Science in China Series D: Earth Sciences © 2008

www.scichina.com earth.scichina.com www.springerlink.com

SCIENCE IN CHINA PRESS

Springer

Seismic anisotropy of upper mantle in Sichuan and adjacent regions CHANG LiJun, WANG ChunYong† & DING ZhiFeng Institute of Geophysics, China Earthquake Administration, Beijing 100081, China

Based on the polarization analysis of teleseismic SKS waveform data recorded at 94 broadband seismic stations in Sichuan and adjacent regions, the SKS fast-wave direction and the delay time between the fast and slow shear waves were determined at each station using the grid searching method of minimum transverse energy and the stacking analysis method, and the image of upper mantle anisotropy was acquired. The fast-wave polarization directions are mainly NW-SE in the study area, NWW-SEE to its northeast and NS to its west. The delay time falls into the interval [0.47 s, 1.68 s]. The spatial variation of the fast-wave directions is similar to the variation of GPS velocity directions. The anisotropic image indicates that the regional tectonic stress field has resulted in deformation and flow of upper mantle material, and made the alignment of upper mantle peridotite lattice parallel to the direction of material deformation. The crust-upper mantle deformation in Sichuan and adjacent regions accords with the mode of vertically coherent deformation. In the eastern Tibetan Plateau, the crustal material was extruded to east or southeast due to SE traction force of the upper mantle material. The extrusion might be obstructed by a rigid block under the Sichuan Basin and the crust has been deformed. After a long-term accumulation of tectonic strain energy, the accumulative energy suddenly released in Yingxiu town of the Longmenshan region, and Wenchuan MS8.0 earthquake occurred. upper mantle anisotropy, SKS wave, fast-wave polarization direction, lithospheric deformation, Wenchuan MS8.0 earthquake

Sichuan and adjacent regions are situated in the tectonically-complicated eastern Tibetan Plateau. It belongs to three tectonic domains: the Tethyan-Himalayan tectonic domain in the west, the Pacific Ocean tectonic domain in the east, and the Paleo-Asiatic tectonic domain in the north[1]. The study area is divided by the boundary of Beichuan-Wenchuan-Kangding-Xiaojinhe into the eastern part (the Yangtze block) and the western part (the Songpan-Garze block and the Sanjiang (Jinsha, Lancang and Nujiang Rivers) fold system) (Figure 1). Sichuan and adjacent regions are located in the central segment of the North-South Seismic Belt, where numerous strong earthquakes had occurred. Wenchuan MS8.0 earthquake of May 12, 2008 caused serious injuries and deaths of the peoples and losses of property. The Tibetan Plateau is the result of collision between the Indian and Eurasian plates, which began at about 50

Ma, and has resulted in the convergence of more than 2000 km, and the uplift of the plateau to elevation of 4―5 km[2]. The Songpan-Garze block and the Longmenshan fault zone in the eastern margin of the Tibetan Plateau have experienced the strong crustal deformation and faulting. Recently, some papers respectively presented the evolution models of the Tibetan Plateau, i.e. the ductile flow in lower crust[3] and the vertically coherent deformation[4,5]. Ductile flow in the lower crust implies that the lower crust is weak, where material flows in ductile pattern. Vertically coherent deformation Received July 28, 2008; accepted October 7, 2008 doi: 10.1007/s11430-008-0147-8 † Corresponding author (email: [email protected]) Supported by the National Natural Science Foundation of China (Grant Nos. 40334041 and 40774037), the Special Project for the Fundamental R & D of Institute of Geophysics, China Earthquake Administration (Grant No. DQJB06B06) and the Special Program of the Ministry of Science and Technology of China (Grant No. 2006FY110100)

Sci China Ser D-Earth Sci | Dec. 2008 | vol. 51 | no. 12 | 1683-1693

Figure 1 Regional geologic setting of the study area and location of stations. F1, Longmenshan fault; F2, Xianshuihe fault; F3, Anninghe-Zemuhe fault; F4, Xiaojiang fault; F5, Huayingshan fault; F6, Litang-Dewu fault; F7, Jinshajiang fault; F8, Lancangjiang fault; F9, Nujiang fault; F10, Zhufeng fault; F11, East Kunlun fault; F12, Wudou-Diebu fault; F13, Zhouqu-Liangdang fault; F14, West Qinling North Margin fault.

