Journal of Geodynamics 63 (2013) 5461

Contents lists available at SciVerse ScienceDirect

Journal of Geodynamics
journal homepage: http://www.elsevier.com/locate/jog

Tide-factor anomalies from observations of well level in the Sichuan Province

prior to the great Wenchuan earthquake of 2008
Chenglong Liu a,b,c , Guangcai Wang b,c, , Zheming Shi b , Yinju Xu b , Huina Fang b , Jinwei Wang b
a
Institute of Geology, China Earthquake Administration, Beijing 100029, China
b
MOE Key Laboratory of Groundwater Circulation and Evolution, School of Water Resources & Environment, China University of Geosciences, Beijing 100083, China
c
State Key Laboratory of Biogeology and Environmental Geology, China University of Geosciences, Beijing 100083, China

a r t i c l e i n f o a b s t r a c t

Article history: We systematically calculated the tide-factors of M2 and O1 waves of the earth tide using water level data
Received 3 August 2011 from 16 wells in the Sichuan Province acquired 2 years prior to the great Wenchuan earthquake of 2008.
Received in revised form 16 July 2012 Our purpose was to extract anomalies in the tide-factors and analyze their variations in time and space.
Accepted 23 September 2012
The results indicate that these anomalies are related to the Wenchuan quake of 2008. The numbers of
Available online 2 October 2012
wells that exhibited anomalies of the O1 wave tide-factor increased gradually as time approached the
occurrence of the quake. These anomalies were distributed in space in a discontinuous manner. In terms of
Keywords:
such signals, it is possible to exclude the interferences from other factors such as rainfall and exploitation
Water level
Tide-factor
to some extent when we attempt to use anomalous well level variations to predict future earthquakes.
Great Wenchuan earthquake 2012 Elsevier Ltd. All rights reserved.

1. Introduction
Anomalous variations of groundwater level have been reported
to appear prior to earthquakes (e.g., Chia et al., 2008; King et al.,
2006). Such anomalies are, however, usually very complicated
because of the effects of rainfall, atmosphere, solid tide and arti-
cial interference (Kmpel, 1991; Matsumoto et al., 2003; Woith
et al., 2011). Some researchers regard solid tide as contributing to
noise in water level data and make corrections to it (Cutillo and
Ge, 2006; Reoloffs et al., 2003). Several studies analyzed the rela-
tionship between tidal changes derived from well water levels and
earthquakes and found that anomalies of tidal variations seemed to
occur in the well levels prior to some quakes (Igarashi and Wakita,
1991; Zhu et al., 1997). The tidal effects in the well water level are
primarily the responses to the volumetric strain of the aquifer in the
subsurface, which are usually not subjected to interference by rain-
fall, discharge and other factors. Therefore, to some extent, we can
study tide anomalies from well level observations prior to earth-
quakes without the inuence of these sources of noise (Akita and
Matsumoto, 2004; Li and Jiao, 2001; Riedel et al., 2010; Taniguchi,
2002).
There are several studies on the coseismic responses to the
great Wenchuan earthquake of 2008 using well water level data
Corresponding author at: MOE Key Laboratory of Groundwater Circulation and
Evolution, School of Water Resources & Environment, China University of Geo-
sciences, Beijing 100083, China.
E-mail addresses: wanggc@pku.edu.cn, wanggc@263.net (G. Wang).
from mainland China (Yan, 2009a; Yang et al., 2008). A number
of researchers have suggested that water level anomalies occurred
before this event at the following wells in the Sichuan Province:
Beichuan, Pujiang 11, Xiaojin Chuan 07, Deyang Chuan 08, Luzhou
Chuan 13 and Qionglai Chuan 22 (Cheng et al., 2011; Lan and Chi,
2010; Liu et al., 2011; Yan, 2009b). Most of these studies focused on
dynamic changes in water level. Because of the interference from
rainfall, water exploitation and other factors, the correct method
for the evaluation of these anomalies is controversial.
This work attempts to study the tidal effects of micro dynamic
variations in the water level. We systemically analyzed the vari-
ations of tidal-factors of M2 and O1 waves in the water levels of
16 wells in the Sichuan Province in the 2 years prior to the 2008
Wenchuan event. Our purpose is to extract anomalies of the tidal-
factors and analyze their variations in time and space, migration
in space and its relation to the fault. The major goal is to exam-
ine whether observable tidal anomalies, represented by minute
changes in the well water level, exist prior to large earthquakes.
2. Seismotectonic setting
On 12 May 2008, the M8.0 Wenchuan earthquake (with epi-
center geographic coordinates of 30.986 N, 103.364 E) struck
Wenchuan, Beichuan, and Qingchuan Counties, Sichuan Province,
China. The earthquake event took place beneath the steep east-
ern margin of the Tibetan plateau, a tectonically active region with
frequent major quakes that occur on active faults. The earthquake
generated 240 km rupture zones along the Longmenshan fault, with
a maximum displacement of 6.5 0.5 m (Xu et al., 2009). The event

