- Email: [email protected]
Detection of a half-microgal coseismic gravity change after the Ms7.0 Lushan earthquake
Geodesy and Geodyoamics
2013,4(3) :7-11
http :llwww. jgg09. com Doi:l0.3724/SP.J.l246.2013.03007
Detection of a half-microgal coseismic gravity change after the Ms7. 0 Lushan earthquake Wei Jin 1 "2 "3 , Zhao Bin 1 "2 , Tan Hongho 1 "2 , Yu Dan4 , Shen Chongyang 1 "2 and Li Hui 1 "2 1
&y LolJoratory of Earthquake Geodesy, lnstiruu of Sei.smology, China Eart/ujuake Admim..tration, Wuhan 430071 , China
2
Wuhan Base of lnstiruu of Crustal Dynamics, China Eartlujuake AdminUtration, Wuhan 430071 , China
' School 4
of Getxksy and
Geomntics, Wuhan University, Wuhan 430079, China
China Earthquake Networks Center, Beijing 100045, China
Abstract: Because only a small near-field coseismic gravity change signal remains after removal of noise from the accuracy of observations and the time and spatial resolution of the earth's surface gravity observation system, it is difficult to verify simulations of dislocation theory. In this study, it is shown that the GS15 gravimeter, located 99.5 km from the epicenter of the Ms7. 0 Lushan earthquake on April 20, 2013 at 08 :04 UTC + 8, showed the iofluence of the earthquake from 2013-04-16 to 2013-04-26 after a time calibration, tide corrections, drift correction, period correction and relaxation correction were applied to its data. The post -seismic
relaxation process of the spring in the gravimeter took approximately 430 minutes and showed a 2. 5
X
10 -• ms _,
gravity change. Mter correcting for the relaxation process, it is shown that a coseismic gravity change of
approximately +0.59 ±0.4 xlO-'ms- 2 was observed by the GS15 gravimeter; this agrees with the simulated gravity change of approximately 0. 31
X
10 -• ms _,. The rate of the coseismic gravity change and the coseismic
vertical displacement, as measured by one-second and one-day sampling interval GPS units, is also consistent
with the theoretical rate of change. Therefore, the GS15 gravimeter at the Pixian Station observed a coseismic gravity change after the Ms7. 0 Lushan earthquake. This and similar measurements could be applied to test and confirm the theory used for these simulations.
Key words: GS15 gravimeter; coseismic gravity change; the Ms7. 0 Lushan earthquake
the coseismic gravity changes caused by powerful earth-
1
Introduction
quakes (the 2003 Mw7. 9 Tokachi-oki earthquake; the 2007 Mw9. 3 Sumatra earthquake; the 2010 Mw8. 8
A powerful earthquake always results in a change in the
Chile earthquake) have been acquired by high-preci-
gravity potential field of the earth that can be observed
sion , time-variable gravity measurements from networks
as a small change in the local gravitational acceleration.
of superconducting gravimeters
Such coseismic gravity changes have been observed after
These measurements have been verified by simulations
the 1964 Mw9. 2 Alaska earthquake by mobile relative
based on fmite rectangular dislocation theory and the
gravinretry[IJ and after the 1998 M6. 1 earthquake near
local layered wave velocity structure of the crust and
lwate-san Mountain by ai3solute gravimetry[']. Recently,
upper mantle, which calculate the surface coseismic
and
GRACEc•.•l.
and post-seismic gravity changes[SJ. There are several Received :2013 -05 -09 ; Accepted :2013 -05-28 Corresponding author: Wei Jin, E-mail: pierce212@ 163. com. This work is supported by the National Natural Science Foundation of China ( 41204058) and the Rwming Foundation of the Gravity Network Center of China (201301008).
difficulties in validating the accuracy of these simulations using the current observation system , including
the low precision and low temporal resolution of both the vertical displacement measured by GPS and the
8
Geodesy and Geodynamics
Vol.4
time-variable gravity measured by GRACE. These are
d 1 are the constant and linear terms of the residual
largely inadequate to use at high temporal resolution for
gravity field, respectively, and the coefficients
checking the fast coseismic gravity changes that occur
bj describe the residual diurnal and semidiurnal wave
during earthquakes. Furthermore, due to the powerful
signals of the residual gravity, respectively. The next
shock of earthquakes, there are few near-field, ground-
term corrects for any number ( N,) of offsets that have
that can confmn the existence of
magnitudes g and intervals T,. The post-seismic gravi-
physical phenomena suggested by observations of co-
ty change is modeled as a mte of change hj and an ex-
seismic gravity changes.
ponential decay with magnitude kj over the selected
based gravimeters
On April20, 2013 at 08 :02 :46 UTC +8, a Ms7. 0
aj
and
earthquake time intervals Tloj and T0 • The measurement
(Mw6. 6) earthquake occurred in Lushan, in the Si-
errors v are initially assumed to be independent , iden-
chuan Province. The GS15 gravimeter located approxi-
tically distributed and mndom, with E(v) =0.