indicates the crust-mantle deformation is coherence, which plays a significant role in the study of evolution of the Tibetan Plateau. The observations of Lev et al.[6] in eastern Tibet support the model that allows differential movement of upper crust relative to lithospheric mantle. In the northern part of the region the relationships are more ambiguous due to limited data, and they infer that coherent deformation of the crust and mantle lithosphere cannot be excluded. Seismic anisotropy is an effective method for understanding of the crust and upper mantle deformation. The anisotropy can be used to understand the intraplate deformation features, and the deformation status of lithospheric mantle related to the plate tectonic movements. The anisotropy in the upper crust is generally considered as the indication of the orientation alignment of numerous cracks under the action of stress, whereas the anisotropy in the mid-lower crust is mainly determined by the lattice-preferred orientation of anisotropic mineral (such as biotite and hornbilende)[7]. Some studies con1684

cluded that the crustal anisotropic fast-wave direction is basically consistent with the strike of the active faults, and is related with the regional horizontal principal compressive stress, which is conducive to the study of crustal movement and earthquake monitoring[8,9]. The mantle anisotropy is generally considered to be caused by the lattice-preferred orientation of olivine crystals[10]. The plate motion is a direct cause resulting in the mantle deformation. The size and direction of the mantle anisotropy strongly depend on the velocity and direction of the plate motion. Many anisotropic studies have been done that are related to the mantle deformation beneath the Tibetan plateau, and carried out a lot of relevant anisotropic stud― ies[11 17]. The SKS wave splitting results were obtained at some seismic stations in Sichuan and adjacent re― gions[15 17], but there are no more broadband seismic stations. Up to now, 52 broadband seismic stations have been deployed and operated in Sichuan, with relatively even station distribution and high data quality, facilitat-

CHANG LiJun et al. Sci China Ser D-Earth Sci | Dec. 2008 | vol. 51 | no. 12 | 1683-1693

ing the studies on the upper mantle anisotropy. Combined with the geological, geodesic and geophysical data (GPS, crustal stress field, Quaternary fault slip rates, etc.), the upper mantle anisotropies are used to discuss the continental dynamic issues, such as the mechanism of continental strong earthquakes and the coupling of crust and mantle.

1 Seismic data The observational data come from the digital records at 84 permanent broadband seismic stations and 10 temporary broadband digital seismic stations in Sichuan and adjacent regions (Figure 1 and Table 1). The teleseismic events were collected at distance between 85° and 110° and with magnitude MS>6.0 for shallow earthquake (focal depth 5.5 for deep earthquake (focal depth >150 km). Sixty teleseismic events were collected in this study (Figure 2). Most events occurred in the Fiji-Tonga area in South Pacific Ocean. On the whole, the azimuthal distribution of events meets the requirements of this study.

Figure 2

Epicenters of events (black dots) used in the study.

2 Analysis method The SKS splitting analysis method includes two steps. First, we use the SC method[10], i.e. the grid searching method of the minimum transverse energy, to measure the anisotropic parameters and their errors from individual event with different azimuth. Then, a group of

anisotropic parameters from all events at a station are stacked by the splitting parameters method similar to that presented by Vinnik et al.[18]. The stacking analysis is an effective method to increase the accuracy of SKS splitting measurement at a station[19]. In this study, the splitting parameters of teleseismic SKS waves of each station were obtained by use of the stacking analysis method. When the epicentral distances at 85°―110°, SKS wave travels along a ray path nearly-vertically incident on the ground. Assuming that the media is isotropic, when SKS wave penetrates the core-mantle boundary, it is radially polarized, with radial component (SV wave), and without transverse component (SH wave). When there are anisotropic layers in mantle, SKS splits into two phases with different velocities (i.e., fast and slow waves) traveling at directions orthogonal with each other, and then the particle motion changes from the originally linear polarization to elliptic polarization. The parameter pair (ϕ, δt) is generally used to describe seismic anisotropy, where ϕ denotes the fast-wave polarization direction, and δt is the delay time between fast and slow waves. We can describe S wave splitting using the parameters of (ϕ, δt). Theoretically, the transverse component is zero before the SKS wave propagates through anisotropic layer. If we can find the pair (ϕ, δt) which corresponds with the minimum horizontal transverse component, the anisotropy parameters of station are the pair (ϕ, δt). The SC method is a method of measuring the anisotropic parameters (ϕ, δt) beneath seismic stations based on the theory of the minimum transverse energy, and the errors of the pair (ϕ, δt) were determined by F-test, where 95% confidence was used to the error estimation. For the stacking analysis, we first used the SC method to measure the anisotropic parameters and their errors determined by individual event with different azimuth. Then, a group of anisotropic parameters from all events at the station are stacked. Figure 3 shows an example of SKS splitting analysis for Fiji earthquake (21: 04 on 2007-10-16 with Mw6.6) at station MNI. In Figure 3(a), the distinct transverse component denotes the SKS-wave splitting when it propagates through an anisotropic medium, and the particle motion is elliptical (Figure 3(c)). Contour plot (Figure 3(e)) of transverse energy on the corrected SKS transverse component was computed through the entire range (ϕ, δt). From Figure 3(e), we can find the minimum (denoted by star), and