0264-3707/$ see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jog.2012.09.005
 C. Liu et al. / Journal of Geodynamics 63 (2013) 5461 55

Fig. 1. Map showing the locations of the wells for water level observations in the Sichuan Province.

ruptured the Longmenshan. Reverse and right-slip components
were of comparable magnitude along the southwestern portion of
the rupture, but right-slips dominated the northeastern portion.
This event occurred within the context of a long-term uplift and
eastward enlargement of the Tibetan plateau, which is hampered
by the rigid Sichuan Basin in the east. Such a context in geodynam-
ics is associated with the far-eld effect of the IndiaAsia collision
(Burchel et al., 2008; Zhang et al., 2008).
The study area is located in the Sichuan Province of China,
which has a geographic coordinate of 97109 E, 2535 N. The
study area includes the Sichuan Basin and the eastern Tibetan
plateau. Rocks exposed in the Sichuan Basin are mainly of Cre-
taceous and, locally, of Jurassic age. Eocene rocks are folded and
exposed in active anticlines that form ridges in the southwestern
part of the basin and are unconformably overlain by Pleistocene
conglomerates and sandstone. Between the folds, Eocene and
56 C. Liu et al. / Journal of Geodynamics 63 (2013) 5461
Quaternary strata are generally subhorizontal. Quaternary strata
are present only in the southwestern part of the Sichuan Basin,
southwest of the Longchuan anticline. These strata are typically
<100 m thick and are ponded behind the rising anticlines (Burchel
et al., 2008). Longmenshan fault, which is the seismogenic fault of
the great Wenchuan quake, marks the tectonic contact between
the low-elevation Sichuan Basin to the east and the high-elevation
eastern Tibetan plateau to the west (Burchel et al., 1995). Global
positioning system (GPS) measurements indicate that shorten-
ing across the eastern margin of the Tibetan plateau appears to
have been relatively minor (<3 mm/year) (Zhang et al., 2004).
In addition to the Longmenshan thrust belt, other active faults
in the study area include the Xianshuihe and Anninghe faults.
The northwest-striking Xianshuihe fault is a sinistral strike-slip
fault that dates from at least the Quaternary period (Allen et al.,
1991). The Xianshuihe fault accommodates a left-lateral strike-
slip motion related to the clockwise rotation around the Eastern
Himalayan Syntaxis, and geodetic studies indicate a cumulative
displacement rate of 10 mm/year relative to the South China
Block (Larson et al., 1999; Zhang et al., 2004). The northsouth
trending Anninghe fault is also a sinistral strike-slip fault sit-
uated along the southwest edge of the Yangtze craton. Due to
the complex tectonic setting and active deformation, the study
region also shows the highest level of seismicity in mainland China
(Li et al., 2009). Considering the past century as an example, an
M7.3 event occurred at Diexi in August 1933 (Kan et al., 1977),
and an M7.2 doublet event occurred at Songpan in August 1976
(Molnar and Deng, 1984), both located only approximately 100 km
northwest of the Longmenshan fault system. Recent studies have
shown that the minimum recurrence rate of a major earthquake is
every 23003000 years, while the recurrence interval may be only
10001200 years for the Beichuan-Yingxiu fault (Lin et al., 2010;
Ran et al., 2010).
3. Observation wells
Two years before the Wenchuan quake, the underground water
levels of 16 wells were under observation. The locations of these
wells are shown in Fig. 1. The descriptions of these wells are listed
in Table 1.
Most of the observational wells are situated on the NE-
trending Longmenshan and Huayingshan faults and the NS-striking
Anninghe fault. The distance of the wells to the epicenter of the
2008 Wenchuan M8.0 event ranges between 70 km (Chuan 22 at
Qinglai) and 530 km (Chuan 06 and Chuan 18 at Huili) (see Table 1).
All of the wells penetrated the aquifers and were cased and
screened in order to monitor the groundwater in the conned
aquifers. There are two types of wells, artesian and non-artesian
wells. The well water level in the artesian wells (in which water
can freely ow out of the well) is referred to as the dynamic level,
while the well water level in the non-artesian wells is termed the
static level. All dynamic and static well water levels were continu-
ously recorded by an LN-3A digital instrument, with an accuracy of
1 mm and sampling frequency of 1 min. LN-3A digital instruments