mately 99. 5 km from the epicenter, at the Pixian seis-
The p-wave arrival time (2013.()4-20 08:03:04.67
mic station, captured a complete record of the earth-
UTC + 8 ) identified in the Pixian Station seismometer
quake at a one-minute sampling interval 1•l • In this
record is used to calibrate the site' s GS15 gravimeter.
study, we analyzed the seismic record from April 16 to
Several years of synthetic tides are determined from this
26 and applied time calibmtions, tide and drift correc-
gravity data record. An automatic pikes processing
tions and corrections by period and relaxation analysis.
method is used to correct the sudden jumps in signal
The gravimeter data were used to check and confmn the
caused by the electronics and the far-field earthquake 1'l.
near-field coseismic gravity change occurring as a result
The tide , air pressure , drift and period corrections and
of the Ms7. 0 Lushan earthquake.
the residual gravity are shown in figure 1. Gravimeter records are susceptible to spurious sig-
nals from the step and sticking bar; these signals can
2 Data processing methods
strongly influence the long-term drift behavior. Figure
The gravity data were sampled at one-minute intervals by the GS15 gravimeter at the Pixian seismic station. The data were influenced by changing air pressures and solid tides but reflect timing errors from the data acqui-
sition system, errors in the residual diurnal and semidiumal wave signals corrected by tide theory, drift of the spring in the gravimeter, coseismic gravity changes and }X)St-seismic relaxation effects [7 ]. To simulate these er-
1 shows that the GS15 gravimeter has no such instrumentation problems. After subtracting the influence of the gravity tide ( ±200 spring ( -100 x ( ± 4 x 10 -• ms -
2
xto-• ms- 2 ) ,
to-• ms ) ,
2
)
the drift of the
and the air pressure
a gravity anomaly of approximately 2
(3 -4) x 10 -• ms - remains in the period correction before and after the Ms7. 0 Lushan earthquake. The period correction comprises the error in the time system, the synthetic tide model, the coefficient of the
ror sources , the following model is used :
scale factor, the nonlinear drift of the spring and the influence of the earthquake.
3
Calculation and analysis of the coseismic gravity changes
N,
L hjH(t; -
When the spring in the GS15 gravimeter receives seis-
Tlif)t; +
j=l
mic waves , the original dynamic equilibrium of the
N,
L kjexp[- (t; - r.)lrj]H(t; - T,)t; +v; j=l
(1)
spring suddenly breaks , and the spring undergoes a relaxation process. In this study, the relaxation process
is simulated with the exponential portion of equation 1. Here, t, for i = 1· · · N are the time intervals in units of
This relaxation analysis , shown in figure 2 , finds an
minutes, g~ ( t;) is the residual gravity corrected by
optimal
tidal theory , H is the Heaviside step function , d0 and
proximately 0. 40 x 10 -• ms - 2 •
T
of 430 minutes and a residual error of ap-
No.3
Wei Jin,et al. Detection of a half-microgal coseismic gravity change after the Ms7. 0 Lushan earthquake
9
§ -440! "£1;;;'-460 ~ Ei -480 8. -500
~ ~-520 ~ ~ -540
Figure 1
The tidal, air pressure, drift and period corrections for the GS15 gravimeter at Pixian (The pressure admittance is -3.79
X
10 -s ms - 2 mbar. Over the 10-day period, the pole tide is linear)
0.49 .r 0.48 ·8 o.47 '6 0.46 ~ 0.45 § 0.44 " 0.43 The confidence interval of 90% ] 0.42 . 470 390 I :9 0.41 1+-'-'~ 0.40 I 0.39 0.38L___L_~--~--~~L_~L_~--~--_L_ __L~ 0 100 200 300 400 500 600 700 800 900 1000 Relaxation time (miniute)
Figure 2
Using the relaxation time ( 430 minutes) and the pwave arrival time (2013-04-20 00:04:00) as constraints[9J, the entire coseismic and post-seismic processes observed by near-field ground-based gravimeters are simulated from the residual gravity record from April 16 to 23 , which is shown in figure 3. The post-seismic relaxation process indicated by the springs of such gravimeters is approximately + 2. 5 X 10 - 8 ms - 2 • The post-seismic gravity
The relaxation time fitting result for the Lushan
change exhibits a negative trend, as before. The coseismic
earthquake observed at Pixian
gravity change is approximately (0.59 ±0.40) xl0- 8 ms- 2
2.0 1.5
~ 1.0
~
0.5
~
f
0
-~ -0.5
a
-1.0
-1.5 -2.0L---~--~--~~~~uw~~~~~--~~~--~--~----~~~~
,(,'t~ ,(,'t\1. •1't~ •'1-t\'l. •'i>'tf$! •'0-t\'l. •9'tl;)l:l •9-t\'l. ,~;~'tl:l\;1 ,~;~'t\1. ,\'t~;~l:l ,\'t\1. ,'J;'tl;)l:l ,'J;'t\1. ,~'t~;~l:l ~·~·~· ~·~· ~·~· ~·~ ~~ ~~~ ~~ Time (ddThh)
Figure 3
Coseismic gravity change in the Lushan earthquake at Pixian ( The blue line is the gravity change at Pixian Station, the magenta line is the relaxation process and the red line is the simulation result for the coseismic gravity change in the Lushan earthquake at Pixian Station)
10
Geodesy and Geodynamics
Vol.4
approximately 0. 5 X 10- 8 ms - 2 in the Pixian region
for the Lushan earthquake observed at Pixian. The coseismic vertical displacement measured by
after the Ms7. 0 Lushan earthquake. In cases where the
high-frequency GPS and gravity changes measured by
observed value and the precision are of similar magni-
the GS15 gravimeter during the earthquake ( from
tude, we used the vertical displacement of the one-
2013-04-20: 08 : 00 : 00 to 08 : 06 : 37 UTC + 8) are
second sampling interval GPS because its precision is
shown in figure 4. These two measurements sample the
higher than that of the daily sampling interval GPS.
