Tectonophysics 679 (2016) 211–217

Contents lists available at ScienceDirect

Tectonophysics

j ourna l homepage: www.e lsev ie r .com/ locate / tecto

Co-seismicwater level changes in response tomultiple large earthquakesat the LGH well in Sichuan, China

Guijuan Lai a,⁎, Changsheng Jiang a, Libo Han a, Shuzhong Sheng b, Yuchuan Ma c

a Institute of Geophysics, China Earthquake Administration, Beijing 100081, Chinab Institute of Disaster Prevention, Hebei, Sanhe 065201, Chinac China Earthquake Networks Center, Beijing 100045, China

⁎ Corresponding author.E-mail address: laigj@cea-igp.ac.cn (G. Lai).

http://dx.doi.org/10.1016/j.tecto.2016.04.0470040-1951/© 2016 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 12 January 2016Received in revised form 12 April 2016Accepted 28 April 2016Available online 10 May 2016

We examined the water level data at the LGHwell in Sichuan, China, from December 2007 to July 2015 and theirresponses to multiple large earthquakes with seismic energy densities greater than 10−4 J/m3. Co-seismic waterlevel declineswere observed in response to eleven earthquakes out of twelve in the farfield, and co-seismicwaterlevel increasewas observed in one nearfield case. The water level declines in the farfield showed a linear relationwith the common logarithmof the seismic energy densities, whereas thewater level increase in the nearfield fellaway from this relation, indicating that the farfield responses and the nearfield response were produced by dis-tinct mechanisms. We used the phase shift of tidal responses as a proxy for permeability and found that perme-ability enhancements were observed both in the farfield and nearfield. The co-seismic water level declines inresponse to the distant earthquakes could be explained by permeability enhancements caused by the passageof seismic waves through the mobilization of colloidal particles; the co-seismic water level increase in responseto the nearfield case could be caused both by the compression of the static stress and by the seismic waves.

© 2016 Elsevier B.V. All rights reserved.

Keywords:Water levelSeismic energy densityStatic stressDynamic stressPermeability enhancement

1. Introduction

Large earthquakes have long been reported to cause various hydro-logical responses, such as variations in the well water level (e.g., Liuet al., 1989; Roeloffs, 1996; Manga and Wang, 2007; Chia et al., 2008;Yan et al., 2014; Kinoshita et al., 2015), water temperature (Mogiet al., 1989; Shi et al., 2007; Wang et al., 2013; Shi and Wang, 2014;Ma, 2015) and stream discharge (Manga et al., 2003; Montgomeryand Manga, 2003; Wang et al., 2004; Mohr et al., 2012). Significantadvances have been made in the mechanisms of water level changesduring earthquakes.

Earthquakes can cause changes in static stress and dynamic stress;both decay with distance, but the static stress decays more quicklythan dynamic stress (Lay and Wallace, 1995). In the farfield (manyfault lengths), the static strain is too small to explain the co-seismicsteps in the well water level. Brodsky et al. (2003) propose that thewater level steps can be caused by dynamic stress (seismic waves)that enhances permeability through the mobilization of colloidal parti-cles in pores and fractures of aquifers. The hypothesis is supported bysubsequent field observations (Doan et al., 2006; Wang et al., 2009;Xue et al., 2013; Shi et al., 2015) and laboratory experiments(Elkhoury et al., 2006, 2011; Manga et al., 2012; Candela et al., 2014,

2015), though shaking-induced permeability decrease is also observedin experimental study (Liu andManga, 2009). However, in the nearfield(within 1–2 fault lengths of the ruptured fault), the effects of dynamicstress and static stress are more difficult to differentiate. In somecases, co-seismic water level changes are thought to be caused by staticstrain, in which thewater level rises in contraction zones and falls in di-latation zones (Wakita, 1975; Quilty and Roeloffs, 1997; Jonsson et al.,2003; Shibata et al., 2010; Zhang and Huang, 2011; Shi et al., 2013).On the other hand, in other cases, both the pattern and magnitude ofco-seismic water level changes are observed to not be coincident withporoelastic strain (Koizumi et al., 2004; Shi et al., 2014), indicatingthat dynamic strain can also change the water levels.

In thiswork, we examined thewater level data at the LGHwell, fromDecember 2007 to July 2015, and their responses to multiple largeearthquakes both in the farfield and in the nearfieldwith seismic energydensities greater than 10−4 J/m3. We use the phase shift of tidal re-sponses as a proxy for the permeability and analyze its variations afterearthquakes, which would help us better understand the mechanismsof co-seismic water level changes at the well.