CHANG LiJun et al. Sci China Ser D-Earth Sci | Dec. 2008 | vol. 51 | no. 12 | 1683-1693

1685

Table 1 Code AXI BTA BYD BZH DFU EMS GDS GZA GZI HLI HMS HSH HWS HYS JJS JLI JLO JMG JYA LBO LGH LTA LZZ MBI MDS MEK MGU MLI MNI MXI PGE PWU QCH REG RTA SMI SMK SPA WCH WMP XCH XCO XHA XJI XJP XSB YGD YJI YYC YYU YZP ZJG CHX DBU MIX

Station Anxian Batang Bingyidi Bazhong Daofu Emeishan Guodashan Guza Ganzi Huili Huamashi Heishui Hanwangshan Hongyashan Jinjisi Junlian Jiulong Jianmenguan Jingyan Leibo Luguhu Litang Laozhaizi Mabian Mendingshan Maerkang Meigu Muli Mianning Maoxian Puge Pingwu Qingchuan Ruoergai Rangtang Shimian Shimenkan Songpan Wenchuan Wumaping Xiangcheng Xichong Xuanhan Xiaojin Xianjiaping Xuanshengba Youguanding Yajiang Yuanyichang Yanyuan Youzhaping Zhongjiagou Chengxian Diebu Minxian

Region B C B B C B C C C C B B B B B B C B B B C C B B B B B C C B C B B A B C C B B B C B B B C C B C C C B B A A A

Splitting parameters for SKS phase in the study areaa)

ϕ(°) 118.5 129.7 142.9 106.7 159.4 110.4 12.2 147.8 170.1 76.5 135.1 122.6 152.2 128.6 115.8 123.0 165.7 109.2 128.7 160.0 43.2 96.9 123.8 139.5 173.9 144.4 152.2 175.7 173.7 101.6 1.7 112.1 43 123.5 155.1 109.9 132.8 72.3 11.6 131.3 2.0 100.9 109.8 141.7 168.5 162.4 121.5 86.2 140.1 129.2 2.9 119.9 115.9 114.5 119.7

Δϕ(°) 6.9 5.4 4.9 6.7 7.6 7.3 3.7 6.6 4.1 8．1 6.7 2.6 6.5 3.7 3.0 3.4 9.3 2.9 3.7 5.1 3.7 5.3 1.1 2.0 10.8 6.9 8.6 21.4 5.4 6.9 5.3 5.1 3.7 2.3 6 2.0 9.9 5.8 10.8 2.5 5.9 2.8 4.7 4.5 4.5 8.3 9.3 3.2 5.5 4.9 19.3 3.5 1.8 3.1 3.4

δ t(s) 0.95 0.62 0.57 0.90 0.85 0.62 0.60 0.65 1.50 0.57 0.87 1.32 0.58 0.67 0.68 0.95 0.76 0.72 0.62 0.68 1.04 0.94 1.41 0.73 0.71 0.89 0.67 0.88 0.75 0.66 0.67 0.88 1.01 1.43 1.16 0.90 0.79 0.88 0.86 0.87 0.94 0.59 0.65 0.65 0.50 0.78 1.08 1.38 0.92 0.63 0.64 1.36 1.03 1.16 0.70

Δδt (s) 0.13 0.10 0.05 0.13 0.08 0.15 0.13 0.19 0.08 0.06 0.33 0.28 0.05 0.10 0.12 0.13 0.10 0.05 0.10 0.05 0.21 0.13 0.25 0.05 0.26 0.10 0.08 0.37 0.08 0.18 0.05 0.11 0.13 0.11 0.07 0.13 0.32 0.15 0.42 0.05 0.12 0.07 0.08 0.12 0.01 0.06 0.21 0.10 0.11 0.12 0.10 0.14 0.13 0.13 0.13

No. 7 10 9 11 9 9 4 11 13 13 8 7 12 11 9 10 13 12 8 11 9 13 5 12 9 11 7 5 10 3 15 13 11 14 18 11 7 10 6 11 11 12 14 8 12 9 6 17 8 9 3 3 9 16 3