same region. The coseismic vertical displacement is
The vertical displacement changes exhibit negative
approximately - ( 1. 1 ± 2. 1 ) mm , as determined by
trends, while the simulated coseismic gravity change
piecewise polynomial fitting.
and that observed by the GS15 gravimeter both exhibit
The statistics of the coseismic gravity change meas-
positive trends. The coseismic gravity change observed
ured by the GS15 gravimeter, the simulated gravity
by the GS15 gravimeter shows local changes, while the
change, and the vertical displacements of the GPS with
simulation captures the regional change. Including the
sampling intervals of one day and one second are shown
coseismic gravity change indicated by the GPS meas-
in table 1. The coseismic gravity changes from the GPS
urements using the theoretical gravity gradient, all
measurements with a theoretical gravity gradient of
gravity changes are approximately 0. 5
-300 x 10-8 ms - 2 m -I are approximately 0. 33 to 0. 63 mm.
Thus, the GS15 gravimeter at Pixian detected a half-
Table 1 shows that both the ground-based coseismic
microgal coseismic gravity change after the Ms7. 0
gravity change and the simulated gravity change are
Lushan earthquake.
.8
8
~r~~~tM:P~~mro-___;_-i'W1 °· 6 ~
]'7 ';::6
0.4 'b
0.2 :::;
5 4 ~----------~--------~~~~_r
]
t
0
~
-0.2 -5
3 ~ 2
~
10- 8 ms - 2 •
1.0 0.8
9
I
X
-
1 0
The vertical displacement of GPS The coseismic vertical displacment of GPS The Lushan earthquake The coseismic gravity change of GSIS
-1 00:00 00:24 00:48 01:12 01:36 02:00 02:24 02:48 03:12 03:36 04:00 04:24 04:48 05:12 05:36 06:00 06:24
-0.4·~ -0.6 -0.8 -1.0
a
Time (second)
Figure 4
Coseismic vertical displacement and gravity changes from high-frequency GPS and the GS15 gravimeter for the same region as shown in figure 3 (The blue line is the vertical displacement indicated by the 1 Hz GPS during the Lushan earthquake, the black line is the model of the coseismic vertical displacement indicated by the 1 Hz GPS, the red line is the model of the coseismic gravity, and the magenta line indicates the occurrence time of the Ms7. 0 Lushan earthquake)
Table 1
Statistics of the coseismic gravity and displacement changes of the Ms7. 0 Lushan earthquake, as measured at Pixian Station.
Obseration type
GPS GS15 Simulation result
Longitude (oE)
103.7 103.8
Latitude (oN)
30.9 30.9
Sampling interval (second)
86400 60
Coseismic vertical displacement ( mm)
Coseismic gravity change (10- 8 ms- 2 )
value
precision
-1.1
± 1.40
0.33
value
-2.1
±3.11
0.63 0.59
precision
±0.40
0.31
Note: The coseismic gravity change with the dislocation model of the earthquake source is simulated by the fault model used in seismic wave inversion offered by Institute of Geophysics, China Earthquake Administration [!OJ • Following the assumption that coseismic gravity changes are caused entirely by vertical displacement, the simulation of the gravity changes from the GPS data uses a gravity gradient of -300 X 10 -s ms - 2 m _, [u].
No.3
11
Wei Jin,et al. Detection of a half-microgal coseismic gravity change after the Ms7. 0 Lushan earthquake
model can be validated with this method in the future.
