Earthquake Reaearch in China  2017, Vol. 31 Issue (2): 269-281
Analysis of Sacks Body Strain Interference at Beida No.200 Station in Changping and Earthquake Case Study
Li Hangu1, Hu Lan2, Wu Lijun1
1. Institute of Earthquake Science, China Earthquake Administration, Beijing 100036, China;
2. Shaoyang Seismic Station, Earthquake Administration of Hunan Province, Shaoyang 422000, Hunan, China
Abstract: The relationship between Sacks body strain deformation at Beida No. 200 station in Changping and tidal solids, atmospheric pressure and water level is analyzed in this paper. Sacks body strain deformation data before the MS 8.0 Wenchuan earthquake is studied based on the analysis of the interference. The short-impending anomaly of the body strain deformation is considered to be reliable. The anomaly characteristics conclude: (1) The trend anomaly as extensional change of the body strain deformations on a quasi 1 year time scale before the Wenchuan earthquake was recorded, and the accumulative amount was about 4000×10-9. Correspondingly, the short-term precursor of earthquake was manifested as an extensional abrupt change. (2) The extensional intermittent anomalous abrupt change was recorded by body strainmeters between March 1 and May 7 in 2008. (3) Four compressional abrupt changes were recorded in the intermittent distortions recorded between April 13 and May 11. (4) High frequency components were increased in the distortion process in May 1 to 3, 5, 7, and 9 to 12, caused by slow earthquakes before the Wenchuan earthquake according to wavelet analysis. The abnormal phenomena are summarized and the mechanics discussed in this paper. Strain solid tide distortions in body strain observations, the continuous repeated extensional and compressional abrupt changes accompanying these distortions, and the increase of high frequency components can be regarded as the index of short term and impending earthquake prediction, based on analysis of interference factors such as air pressure and water level.
Key words: Body strain     The solid tide distortion     Slow earthquake     Short imminent anomalies

INTRODUCTION

Body strain observational data contains crustal deformation, the solid tide influence from the sun and the moon and earthquake precursor information. Extracting earthquake precursor information has long been a concern of seismologists. Many scholars have analyzed and discussed strain body interference factors and instrument performance. Gao Fuwang et al. (2004) theoretically studied interference factors influencing the Sacks body strain meter, pointing out that the effect of atmospheric pressure on overlying strata and pore pressure changes caused by underground water level and irrigation and drainage are the main interference factors. Wu Lijun et al. (2013) have performed interference analysis using the observational data of body strain, atmospheric pressure and water level at Dongsanqi station, and gained a more thorough grasp of the interference factors causing body strain changes at the station. Tian Tao et al.(2014a, b), using the combined three-layer thick wall cylinder model, provided the mutual transition relationship between different types of body strain, analyzed the influences of the change of model parameters on three types of body strain observations, and concluded that in the environment of the same cement sheath and formation elastic parameters, the Sacks-Evertson body strain meter is more sensitive than the TJ-2 series. Additionally, some scholars have carried out relevant research on earthquake cases on the basis of interference analysis using body strain observation data and extracted earthquake precursor information. Li Haixiao et al. (2005) discussed anomaly characteristics of body strain data from Huailai station appearing before earthquakes and its earthquake-reflecting effect according to the anomaly pattern of original body strain observation data. Chen Qilin et al. (2002), through the analysis of body strain data from Liyang seismic network station, found that some precursory anomalies were observed from body strain data at Liyang station before and after the moderate-strong earthquakes in the South Yellow Sea and strong earthquakes in Taiwan, and figured out the anomaly characteristics of body strain observation data from Liyang station before strong earthquakes in the South Yellow Sea and Taiwan. Song Zhiying et al. (2009) have done a preliminary analysis of anomalous changes appearing at the volumetric strain meter at the Xiyang seismic station in Shanxi before the Wenchuan MS 8.0 earthquake in Sichuan on May 12, 2008, and the results indicated that anomalous changes appearing from March 29 completely accorded with the mutation characteristics shown before and during the rock rupture in experimental research. Wang Zhaihua and Teng Haitao (2012) mainly studied anomaly trends and partial medium-and short-term anomalies observed by observation stations near the epicenter before the Xinyuan-Hejing MS 6.6 on June 30, 2012, and concluded that medium-short-term anomalies mainly appeared within a range of 200km-300km and some short term anomalies were repetitive. In this article, in combination with observation data of atmospheric pressure and static water level from the Institute of Crustal Dynamics, we analyzed the body strain observed at the Beida No.200 station in Changping before the Wenchuan MS 8.0 earthquake on May 12, 2012, believing that a summary up of pre-earthquake anomalies can provide a certain reference for the practice of earthquake prediction.