Remark 1 1, [16] 1 1 1, [16] 1 1, [16] 1, [16] 1, [16] 1 1, [16] 1 1, [16] 1 1, [16] 1 1 1 1 1 1 1, [16] 1, [16] 1 1 1, [16] 1 1 1 1 1 1 1, [17] 1, [17] 1, [16,17] 1 1 1, [16,17] 1 1 1 1 1 1 1, [16] 1 1, [16] 1, [16] 1 1 1, [16] 1, [16] 2, [17] 2, [17] 2, [17]

(To be continued on the next page) 1686

CHANG LiJun et al. Sci China Ser D-Earth Sci | Dec. 2008 | vol. 51 | no. 12 | 1683-1693

(Continued) Region

ϕ(°)

Δϕ(°)

δ t(s)

Code

Station

MAQI

Maqin

A

131.8

3.0

0.82

MQU

Maqu

A

131.2

5.9

1.57

Δδt (s)

No.

Remark

0.10

4

2, [16,17]

0.20

8

2

TSH

Tianshui

A

121.3

3.3

1.12

0.24

5

2, [17]

WEX

Wenxian

A

97.9

6.0

1.1

0.18

6

2

WUD

Wudu Zhouqu

A

96.5

9.4

0.83

0.14

11

2, [17]

A

116.1

4.4

1.24

0.24

4

2

ZHQ ANK

Ankang

A

107.2

4.3

0.57

0.11

7

3

HZH

Hanzhong

A

46.6

4.4

0.70

0.13

9

3

LUY

Lueyang

A

110.3

13.5

0.83

0.16

2

3

MAX

Mianxian

A

73.1

11.4

0.50

0.08

7

3

TAB

Taibai

A

105.8

3.8

0.85

0.09

13

3

XIX

Xixiang

A

60.8

9.5

0.58

0.05

11

3

ZOZ

Zhouzhi

A

101.2

3.6

1.23

0.10

9

3

CHS

Changshou

B

141.8

7.7

0.57

0.10

5

4

FUL

Fuling

B

151.1

12.3

1.18

0.44

4

4

ROC

Rongchang

B

126.1

8.3

0.83

0.10

4

4

SHZ

Shizhu

B

169.4

6.1

0.72

0.20

5

4

SNB

Nanbin

B

100.7

2.2

0.89

0.08

7

4

WAZ

Wanzhou

B

138.4

10.6

0.82

0.25

2

4

WRC

Ruichi

B

140.8

5.6

0.47

0.11

5

4

WUL

Wulong

B

132.9

0.65

0.39

1

4

HEQ

Heqing

C

132.1

0.76

0.10

23

5, [15,16]

20 3.3

LIJ

Lijiang

C

8.2

2.6

1.68

0.13

5

5, [15,16]

YSH

Yongsheng

C

175.4

5.5

0.72

0.10

16

5, [15,16]

ZHD

Zhongdian

C

178.0

4.9

0.82

0.20

7

5, [15,16]

ZOT

Zhaotong

C

134.2

1.9

1.08

0.11

12

5, [15,16]

CD2

Chengdu

B

124.9

2.4

1.00

0.12

12

6, [15,16]

GYA

Guiyang

B

160.5

6.8

0.52

0.04

18

6, [16]

PZH

Panzhihua

C

108.7

4.9

1.18

0.21

15

6, [15,16]

DGN

Daguan

C

142.8

6.6

0.59

0.12

4

7, [15,16]

JCH

Jianchuan

C

98.0

10.0

0.88

0.20

3

7, [15,16]

NLG

Ninglang

C

18.0

4.6

1.35

0.08

2

7, [15,16]

DAB

Danba

B

142.1

3.9

1.03

0.15

4

8, [16]

DNB

Dengba

C

139.7

4.6

0.53

0.21

9

8, [16]

HLS

Honglashan

C

172.7

14.4

0.87

0.16

2

8, [16]

HNI

Heni

C

175.5

3.6

1.24

0.08

10

8, [16]

LUH

Luhuo

C

129.1

7.7

1.15

0.27

6

8, [16]

MNK

Mangkang

B

137.2

5.6

0.86

0.21

8

8, [16]