4
Conclusion and discussion References
Tills paper demonstrates that the drift of creep characteristics can be corrected not only by a time system cal-
[ I ]
ibration and the traditional correction method["], but also by a conducting period and relaxation analysis of
JZ071 i002]ffl451. [2 ]
the spring gravimeter. The traditional method finds that ms -
•
Simulated with the Heaviside step function and a
2001, 28(15)' 2979-2981.
[3 ]
X
to-'
2004, 306: 476-478. doi:l0.1126/science.l101875.
2
ms - ,
[4 ]
which agrees with the one-day and one-second GPS
quake. Chineoe !. Geophy•. 2011, 54(7), 1745 -1749. doi,
The data record of the GS15 gravimeter at the Pixian
10.3969/j. issn. 0001-5733. 2011.
[5 ]
rn.
007. (in Chinese)
Shen Chongyung, Li Hui and Tan Hongbo. Simulation of coseismic gravity changes and defonnati.on effect of wenchan Ms8. 0 earth-
drift, solid earth tide and air pressure and is also affected by a number of important and inevitable error
Zbou X, Stm W K and Fu G Y. Gravity satellite GRACE detects
coseismic gravity changes caused by 2010 Chile Mw8. 8 earth-
records and the simulated result for the same region.
Station is sensitive to a number of factors related to
Yuichi lmanishi , et al.. A Network of Supereonducting Gravimeters Detects Submicrogal Coseismic Gravity Changes. Science,
relaxation function, the coseismic gravity change measured
by the GS15 gravimeter is ( 0. 59 ± 0. 40)
Tanaka Y, Okubo S, Macbida M, et al. First detection of absolute gravity change caused by earthquake. Geophys. Res. Lett.
the residual gravity change is approximately ± 1 x 10 -• 2
B=es D. Gmvity Change during the AJa.ka earthquake. !. Geophyo. Reo., 1966, 71(2),451-457. doi, 10.1029/
quake. Geodesy aod Geodyoamic•. 2008 , 28 ( 5) , 6 -12. [ 6 ]
sources , such as error in the time system , the synthetic
Wei Jin, Hao Hongtao, et al. Gravity disturbance of high frequency at Chengdu seismostation before Wenchuan Ms8. 0 earth-
tide model, the coefficient of the scale factor and the
quake. Journal of Geodesy and Geodynamic•. 2009, 29 ( oopp) ,
nonlinearity of the spring drift. The period correction
15 -19. (in Chinese)
shows that there was a ( 3 - 4) X 10 -• ms - 2 gravity
[7 ]
anomaly before and after the Ms7. 0 Lushan earthquake. Mter the earthquake, the residual gravity took
fornia: SanDiego. [ 8 ]
approximately 430 minutes to relax to a gravity change 8
of approximately 2. 5 x 10 - ms -
2
•
of ( 0. 59 ± 0. 40)
X
present-day crustal deformation in China. Beijing, 2009, Institute
[9 ]
Zhang Xiaohong, Guo Fei, Guo Bofeng, et al.
Coseismic
displacement monitoring and wave picking with high-frequency
10- 8 ms - 2 was recorded at Pixian
GPS. Chioeoe !. Geophy,,, 2012, 55(6), 1912-1918. doi,
by the GS15 gravimeter during the Ms7. 0 Lushan earthquake.
Wang Ming. Analysis of CPS data with high precision & study on
of Geology, China Earthquake Administration. (in Chinese)
With both periodic
and relaxation corrections, a coseismic gravity change
Rosanne Nikolaidis. Observation of geodetic and seismic deforma-
tion with the global positioning system. 2002, University of Cali-
10.6038/ j. ;.,., 0001-5733. 2012. 06. 012. (in Chinese) [ 10]
Tan Hongbo, Shen Chongyang, Li Hui, et al. Simulation of post-
The simulated coseismic gravity change, as well as
seismic gravity change and defOl'liUIIion of the Wenchuan earth-
the change measured by the GS15 gravimeter, was
quake based on viscoelsati.c layered falf-space mode. Acta Seis-
approximately 0. 5
X
10 -• ms - 2 • The coseismic vertical
mologica Sinica. 2009 , 31 ( 5 ) : 491 - 505.
[11]
WeiJin, LiHui, LiuZiwei, etal. Observationofsuperconduct-
displacements simulated with one-day and one-second
ing gravimeter at Jiufeng seismic station. Chinese J. Geophys,
time interval GPS records were approximately - 1 to
2012,55(6), 1894-1902. doi, 10. 6tl8/ j ...... 0001-5733.