2. Well setting and data

Lugu Lake is located at the junction of Sichuan and Yunnan province.It is a rift plateau lake and the elevation is 2690m. The LGHwell is locat-ed at 100.85° E, 27.73° N, which is ~3.6 km to the east of the Lugu Lake(Fig. 1a). The LGHwell is in the vicinity of the region between the south

Fig. 1. (a) Locations of the LGHwell and the Lugu Lake fromGoogle Earth; (b) Locations ofthe LGH well (blue filled triangle) and the nearby faults in the Sichuan–Yunnan block;(c) Lithologic log of the LGH well. Perforated sections of the casing are located at102.00–106.70 m, 122.13–133.60 m and 179.26–183.72 m, where highly conductivefractures exist. The water level sensor (rectangle) is located at a depth of 9 m. (Forinterpretation of the references to color in this figure legend, the reader is referred tothe web version of this article.)

212 G. Lai et al. / Tectonophysics 679 (2016) 211–217

boundary (the Zhongdian fault zone) of the Sichuan and Yunnan activeblock and its secondary block boundary (NE-trending Xiaojinhe–Lijiang–Jianchuan fault zone) (Fig. 1b). These faults show Holocene

activity, and many historical strong earthquakes occurred in this region(Wen and Yi, 2003).

The lithologic log of the LGH well is shown in Fig. 1c. At depths of0.0–71.52 m, the borehole and casing diameters are 219 mm and146 mm, respectively; and at depths of 71.52–200.07 m, the boreholeand casing diameters are 130mmand 127mm, respectively. Perforatedsections of the casing are located at 102.00–106.70m, 122.13–133.60mand 179.26–183.72 m, where highly conductive fractures exist. TheQuaternary marine sediments, which are made of breccia, mudstone,fine sandstone and sandstone, are above 61 m, and the Permian lime-stones are distributed at depths of 61–200.07 m.

Thewater level at the LGHwell has been observed by a LN-3A instru-ment with a 1-min sampling interval since December 2007, with thesensor located at a depth of 9 m. The water level data are transmittedto the Data Management Center of Xichang, in Sichuan, by remotedata transmission of CDMA, and then sent to the China Earthquake Net-works Center (Chen et al., 2011). The LGH station is a permanent verybroadband station that is ~2 km away from the LGH well. The groundvelocity is recorded by a CMG-3T seismometer with corner frequency120 s–50 Hz, and the collector type is EDAS-24IP (Wang et al., 2014).

3. Co-seismic water level changes vs seismic energy density

Previous studies found that sustainedwater level changes are gener-ally observedwhen the seismic energy density is greater than 10−4J/m3

(Wang and Chia, 2008; Manga et al., 2012). During the observation ofthe water level at the LGHwell, strong earthquakes occurred on a num-ber of occasions, and there were a total of fourteen earthquakes withseismic energy densities greater than 10−4J/m3 at the well (Fig. 2,Table 1), including two earthquakes in the nearfield (the MW 7.9Wenchuan earthquake on May 12, 2008 and the MW 5.6 Ninglangearthquake on June 24, 2012) and 12 earthquakes in the farfield. Thetime (date), magnitude and epicentral distance (calculated from theepicenter coordinate) for the fourteen earthquakes were obtainedfrom the results of Global CMT (Dziewonski et al., 1981; Ekströmet al., 2012; Table 1).

We collated the ground velocity waveforms for the fourteen earth-quakes at the LGH station (Data Management Centre of China NationalSeismic Network, 2007; Zheng et al., 2010; Fig. S1 in the supplement),and found that only the waveforms for the farfield earthquakes wereavailable. For the twelve farfield earthquakes,we focused on the verticalcomponent of the ground velocity waveforms. We first demeaned eachwaveform and then picked out the peak ground velocity vz. The seismicenergy density e is determined by (Lay and Wallace, 1995; Wang,2007):

e≈12ρvz2 ð1Þ

where ρ is the rock density and we assume it to be 2.64 g/cm3 for thelimestone at the LGH well. The calculated seismic energy densitybased on the seismic waveforms is shown in Table 1.

