Geodesy and Geodynamics 2011 ,2(1) :47-54

http://www. jgg09. com

Doi:10.3724/SP.J.1246.2011.00047

Influence of crustal layering and thickness on co-seismic effects of Wenchuan earthquake

Tan Hongbo1'2 ,Shen Chongyang1

'2

, Xuan Songhai1'2 and Yang Guangliang1

'2

1 lnstirute of Seis11Wlogy, CEA, Wuhan 430071, Chma

'CrustaJal MovemenJ.laborator, Wuhan 430071, Chma

Abstract: Using the PSGRN/PSCMP software and the fault model offered by USGS and on the basis of finite

rectangular dislocation theory and the local layered wave velocity structures of the crust-upper-mantle, the in­

fluences of crustal layering and thickness on co-seismic gravity changes and deformation of Wenchuan earth­

quake have been simulated. The results indicate that: the influences have a relationship with the attitude of

faults and the relative position between calculated points and fault. The difference distribution form of simula­

ted results between the two models is similar to that of co-seismic effect. For the per centum distribution, it'

s restricted by the zero line of the co-seismic effects obviously. Its positive is far away from the zero line. For

the crustal thickness, the effect is about 10% - 20%. The negative and the effect over 30% focus around the

zero line. The average influences of crustal layering and thickness for the E-W displacement, N -S displace­

ment, vertical displacement and gravity changes are 18.4% ,18. 0% ,15. 8% and 16. 2% respectively, When

the crustal thickness is 40 km, they are 4. 6% ,5. 3% ,3. 8% and 3. 8%. Then the crustal thickness is 70

km, the average influences are 3. 5%, 4. 6% ,3. 0% and 2. 5% respectively.

Key words: Wenchuan earthquake ; co-seismic effects ; crustal layering; crustal thickness ; fmite rectangular

dislocation theory

1 Introduction

Wenchuan Ms8. 0 earthquake took place at Longmen­

shan fault zone , which is the secondary fault zone in

the middle of the most active N-S seismic belt in Chi­

na [ ll . The Longmenshan fault zone is located at the

joint of Songpan -Ganzi orogenic zone and Y angzi

block, which is the composition of the east borderline

of the Qinghai-Tibet plateau together with Minshan up­

lift. It 's quite different in the structure of crustal

Received ,2010·12-25; Acoepted ,2011.01-24

Corresponding author: Tel. 13476091954; E-mail: thbhong@ 163. com

This study was supported by the National Natural Science Foundation of

China ( 40574012 ) and the Earthquake Science Joint Foundation

(A07030)

layering and thickness of different areas because of the

nappe structure environment[2 -s]. In the northwest,

the crustal density of the Qinghai-Tibet plateau is smal­

ler and the crustal thickness can reach to 70 km,

which is thinning eastwardly. While in the southeast,

crustal thickness of Sichuan basin is about 40 km. The

local crustal structure of Longmenshan fault zone pro­

vides the basis for simulation study by using dislocation

theory.

Many scientists have researches on co-seismic effects

by using dislocation theory at present[•-•l. Their stud­

ies show that the influence of layered structure is larger

than other factors of medium parameters in dislocation

simulation, but their conclusions are not identical, and

don ' t have specific description about the influence dis­

tribution. They also don' t consider the factor of crustal

48 Geodesy and Geodynamics Vol. 2

thickness and the influence on co-seismic gravity chan­

ges caused by every factor. In layered elastic-viscoe­

lastic half space, the surface deformation and gravity

changes caused by VVenchuan earthquake have been

simulated by Tan Hongbo[ 10J. The results between

measuring and simulating are different , which may be

related to the crust medium parameter selecting. Based

on that, the influence of VV enchuan co-seismic deform­

ation and gravity changes caused by crustal layering

and thickness will be simulated by using the PSGRN/

PSCMP software[ 11 J. And the distribution of percentage

will be elaborated too. It may offer some gists for ex­

plaining the mechanism of VV enchuan earthquake prep­

aration.

2 Fault model and layered medium model

This model is inversed by Chen Ji ( USGS ) , of which

the fault plane is inconsistent with the slip direction

( Fig. 1 ) . Red star represents the epicenter, the color

represents slippage value, and the arrowhead repre­

sents the moving direction of top-wall opposite to the

footwall. Isoline is the rupture start time. The optimum

fault parameters are as follows: strike is 229°, dip an­

gle is 33 °, 21 X 8 blocks along strike and dip, every

sub-fault is 15 km X 5 km.