- 2 mm. The ratio of the coseismic gravity changes to the vertical displacement is consistent with the theoretical ratio. In conclusion, the GS15 gravimeter at the Pixian station detected a half-microgal coseismic gravity change in the Ms7. 0 Lushan earthquake, and the dislocation
2012. 06. 010. (in Chinese) [12]
Wei lin, Lin Gaoohuan, et al. Analysis of non-linear drift of spring gravimeter. Journal of Geodesy and Geodynamics, 2012, 32(5), 137-142. (in Chin""')
2013,4(3) :7-11
http :llwww. jgg09. com Doi:l0.3724/SP.J.l246.2013.03007
Detection of a half-microgal coseismic gravity change after the Ms7. 0 Lushan earthquake Wei Jin 1 "2 "3 , Zhao Bin 1 "2 , Tan Hongho 1 "2 , Yu Dan4 , Shen Chongyang 1 "2 and Li Hui 1 "2 1
&y LolJoratory of Earthquake Geodesy, lnstiruu of Sei.smology, China Eart/ujuake Admim..tration, Wuhan 430071 , China
2
Wuhan Base of lnstiruu of Crustal Dynamics, China Eartlujuake AdminUtration, Wuhan 430071 , China
' School 4
of Getxksy and
Geomntics, Wuhan University, Wuhan 430079, China
China Earthquake Networks Center, Beijing 100045, China
Abstract: Because only a small near-field coseismic gravity change signal remains after removal of noise from the accuracy of observations and the time and spatial resolution of the earth's surface gravity observation system, it is difficult to verify simulations of dislocation theory. In this study, it is shown that the GS15 gravimeter, located 99.5 km from the epicenter of the Ms7. 0 Lushan earthquake on April 20, 2013 at 08 :04 UTC + 8, showed the iofluence of the earthquake from 2013-04-16 to 2013-04-26 after a time calibration, tide corrections, drift correction, period correction and relaxation correction were applied to its data. The post -seismic
relaxation process of the spring in the gravimeter took approximately 430 minutes and showed a 2. 5
X
10 -• ms _,
gravity change. Mter correcting for the relaxation process, it is shown that a coseismic gravity change of
approximately +0.59 ±0.4 xlO-'ms- 2 was observed by the GS15 gravimeter; this agrees with the simulated gravity change of approximately 0. 31
X
10 -• ms _,. The rate of the coseismic gravity change and the coseismic
vertical displacement, as measured by one-second and one-day sampling interval GPS units, is also consistent
with the theoretical rate of change. Therefore, the GS15 gravimeter at the Pixian Station observed a coseismic gravity change after the Ms7. 0 Lushan earthquake. This and similar measurements could be applied to test and confirm the theory used for these simulations.
Key words: GS15 gravimeter; coseismic gravity change; the Ms7. 0 Lushan earthquake
the coseismic gravity changes caused by powerful earth-
1
Introduction
quakes (the 2003 Mw7. 9 Tokachi-oki earthquake; the 2007 Mw9. 3 Sumatra earthquake; the 2010 Mw8. 8
A powerful earthquake always results in a change in the
Chile earthquake) have been acquired by high-preci-
gravity potential field of the earth that can be observed
sion , time-variable gravity measurements from networks
as a small change in the local gravitational acceleration.
of superconducting gravimeters
Such coseismic gravity changes have been observed after
These measurements have been verified by simulations
the 1964 Mw9. 2 Alaska earthquake by mobile relative
based on fmite rectangular dislocation theory and the
gravinretry[IJ and after the 1998 M6. 1 earthquake near
local layered wave velocity structure of the crust and
lwate-san Mountain by ai3solute gravimetry[']. Recently,
upper mantle, which calculate the surface coseismic
and
GRACEc•.•l.
and post-seismic gravity changes[SJ. There are several Received :2013 -05 -09 ; Accepted :2013 -05-28 Corresponding author: Wei Jin, E-mail: pierce212@ 163. com. This work is supported by the National Natural Science Foundation of China ( 41204058) and the Rwming Foundation of the Gravity Network Center of China (201301008).
difficulties in validating the accuracy of these simulations using the current observation system , including
the low precision and low temporal resolution of both the vertical displacement measured by GPS and the
8
Geodesy and Geodynamics
Vol.4
time-variable gravity measured by GRACE. These are
d 1 are the constant and linear terms of the residual
largely inadequate to use at high temporal resolution for
gravity field, respectively, and the coefficients
checking the fast coseismic gravity changes that occur
bj describe the residual diurnal and semidiurnal wave
during earthquakes. Furthermore, due to the powerful
signals of the residual gravity, respectively. The next
shock of earthquakes, there are few near-field, ground-
term corrects for any number ( N,) of offsets that have
that can confmn the existence of
magnitudes g and intervals T,. The post-seismic gravi-
physical phenomena suggested by observations of co-
ty change is modeled as a mte of change hj and an ex-
seismic gravity changes.
ponential decay with magnitude kj over the selected
based gravimeters
On April20, 2013 at 08 :02 :46 UTC +8, a Ms7. 0
aj
and
earthquake time intervals Tloj and T0 • The measurement
(Mw6. 6) earthquake occurred in Lushan, in the Si-
errors v are initially assumed to be independent , iden-
chuan Province. The GS15 gravimeter located approxi-
tically distributed and mndom, with E(v) =0.