There is a statistical empirical relation between the seismic energydensity e, the surface wave magnitude M and the epicentral distanceD For California (Wang, 2007):

log10e ¼ −3 log10Dþ 1:5M−4:2: ð2Þ

Wemade a similar regression analysis for the LGHwell and the scal-ing is as following:

log10e ¼ −2 log10Dþ 1:3M−6:1: ð3Þ

The correlation coefficient R2 equals 0.68 and the P-value is 0.006.We calculate the approximate seismic energy density for theWenchuanand Ninglang earthquake based on Eq. (3), which is 0.139 J/m3 and0.044 J/m3, respectively.

Fig. 2. Locations of the LGH well (yellow filled triangle) and the fourteen earthquakes with seismic energy density greater than 10−4 J/m3 at the well. “Beach ball” shows the lowerhemisphere projection of focal mechanism (from Global CMT) for each earthquake, and red, green and blue colors indicate thrust, strike-slip and normal fault earthquakes,respectively. Black lines indicate the location of surface faults on the Chinese mainland (Deng et al., 2006). (For interpretation of the references to color in this figure legend, the readeris referred to the web version of this article.)

213G. Lai et al. / Tectonophysics 679 (2016) 211–217

The water level (the distance between the wellhead and thewater surface inside the well) observation records at the LGH wellare shown in Fig. 3, where the red dashed vertical line correspondsto the time of the fourteen earthquakes. The records one day beforeand after the time of the earthquakes for three representative co-seismic responses (water level increase for the Wenchuan earth-quake, no observed response for the Yao'an earthquake and waterlevel decline for the Tohoku earthquake) are shown in Fig. 4, andthe co-seismic water level changes in response to the fourteen earth-quakes are shown in the supplementary materials (Fig. S2). Becauseof the lack of co-seismic data for the Ninglang earthquake due to bro-ken instruments, we will not perform any analysis for this earth-quake in this work.

The sign and magnitude of the available co-seismic water levelchanges in response to the earthquakes are also shown in Table 1. The

Table 1Basic information for the fourteen earthquakeswith seismic energydensity greater than10−4 J/mphase shift changes. “–”means no change or the change is not clear; “NaN”means lack of data. Bdue to instruments, we include “NaN”. The occurrence time of the 3rd and 4th earthquakes, anphase changes for these four earthquakes.

Eventnumber

Earthquakes Date (UTC) Magnitude(from Global CMT)

1 Wenchuan, China 2008/05/12 MW 7.9/MS 8.12 Yao'an,China 2009/07/09 MW 5.6/MS 5.73 Yushu, China 2010/04/13 MW 6.9/MS 7.04 Tohoku, Japan 2011/03/11 MW 9.1/MS 9.05 Myanmar 2011-03-24 MW 6.8/MS 7.16 off the west coast of northern Sumatra 2012-04-11 MW 8.6/MS 8.67 off the west coast of northern Sumatra 2012-04-11 MW 8.2/MS 8.28 Ninglang, China 2012-06-24 MW 5.6/MS 5.59 Myanmar 2012-11-11 MW 6.8/MS 6.810 Lushan, China 2013-04-20 MW 6.6/MS 6.811 Deqin, China 2013-08-31 MW 5.7/MS 5.612 Ludian, China 2014-08-03 MW 6.2/MS 6.213 Kangding, China 2014-11-22 MW 6.1/MS 5.914 Lamjung, Nepal 2015-04-25 MW 7.9/MS 7.9

water level increase observed at the LGHwell is shown to be a responseto theMW 7.9 Wenchuan earthquake on May 12, 2008 in the nearfield,and the water level declines are observed to be responses to elevenearthquakes in the farfield. No co-seismic change is observed duringtheMW 5.6 Yao'an earthquake.

The observed co-seismic water level changes varying with thecommon logarithm of seismic energy density are shown in Fig. 5. Forthe eleven earthquakes in the farfield, we fit the co-seismic waterlevel decline Δh with the common logarithm of the seismic energydensity e, and obtained a linear relationship:

Δh ¼ −0:0623 log10e−0:2120: ð4Þ

The correlation coefficient R2 equals 0.62 and the P-value is 0.004,indicating a linear relation between the co-seismic water level decline

3 at the LGHwell, aswell as the corresponding co-seismicwater level, amplitude ratio andecause the datawithin severalmonths after the 8th, 9th, and 11th earthquakes aremissingd the 5th and 6th earthquakes is too close to identify the exact co-seismic amplitude and

Epicentraldistance (km)

Co-seismicwater levelchanges (m)

Amplituderatio change(×10−7 1/m)

Phase shiftchange (°)

Seismic energydensity (J/m3)