1 AN OVERVIEW OF OBSERVATION WELLS AND DATA USAGE

The well for body strain observation at the Beida No.200 station in Changping, located at the end of the Nankou-Sunhe fault, with a depth of 320m and Quaternary thickness of 83m, is composed of an piedmont aggraded sandy soil layer, with underlying Simian dolomite and limestone where Karst fissures are relatively developed, showing good condition of groundwater recharge and runoff.

Because the atmospheric pressure and well water level observation data are damaged as a result of aging, in order to ensure the comprehensiveness of data analysis, we used atmospheric pressure and well water level observation data from the nearby Changping station for reference analysis in this article. Two wells, which are only 1.97km apart, both lie in the intersection of the NE-striking Nankou piedmont active fault and the NW-striking Nankou-Sunhe active fault (Fig. 1), and it was found in the course of study that the well water level at Changping station had a good correlation with body strain observations at Beida No.200 station in Changping. In addition, according to the results of previous studies 2, 3, from Qikongqiao and Huzhuang in the north, south to Daguanmen and Shipaifang, paleo-river channels develop in NNE direction, and the two wells both lie between the river channels, and a large accumulation of gravel layers near the observation wells also supports this point of view well. Therefore, the analysis of water level observation at Changping station has a certain reference value for the interference analysis of body strain data at the Beida No.200 station in Changping.

2 Xiao Shurong et al. A preliminary study of seepage in the Shisanling (Ming Tombs) Reservoir. Teaching and Research Office of China University of Geosciences, 1959.

3The First Team of Hydrogeology, Geological Bureau of Beijing. A report of hydrogeology and seepage of Dagongmen paleo-river channel of the Shisanling (Ming Tombs) Reservoir in Beijing, 1971.

 Fig. 1 Seismic station location and geological fault distribution
2 ANALYSIS OF RELATED INTERFERENCE 2.1 Fundamental Analysis of Body Strain Observation Data

The daily mean value is used to analyze the change trend of body strain observation at the station, and its basic annual variation is as follows: reaching its peak value in March, dropping to its lowest in July, with gentle fluctuation from August-November and rising in December.

Linear fitting was done with data during 2002-2009 using fitting formula Y=1.12X+414.94. We got the fitted correlation coefficient R=0.97, and the change trend of observation data we obtained showed a high degree of correlation. Fig. 2 is the residual curve after removing the trend term, from which we can see that the body strain dropped more sharply (suffering tension force) in May, 2007, showing distortion of annual variation, again dropping sharply (accelerated tension) after the Wenchuan MS 8.0 earthquake, then turning flat, and then increased rapidly from August 8, 2008, and finally recovered and recombined with the fitting curve in December 2009. Comparative analysis of annual variation curve during 2002-2009 shows that annual variation after March, 2009 returned to normal, and the observation curves of two years are basically in coincidence.

 Fig. 2 The daily mean value body strain curve fitting
2.2 Influences of Solid Tide on Body Strain and Well Water Level

The solid tide is caused by changes of gravitational attracting force produced by the moon and the sun. The minimum resolutions of Sacks body strain meter and water level sensor are respectively set to be 10-9 and 0.001mm, which are both high-precision observation instruments. Therefore, the body strain meter can record clear solid tide changes, while well water level tidal observations are greatly influenced by the taquifer structure. Tidal changes at the Beida No.200 station in Changping are obvious. Original observation data of body strain and well water level is filtered the using Berzev filtering method, by which we can get diurnal and semi-diurnal wave curves, and then a regression analysis is done based on theoretical solid tide of body strain to get the respective correlation coefficient R and regression coefficient b.