There are lots of results about deep seismic sounding

tomography , natural and artificial earthquake in the ar-

f S. h d y . [5,13-15] Th . ea o 1c uan an unnan provmce . eu re-

sults indicate that the crust-upper mantle appears to be

layered. But the parameters of the wave velocity layered

]10

! 20

-240

I

0 180

Strike=229 deg • km

-180 -120 -60

I I

360 540 I

720

IIIII! em 900

0

Figure 1 Cross section of the fault including the

movement distribution (from USGS)

model are different from each other by dissimilar way

and data. Overall consideration, we choose the model

as table 1 , where the density is gained from the Nafe­

Drake density-velocity experience transform formular­

y[16l , and the viscosity modulus is from the result of

Tan Kai[l?J.

For better analyzing the influence of crustal layering,

we design a homogeneous crust model (Tab. 2) based

on the weighted average of crust parameters in each

layer thickness. Obviously, the difference between

simulant results of the two models can reflect the influ­

ence of crustal layering[JsJ. VVe choose two models as

examples, one is 70 km crustal thickness of the north­

west Qinghai-Tibet plateau and the other is 40 km crus­

tal thickness of the southeast Sichuan basin. Compa­

ring with the homogeneous crust model of which the

crustal thickness is 57 km, co-seismic influence caused

by different crustal thickness will be quantitatively sim­

ulated and analyzed.

Table 1 Layered structure of the crust-upper-mantle based on the average velocity

h(km) VP ( km/s)

0-20 5.98

20-40 6.55

40-57 6.83

57-00 7.75

Table 2

h(km) VP ( krnls)

0.0-57.0 6.4335

57.0-00 7.7500

V, (km/s) p(kglm3) 71(Pas)

3.45 2679.0 00

3.78 2835.0 00

3.84 2977.0 4.0E + 19

4.35 3175.0 1. OE + 19

Homogeneous structure of the crust-upper-mantle

V,(krnls) p(kglm3) 7J(Pas)

3.6821 2822.6 O.OE +00

4.3500 3175.0 1. OE + 19

a

1.0

1.0

0.0

0.0

a

1.000

0.000

notes: Where h is the depth, VP is the P wave velocity, V, is the S wave velocity, p is density, 7J is the viscosity modulus, a is the relax ratio, a = 0 means

complete elastic medium, a = 1 means Maxwell viscoelastic medium.

No.1 Tan Hongbo, et al. Influence of crustal layering and thickness on co-seismic effects of Wenchuan earthquake 49

3 The influence of surface deforma­tion and gravity changes caused by crustal layering

The pictures of the co-seismic deformation and gravity

changes are similar between the homogeneous crust

model and layered crust model. For the homogeneous

model , the radiation range of the pictures is larger. As

shown in Figure 2 ( a) , we can see that : the outline is

similar to the picture of co-seismic E-W horizontal dis­

placement, which are divided into northwestern and

southeastern parts against the rupture zone. The iso­

lines are antisymmetric : maximum positive difference

can reach to 55 mm; maximum negative difference can

35

®

Xining ® Lanzhou

60mm 50 40 30 20 10

-hnm -2 -3

reach to - 85 mm. Most of the difference focus on the

fault around, and decays in the far field. The differ­

ence of N -S horizontal displacement ( Fig. 2 ( c ) ) is

similar to the N -S horizontal displacement in outline ,

but present to be negative in the southwest. So, the

picture comes to be four quadrants positive and nega­

tive distribution: the maximum positive difference can

reach to 25 mm; the maximum negative difference can

reach to - 45 mm. The difference of vertical displace­

ment ( Fig. 2 ( e) ) and gravity changes ( Fig. 2 ( g) ) fo­

cus on the fault around. It's not obvious in four quad­

rants distribution comparing with vertical displacement

and gravity changes, but presents a positive and nega­

tive distribution along fault, and decays fast in the near

field.