440 0.020 −0.4 18.9 1.39 × 10−1

248 – −0.3 1.2 3.086 × 10−4

732 −0.056 −1.7 3.5 1.24 × 10−2

4027 −0.174 – – 1.95 × 10−2

778 −0.066 – – 4.2 × 10−3

2945 −0.108 – – 9.2 × 10−3

3119 −0.002 – – 2.3 × 10−3

16 NaN NaN NaN 4.38 × 10−2

734 −0.040 NaN NaN 1.1 × 10−3

354 −0.103 −0.5 4.0 6.4 × 10−3

150 −0.015 NaN NaN 1.973 × 10−4

252 −0.026 −0.1 1.1 2.6 × 10−3

293 −0.002 −0.2 0.2 7.567 × 10−4

1586 −0.121 −1.5 6.6 4.58 × 10−2

2008 2009 2010 2011 2012 2013 2014 2015

2.6

2.8

3

3.2

3.4

3.6

3.8

4

4.2

4.4

4.6

Time (year)

Wat

er le

vel (

m)

1 2 3 4 5 67

8 9 10 11 12 13 14

Fig. 3.Water level (the distance between the wellhead and the water surface inside the well) records at the LGH well from December 1, 2007 to July 1, 2015. Red dashed vertical linesindicate the time of the fourteen earthquakes which are described in Table 1. (For interpretation of the references to color in this figure legend, the reader is referred to the webversion of this article.)

214 G. Lai et al. / Tectonophysics 679 (2016) 211–217

and the seismic energy density in the farfield. However, the water levelchange due to the Wenchuan earthquake in the nearfield falls awayfrom this relation (Fig. 5), indicating that the farfield responses andthe nearfield response were produced by distinct mechanisms.

4. Discussion

4.1. Static strain changes

The poroelastic strain at the LGH well caused by the Wenchuanearthquake in the nearfield has been calculated by Zhang andHuang (2011) and Shi et al. (2013).We also calculate the static strainbased on the Okada model (Lin and Stein, 2004; Toda et al., 2005),which turns out to be about −3 × 10−8 (positive for dilatation,Fig. 6) and close to their results. The amplitude response at theLGH well before the Wenchuan earthquake is approximately8.6 × 10−7/m (Fig. 7), thereby the predicted co-seismic water levelrise for this earthquake is 0.035m. The sign is consistent with the ob-served water level change, however, themagnitude is larger than theobserved rise of 0.020 m.

The poroelastic strains at the LGH well caused by the other twelveearthquakes in the farfield are less than 10−10, too small to explainthe co-seismic water level declines range from a few centimeters to adozen centimeters, indicating that the effects of static strain on the co-seismic water level change in the farfield are negligible, and the waterlevel changes can only be caused by dynamic strain.

4.2. Permeability enhancement

4.2.1. Tidal response methodThe amplitude andphase responses of thewater level to Earth's tides

have been used as measurements of an aquifer's specific storage andpermeability, respectively (Doan et al., 2006; Elkhoury et al., 2006;Hsieh et al., 1987; Xue et al., 2013). Tidal volumetric strains imposedby the Earth's tides are calculated using the “SPOTL” program (Agnew,1997, 2012; Berger et al., 1987). The tidal responses are obtained inthe time domain by the least squares fitting method (Hsieh et al.,1987); we included O1, K1, M2 and S2 in the tidal analysis. Considering

that the M2wave has the strongest energy and a relatively low contam-ination by barometric pressure, we focus on theM2wave in thiswork asmost studies also do.

The data are processed as described by Lai et al. (2014). Herewe onlyprovide a brief summary: first, we interpolate, demean and detrend thedata; second, we filter the water level and synthetics between 10 h and30 h using a 2-pass second-order Butterworth band-pass filter; third,we divide the data into different segments, with each segment being29.5 days, to distinguish the M2 wave from the S2 wave and to reducethe spectral leakage (Doan et al., 2006), and overlap the segments by2.95 days; finally, we apply the least squares fitting procedure to eachsegment.

4.2.2. Tidal response observationsThe amplitude and phase responses at the LGH well are shown in

Fig. 7. Both phase leads (positive phase) and lags (negative phase) areobserved. As mentioned by previous authors (Hsieh et al., 1987;Roeloffs, 1996; Elkhoury et al. 2006; Lai et al., 2014), the phase shift isa combination of the phase lag due to the borehole storage effect andthe phase lead due to the water table drainage, and a phase shiftincrease implies a permeability increase.

