan fault '- 96oE 98oE lOOoE 102oE 104oE 106oE 108oE llOoE

Figure 3 Horizontal movement velocity field in northeastern margin of Qinghai-Tibet block No. 1 Duxin Cui, et al. Effect of Kunlun Ms8. 1 earthquake on crustal deformation in northeastern edge region of Qinghai-Tibet plateau 39

Lake , where the surface extension rate reaches to - 5 Ningxia area. By companng with Fig. 3 and Fig. 4, 8 X 10 - I a. The areas of north to Qilian Mountain and which show patterns without deducting the co-seismic Hexi Corridor are mainly under weak compression. The deformation, two major differences may be found: ( 1 ) biggest surface extension rate is in Zhangye ( - 5 X The overall crustal-movement direction in the western 10 - 8 I a) . The absolute value of surface extension rate area of the region is opposite to each other , with the 8 1s about 2 X 10 - I a in the southeastern Gansu and former moveing to northeast , and the latter to southeast.

40"N ~ 1 ,J t 1~yu~ 1 ~ ~aotoo / ~4Jlq~ ,- ' / ~- ~~;~~21 ·- . / / ~ Wuhai Dongshen

"-. ~1 ~ :Zhangye • "--7; I , , ~~ . J-. h ' I .Affishkuoqi 38"N -1 ..._ ~ me ap.g J ~ ~ , ' ~ - Yinchuan / e1illlilia (6\: _ '-.: / ' . , Yulin fo , 7 I/,/ ~ ~ ~7 }Vu~e' _ ' ~ ~uzhong { ,., ~ enyu~ - -- >----(lb:~l \ 36"N 19~-~~J;ini ~~~ l ~~ \ Yan·an ~~ \" - ~ ::?::-;_~~ ~ GUY"'" ' Ton~ -_!:~a -D~g»,i \~~ Pingli~ifeng -. ._~ \. \ '\. ' '\. Tongch 34"N ~ ~~ ~ ~ - ._ "\.~ian\h ~ ~ 1 tyn fault ~ '\ ~B agji ~. yd\"eman... Sh eet 2 NorlhsidefaultofHexiCorridor ...... -. ..._ \. ' -.....::::s,_ ~ City 3 Norlhem edge region fault ofQilian Mountain _\.: Sh gzh Xi·an 4 Qilian Mountain fault \ Chen15 ...~an an o --main fault 5 Tuole Mountain-L~ong Ridge fault '\ \ - 6 North side fault of ilian Mountain ~ - velocity vector 7 Huangcheng-Ta'er uanfault H h ~ !Omm/a ~ ~?."~~~=;!J~iiMountain anz on '--Ankan • Earlhqualre IO)Xiangshan Mountain-Rianjing Mountain fault Guangyuan --..._ g --- motive edge zone 11 )Haiyuan fault "- 96"E 98"E lOO"E 102"E 104"E 106"E 108"E llO"E Figure 4 Contour map of horizontal surface expansion rate in northeastern margin of Qinghai-Tibet block ( 2001 - 09 - 2003 - 09)

96"E 98"E 100"E 102"E 104"E 106"E 108"E llO"E

Figure 5 Horizontal movement velocity field deducted co-seismic deformation in northeastern margin of Qinghai-Tibet block 40 Geodesy and Geodynamics Vol. 1

'-=------

40"N

A

38"N --

36"N

Tongren Linxia" Dingxi

JG'an City --main fault -- velocity vector 20mm/a e Earthquake motive edge zone J 96"E 98"E lOO"E 102"E 104"E 106"E 108"E llO"E

Figure 6 Contour map of horizontal surface expansion rate deducted co-seismic deformation in northeastern margin of Qinghai-Tibet block during 2001 -2003

( 2) The former shows mainly an overall compression of velocity contains an element of post-earthquake relaxa­ crustal movement in the western area of the region ; tion. Secondly, the Okada model that we used may not whereas the latter shows mainly extension movement. be sufficiently realistic ,for it assumes the crust to be i­ However, they have two similarities : The velocities are sotropic and elastic , without considering the effect of comparable in Alashan block , and the movements are non-seismogenic faults, which do exist in the region. characterized by crustal shortening. The eastern area of ( 4 ) The appearance of a large compression zone in the region, including mainly southeastern Gansu and Qinghai Lake is due to its location at the border zone Ningxia area, shows clockwise rotation. between the western and the eastern areas of the Gan­

The similarities and dissimilarities between the de­ su -Qinghai block [ !7] • ( 5 ) The appearance of a large ar­ formation figures before and after deducting co-seismic ea of crustal compressive movement in the mid-eastern deformation may be due to several causes : ( 1 ) The op­ area of Hexi Corridor and northern area of Qilian posite movement directions shows that the co-seismic Mountain may be due to the increase of the principal deformation has a large influence in the western area , compressive strain in the northeast direction before the whereas the effect in the eastern and northern areas is Mingle-Shandan Ms6. 1 earthquake. much smaller, on account of the larger distance from The similarity between the pre-and post-earthquake the causative fault. ( 2) The strong compression zone in velocity fields, after deducting the co-seismic deforma­ Delinha is probably caused by co-seismic deformation tion, together with the consistency with regional earth­ of an M s6. 6 earthquake in April 2002. ( 3 ) In the west­ quake activity indicate reasonableness of the above­ ern part of the region , the two figures show that the mentioned calculation approach. crustal movement are roughly identical in direction, but much different in speed. This difference may be due to: 5 Summary Firstly , before the Kunlun earthquake , the regional strain energy was accumulating, whereas after the earth­ 1) The effect of Kunlun Ms8. 1 earthquake on the crus­ quake the strain energy was relieved and the observed tal displacement field in the northeastern-edge region of No. I Iluxin lli, et al.Effect of Kunhm Ms8.1 earthquake oo crustal dd"anmtioo in mrtlreastem ecJw:: "'19m of Qingbal-Tiliet plateau 41