mately 99. 5 km from the epicenter, at the Pixian seis-
The p-wave arrival time (2013.()4-20 08:03:04.67
mic station, captured a complete record of the earth-
UTC + 8 ) identified in the Pixian Station seismometer
quake at a one-minute sampling interval 1•l • In this
record is used to calibrate the site' s GS15 gravimeter.
study, we analyzed the seismic record from April 16 to
Several years of synthetic tides are determined from this
26 and applied time calibmtions, tide and drift correc-
gravity data record. An automatic pikes processing
tions and corrections by period and relaxation analysis.
method is used to correct the sudden jumps in signal
The gravimeter data were used to check and confmn the
caused by the electronics and the far-field earthquake 1'l.
near-field coseismic gravity change occurring as a result
The tide , air pressure , drift and period corrections and
of the Ms7. 0 Lushan earthquake.
the residual gravity are shown in figure 1. Gravimeter records are susceptible to spurious sig-
nals from the step and sticking bar; these signals can
2 Data processing methods
strongly influence the long-term drift behavior. Figure
The gravity data were sampled at one-minute intervals by the GS15 gravimeter at the Pixian seismic station. The data were influenced by changing air pressures and solid tides but reflect timing errors from the data acqui-
sition system, errors in the residual diurnal and semidiumal wave signals corrected by tide theory, drift of the spring in the gravimeter, coseismic gravity changes and }X)St-seismic relaxation effects [7 ]. To simulate these er-
1 shows that the GS15 gravimeter has no such instrumentation problems. After subtracting the influence of the gravity tide ( ±200 spring ( -100 x ( ± 4 x 10 -• ms -
2
xto-• ms- 2 ) ,
to-• ms ) ,
2
)
the drift of the
and the air pressure
a gravity anomaly of approximately 2
(3 -4) x 10 -• ms - remains in the period correction before and after the Ms7. 0 Lushan earthquake. The period correction comprises the error in the time system, the synthetic tide model, the coefficient of the
ror sources , the following model is used :
scale factor, the nonlinear drift of the spring and the influence of the earthquake.
3
Calculation and analysis of the coseismic gravity changes
N,
L hjH(t; -
When the spring in the GS15 gravimeter receives seis-
Tlif)t; +
j=l
mic waves , the original dynamic equilibrium of the
N,
L kjexp[- (t; - r.)lrj]H(t; - T,)t; +v; j=l
(1)
spring suddenly breaks , and the spring undergoes a relaxation process. In this study, the relaxation process
is simulated with the exponential portion of equation 1. Here, t, for i = 1· · · N are the time intervals in units of
This relaxation analysis , shown in figure 2 , finds an
minutes, g~ ( t;) is the residual gravity corrected by
optimal
tidal theory , H is the Heaviside step function , d0 and
proximately 0. 40 x 10 -• ms - 2 •
T
of 430 minutes and a residual error of ap-
No.3
Wei Jin,et al. Detection of a half-microgal coseismic gravity change after the Ms7. 0 Lushan earthquake
9
§ -440! "£1;;;'-460 ~ Ei -480 8. -500
~ ~-520 ~ ~ -540
Figure 1
The tidal, air pressure, drift and period corrections for the GS15 gravimeter at Pixian (The pressure admittance is -3.79
X
10 -s ms - 2 mbar. Over the 10-day period, the pole tide is linear)
0.49 .r 0.48 ·8 o.47 '6 0.46 ~ 0.45 § 0.44 " 0.43 The confidence interval of 90% ] 0.42 . 470 390 I :9 0.41 1+-'-'~ 0.40 I 0.39 0.38L___L_~--~--~~L_~L_~--~--_L_ __L~ 0 100 200 300 400 500 600 700 800 900 1000 Relaxation time (miniute)
Figure 2
Using the relaxation time ( 430 minutes) and the pwave arrival time (2013-04-20 00:04:00) as constraints[9J, the entire coseismic and post-seismic processes observed by near-field ground-based gravimeters are simulated from the residual gravity record from April 16 to 23 , which is shown in figure 3. The post-seismic relaxation process indicated by the springs of such gravimeters is approximately + 2. 5 X 10 - 8 ms - 2 • The post-seismic gravity
The relaxation time fitting result for the Lushan
change exhibits a negative trend, as before. The coseismic
earthquake observed at Pixian
gravity change is approximately (0.59 ±0.40) xl0- 8 ms- 2
2.0 1.5
~ 1.0
~
0.5
~
f
0
-~ -0.5
a
-1.0
-1.5 -2.0L---~--~--~~~~uw~~~~~--~~~--~--~----~~~~
,(,'t~ ,(,'t\1. •1't~ •'1-t\'l. •'i>'tf$! •'0-t\'l. •9'tl;)l:l •9-t\'l. ,~;~'tl:l\;1 ,~;~'t\1. ,\'t~;~l:l ,\'t\1. ,'J;'tl;)l:l ,'J;'t\1. ,~'t~;~l:l ~·~·~· ~·~· ~·~· ~·~ ~~ ~~~ ~~ Time (ddThh)
Figure 3
Coseismic gravity change in the Lushan earthquake at Pixian ( The blue line is the gravity change at Pixian Station, the magenta line is the relaxation process and the red line is the simulation result for the coseismic gravity change in the Lushan earthquake at Pixian Station)
10
Geodesy and Geodynamics
Vol.4
approximately 0. 5 X 10- 8 ms - 2 in the Pixian region
for the Lushan earthquake observed at Pixian. The coseismic vertical displacement measured by
after the Ms7. 0 Lushan earthquake. In cases where the
high-frequency GPS and gravity changes measured by
observed value and the precision are of similar magni-
the GS15 gravimeter during the earthquake ( from
tude, we used the vertical displacement of the one-
2013-04-20: 08 : 00 : 00 to 08 : 06 : 37 UTC + 8) are
second sampling interval GPS because its precision is
shown in figure 4. These two measurements sample the
higher than that of the daily sampling interval GPS.