We use the first value after an earthquake to subtract the lastvalue before an earthquake to then obtain the co-seismic amplitudeand phase changes for the earthquakes. The amplitude and phasechanges for earthquakes that allowed unambiguous analysis arelisted in Table 1. For all these earthquakes, phase increased afterthe earthquakes, indicating that permeability was enhanced afterthe earthquakes.

We fit the co-seismicwater level changeΔhwith the available phasechangeΔφ for the six earthquakes in the farfield (Fig. 8), and obtain thefollowing relationship:

Δh ¼ −0:0190Δφ−0:0031: ð5Þ

The correlation coefficient R2 is as high as 0.92 and the P-value is0.011, indicating that the water level declines are highly related to thephase increases. Because the phase increases are proxies of permeabilityenhancements (Hsieh et al., 1987; Roeloffs, 1996; Lai et al., 2014), we

2008/05/12 2008/05/13

3.52

3.53

3.54

3.55

3.56

3.57

3.58

3.59

1

Wat

er le

vel (

m)

2009/07/09 2009/07/10

3.53

3.54

3.55

3.56

3.57

3.58

3.59

2

Wat

er le

vel (

m)

2011/03/11 2011/03/12

3.4

3.45

3.5

3.55

3.6

4

Wat

er le

vel (

m)

Fig. 4. Three representative co-seismic water level changes (water level increase for theWenchuan earthquake, no observed response for the Yao'an earthquake and water leveldecline for the Tohoku earthquake). The co-seismic water level changes in response tothe fourteen earthquakes are shown in the supplement. The red dashed vertical linesshow the time of those earthquakes. (For interpretation of the references to color in thisfigure legend, the reader is referred to the web version of this article.)

-4 -3 -2 -1 0-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

log10(e) (J/m3)

Wat

er le

vel c

hang

e (m

)

Fig. 5. The observed co-seismic water level changes varying with the common logarithmof seismic energy density from Table 1. Blue circles and red circle indicate co-seismicresponses in the farfield and nearfield, respectively. (For interpretation of the referencesto color in this figure legend, the reader is referred to the web version of this article.)

Longitude (degree)

Latit

ude

(de

gree

)

Dilatation

98 100 102 104 106 10826

27

28

29

30

31

32

33

34

35

36

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1x 10-7

Fig. 6. The calculated co-seismic poroelastic strain based on the Okada model (Lin andStein, 2004; Toda et al., 2005) caused by the Wenchuan earthquake at depth of 0.2 km,positive for dilatation. Yellow triangle indicates that the location of the LGH well is in acompressional quadrant. (For interpretation of the references to color in this figurelegend, the reader is referred to the web version of this article.)

215G. Lai et al. / Tectonophysics 679 (2016) 211–217

speculate that thewater level declines are caused by tele-seismic wavesthat enhance the permeability through themobilization of colloidal par-ticles (Brodsky et al., 2003).

Because the water level observation instruments at the LGH wellhappened to be out of operation for several months after the Ninglangearthquake, the permeability change caused by this local earthquakeremains unknown. Fortunately, the water level data before and afterthe Wenchuan earthquake are available, which supply an opportunityfor us to analyze the response in the nearfield. From Fig. 6 and Table 1,we can see clearly that the phase increased by approximately 19°after the Wenchuan earthquake, and it has not fully recovered so far,indicating permanent permeability enhancement was caused by thisearthquake.

Fig. 7.Amplitude and phase responses over time for the LGHwell at the frequency ofM2wave fromDecember 3, 2007 to July 1, 2015. The amplitude response is the amplitude ratio of thetheoretical volumetric strain of the M2 Earth tides over water level. Negative phases mean the water level lags behind the dilatational strain of the Earth tides, whereas positive phasesmean the water level leads. The red dashed vertical lines show the time of the fourteen earthquakes. (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

216 G. Lai et al. / Tectonophysics 679 (2016) 211–217

We carefully checked the water level data following the Wenchuanearthquake, and found a significant rise in the water level 2 min afterthe earthquake, and a total rise of 0.027 m 5 min after the earthquake.During the following 2 min, the water level decreased by 0.01 m, andthen increased approximately 0.003 m. We speculate that the waterlevel increase during the first 4 min was caused by the compression ofthe static stress. Compression or squeezing can change the aperture ofpre-existing cracks (that cannot recover) in the aquifer, resulting inthe observed phase lead after the earthquake, whereas seismic wavesmay also remove some colloidal particles (that can recover). Both staticand dynamic stress may contribute to the permeability enhancementand cause a temporal decrease in the water level.