Qinghai-Tibet plateau is remarkable. The most affected gram version 4.10. Cambridge: Massachusetts Institute of Technol­ place is in the western area of the region, including Qil­ ogy,2000. [2] Dong D L. QOCA software online manu. http://gipsy. jpl. nasa. ian Mountain , south of Altyn fault and west of Delinha. gov/qoca. The effect on displacement is about 20 - 50 mm. It re­ [ 3] Okada and Yoshimitsu Y. Surface deformation due to shear and duces to 3 - 5 mm in Hexi Corridor and north to Alas­ tensile faults in a baH-space. Bull. Seismol. Soc. Am. ,1985, 75 han blocks , and it is much smaller, no more than 3 (4),1135-1154. mm, in southeast Gansu (east of Xinin-Wuwei) and [ 4] Pollitz F F. Post-seismic relaxation theory on the spherical Earth. Bull. Seismol. Soc. Am. ,1992,82:422 -453. Linxia. The effect on the velocity field is as much as 20 [ 5] Pollitz F F. Gmvitational viscoelastic postseismic relaxation on a mm/a during 2001 -2003. The spatial distribution of layered spherical Earth. J. Geophysics Res. , 1997 , 102 : 17 921 - the effect on velocity is in proportion with the effect on 17 941. displacement. [ 6] King R W wtd Bock Y. Documentation for the GAMIT GPS Wl a­ 2) The earthquake fault parameters are obtained by nalysis software version 10. 3. Cambridge: Massachusetts Institute of Teclmology ,2000. inversion of near-field deformation data. By using these [7] Wwt Yongge,et al. Co-seismic slip distribution of the 2001 west of parameters, we have modeled the far-field co-seismic Kunlun mountain pass earthquake inverted by GPS and leveling displacement, and then deducted effect of co-seismic data. Sei•milogy ond Geology, 2004,26 ( 3 ) , 397 - 404. ( in Chi­ displacement from the observed far-field velocity field. nese) The velocity field calculated by this method has greatly [ 8] Chen Jie, et al. Surface rupture zones of the 2001 earthquake Ms8. 1 west of Kunlwt Pass in northern Qinghai-Tibet Plateau. reduced the effect of co-seismic deformation. The result Quatemruy Scienoes,23(6) ,629-639. (in Chin...,) is helpful in determining earthquake occurrence tend­ [ 9] Xu Xiwei, et al. Characteristics features of the surface ruptures of ency , and may serve as a possible way of study GPS da­ the Ku Sai Hu(Kunlw>aium)emthquake(M.8.1) ,northern Tibet­ ta when data is insufficient after a strong earthquake. "' plateau , China. Seismology ....! Geology , 2002 , 24 (I ) , I - 13. 3 ) The velocity field after deducting the effect of co­ ( in Chinese) seismic displacement is generally similar to the velocity [ 10] Department of Earthquake Monitoring and Prediction, CEA. The 2001 Ms8. 1 large earthquake, west of the Kunlwt pass. Beijing: field before the earthquake. Before deducting co-seis­ Seismological Press ,2002. (in Chinese) mic displacement, the regional strain field shows mostly [ 11 ] Shwt Xinjian ,Li. Jianhua and Ma Chao. Application of SPOT and extension, whereas after deducting co-seismic displace­ IKONOS images to the study of the west of lrunlunshan pass ment , the strain field shows mainly compression. Tills earthquake(M.8. I). Seismology and Geology ,2005,27 (I) , 147 result is consistent with the earthquake activity in this -152. (in Chinese) [ 12 ] Zhang Xiaoliang, et al. Simulation and analysis of co-seismic de­ region,including the occurrence of the Shandan-Miule fonnation field of Ms8.1 earthquake in western kunlun mountain Ms6. 1 and M.S. 8 earthquakes in October 2003. pass. Journal of Geodesy and Geodynamics,2007 ,27 (2) : 16 - 4) In the inversion process , the accuracy of the earth­ 20. (in Chinese) quake fault parameters used is critical. If the error IS [ 13] Tan Kai, Wang Qi and Shen chunyang. Using geodetic data to in­ large, the reliability of calculation is reduced. verse co-seismic dislocation of 2001 Kunlwt earthquake. Journal of Geode.,. ....! Geodynami.,., 2004,24 ( 3) , 47 -50. (in Chi­ 5) The Okada model is derived in assumption of uni­ n""') form elastic half space , while in reality the crust often [ 14] Yang Yunfeng,Su zhi:nm and Hu Jinyan. A new edge-holding al­ has faults , which may accommodate some of the co - gorithm of image interpolation. Journal of Image wtd Gmphic, seismic displacement. This may be maiuly responsible 2005,10(10) ,1249-1251. (in Chine•e) for the inconsistency between the calculated and ob­ [ 15] Yang Shaomin, et al. Contemporary horizontal tectonic defonnation fields in China invemed from GPS observations. Journal of Geode­ served values. 'Y and Geodynamic•,2002,22( I) ,68 -72. (in Chinese) 6) The study did not consider the effect of post-earth­ [ 16] Tao Benzao. Adjustment of free net and analysis of defonnation. quake deformation. Thus further research is needed. Beijing: Surveying and Mapping Press ,2001. (in Chinese) [17] Cui Duxin, et al. Detennination of movement boundary belt of References Gansu-Qinghai block. Journal of Geodesy and Geodynamics, 2007 ,27( I) ,24-30. (in Chinese)

[I] Hening T A. Global Kalman fdter VLBI ....! GPS on onall"i' pro-