35"N

®

Xining ., Lanzhou

80%

70

60

50

40

30

20

15%

10

31 -4 31 "N

103"E 107"E (a) Difference ofE-W displacement

Xian

31 "N

27"N'--------'--------'---------'---'--------'--------' 99"E 103"E 107"E

(c) Difference ofN-S displacement

-5 -10 -20

® Chongqing

-30 -40 -50 -60 -70 -80

-85mm 27"N'------'-..,.,.,_--'------'-----'----'---------'

30mm 25 20 15

110 5 4

-1mm

l -2 I -3

-4 31 -5

-10 -15 -20 -25 -30 -35 -40

99"E 103"E 107"E (b) Difference percentage of E-W displacement

-45~7 N L __ __._ _ __,__.___._-'Sl"-'-----__JJ_-----'

99"E 103"E 107"E (d) Difference percentage ofN-S displacement

·10

-20

-30

-40

-50

-60%

80%

70

60

50

40

30

20

15

10

0%

-10

-20

-30

-40

-50

-60

-70

-80

-90%

50 Geodesy and Geodynamics Vol. 2

35"N

31 "N

@

Xining ., Lanzhou

Xian®

@

Chongqing

60mm

50

40

30

20

10

1mm

-1

-2

-3

-4

-5

-10

-20

-30

35"N

80"/o

70

60

50

40

30

20

15

10"/o

-10

-20

-30

-40

-50

-60

27"N '-----------'--------'-------'-----'-----_J__-__j -35mmZ7"N c__ __ _J._ __ _j_ __ ___~_-=__lf.!i~L--__j -70%

99"E 103"E 107"E

(e) Difference of vertical displacement

@

Lanzhou

2411Gal

22

20

99"E 103"E

(f) Difference percentage of vertical displacement

@

Lanzhou

80%

35"N 18 35"N

70

60

50

40

30

20

15

10"/o

31"N

@

Chongqing

27"NL_ ____ L_ ____ L_ ____ ~~--~--~

99"E 103"E 107"E

(g) Difference of gravity changes

16

14

12

10

8

5l1Gal

-1

-2

-3

-411Gal

31 "N -10

-20

-30

-40

-50

-60

27 "N L__ ____ _L_ ____ ~----__L_L----'-',.___,.--'-__ __j -70%

99"E 103"E 107"E

(h) Difference percentage of gravity changes

Figure 2 Co-seismic effect difference and its percentage between the layered crust model and the

homogenous crust model( gravity unit:10- 8 ms -z; deformation unit: mm; percentage unit: %)

Percentage of co-seismic effect difference between

homogeneous crust model and layered crust model is

the ratio of the co-seismic effect difference and the sim­

ulation results of homogeneous crust model (wipe off a

few ratio which are bigger than 1 ) . The right in Figure

2 shows that co-seismic effect difference and its per­

centage are closely related to each other. Most of the

percentage is positive and the value is between 10%

and 20% . While negative and big one are near the

positive and negative transforming area. The More

close to the zero line , the larger the difference percent­

age is. Based on the close relationship between co­

seismic effect and fault attitude , or the relative position

between the simulated points and fault[ 19'20 l , we can

infer that the influence of co-seismic effect caused by

crustal layering has a relationship with fault attitude,

and the relative position between the simulate points

and fault.

By comprehensive consideration, using absolute val­

ue and taking an average of the difference percentage ,

we can get that: the influence of the E-W horizontal

displacement caused by crustal layering is 18. 4% ; for

the N -S horizontal displacement, E-W and N -S vertical

displacement and gravity changes, they are 18. 0% ,

15.8% and 16.2% respectively.

No.1 Tan Hongho, et al. Influence of crustal layering and thickness on co-seismic effects of Wenchuan earthquake 51

4 The influence of surface deforma­tion and gravity changes caused by crustal thickness

Difference distribution of surface deformation and gravi­

ty changes among the 40 km, 70 km, and 57 km crus­

tal thickness is showed in Figure 3. For co-seismic

effect difference of 40 -57 km crust model ( left in Fig.

3) , the pictures of E-W horizontal displacement are

similar to the difference of E-W displacement: the iso­

lines are left-right two quadrants antisymmetric distri­

bution. The maximum positive change is 26 mm, while

the maximum negative change is - 20 mm. The extre­

mum appears 0. 2 o beside the fault , and decays against

the fault. The difference of N-S horizontal displacement

is four quadrants distribution: on the left, it's positive

difference, and the extremum is 12 mm · on the right ' '

it' s negative difference, and the extremum is - 11

mm. The difference of vertical displacement and gravi­

ty changes focus on the rupture area , decaying against

the fault, and the maximum positive difference can

reach to 18 mm and 3. 5 X 10 - 8 ms - 2; while the maxi­

mum negative difference can reach to - 30 mm and

-7 X 10 - 8 ms - 2• For co-seismic effect difference of 70

-57 km crust model ( right in Fig. 3 ) , it ' s similar in

outline , but smaller in the radiation range , and reverse

in the change value.