0 2 4 6 8 10-0.2

-0.18

-0.16

-0.14

-0.12

-0.1

-0.08

-0.06

-0.04

-0.02

0

Phase ( )

Wat

er le

vel c

hang

e (m

)

Fig. 8. Co-seismic water level declines in response to distant earthquakes are found to behighly related with the phase increase, which are proxies of permeability enhancements.

5. Conclusions

The water level at the LGHwell can respond to multiple large earth-quakes with seismic energy densities greater than 10−4J/m3. Increasingwater levels were observed to be caused by the Wenchuan earthquakein the nearfield, and decreasingwater levelswere observed to be causedby eleven earthquakes out of twelve in the farfield.

Permeability enhancements were observed both in the farfield andnearfield. However, the mechanisms should be distinct. The co-seismic water level declines in response to the distant earthquakescould be explained by permeability enhancements caused by the pas-sage of seismic waves through the mobilization of colloidal particles;the co-seismic water level increase in response to the nearfield casecould be caused both by the compression of the static stress and bythe seismic waves.

Acknowledgments

This work made use of GMT and MATLAB software. Groundwaterlevel data are from China Earthquake Networks Center. Waveformdata for this study are provided by Data Management Centre of ChinaNational Seismic Network at Institute of Geophysics, China EarthquakeAdministration (SEISDMC, doi:10.11998/SeisDmc/SN). The authors sin-cerely thank Prof. Chi-yuen Wang and the other anonymous reviewerfor their constructive comments and suggestions. We thank Prof.Emily Brodsky for sharing the code for calculating the tidal response,and thank Dr. Lihua Fang for collating the basic information of the LGHwell. This work was supported by Institute of Geophysics, China Earth-quake Administration (No. DQJB14B03).

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.tecto.2016.04.047.

217G. Lai et al. / Tectonophysics 679 (2016) 211–217

References

Agnew, D.C., 1997. NLOADF: a program for computing ocean-tide loading. J. Geophys. Res.102, 5109–5110.

Agnew, D.C., 2012. SPOTL: some programs for ocean-tides loading. Technical Report.Scripps Institution of Oceanography, La Jolla, CA.

Berger, J., Farrell, W., Harrison, J.C., Levine, J., Agnew, D.C., 1987. ERTID 1: A Program forCalculation of Solid Earth Tides. Publication of the Scripps Institution of Oceanogra-phy, La Jolla, CA.

Brodsky, E.E., Roeloffs, E.A.,Woodcock, D., et al., 2003. Amechanism for sustained ground-water pressure changes induced by distant earthquakes. J. Geophys. Res. 108 (B8),2390. http://dx.doi.org/10.1029/2002JB002321.

Candela, T., Brodsky, E.E., Marone, C., et al., 2014. Laboratory evidence for particle mobili-zation as a mechanism for permeability enhancement via dynamic stressing. EarthPlanet. Sci. Lett. 392, 279–291.

Candela, T., Brodsky, E.E., Marone, C., et al., 2015. Flow rate dictates permeability enhance-ment during fluid pressure oscillations in laboratory experiments. J. Geophys. Res.120 (4), 2037–2055.

Chen, X.-F., Wang, D.-S., Xu, J.-M., et al., 2011. Removing interferences existed in observa-tion of groundwater table in the seismic station of Lugu Lake. Earthquake Eng. Si-chuan 138, 34–37 (in Chinese).

Chia, Y., Chiu, J.J., Chiang, Y., et al., 2008. Spatial and temporal changes of groundwaterlevel induced by thrust faulting. Pure Appl. Geophys. 165 (1), 5–16.

Data Management Centre of China National Seismic Network, 2007. Waveform Data ofChina National Seismic Network. Institute of Geophysics, China Earthquake Adminis-tration http://dx.doi.org/10.11998/SeisDmc/SN (http://www.seisdmc.ac.cn.).

Deng, Q.-D., Zhang, P.-Z., Ran, Y.-K., 2006. Distribution of Active Faults in China (1:4000000). Science Press, Beijing.

Doan, M.L., Brodsky, E.E., Prioul, R., et al., 2006. Tidal Analysis of Borhole Pressure— A Tu-torial, Schlumberger Research Report.

Dziewonski, A.M., Chou, T.-A., Woodhouse, J.H., 1981. Determination of earthquake sourceparameters from waveform data for studies of global and regional seismicity.J. Geophys. Res. 86, 2825–2852. http://dx.doi.org/10.1029/JB086iB04p02825.