have been widely used to record the water level in the China Seis-
mic Subsurface Fluids Observation Network. The instrument has a
resolution of 1 mm, a stability of 0.25% FS/half year, and a measur-
ing range of 10 m. The instruments (probes) were usually placed at
a depth of several meters below the water tables, depending on the
historical water level uctuation. For water level measurements in
the artesian wells, a special pipe was installed at the well mouths
to lengthen the wells, and a lateral outlet was set up in the lower
part of the pipe, so that the water level could be measured as in the
non-artesian wells.
4. Extraction of tide-factor anomalies
The groundwater level in a closely conned aquifer can change
even if the deformation in the crust is small (Bodvarsson, 1970). The
response sensitivity of water level changes to the crust deforma-
tion is called strain sensitivity. Water level changes caused by the
earths tides are usually known as tide-factor, which is calculated as
the ratio between the measured amplitude of water level changes
(caused by the earths tide) and the theoretical gravity tidal ampli-
tude. The value of the tide-factor does not change much when the
stressstrain eld is steady. Thus, we can detect the stressstrain
anomalies by analyzing the tide-factor. In this paper, we use the
T Tide program to obtain the tide-factor (Pawlowicz et al., 2002).
We have chosen well water level data from the 2 years prior to
and post the Wenchuan event, i.e., from 0:00 h on 1 May 2006 to
23:00 h on 31 December 2008. Before conducting the tidal analysis
(Woith et al., 2011), we rst imputed the missing data using the
triple spline interpolation method and then removed the inuence
of barometric pressure from the raw data using a regression analy-
sis (Nathanial and Todd, 2007; Quilty and Roeloffs, 1991). We used
the hourly water level data of each well to conduct harmonic anal-
yses in order to calculate one tide-factor from a dataset of 30 days.
The subsequent sliding of this dataset backward by 2 days yields a
new tide-factor. This procedure was repeated until 31 December
2008. The effects of barometric pressure uctuations on the water
levels of the conned wells were further removed
This process began on 1 December 2006 for well 5-Zk1, 9 May
2007 for well Beichuan, and 1 December 2007 for well Lugu lake.
The purpose of choosing a time span of 2 years was to extract
intermediate- and short-term anomalies. We calculated the tide-
factors of two primary subwaves, i.e., the semi-diurnal wave (M2
wave) and diurnal wave (O1 wave) of the Moon. Next, we com-
puted their average values (), variance and standard deviations
(). In terms of these values, we determined the baseline as two
standard deviations (2) above the average value ( 2). The
deviations exceeding this line were viewed as anomalies. Taking
the Chuan 02 well as an example, the average tidal-factor value
of M2 was 1.55; the upper and lower limits were 2.05 and 1.05,
respectively. As shown in Fig. 2, the values of M2 were abnormal
in April, May and June of 2008. The value of O1 was also abnormal
during the same periods (Fig. 2). Similar methods have been used
for data processing for all of the other wells. The occurrence times
of tide-factor anomalies from M2 waves and O1 waves at each well
are shown in Tables 2 and 3, respectively.
5. Temporalspatial distribution of tide-factor anomalies
Fig. 3 shows the temporary distribution of the number of wells
bearing M2 and O1 wave tide-factor anomalies from June 2006 to
December 2008.
Fig. 3 shows that the total numbers of wells with M2 and O1 wave
tide-factor anomalies in May 2008 were seven and ten, respec-
tively. Among these wells, the numbers of wells before and after
the M8.0 earthquake event with M2 wave anomalies were one and
six (12 May 2008), respectively. There were ve wells with O1 wave
anomalies before and after the earthquake event.
As shown in Fig. 3, in August 2006, prior to the M8.0 Wenchuan
earthquake, M2 wave tide-factor anomalies appeared in four wells,
which is the most that has been seen. The next was July 2006 with
three wells. In the rest of time prior to the event, for 7 months two
wells had anomalies, for 8 months one well had an anomaly, and
for other 7 months, no anomalies occurred in any well.
For the O1 wave, in April 2008, prior to the Wenchuan earth-
quake, six wells appeared with tide-factor anomalies, which was
the highest number of anomalous wells observed concurrently,
Table 1
Descriptions of observation wells in the Sichuan Province.