same region. The coseismic vertical displacement is
The vertical displacement changes exhibit negative
approximately - ( 1. 1 ± 2. 1 ) mm , as determined by
trends, while the simulated coseismic gravity change
piecewise polynomial fitting.
and that observed by the GS15 gravimeter both exhibit
The statistics of the coseismic gravity change meas-
positive trends. The coseismic gravity change observed
ured by the GS15 gravimeter, the simulated gravity
by the GS15 gravimeter shows local changes, while the
change, and the vertical displacements of the GPS with
simulation captures the regional change. Including the
sampling intervals of one day and one second are shown
coseismic gravity change indicated by the GPS meas-
in table 1. The coseismic gravity changes from the GPS
urements using the theoretical gravity gradient, all
measurements with a theoretical gravity gradient of
gravity changes are approximately 0. 5
-300 x 10-8 ms - 2 m -I are approximately 0. 33 to 0. 63 mm.
Thus, the GS15 gravimeter at Pixian detected a half-
Table 1 shows that both the ground-based coseismic
microgal coseismic gravity change after the Ms7. 0
gravity change and the simulated gravity change are
Lushan earthquake.
.8
8
~r~~~tM:P~~mro-___;_-i'W1 °· 6 ~
]'7 ';::6
0.4 'b
0.2 :::;
5 4 ~----------~--------~~~~_r
]
t
0
~
-0.2 -5
3 ~ 2
~
10- 8 ms - 2 •
1.0 0.8
9
I
X
-
1 0
The vertical displacement of GPS The coseismic vertical displacment of GPS The Lushan earthquake The coseismic gravity change of GSIS
-1 00:00 00:24 00:48 01:12 01:36 02:00 02:24 02:48 03:12 03:36 04:00 04:24 04:48 05:12 05:36 06:00 06:24
-0.4·~ -0.6 -0.8 -1.0
a
Time (second)
Figure 4
Coseismic vertical displacement and gravity changes from high-frequency GPS and the GS15 gravimeter for the same region as shown in figure 3 (The blue line is the vertical displacement indicated by the 1 Hz GPS during the Lushan earthquake, the black line is the model of the coseismic vertical displacement indicated by the 1 Hz GPS, the red line is the model of the coseismic gravity, and the magenta line indicates the occurrence time of the Ms7. 0 Lushan earthquake)
Table 1
Statistics of the coseismic gravity and displacement changes of the Ms7. 0 Lushan earthquake, as measured at Pixian Station.
Obseration type
GPS GS15 Simulation result
Longitude (oE)
103.7 103.8
Latitude (oN)
30.9 30.9
Sampling interval (second)
86400 60
Coseismic vertical displacement ( mm)
Coseismic gravity change (10- 8 ms- 2 )
value
precision
-1.1
± 1.40
0.33
value
-2.1
±3.11
0.63 0.59
precision
±0.40
0.31
Note: The coseismic gravity change with the dislocation model of the earthquake source is simulated by the fault model used in seismic wave inversion offered by Institute of Geophysics, China Earthquake Administration [!OJ • Following the assumption that coseismic gravity changes are caused entirely by vertical displacement, the simulation of the gravity changes from the GPS data uses a gravity gradient of -300 X 10 -s ms - 2 m _, [u].
No.3
11
Wei Jin,et al. Detection of a half-microgal coseismic gravity change after the Ms7. 0 Lushan earthquake
model can be validated with this method in the future.
4
Conclusion and discussion References
Tills paper demonstrates that the drift of creep characteristics can be corrected not only by a time system cal-
[ I ]
ibration and the traditional correction method["], but also by a conducting period and relaxation analysis of
JZ071 i002]ffl451. [2 ]
the spring gravimeter. The traditional method finds that ms -
•
Simulated with the Heaviside step function and a
2001, 28(15)' 2979-2981.