Ekström, G., Nettles, M., Dziewonski, A.M., 2012. The global CMT project 2004–2010: cen-troid-moment tensors for 13,017 earthquakes. Phys. Earth Planet. Inter. 200-201, 1–9.http://dx.doi.org/10.1016/j.pepi.2012.04.002.

Elkhoury, J.E., Brodsky, E.E., et al., 2006. Seismic waves increase permeability. Nature 441(7097), 1135–1138.

Elkhoury, J.E., Niemeijer, A., Brodsky, E.E., et al., 2011. Laboratory observations of perme-ability enhancement by fluid pressure oscillation of in situ fractured rock. J. Geophys.Res. 116, B02311. http://dx.doi.org/10.1029/2010JB007759.

Hsieh, P.A., Bredehoeft, J.D., Farr, J.M., 1987. Determination of aquifer transmissivity fromearth tide analysis. Water Resour. Res. 23 (10), 1824–1832.

Jonsson, S., Segall, P., Pedersen, R., et al., 2003. Postearthquake ground movements corre-lated to pore-pressure transients. Nature 424, 179–183.

Kinoshita, C., Kano, Y., Ito, H., 2015. Shallow crustal permeability enhancement in CentralJapan due to the 2011 Tohoku earthquake. Geophys. Res. Lett. http://dx.doi.org/10.1002/2014GL062792.

Koizumi, N., Lai, W.-C., Kitagawa, Y., et al., 2004. Comment on “Coseismic hydrologicalchanges associated with dislocation of the September 21, 1999 Chichi earthquake,Taiwan” by Lee M et al. Geophys. Res. Lett. 31. http://dx.doi.org/10.1029/2004GL019897.

Lai, G.-J., Ge, H.-K., Xue, L., et al., 2014. Tidal response variation and recovery following theWenchuan earthquake from water level data of multiple wells in the nearfield.Tectonophysics 619-620, 115–122. http://dx.doi.org/10.1016/j.tecto.2013.08.039.

Lay, T., Wallace, T.C., 1995. Modern Global Seismology. Academic Press, New York.Lin, J., Stein, R.S., 2004. Stress triggering in thrust and subduction earthquakes and stress

interaction between the southern San Andreas and nearby thrust and strike slipfaults. J. Geophys. Res. 109, B02303.

Liu, L.-B., Roeloffs, E., Zheng, X.-Y., 1989. Seismically induced water level fluctuations inthe Wali well, Beijing, China. J. Geophys. Res. 94 (B7), 9453–9462. http://dx.doi.org/10.1029/JB094iB07p09453.

Liu, W., Manga, M., 2009. Changes in permeability caused by dynamic stresses in fracturedsandstone. Geophys. Res. Lett. 36, L20307. http://dx.doi.org/10.1029/2009GL039852.

Ma, Y., 2015. Earthquake-related temperature changes in two neighboring hot springs atXiangcheng, China. Geofluids http://dx.doi.org/10.1111/gfl.12161.

Manga, M., Brodsky, E.E., Boone, M., 2003. Response of streamflow to multiple earth-quakes. Geophys. Res. Lett. 30 (5), 1214.

Manga, M., Wang, C.Y., 2007. Earthquake hydrology. In: Kanamori, H. (Ed.), Treatise onGeophysicsEarthquake Seismology 4(10). Elsevier, Boston, Mass, pp. 293–320.

Manga, M., Beresnev, I., Brodsky, E.E., et al., 2012. Changes in permeability caused by tran-sient stresses: field observations, experiments, and mechanisms. Rev. Geophys. 50(2), RG2004. http://dx.doi.org/10.1029/2011RG000382.

Mogi, K., Mochizuki, H., Kurokawa, Y., 1989. Temperature changes in an artesian spring atUsami in the Izu Peninsula (Japan) and their relation to earthquakes. Tectonophysics159 (1), 95–108. http://dx.doi.org/10.1016/0040-1951(89)90172-8.

Mohr, C.H., Montgomery, D.R., Huber, A., et al., 2012. Streamflow response in small uplandcatchments in the Chilean coastal range to the MW 8.8 Maule earthquake on 27 Feb-ruary 2010. J. Geophys. Res. Earth Surf. 117, F02032.

Montgomery, D.R., Manga, M., 2003. Streamflow and water well responses to earth-quakes. Science 300 (5628), 2047–2049.