Due to environmental interference (weather, water and rainfall, etc.) and some unknown reasons, R and b value change with seasons throughout the year. In order to describe the influences of solid tide on body strain and well water level in normal circumstances, we select data in February with minimum external interference and complete observations each year for analysis. The results are listed in Table 1, where the computational accuracy of M2 wave tidal factor=mean square error/tidal factor. It can be seen that body strain and well water level tidal observation both show high accuracy, indicating that instrument operation and observation system have high reliability and can withstand continuous observations for a long time. If the reliability of borehole strain observation data can be guaranteed, it can be directly used without complex inversions in the quantitative study of changes of the crustal stress field (Li Jianchun, 1989). However, it should be pointed out that environmental interference is a problem that an instrument itself is unable to overcome.

Table 1 The relationship between body strain, well water level and solid tide
2.3 Correlation Analysis of Atmospheric Pressure and Body Strain

According to elastic mechanics, a borehole is acted upon by the direct effect of atmospheric pressure in the vertical direction and by the effect of lateral confining pressure in the horizontal direction. According to the calculation results of Su Kaizhi et al. (2003), the interference coefficient of pressure on body strain observation under simple conditions is expressed as follows:

 $b = \frac{{\frac{{\mu \left({4 - 18\mu } \right)}}{{1 - \mu }} + 0.9 - 2\mu }}{E}$ (1)

Where, b denotes atmospheric pressure influence coefficient, E rock elastic modulus, μ Poisson's ratio of rock. In general, body strain increases with the increase of atmospheric pressure (probe contraction), decreases with the decrease of atmospheric pressure (probe inflation), which is positively related.

Body strain observations are always affected by atmospheric pressure, therefore, the interference analysis of atmospheric pressure is crucial for distinguishing body strain anomalous changes. We analyzed the hourly readings of atmospheric pressure from 2002-2010. The pressure changes between 990-1050hPa, reaching the lowest value in early July and the peak value in December, presenting a clear annual variation and no trend could be seen, which is not strongly correlated with body strain observations in terms of annual variation and change trend. Zhang Lingkong et al. (2012a)made a more detailed study on the causes of annual changes of body strain at this seismic station, and the results showed that annual average variations of body strain and water level are respectively 9.3 and 14.2 times the effect of atmospheric pressure, thus interference from atmospheric pressure can be ignored in annual dynamic analysis.

Body strain has a good relativity with short-term and daily variation of atmospheric pressure (Fig. 3). We made a daily correlation analysis, and the statistical results during 2002-2009 showed mainly positive correlation, which accounts for 82.1%, while negative correlation occupied 17.9%. 42.7% of the results had correlation coefficient greater than 0.5, indicating that short-term change and daily variation of body strain observation is greatly influenced by atmospheric pressure.

 Fig. 3 Body strain and air pressure curve during March 2005 and the correlation coefficient analysis
2.4 Correlation Analysis of Water Level and Body Strain

Body strain observations are also affected by water level (Wu Lijun et al., 2013; Wang Mei, 2012; ZhangLingkong, 2005; Qin Shuanglong et al., 2014; Chen Dongbai et al., 2012). According to intensive study and field investigation of hydrogeological conditions near the observation well at the Beida No.200 seismic station in Changping done by Yang Huinian (1991), it is believed that the sharp rise of water level from October, 1988-April, 1990 is a normal dynamic caused by seepage and water recharge when taking water from Baihebao Reservoir in Yanqing county to the Shisanling (Ming Tombs) Reservoir and passing through the paleo-river channels, but not precursory anomalies before the Datong-Yanggao MS 6.1 earthquake swarm in October, 1989. According to water level observation curves during January, 2002 and the Wenchuan earthquake, there were no significant changes before the Wenchuan earthquake, while the water level rose sharply after the earthquake. Many investigations revealed that since November, 2014, filling of the Shisanling (Ming Tombs) Reservoir has adopted a piping method by taking water from the Shahe Reservoir, at a frequency of about (250-260) days/a. The amount of water diversion is basically fixed without much change, and both pipes and the reservoirs are closed guarded against leakage, therefore the leakage effect caused by water diversion for the Shisanling (Ming Tombs) Reservoir and the reservoir itself can be eliminated.