YID

Yidun

C

5.7

3.7

0.61

0.08

6

8, [16]

a) ϕ is fast-wave direction, Δϕ is the error of the fast-wave direction, δt is delay time, Δδt is the error of the delay time. Remark: 1, Sichuan regional digital seismograph network (52 stations from January 2007 to June 2008); 2, Gansu regional digital seismograph network (9 stations from July 2007 to June 2008); 3, Shaanxi regional digital seismograph network (7 stations from July 2007 to June 2008); 4, Chongqing regional digital seismograph network (8 stations from July 2007 to June 2008); 5, Yunnan regional digital seismograph network (5 stations from January 2002 to June 2008); 6, China National Digital Seismograph Network (3 stations from January 2003 to June 2008); 7, Sino-US temporary seismic observation in Yunnan (3 stations from November 2000 to April 2002); 8, Sino-US temporary seismic observation in western Sichuan and eastern Tibet (7 stations from July 2004 to April 2006). No. is event number used in SKS measurement. CHANG LiJun et al. Sci China Ser D-Earth Sci | Dec. 2008 | vol. 51 | no. 12 | 1683-1693

1687

Figure 3 An example of SKS splitting analysis. Original waveform (a) and particle motion (c) of the SKS phase in the radial-transverse coordinate system, corrected waveform (b) and particle motion (d) of the corrected SKS phase in the radial-transverse coordinate system, and contour plot of transverse energy (e).

the corresponding (ϕ, δt) is the anisotropic result of the station. The transverse component of corrected SKS phase (Figure 3(b)) is not distinct, and the particle motion (Figure 3(d)) of the corrected SKS phase is linear. Hence, the S wave splitting parameters are credible. Suppose that there are N events with different back-azimuths at a station. For ith event (i = 1,…, N), parameter pair (ϕ, δt) corresponds to its transverse energy Et(ϕ, δt)i. Each transverse energy Et(ϕ, δt)i is normalized with the minimum transverse energy, and then we sum Et(ϕ, δt)i by Et(ϕ, δt) =ΣEt(ϕ, δt)i. The mini1688

mum transverse energy of Et(ϕ, δt) is obtained by using the grid searching, and the corresponding parameter pair is the optimal anisotropic parameter pair beneath the station. The error estimation is obtained in a similar way. For SKS waves with different azimuths and different signal-noise ratios, this method increases the reliability of the splitting result, especially for the stations with a larger background noise. Figure 4 shows that the individual SKS splitting result of two teleseismic events (12:37 on 2007-08-28 with M w 6.1 and 09:47 on 2007-09-30 with Mw6.7), which were recorded at station

CHANG LiJun et al. Sci China Ser D-Earth Sci | Dec. 2008 | vol. 51 | no. 12 | 1683-1693

Figure 4 Transverse energy contour of SKS splitting analysis at station GZI from two teleseismic events (a) and (b), and the transverse energy contour of the result after stacking (c). The asterisk denotes the position of optimal parameter pair.

GZI, is compared to the result by using the stacking analysis. Obviously, the error of the stacked result decreases significantly, compared with the errors of the individual event.

3 Results of SKS splitting measurement and analysis The SKS splitting results of 94 stations in Sichuan and adjacent regions are listed in Table 1, where the splitting

parameters of 50 stations were recently obtained. The ― results of the rest stations were published[15 17], where we re-measure the results at some stations that have new seismic records. Figure 5 shows the scatter of shear-wave splitting parameters by stations. In general, the permanent stations have been kept in operation under the good observational environment with low background noise, hence lots of available teleseismic events were obtained, and thus the SKS splitting results have higher quality. Table 1 shows the results at most stations are good, in which the errors of azimuth are less than 10°, and the errors of delay time are less than 0.2 s. Despite the adverse observational environment at the temporary stations, we have collected many teleseismic records available for SKS splitting analysis at those stations in the observation of two years or so. Besides, the involvement of the stacking analysis method enables the measurements with good quality at most stations. Based on the polarization analysis from teleseismic SKS data recorded at the 94 stations in study area, we plot an anisotropic image of the upper mantle shown in Figure 6. The fast-wave polarization direction is mostly NW-SE. To be specific, it is NWW-SEE in the northeast of the study area, NW-SE in the middle, and NS in the west of the study area, which correspond to areas A, B, and C, respectively. However, there are some exceptions. The fast-wave directions at three stations (JCH, PZH and HLI) at the latitude 26.5°N and at two stations (LIT and YJI) at the latitude 30°N west of the study area are EW. The fast-wave directions at stations QCH, HZH, MAX and MIX in the east segment of the Longmenshan fault zone are NE or NEE, and the directions at stations YZP and WCH in the west segment are near NS. For the study area, the delay time between fast and slow waves falls into the interval [0.47 s, 1.68 s], and it is about 1.0 s at most stations. In the middle part, the average delay time at 9 stations to the west of the Longmenshan fault zone is 1.11 s, while it is 0.77 s at 14 stations to the east of the Longmenshan fault zone, which is less than that to the west.