Quilty, E.G., Roeloffs, E.A., 1997. Water-level changes in response to the 20 December1994 earthquake near Parkfield, California. Bull. Seismol. Soc. Am. 87 (2),310–317.

Roeloffs, E.A., 1996. Poroelastic techniques in the study of earthquake-related hydrologicphenomena. Adv. Geophys. 37, 135–195.

Shi, Y.L., Cao, J.L., Ma, L., 2007. Tele-seismic coseismic well temperature changes and theirinterpretation. Acta Seismologica Sinica 29 (3), 265–273.

Shi, Z.-M., Wang, G.-C., Liu, C.-L., 2013. Co-seismic groundwater level changes induced bythe May 12, 2008 Wenchuan earthquake in the near field. Pure Appl. Geophys. 170(11), 1773–1783.

Shi, Z.-M., Wang, G.-C., 2014. Hydrological response to multiple large distant earthquakesin the Mile well, China. J. Geophys. Res. Earth Surf. 11, 2448–2459.

Shi, Z.-M., Wang, G.-C., Wang, C.-Y., et al., 2014. Comparison of hydrological responses tothe Wenchuan and Lushan earthquakes. Earth Planet. Sci. Lett. 391, 193–200.

Shi, Z.-M., Wang, G.-C., Manga, M., et al., 2015. Mechanism of co-seismic water levelchange following four great earthquakes — insights from co-seismic responsesthroughout the Chinese mainland. Earth Planet. Sci. Lett. 430, 66–74.

Shibata, T., Matsumoto, N., et al., 2010. Linear poroelasticity of groundwater levelsfrom observational records at wells in Hokkaido, Japan. Tectonophysics 483(3), 305–309.

Toda, S., Stein, R.S., Richards-Dinger, K., et al., 2005. Forecasting the evolution of seismicityin southern California: animations built on earthquake stress transfer. J. Geophys. Res.110, B05S16.

Wakita, H., 1975. Water wells as possible indicators of tectonic strain. Science 189,553–555.

Wang, 2007. Liquefaction beyond the near field. Seismol. Res. Lett. 78 (5), 512–517.Wang, C.-Y., Chia, Y., 2008. Mechanism of water level changes during earthquakes: near

field versus intermediate field. Geophys. Res. Lett. 35 (12), L12402.Wang, C.-Y., Manga, M., Dreger, D., et al., 2004. Streamflow increase due to rupturing of

hydrothermal reservoirs: evidence from the 2003 San Simeon, California, earthquake.Geophys. Res. Lett. 31 (10), L10502.

Wang, C.-Y., Chia, Y., Wang, P., et al., 2009. Role of S waves and love waves in coseismicpermeability enhancement. Geophys. Res. Lett. 36, L09404. http://dx.doi.org/10.1029/2009GL037330.

Wang, C.-Y., Wang, L.-P., Manga, M., et al., 2013. Basin-scale transport of heat and fluid in-duced by earthquakes. Geophys. Res. Lett. 40, 3893–3897. http://dx.doi.org/10.1002/grl.50738.

Wang, W., Wu, J., Fang, L., et al., 2014. S wave velocity structure in southwest China fromsurface wave tomography and receiver functions. J. Geophys. Res. Solid Earth 119,1061–1078. http://dx.doi.org/10.1002/2013JB010317.

Wen, X.-Z., Yi, G.-X., 2003. Re-zoning of statistic units of seismicity in Sichuan–Yunnan re-gion. J. Seismol. Res. 26, 1–9 (sup.). (in Chinese).

Xue, L., Li, H.-B., Brodsky, E.E., et al., 2013. Continuous permeability measurements recordhealing inside theWenchuan earthquake fault zone. Science 340 (6140), 1555–1559.http://dx.doi.org/10.1126/science.1237237.

Yan, R., Heiko, W., Wang, R.-J., 2014. Groundwater level changes induced by the 2011Tohoku earthquake in China mainland. Geophys. J. Int. 199 (1), 533–548.

Zhang, Y., Huang, F.-Q., 2011. Mechanism of different coseismic water-level changes inwells with similar epicentral distances of intermediate field. Bull. Seismol. Soc. Am.101 (4), 1531–1541.

Zheng, X.F., Yao, Z.X., Liang, J.H., et al., 2010. The role played and opportunities providedby IGP DMC of China National Seismic Network inWenchuan earthquake disaster re-lief and researches. Bull. Seismol. Soc. Am. 100 (5B), 2866–2872. http://dx.doi.org/10.1785/0120090257.