Water level data during 2002-2010 was processed by a continuous method and the daily mean value was calculated. The overall changes show that the water level was in a state of decline from 2002 to June, 2008, rising from August, 2008 to early March, 2009, and then resumed its decline and annual change. The basic state of annual change roughly shows the peak value in March and lowest value in early December.

The water level daily mean value from 2002 to 2006 was taken for linear fitting. We got the fitting formula Y=-0.01X+3.61 and correlation coefficient R=0.98, revealing a high degree of linear correlation, thereby eliminating the change trend of observation data. The downtrend of the water level turned flat from June 15, 2008 to July 30, rose rapidly and then returned to the normal state of annual change and trend after March 17, 2009.

Change trend, annual state and time of turning points of body strain and water level are basically consistent. Body strain values varied only from May, 2007 to May, 2008, but later the changes were identical.

Correlation analysis of daily mean value of body strain and water level from 2002 to 2009 was done, including annual correlation analysis and monthly correlation analysis. We calculated the annual correlation coefficient for annual correlation analysis, and the results showed a positive correlation. It can be seen from Fig. 4(a) that correlation coefficients in the year of 2004 and 2008 were around 0.75, the rest were greater than 0.9, reaching 0.99 in 2007, indicating that body strain is strongly correlated with the annual state of water level. The time window for monthly correlation analysis is 30 days. We carried out monthly correlation analysis and sliding monthly correlation analysis. The results showing positive correlation are 88.54%, and a negative correlation appeared mainly in the second half of the year, twice in October, November and December and once in February, March, July and September. From November, 2006 to October, 2008, the results show positive correlation in the two consecutive years (24 months).

 Fig. 4 Body strain and water level correlation coefficient analysis results (a)Annual correlation coefficient; (b) Monthly correlation coefficient

According to results of sliding monthly correlation (Fig. 4(b)), negative correlation appears more often from early March to early April and from October to December, which happen to be the time for a turn of annual change when the velocities of the body strain and water level change slowly. The most obvious differences are that the period of negative correlation were advanced to February and September in 2004 and a short-term negative correlation only appeared from October to November in 2007.

3 EARTHQUAKE CASE STUDY

The Beida No.200 station in Changping has a strong ability for co-seismic recording and response to precursor anomalies of far-field earthquakes. Song Mo et al. (2010) analyzed the earthquake recording ability of the seismic station for 167 MS≥6.0 earthquakes occurring all round the world during April, 1994-November, 2002, and concluded that the response distance of the station is 1077km to MS 6.0 earthquakes and 5123km to MS 7 earthquakes. Yi Zhigang et al.(2005) studied the MS 8.7 and MS 8.5 earthquake that occurred in the northwest waters of Sumatra in Indonesia respectively on December 26, 2004 and March 29, 2005 using hourly readings of observation data at the station (the distances from the station to epicenters of the two earthquakes are respectively 5324km and 5104km), and found that short-term changes before the earthquakes showed distortion of strain solid tide waves and the change trend broke the existing trend of strain measurements. In addition, Wu Lijun et al. (2015), using the same type of Sacks body strain observation data from the Erzhangying and the Tiantang seismic station from January, 2011 to April, 2014, studied 80 earthquakes with MS≥7.0 which occurred all over the world during the same period of time, believing that the two stations could respond well to earthquake and precursor anomalies of MS≥7.0 earthquakes with epicentral distances within 8000km, which is consistent with the analysis results of co-seismic recording ability obtained by Song Mo et al.(2010) (The response distances of the two stations to MS 7.0 earthquakes are respectively 9230km and 15926km). This provides the basic idea for the study of far-field earthquakes using observation data from the station.

We studied the Wenchuan MS 8.0 earthquake on May 12, 2008. This earthquake released enormous energy which triggered intensive tectonic activities. The Beida No.200 seismic station in Changping is about 1614km away from the epicenter of the Wenchuan earthquake. Through the analysis and discussion of anomalous changes of Sacks body strain at the station before the earthquake, it is believed that some significant earthquake precursors were obtained.