For better understanding the influence of surface de­

formation and gravity changes caused by crustal thick­

ness, difference per centum has been calculated by u-

" 26mm

Xining u 22

18

35•N 14

10

6

4

3

2

1mm

0

31 ON -1

-2

-3

-4

-5

-8

-12

40km-57km -16

sing the similar way which Figure 2 has used. The dis­

tribution has similar characteristics with Figure 2 :

difference per centum is restricted by the co-seismic

zero line, where negative and large value focus on. U­

sing absolute value and taking an average of the differ­

ence percentage , we can get that : for the difference

pictures of 40 - 57 km crustal thickness model differ-'

ence per centum of the E-W displacement , N -S dis­

placement, vertical displacement and gravity changes

are 4. 6% , 5. 3% , 3. 8% and 3. 8% respectively;

while for 70 -57 km crustal thickness model, they are

3. 5% ,4. 6% ,3. 0% and 2. 5% respectively.

Conclusively, co-seismic effect difference of the 40

-57 km model is bigger than that of the 70 - 57 km

model. The radiation range of the E-W and N-S dis­

placement difference is wider, and decays slowly a­

gainst the fault. The radiation range of the vertical dis­

placement difference and gravity changes difference is

smaller , and decays fast. Surface deformation and

gravity changes caused by W enchuan earthquake are

increasing when the crustal thickness increases. The

influence caused by crustal thickness is biggest on the

N -S displacement , then the vertical displacement and

E-W displacement, smallest on the gravity changes.

5 Discussion and conclusions

1 ) Crustal layering has great influences to the co­

seismic deformation and gravity changes: for W enchuan

earthquake , the influences to E-W displacement , N -S

displacement, vertical displacement and gravity changes

" 8mm

Xining ®- 6 J4lzll u

35° 4

31

70km-57km

Omm

-1

-2

-3

-4

-5

-6

-8

21•N -20mm27 -!Omm

99•E 103•E 107•E 99•E 103•E 107•E

(a) Difference ofE-W displacement

52

® Xining

40km-57km

Geodesy and Geodynamics

12mm

10

4

Omm

-1 31 °N -2

-3

4

-5

-7

-9

27oN '----'----.JL----'----'-------' -llmm 27oN ~..._ __ l...J.._ __ ----1.1...-----'-------'---...J

99oE

31 ON

27oN 99oE

35°N

31 °N

27oN 99oE

Kangding

®

Lanzhou

40km-57km

103oE

®

Lanzhou

103°E

99oE

(b) Difference ofN-S displacement

lOTE

18mm 14 10

-1mm -2 -3 4

-5 -6 -10 -14 -18 -22 -26

31 °N

-30mm 27°N

99°E

(c) Difference of vertical displacement

3.5Gal 3.0 2.5

2.0 1.5 35°N

1.0

0.5 0.0 -0.5 -1.0

-1.5

-2.011Gal -2.5 -3.0 31 °N

-3.5 -4.0

® -4.5 Chongqing

-5.0

-5.5 -6.0 -6.5

-7.0 11Ga!z7oN

lOTE 99oE

(d) Difference of gravity changes

®

Lanzhou

70km-57km

103°E

<!' Lanzhou

70km-57km

103°E

107oE

® Chong g

107°E

® Chongqing

107oE

Figure 3 Co-seismic effect difference among crustal thickness of 40 km , 70 km and 57 km

(gravity unit: 10 - 8 ms - 2; deformation unit: mm)

Xian"

Vol. 2

4.5mm

3.5

2.5

1.5

1.0

0.5

0.0

-O.Smm

-1.0

-2.0

-3.0

-4.0

-5.0

-5.5mm

13mm

11

4

1mm

-1

-2

-3

4

-5

-7mm

3.011 Gal

2.6

2.2

1.8

1.4

1.0

0.6

0.411 Gal

0.2

0.0

-0.2

-0.4

-0.8

-10

-1.411 Gal

No.I Tan Hongbo, et al. Influence of crustal layering and thickness on co-seismic effects of W enchuan earthquake 53

are 18. 4% , 18. 0% , 15. 8% and 16. 2% respective­

ly, which are bigger than the results analyzed by Deng

MinglP1• She got the per centum by comparing the

maximum co-seismic difference of the two models with

the maximum co-seismic effects. That is not compre­

hensive , because the maximum influence per centum is

no around the fault. Our results are more similar to

Pollitz1' 1 , Sun and Okubo' s171 , but with some value

over 100% , because the position of the co-seismic

effect zero lines simulated by different models are no

the same. So , the difference per centum is singular

when closing to the zero line. Generally speaking, the

inlluence distribution caused by crustal layering is reli­

able.