Well name Geographical Level type Geological setting Water Water Well depth Casing depth Permeability Aquifer lithology Distance (km)
location temperature ( C) chemical type (m/d) type

Chuan 2 Wajiao, Dynamic Xianshuihe fault 3841 HCO3 K + Na 501.17 25.97 Fissured conned 210
Shimian water in granite
Chuan 3 Taihe, Xichang Static middle Anninghe 20 HCO3 Na 765.60 105 0.01030.345 Fissured conned 360
fault water in granite
Chuan 5 Renhe Dynamic Yalongjiang fault 24 Cl Na 200.06 19.54 0.14 Fissured conned 520
water in granite
Chuan 6 Huili Static intersection of 20 HCO3 K + NaMg 600.26 14.90 0.0135 Fissured conned 530
Mopanshan-Xigeda water in
fault and Ninghui sedimentary rock

C. Liu et al. / Journal of Geodynamics 63 (2013) 5461

fault
Chuan 7 Xiaojin Static near top of Jintang 15 HCO3 MgCa 100.53 15.82 0.177 Fissured conned 90
arc-like structure water in sandstone
Chuan 8 Deyang Dynamic southeast of Cl Na 3072.00 1937.03 Fissured conned 110
Longmenshan fault water in sandstone
Chuan 10 Jiangyou Static southeast of 4076.50 4075.20 Fissured conned 160
Longmenshan fault water in sandstone
Chuan 11 Pujiang Static Longmenshan fault 1688.50 108.34 Fissured conned 80
water in sandstone
Chuan 12 Nanxi Static Huayingshan fault 20 Cl K + Na 105.54 6.10 0.41 Fissured conned 270
water in sandstone
Chuan 13 Luzhou Static Huayingshan fault 21 HCO3 K + NaCa 300.60 6 0.62 Fissured conned 310
water in sandstone
Chuan 18 Huili Static intersection of 20 HCO3 K + Na 523.28 10 0.1457 Fissured conne 530
Mopanshan-Xigeda water
fault and Ninghui
fault
Chuan 20 Chengguan, Dynamic north section of 20 HCO3 CaMg 400.14 79.34 Fissured conned 220
Shimian Anninghe fault water in granite
Chuan 22 Qionglai Static ank of 19 HCO3 NaCa 175.78 18.5 Fissured conned 70
Sanhechang water in sandstone
anticline
5-ZK1 Static middle section of 24 HCO3 NaMg 760 Unknown 360
Anninghe fault
Beichuan Beichuan Static Longmenshan fault 101.14 39 Fissured conned 140
water
Lugu lake Lugu lake Static HCO3 NaCa 200.07 74.12 0.135 Fissured conned 430
water