Before 2007, there was a good correlation between body strain and water level in both change trend and annual variation. However, the correlation significantly decreased around 2008. It can be seen intuitively from Fig. 5 that the body strain value deviated from normal trends from 2007 to 2008, while the water level appeared normal.

 Fig. 5 The curves of body strain and water level

We analyzed body strain and atmospheric pressure observation data before the Wenchuan MS 8.0 earthquake. It is found that daily deformation of body strain was basically normal from January to February, which fluctuated with the change of atmospheric pressure in a short period of time, and from March and the MS 8.0 earthquake, multiple sudden drops of body strain value appeared (sudden change of tension), while no mutation of atmospheric pressure appeared, among which, changes on April 6-9 and 15-19 were reversed from the short-term atmospheric pressure change. Table 2 lists the correlation coefficients between body strain and atmospheric pressure. It can be seen that in the selected period of time, when body strain and atmospheric pressure synchronously changed, daily correlation coefficients showed negative correlation for only one day, between 0.6-0.9, indicating good correlation, and during the mutation periods with unsynchronized change, correlation coefficients lay between 0.5-0.7 for only 4 days, showing poor correlation.

Table 2 Statistics of correlation coefficients between body strain and atmospheric pressure before the Wenchuan earthquake

Fig. 6 provides the original curves (Fig. 6(b)) and data curves (Fig. 6(a)) of body strain after eliminating the influence of atmospheric pressure and water level at 02:00 p.m. during January 1, 2008-May 12, 2008. It can be seen that there have been sharp data mutations since March 2 (there is a lack of data during March 19-27), which are mainly extensional changes.

 Fig. 6 Influence of air pressure and water level after elimination of body strain original curve

The above analysis shows that the Beida No.200 seismic station in Changping recorded intermittent mutation anomalies of extensional type before the Wenchuan MS 8.0 earthquake, which appeared during March 1 to May 7, 2008. Intermittent distortions were recorded from April 13 to May 11, 2008, after 4 distortions for which reverse (compressive) mutations appeared, with anomalous variation of -1020×10-9 and reverse change of 125×10-9.

Fig. 7 provides the original curve of body strain observations at 02:00 p.m. from May 1 to 12 and its wavelet analysis results. The analysis results show that in the distortion process a large amount of high-frequency components appeared on May 1-3, 5, 7 and 9-12, particularly obvious at the 6th order, which appeared relatively late over a period of 64-128 minutes, and the high-frequency components were evident only at the 6th order on May 5 and May 7. According to amplitudes at various orders, we can conclude that the preferred period for these high-frequency components is between 16-128 minutes, indicating a wider range. This is consistent with the slow earthquake phenomenon in both time and form discovered by many scholars (Zhang Shuliang et al., 2009; Liu Hongbin et al., 2012; Zhang Xi et al., 2012) using long-period deformation observation before the Wenchuan earthquake.

 Fig. 7 Body strain original curve and results of wavelet analysis (a) Original data, (b) Period: 2-4 minutes, (c) Period: 4-8 minutes, (d) Period: 8-16 minutes, (e) Period: 16-32 minutes, (f) Period: 32-64 minutes, (g) Period: 64-128 minutes
4 CONCLUSION AND DISCUSSION

(1) The changes of body strain and water level are well correlated, but anomalous changes of body strain are weakly correlated with water level changes from July, 2007 to 2008. Body strain has a good relativity with short-term and daily changes of atmospheric pressure, indicating that short-term and daily changes of body strain observations are greatly influenced by atmospheric pressure. However, during the period from March, 2008 to the Wenchuan MS 8.0 earthquake, multiple sudden changes and distortions of body strain appeared successively, which has no correlation with short-term changes of atmospheric pressure. According to the above analysis, we believe that short-impending anomalies of body strain observed before the Wenchuan earthquake at the Beida No.200 seismic station in Changping are reliable.

(2) A one-year scale of body strain trend anomaly was recorded before the Wenchuan earthquake, characterized by extensional changes, with a cumulative amount of approximately 4000×10-9. Ten months after the earthquake it showed reverse (compressive) recovery, and then returned to normal. Correspondingly, short-impending precursor anomalies also show extensional mutations.