4 Discussion and conclusions 4.1 Localization and thickness of anisotropic layer McNamara et al.[11] analyzed the splitting of PS converted wave on the Moho beneath the Tibetan Plateau, and obtained the delay time of 0.17―0.26 s caused by

CHANG LiJun et al. Sci China Ser D-Earth Sci | Dec. 2008 | vol. 51 | no. 12 | 1683-1693

1689

Figure 5 Results of anisotropy measurements of SKS splitting by stations. Radius of concentric circle denotes the delay time between fast and slow waves. Radius of the inner circle is 1s, outward circle is 2 s. The line direction denotes the fast-wave direction, the line length denotes the delay time.

1690

CHANG LiJun et al. Sci China Ser D-Earth Sci | Dec. 2008 | vol. 51 | no. 12 | 1683-1693

Figure 6 Anisotropic image of upper mantle beneath Sichuan and adjacent regions. The orientations and length of thick line indicate the fast-wave direc― tion and the delay time, respectively. Arrows denote geodetically measured surface velocities relative to the fixed Eurasian reference frame[27 29]. A, B and C are the regions with consistent fast-wave direction.

the crustal anisotropy, which is only about 1/5 of the total delay time of SKS splitting. The analysis of the contents of olivine and other mineral in upper mantle and laboratory measurement of olivine lattice inferred anisotropy degree of 0.04[11], in this case, delay time of 1s corresponds to thickness of 115 km. Thus we deduce the anisotropic thickness of upper mantle in the study area is about 54－193 km based on the delay time of 0.47－1.68 s in this study, and the average thickness is about 120 km. There exists a large variation of the thickness of anisotropic layer, which implies that the inside lithospheric deformation is asymmetrical beneath the study area. Other geophysical results show that the structures of the crust and upper mantle are obviously ― laterally inhomogenous[20 23]. According to the surface

wave anisotropy beneath China continent, there exists a stronger anisotropy at the depth of 70－150 km[24]. The lithosphere thickness of the Tibetan Plateau is about 160 －220 km, and the thickness of the Yangtze block is about 170 km[25]. Hence, we suggest that the anisotropy in Sichuan and adjacent regions mainly comes from upper mantle lithosphere. 4.2 Causes of anisotropy Anisotropy in the upper mantle is a result of strain- induced lattice preferred orientation of olivine crystals, and it reflects the past and present internal deformation of the subcontinental upper mantle by tectonic episodes. Hence, we can infer kinematic models of the medium of the earth’s interior[17]. For stable continental regions, it is

CHANG LiJun et al. Sci China Ser D-Earth Sci | Dec. 2008 | vol. 51 | no. 12 | 1683-1693

1691

interpreted as “fossil” anisotropy, whereas the anisotropy reflects present-day tectonic activities in tectonic active continental regions. The strike directions of major faults in the northeast and west parts of the study area are SEE and NS, respectively[26], where the fast-wave direction is consistent with the strike of the major faults. The fast-wave direction in the anisotropy is close to the trend of the velocity ― field of GPS measurements[27 29]. These indicate the identical deformation models of upper crust and upper mantle and the presence of vertically coherent deformation of crust-mantle in study area. The fast-wave directions of three stations (JCH, PZH and HLI) at the latitude 26.5°N are EW, which are perpendicular to the surface tectonic strike, this could be related to the Burma back-arc extension. It is similar with the results of Lev et ― al.[6]. Some studies[30 33] also suggested that the geophysical parameters (such as crustal thickness, Bouguer gravity anomaly, tectonic stress directions and crust-mantle velocity structure) vary significantly in this zone. The fast-wave directions of EW at stations LIT and YJI at the latitude 30°N are special phenomena, where the cause remains unknown. The anomalous fast-wave directions at stations QCH, HZH, MAX and MIX in the east segment and at stations YZP and WCH in the west segment of the Longmenshan fault zone are probably related to the local fault strike. 4.3 SKS splitting, mantle deformation field and Wenchuan MS8.0 earthquake The fast-wave direction in the area B in Figure 6 is consistent with the trend of the velocity field of GPS measurements. The vertically coherent deformation of the crust and mantle lithosphere simultaneously exists in the Songpan-Garze block and the Yangtze block. Although GPS data show no significant convergence between the eastern margin of the plateau and the Sichuan Basin, the strain calculation results[27] indicate that there is a significant compression strain rate component of 10.5 ± 2.8