Note: Distance is from a well to the epicenter of the 2008 Wenchuan M8.0 event; level type: the dynamic indicates that the water table is higher than the ground surface, and water can automatically ow out of the well; static
indicates that the water table is lower than the ground surface, and water cannot freely ow out of the well.

57
58 C. Liu et al. / Journal of Geodynamics 63 (2013) 5461

Fig. 2. Changes in M2 and O1 with anomalies in the Chuan 02 well.

accounting for 40% of the total number of observational wells. The
next highest occurred in March and 111 May of 2008, where
ve wells had anomalies, representing 33% of the total. During
5 months, i.e., October 2006, May and June of 2007 and February
and May of 2008, anomalies occurred in two wells. For 2 months,
two wells exhibited anomalies, and for 7 months, only one well
exhibited an anomaly. No anomaly was present in the 4 months
period.
The M8.0 Wenchuan earthquake occurred on 12 May 2008. The
number of the wells that exhibited anomalies of the tide-factor
(especially that of the O1 wave) increased gradually as the time
approached the occurrence of the quake. The numbers of wells with
O1 wave anomalies were three, ve, six, and ve in February, March,
April and 111 May of 2008, respectively, forming the peak of well
anomalies during the 2 years. The numbers of wells with anoma-
lies remained high for 2 months following the event. There were
six wells each with M2 wave anomalies in 1331 May and June
of 2008; the numbers of wells with O1 wave anomalies were ve
and six in 1331 May and June of 2008, respectively. Eventually,
the numbers of wells with anomalies decreased gradually. As the
tide-factor was closely related to the properties of aquifer materials
(Elkhoury et al., 2006; Rhoads and Robinson, 1979), the evolution
of the above anomalies may suggest that aquifer materials suffer
changes from the processes of development and the occurrence of
earthquakes.
Table 4 and Fig. 4 show the occurrence times of O1 wave tide-
factor anomalies in relation to faults. The May 2008 data were
truncated on the 11th of the month, the day on which the large
event occurred. The data collected after this date were excluded
from the analysis.
As stated previously, most of the observational wells were
located on faults. Tide-factor anomalies were primarily present at

Table 2
Occurrence times of M2 wave tide-factor anomalies in wells.

Well JuneDecember 2006 JanuaryDecember 2007 JanuaryDecember 2008

6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12

Chuan 02
Chuan 03
Chuan 05
Chuan 06
Chuan 07
Chuan 08
Chuan 10
Chuan 11
Chuan 12
Chuan 13
Chuan 18
Chuan 20
Chuan 22
5-ZK1
Beichuan
Lugu lake

Total 2 3 4 1 1 1 0 0 1 0 0 0 0 1 2 1 2 2 2 0 2 1 2 7 6 2 2 4 2 1 2

Note: denotes no data; means anomaly, blank indicates with data but no anomaly; total, total number of wells with anomaly.
 C. Liu et al. / Journal of Geodynamics 63 (2013) 5461 59

Table 3
Occurrence times of O1 wave tide-factor anomalies in wells.

Well JuneDecember 2006 JanuaryDecember 2007 JanuaryDecember 2008

6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12

Chuan 02
Chuan 03
Chuan 05
Chuan 06
Chuan 07
Chuan 08
Chuan 10
Chuan 11
Chuan 12
Chuan 13
Chuan 18
Chuan 20
Chuan 22
5-ZK1
Beichuan
Lugu lake

Total 0 1 1 0 3 1 1 0 0 2 2 3 3 1 1 1 1 2 1 2 3 5 6 10 6 1 3 3 2 0 1

Note: denotes no data; means anomaly, blank indicates with data but no anomaly; total, total number of wells with anomaly.