(3) Wavelet analysis shows that a clear slow earthquake phenomenon was recorded 10 days before the Wenchuan earthquake, appearing intermittently, and the preferred period for these slow earthquakes was between 16-128 minutes.

Borehole strain and stresss hort-impending anomalies are dynamic changes manifested when the seismogenic system approaches a rupture state. This kind of anomaly shows nature of pulse and intermittency. Strain and stress value display anomaly patterns of massive disturbance, step change, pulse, fluctuation, direction deflection of the maximum principal stress and distortions of strain solid tide (Li Haixiao et al., 2005; Pan Zhensheng et al., 2015). With respect to compressional and extensional abrupt changes appearing alternately before the earthquake, it is due to the forces acting on the seismic source body being multi-directional, and different forces that cause far-field stress field changes are not constant, as one falls, another rises. Interaction of forces such as the increase of forces in different directions which offset each other results in alternate extensional and compressional changes of strain value. Distortions of solid tide appearing before an earthquake is a direct expression of imbalance of the regional stress field in source region, and the duration of distortion reflects the time for accelerated stress accumulation. Extensional and compressional abrupt changes appearing in the process of solid tide distortions are significant anomaly characteristics of body strain observation values reflected from rock creep deformation to plastic deformation in the process of imbalance of the regional stress field in the seismic source region. Therefore, on the basis of analysis of interference factors such as atmospheric pressure and water level etc., strain solid tide distortions appearing in the body strain observations, and the appearance of multiple and successive extensional and compressional abrupt changes and the increase of high-frequency components in the process of distortions can be used as indicators for impending earthquake forecast.

This paper has been published in Chinese in the journal of Inland Earthquake, Volume 29, Number 3, 2015.