2 ) The inlluences of co-seismic deformation and

gravity changes caused by crustal layering have a rela­

tionship to the attitude of fault and the relative position

between calculated points and fault. Results of homo­

geneous crust model and crustal layering model have

the biggest difference in the near field, which decays

against the fault with similar outline to co-seismic

effects. The per centum distribution is restricted by the

zero line of the co-seismic effects obviously : most of

the percentage is positive and the value is between

10% and 20%; while negative and big one are near

the positive and negative transforming area.

3 ) Crustal thickness has a little effect on co-seismic

deformation and gravity changes : in the case of 40 km

crustal thickness , the inlluences of E-W displacement,

N -S displacement, vertical displacement and gravity

changes caused by Wenchuan earthquake are 4. 6% ,

5. 3% , 3. 8% and 3. 8% respectively; while in the

case of 70 km crustal thickness, they are 3. 5% ,

4. 6% ,3. 0% and 2. 5% respectively.

4) Comparison between simulated and measured re­

sults

As an example of comparison between simulated and

measured results, we get the results of 3 GPS continu-

ous observation stations which are Pixian( 103. 76°E,

31.91° N), Mianyang ( 104.73° E, 31.44° N) and

Qionglai(103.31°E,30.35°N) (Tab. 3) 1211 • Mian­

yang and Pixian stations are in the east of the surface

rapture zone. The shortest distances between these two

stations and Yingxiou-Beichuan fault are 47 km and 28

km respectively. Qionglai station is in the southeast of

the surface rapture zone. Only judging on vertical di­

rection, Qionglai simulated results is consistent with

Pixian ' s, but reverse for Mianyang station, because

the precision of GPS vertical displacement is too low.

To horizontal displacement, Pixian simulated and

measured results are consistent in terms of direction ,

but great different on magnitude. The differences be­

tween Mianyang simulated and measured results are

7% and 49% on E-W and N-S displacement. Taking

account of crustal layering and thickness inlluence , the

E-W displacement simulated result is credible, but the

difference is biggest on the N-S displacement. For

Qionglai station, the direction is reverse on N -S dis­

placement, and consistent on E-W displacement with

34% difference which is over the range of crustal laye­

ring and thickness inlluence.

Generally speaking, some station 's simulated and

measured results can be inosculated in error range of

analyzed crustal layering and thickness. But in most

cases, it' s not reasonable. Back to the original ques­

tion, there are two factors which cause the difference

between simulated and measured value : medium model

difference and fault model difference. Since two main

factors( crustal layering and crustal thickness) in medi­

um model can not explain the difference , the reason

might be the fault mode difference. The fault model in­

versed by far field earthquake wave is too simple, and

has big difference with the actual situation. Joint inver­

sing Wenchuan earthquake fault model by using gravi­

ty, leveling and GPS data will be the best way to solve

the difference between simulated and measured data.

Table3 Comparison between simulated and measured results( unit: mm)

Pixian Mianyang Q;onglai displacement

measured simulated measured simulated measured .unulated

E-W -563 -189.5 -305 -285 -15 -26

N-S 426 64.5 66 131 -3 10.4

vertical 81 43.9 14 -9.4 28 27

54 Geodesy and Geodynamics Vol. 2

References

[ 1 ] Jiao Qing, et al. Preliminary study on motion characteristics of

Longmenshan fault before and after Ms8. 0 Wenchuan earth­

quake. Journal of Geode'l' ond Geodynamioo, 2008 , 28 ( 4) '7 -

11. (in Chinese)

[2] Teng Jiwen,et al. Deep processes and dynamic responses associat­

ed with the W enchuan M s8. 0 earthquake of 2008. Chinese J.

Geophys. ,2008,51(5) :1385-1402. (in Chinese)

[3 J Yang Haiyan,et al. Crust mantle structure and seismogenic back­

ground of Wenchuan Ms8. 0 earthquake in Western Sichuan area.