1

nstrain/a in the region around the Longmenshan. The SKS splitting analysis indicates that the fast-wave polarization directions of NW-SE in the Yangtze block are consistent with the fast-wave directions at stations in the Songpan-Garze block, but the average delay time to the east of the Longmenshan fault zone is less than that to the west. The Longmenshan thrust nappe belt is the boundary between the Songpan-Garze block and the Yangtze block. Lou et al.[34] have studied the Bouguer gravity anomaly in Sichuan and adjacent regions using wavelet transform, and presented that the crustal density in the Sichuan Basin is higher than that in the Songpan-Garze block. According to crustal velocity structure, the average crustal velocity in the Sichuan Basin is greater than that in the Songpan-Garze block[35]. Therefore, the crust in the Yangtze block (Sichuan Basin) is more rigid than the crust in the Songpan-Garze block. Although the fast-wave directions are consistent with the trends of GPS velocity field in the Songpan-Garze block and Yangtze block, the strength of the crustmantle deformation in the Songpan-Garze block is greater than that in the Yangtze block. It implies that the Yangtze block play the role of the resistance of the lateral extrusion. In the eastern Tibetan Plateau, the crustal material was extruded to east or southeast due to SE traction force of the upper mantle material, and the extrusion might be obstructed by the rigid block under the Sichuan Basin. After a long-term accumulation of tectonic strain energy, the accumulative energy was suddenly released in Yingxiu town of the Longmenshan region, and Wenchuan MS8.0 earthquake occurred. The broadband seismic data of Sichuan Digital Seismograph Network in this study was provided by Lü Z Y and Dai S G. The data of the adjacent regional Network of Sichuan was provided by Data Management Center of China National Seismic Network at Institute of Geophysics, China Earthquake Administration. We are grateful to the help of Paul Silver, Lucy Flesch. We thank two anonymous reviewers for valuable comments and constructive criticism.

crust-mantle coupling in central Asia using GPS, geologic, and shear

Ren J S. The basic characteristics of the tectonic evolution of the continental lithosphere in China (in Chinese). Reg Geol China, 1991, 2: 289－293

2

wave splitting data. Earth Planet Sci Lett, 2005, 238: 248－268 Silver P G. Seismic anisotropy beneath the continents: Probing the

6

depths of geology. Ann Rev Earth Planet Sci, 1996, 24: 385－432 Lev E, Long M D, van der Hilst R D. Seismic anisotropy in Eastern

Molnar P, Tapponnier P. Active tectonics of Tibet. J Geophys Res, 1978, 83: 5361－5376

3

5

Royden L H, Burchfiel B C, King R W, et al. Surface deformation and

Tibet from shear wave splitting reveals changes in lithospheric de-

lower crustal flow in eastern Tibet. Science, 1997, 276: 788－790 4

Flesch M L, Holt W E, Silver P G, et al. Constraining the extent of

1692

7

formation. Earth Planet Sci Lett, 2006, 251: 293－304 Meissner R, Mooney W D, Artemieva I. Seismic anisotropy and

CHANG LiJun et al. Sci China Ser D-Earth Sci | Dec. 2008 | vol. 51 | no. 12 | 1683-1693

8

mantle creep in young orogens. Geophysics, 2002, 149: 1－14 Gao Y, Zheng S H, Zhou H L. Polarization patterns of fast shear wave

22

dawana and Yangtze blocks: An example by wide angel seismic re-

in Tangshan region and their variations (in Chinese). Chin J Geophys,

flection profiling between Menglian and Malong. Sci China Ser

1999, 42(2): 228－232 9

Zheng X F, Chen C H, Zhang C H. Study on temporal variations of

23

Chichi earthquake, Taiwan (in Chinese). Chin J Geophys, 2008, 51(1): 149－157

24

Silver P G, Chan W W. Share-wave splitting and subcontinental

11

McNamara D, Owens T, Silver P G, et al. Shear-wave anisotropy

25

12

beneath the Tibetan Plateau. J Geophys Res, 1994, 99: 13655－13665 Ding Z F, Zeng R S. Observation and study of shear wave anisotropy

26 27

phys Res, 2007, 112, B08416, doi:10.1029/2005JB004120 28

111－116

16

29

Chang L J, Wang C Y, Ding Z F. A study on SKS splitting beneath the

measurements from eastern Tibet and their implications for India/Eurasia intercontinental deformation. J Geophys Res, 2000, (105):

Wang C Y, Chang L J, Lü Z Y, et al. Seismic anisotropy of upper 30

pattern. Sci China Ser D-Earth Sci, 2007, 50(8): 1150－1160 Chang L J, Wang C Y, Ding Z F, et al. Seismic anisotropy of upper Chin J Geophys, 2008, 51(2): 431－438

ocean areas (in Chinese). Sci China Ser D-Earth Sci, 2002, 32(1): 89−100 31

Vinnik L P, Farra V, Romanowicz B. Azimuthal anisotropy in the Earth from observations of SKS at Geoscope and NARS broadband

20

33

1998, 103(B1): 749－771 Wang C Y, Chan W W, Mooney W D. 3-D velocity structure of crust and upper mantle in southwestern China and its tectonic implications.