[3 ]
X
to-'
2004, 306: 476-478. doi:l0.1126/science.l101875.
2
ms - ,
[4 ]
which agrees with the one-day and one-second GPS
quake. Chineoe !. Geophy•. 2011, 54(7), 1745 -1749. doi,
The data record of the GS15 gravimeter at the Pixian
10.3969/j. issn. 0001-5733. 2011.
[5 ]
rn.
007. (in Chinese)
Shen Chongyung, Li Hui and Tan Hongbo. Simulation of coseismic gravity changes and defonnati.on effect of wenchan Ms8. 0 earth-
drift, solid earth tide and air pressure and is also affected by a number of important and inevitable error
Zbou X, Stm W K and Fu G Y. Gravity satellite GRACE detects
coseismic gravity changes caused by 2010 Chile Mw8. 8 earth-
records and the simulated result for the same region.
Station is sensitive to a number of factors related to
Yuichi lmanishi , et al.. A Network of Supereonducting Gravimeters Detects Submicrogal Coseismic Gravity Changes. Science,
relaxation function, the coseismic gravity change measured
by the GS15 gravimeter is ( 0. 59 ± 0. 40)
Tanaka Y, Okubo S, Macbida M, et al. First detection of absolute gravity change caused by earthquake. Geophys. Res. Lett.
the residual gravity change is approximately ± 1 x 10 -• 2
B=es D. Gmvity Change during the AJa.ka earthquake. !. Geophyo. Reo., 1966, 71(2),451-457. doi, 10.1029/
quake. Geodesy aod Geodyoamic•. 2008 , 28 ( 5) , 6 -12. [ 6 ]
sources , such as error in the time system , the synthetic
Wei Jin, Hao Hongtao, et al. Gravity disturbance of high frequency at Chengdu seismostation before Wenchuan Ms8. 0 earth-
tide model, the coefficient of the scale factor and the
quake. Journal of Geodesy and Geodynamic•. 2009, 29 ( oopp) ,
nonlinearity of the spring drift. The period correction
15 -19. (in Chinese)
shows that there was a ( 3 - 4) X 10 -• ms - 2 gravity
[7 ]
anomaly before and after the Ms7. 0 Lushan earthquake. Mter the earthquake, the residual gravity took
fornia: SanDiego. [ 8 ]
approximately 430 minutes to relax to a gravity change 8
of approximately 2. 5 x 10 - ms -
2
•
of ( 0. 59 ± 0. 40)
X
present-day crustal deformation in China. Beijing, 2009, Institute
[9 ]
Zhang Xiaohong, Guo Fei, Guo Bofeng, et al.
Coseismic
displacement monitoring and wave picking with high-frequency
10- 8 ms - 2 was recorded at Pixian
GPS. Chioeoe !. Geophy,,, 2012, 55(6), 1912-1918. doi,
by the GS15 gravimeter during the Ms7. 0 Lushan earthquake.
Wang Ming. Analysis of CPS data with high precision & study on
of Geology, China Earthquake Administration. (in Chinese)
With both periodic
and relaxation corrections, a coseismic gravity change
Rosanne Nikolaidis. Observation of geodetic and seismic deforma-
tion with the global positioning system. 2002, University of Cali-
10.6038/ j. ;.,., 0001-5733. 2012. 06. 012. (in Chinese) [ 10]
Tan Hongbo, Shen Chongyang, Li Hui, et al. Simulation of post-
The simulated coseismic gravity change, as well as
seismic gravity change and defOl'liUIIion of the Wenchuan earth-
the change measured by the GS15 gravimeter, was
quake based on viscoelsati.c layered falf-space mode. Acta Seis-
approximately 0. 5
X
10 -• ms - 2 • The coseismic vertical
mologica Sinica. 2009 , 31 ( 5 ) : 491 - 505.
[11]
WeiJin, LiHui, LiuZiwei, etal. Observationofsuperconduct-
displacements simulated with one-day and one-second
ing gravimeter at Jiufeng seismic station. Chinese J. Geophys,
time interval GPS records were approximately - 1 to
2012,55(6), 1894-1902. doi, 10. 6tl8/ j ...... 0001-5733.
- 2 mm. The ratio of the coseismic gravity changes to the vertical displacement is consistent with the theoretical ratio. In conclusion, the GS15 gravimeter at the Pixian station detected a half-microgal coseismic gravity change in the Ms7. 0 Lushan earthquake, and the dislocation
2012. 06. 010. (in Chinese) [12]
Wei lin, Lin Gaoohuan, et al. Analysis of non-linear drift of spring gravimeter. Journal of Geodesy and Geodynamics, 2012, 32(5), 137-142. (in Chin""')