Chinese J. Geopys. ,2009,52(2) :356-564. (in Chinese)

[ 4 J Lei Jianshe, et al. Fine seismic structure under the Longmenshan

fault zone and the mechanism of the large Wenchuan earthquake.

Chine•e J. Geoph,... ,2009 ,52(2) ,339-345. (in Chineoe)

[5] Xu Zhaofan,et al. Deep geophysical exploration af Longmenshan

and its adjacent area. Journal of Geodesy and Geodynamics ,2008 ,

28(6) ,31 -37. (in Chine•e)

[ 6 J Sun W and Okubo S. Effects of the earth' s spherical curvature

and radial heterogeneity in dislocation studies-fm a point disloca­

tion. Genph,... R. L., 2002,29(12) ,461-464.

[ 7 J Pollitz F F. Co-seismic defm:mation from earthquake faulting in a

layered •pherioal eorth. Geoph,... J.Int. ,1996,125 ,1 -4.

[ 8 J Fu G and Sun W. Surface coseismic gravity changes caused by dis­

locations in a 3-D heterogeneous earth. Geophysical Journal Iner­

natiooal,2008,172(2),479 -503. doi,!O.llll/j. 1365-

246X. 2007. 03684. x.

[ 9 J Den Mingli, Sun Heping and Xu Jianqiao. A sectionalized fault

model of Kunlun Ms8. 1 earthquake in constraint of GPS data.

Journal of Geodesy ond Geodyoamic•. 2008 ,28 ( 4) ,31 - 3. (in

Chinese)

[ 10 J Tan Hongbo, et al. Simulation of post seismic gravity change and

deformation of the Wenchuan earthquake based on viscoelastic

layered half space model. Acta Seismologica Simica, 2009 , 31

(5) ,491-505. (in Chineoe)

[ 11 J Wang R, Francisco Lorenzo-martin and Frank Roth. PSGRN/

PSCMP-a new code for caluclation co-and post-seismic deforma­

tion, geoid and gravity changes based on the viscoelastic - gravi­

tational dislocation theory. Computers and Geosciences ,2006 ,32:

527 -541.

[ 12] Chen Ji, UCSB and Gavin Hayes. NEIC. preliminary result of the

May 12, 2008 Mw 7. 9 eastern Sichuan, China earthquake. ht­

tp :I I earthquake. usgs. gov/ eqcenter/ eqinthenews/us2008ryan/.

200805

[13] Wang Chunyong,et al. Crustal structure beneath the Songpan­

Ganze orogenic belt. Acta Seismologica Simica,2003,25(3) :229

-241. (in Chinese)

[ 14 J Wang C Y, Han W B and Wu J P. Crustal stuacture beneach the

easton margin of the Tibetan plateau and its tectonic implication.

Journal Geophy•icol Reoearoh ,112, B07307 , I -21.

[ 15 J Wu Jianping, Ming Yuehong and Wang Chunyong. Regional wave­

form inversion for crustalal and upper mande velocity structure

below Chuandian region. Chinese Journal of Geophysics. 2006,49

( 5) , 1369 -1376. (in Chineoe)

[ 16 ] Talwandl M, et al. A crustal section across the Puerto Pico

Trench. J. Genph,... Res .• 1959 ,64 ( 10) '1545 -1556.

[17] Tan Kai,Li Jie and Wang Qi. Lithospheric rheological structure

constrained by grodetic data in Altay. Chinese Journal of Geo­

ph,..ics,2007 ,50( 6) ,1713 -1718. (in Chine•e)

[ 18 J Shen Zhengkang, et al. Viscoelastic triggering among large earth­

quakes along the east Kunlun fault system. Chinese Journal of Ge­

ophysic•,2003 ,46( 6) , 786 -795. (in Chineoe)

[ 19 ] Tan Hong bo, Shen Chongyang and Li Hui. Characteristics of

sureface gravity changes caused by a fault or faults dislocation.

Journal of Geode•y and Geodynamics. 2008 ,28( 4) ,54 -62. (in

Chinese)

[20] Tan Hong bo,et al. Characteristics of sureface defonnation caused

by fault dislocation. Journal of Geodesy and Geodynamics. 2009,

29(3) ,42-49. (in Chine..)

[ 21 J Project Group of the Crust Movement Observation Network of Chi­

na. The co-seismic defonnation of the Wenchuan Ms8. 0 earth­

quake based on GPS observation. Science China D, 2008, 38

(10) ,1195-1206. (in Chine•e)