34

Planet Int, 2006, 154: 180－195 Lou H, Wang C Y. Wavelet analysis and interpretation of gravity data in Sichuan-Yunnan region, China (in Chinese). Acta Seismol Sin,

Zhang Z J, Bai Z M, Wang C Y, et al. The crustal structure under Sanjiang and its dynamic implications: Revealed by seismic reflec-

Chin J Geophys, 1987, 30: 476－486 Li C, R D van der Hilst, Toks M N. Constraining P-wave velocity variations in the upper mantle beneath Southeast Asia. Phys Earth

J Geophys Res, 2003, 108 (B9): 2442, doi:10.1029/2002JB001973 21

Xu Z H, Wang S Y, Huang Y R, et al. Directions of mean stress axes in southwestern China deduced from microearthquake data (in Chinese).

Wolfe C J, Silver P G. Seismic anisotropy of oceanic upper mantle: Shear wave splitting methodologies and observations. J Geophys Res,

Ma X Y. Map of Lithospheric Dynamics in China and the Vicinity (in Chinese). Beijing: Geological Publishing House, 1989

32

stations. Bull Sesimol Soc Am, 1989, 79: 1452－1558 19

16215－16228 Teng J W, Zeng R S, Yan Y F, et al. The depth distribution of Moho and tectonic framework in eastern Asian continent and its adjacent

mantle in the northeastern margin of the Tibetan Plateau (in Chinese). 18

32(9): 809－812 Chen Z, Burchfiel B C, Liu Y, et al. Global Positioning System

Yunnan region (in Chinese). Chin J Geophys, 2006, 49: 197－204 mantle in eastern Tibetan Plateau and related crust-mantle coupling 17

Zhang P Z, Shen Z K, Wang M, et al. Continuous deformation of the Tibetan Plateau from global positioning system data. Geology, 2004,

Lü Q T, Ma K Y, Jiang M. Seismic anisotropy beneath Southern Tibet (in Chinese). Acta Seismol Sin, 1996, 18(2): 215－223

15

Gan W J, Zhang P Z, Shen Z K, et al. Present-day crustal motion within the Tibetan Plateau inferred from GPS measurements. J Geo-

corresponding features of upper mantle in Tibetan Plateau and its

14

Ding G Y. Lithospheric Dynamics of China (in Chinese). Beijing: Seismological Press, 1991

Jiang M, Xu Z Q, Hirn A, et al. Teleseimic anisotropy and neighboring areas (in Chinese). Acta Geosci Sin, 2001, 22(2):

Chin J Geophys, 2007, 50(3): 752－759 An M, Shi Y. Lithospheric thickness of the Chinese continent. Phys Earth Planet Int, 2006, 159: 257－266

in Tibetan Plateau (in Chinese). Chin J Geophys, 1996, 39(2): 211—219

Peng Y J, Huang Z X, Su W, et al. Anisotropy in crust and upper mantle beneath China continent and its adjacent seas (in Chinese).

mantle deformation. J Geophys Res, 1991, 96: 16429－16454

13

D-Earth Sci, 2005, 48(11): 1828－1836 Wang C Y, Lou H, Lü Z Y, et al. S wave crustal and upper mantle’s velocity structure in eastern Tibetan Plateau－Deep environment of low crustal flow. Sci China Ser D-Earth Sci, 2008, 51(2): 263－274

shear-wave splitting in the Chiayi area, aftershock zone of 1999

10

Zhang Z J, Bai Z M, Wang C Y, et al. The crustal structure of Gon-

35

2003, 27(5): 515－523 Wang C Y, Han W B, Wu J P. Crustal structure beneath the eastern

tion/refraction profiling between Zhefang and Binchuan,Yunnan. Sci

margin of the Tibetan Plateau and its tectonic implications. J Geophys

China Ser D-Earth Sci, 2005, 48(9): 1329－1336

Res, 2007, 112, B07307, doi:10.1029/2005JB003873

CHANG LiJun et al. Sci China Ser D-Earth Sci | Dec. 2008 | vol. 51 | no. 12 | 1683-1693

1693