REFERENCES
 Chen Dongbai, Li Yuanyuan, Luo Cijian, Ren Jun, Zhang Lu. Influence of impoundment and discharge of Shibianyu reservoir on the digital borehole strain observation in Xi'an Station[J]. Recent Developments in World Seismology, 2012(6): 206–206. Chen Dongbai, Luo Cijian, Li Yuanyuan, Ren Jun, Zhang Lu. Analysis on the interference of borehole strain observation in Xi'an seismic station[J]. Plateau Earthquake Research, 2012, 24(2): 50–54. Chen Qilin, Wang Hao, Zhang Shuigeng, Di Liang, Huang Caizhong. Preliminary study on the relation between abnormal characters of body-strain observation and moderate strong earthquakes[J]. South China Journal of Seismology, 2002, 22(1): 76–79. Gao Fuwang, Li Li, Niu Anfu, Zhang Jing. Study on the identification and elimination of the disturbance in crustal strain record[J]. Earthquake, 2004, 24(S): 90–97. Gao Mingzhi, Ren Yaping, Wang Huasheng, Gao Li, Zhang Cen. The effect of thunderstorm weather to body strain in Xuzhou seismostation[J]. Inland Earthquake, 2014, 28(3): 281–285. Li Haixiao, Song Xiaobing, Zhang Changhui, Li Yuping. Body strain anomaly at Huailai seismological station, Hebei, and Zhangbei Earthquake of M 6.2, Hebei[J]. Earthquake Research in Shanxi, 2005(2): 8–10. Li Jianchun. Achievements Compilation of Sorting out of the Methods of Earthquake Monitoring and Prediction: Gravity, Ground Tilt and Crustal Stress[M]. Beijing: Seismological Press, 1989 Liu Hongbin, Yin Liang, Li Dongsheng, Li Xinjian, Liu Xiaoming, Zou Xiaobo. Analysis on characteristics of precursor waves recorded by component strain observation at Gaotai Seismic Station, Gansu Province[J]. Northwestern Seismological Journal, 2012, 34(3): 299–307. Pan Zhensheng, Xu Yangang, Lai Aijing. Characteristics of borehole strain anomalies in Wushi before earthquakes[J]. Inland Earthquake, 2015, 29(1): 43–49. Qin Shuanglong, Liao Lixia, Chen Ying, Wang Qingping. Discussion on inversion for body strain of aquifer using co-seismic change in Fujian[J]. Inland Earthquake, 2014, 28(4): 353–359. Song Mo, Qiu Zehua, Tang Lei. Earthquake recording ability of borehole strain meter[J]. Journal of Geodesy and Geodynamics, 2010, 30(S2): 31–35. Song Zhiying, Wang Zhenjiang, Lv Yongqing, Li Shuhua, Li Jiaji. Study on body strain characteristics of Xiyang Seismological Station, Shanxi, before Wenchuan, Sichuan, 8.0 earthquake and coseismic effect[J]. Earthquake Research in Shanxi, 2009(2): 19–20. Su Kaizhi, Li Hailiang, Zhang Jun, Li Xiuhuan, Ma Xiangbo. Advances in Borehole Crustal Strain Observation[M]. Beijing: Seismological Press, 2003 Tian Tao, Wang Junfei, Wang Wei, Hao Yongliang. Effect of coupling on the borehole strain observation[J]. Seismological and Geomagnetic Observation and Research, 2014, 35(3/4): 165–173. Tian Tao, Yang Jun, Hao Yongliang, Wang Wei. Impact analysis of the borehole coupling effect on the tide generating force calibration[J]. Inland Earthquake, 2014, 28(4): 312–318. Wang Mei. Study on correlation of digital body strain data with atmosphere and well water level[J]. Journal of Geodesy and Geodynamics, 2012, 22(4): 85–88. Wang Zaihua, Teng Haitao. Deformation anomalies before Xinyuan, Hejing earthquake with MS 6.6[J]. Inland Earthquake, 2012, 26(4): 347–353. Wu Lijun, Chen Zhengwei, Li Hangu, Xu Lei, An Bayaer, Xu Shunqiang. Abnormal phenomena of body strain before large earthquakes[J]. Technology for Earthquake Disaster Prevention, 2015, 10(1): 151–162. Wu Lijun, Zhou Liming, Zhang Zhongwu, Jiang Jianing, Li Hangu, Dai Yaqing. Analysis of Sacks body strain interference at Dongsanqi seismic station and earthquake case study[J]. Journal of Institute of Disaster Prevention, 2013, 15(3): 41–44. Yang Huinian. Analysis on the causes of water level dynamic anomaly in Changping well, Beijing[J]. Seismology and Geology, 1991, 13(4): 306–316. Yi Zhigang, Song Mo, Qiu Zehua. The hourly values of borehole strain observation network of China in the great Sumatra earthquake[J]. Collected Works on Crustal Structure and Crustal Stress, 2005(3): 25–40. Zhang Lingkong, Wang Guangcai, Niu Anfu. Analysis on several factors of periodic air pressure wave affecting crustal strain field[J]. Acta Seismologica Sinica, 2011, 33(3): 351–361. Zhang Lingkong, Wu Lijun, Yang Ying. Comparative study of the interference of mutation pressure generated by thunder storms with volume strain and same well water-level[J]. Earthquake Research in China, 2012, 28(1): 69–77. Zhang Lingkong, Wu Lijun, Yang Ying. Comparative study of the interference of mutation pressure generated by thunder storms with volume strain and same well water-level[J]. Earthquake Research in China, 2012, 28(1): 69–77. Zhang Lingkong. Research on annual variation reason of the Sacks volumetric strain[J]. Earthquake Research in China, 2005, 21(2): 254–259. Zhang Shuliang, Liu Ruichun, Ning Yaling, Tang Leili, Li Bin. Analysis of characteristics of precursor wave of low frequent in Shanxi before Wenchuan MS 8.0 earthquake[J]. Journal of Geodesy and Geodynamics, 2009, 29(6): 35–39. Zhang Xi, Li Ruisha, Jia Peng, Wang Qingliang. Epicentre-source area from disturbance abnormity before Wenchuan Earthquake[J]. Geomatics and Information Science of Wuhan University, 2012, 37(3): 310–313.

1. 中国地震局地震预测研究所，北京 100036;
2. 湖南省地震局邵阳地震台，湖南 邵阳 422000