12

10
number of wells with anomalies

anomaly of m2 wave
anomaly of o1 wave
8

0
06-06 06-08 06-1 0 06-1 2 07-02 07-04 07-06 07-08 07-1 0 07-1 2 08-02 08-04 08-06 08-08 08-1 0 08-1 2
time (year/month)

Fig. 3. Number of wells with anomalies of M2 and O1 waves versus times (year/month) (e.g., 0606 = June 2006).

the NE trending Longmenshan and Huayingshan faults as well as 6. Mechanism of tide-factor anomalies
the NS striking Anninghe fault. The anomalies appeared earlier on
the Huayingshan fault and were notable and faint, without a trend The tidal force refers to the difference between the gravitational
toward enhancement. The latest but strongest anomalies occurred force of the Moon, Sun or another planet on a unit mass body on
on the Anninghe fault, were roughly continuous, and were the the earth and that on a unit mass at the earths center. The tidal
largest in amount and proportion of all anomalous wells. On the force also refers to the composite force of the inertial centrifugal
Longmenshan fault, tide-factor anomalies were also rather conspic- force caused by the Earths motion around the mass center of the
uous, particularly for January to April 2008, less than 5 months prior EarthMoon (Sun) system and the gravitational force of the Moon
to the Wenchuan event. This probably implies that changes in the (Sun). Such tidal forces can generate periodic deformations of solid
stress state of the Longmenshan fault led to variations in volumet- earth, known as solid tide.
ric strain, which further altered the porosity of the rocks, resulting The response of the groundwater level to solid tide is known
in anomalies of tide-factors derived from well level observations. as the tidal effect, which can be derived from observations of well

Table 4
Occurrence times of O1 wave tide-factor anomalies in relation to faults.

Fault 2006 2007 2008 N

6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4

Longmenshan 0 0 0 0 2 0 0 0 0 1 0 0 2 1 1 0 1 1 0 1 2 1 2 6
Anninghe 0 0 0 0 0 0 0 0 0 1 2 3 1 1 1 1 0 1 1 0 1 3 4 6
Zemuhe 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Huayingshan 0 1 1 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 2

Note: N, number of observation wells.

60 C. Liu et al. / Journal of Geodynamics 63 (2013) 5461

Fig. 4. The numbers of abnormal wells of O1 wave changes with time in different faults.

water levels. Mechanistically, the tidal force can produce a bulk of stress. For example, when the aquifer is subjected to extension,
dilation of the aquifer, leading to a change in the well water level. the porosity n becomes larger, volumetric strain increases and the
For an idealized model of the aquifer system, which is innite tide-factor becomes smaller. On the contrary, as the aquifer is com-
in extent and consists of a porous, permeable material, a dilatation pressed, the porosity decreases, volumetric strain decreases and the
of the aquifer will cause a change in water level within the well. tide-factor increases. In summary, tide-factor changes can reect
The change in pore pressure results from the dilatation of the variations of the stress state of an aquifer. Therefore, it is likely
aquifer and can be expressed as (Rhoads and Robinson, 1979): that the variations in the temporalspatial distribution of the tide-
factor derived from well water level data presented previously are

dp =   (1) indicative of changes in the stressstrain in the aquifer during the
((1 n)/Em ) + (n/Ew ) development and occurrence of the earthquake. After the earth-
where dp is the change of pore pressure in the aquifer,  is the quake, the number of tide-factor anomalies gradually decreased.
dilation of the aquifer, n is the porosity of the aquifer, and Em and
Ew are the bulk modulus of the rock matrix and water in the aquifer. 7. Discussion
Eq. (1) can be rewritten as:
(1) There are only a few previous studies concerning the appli-
g cation of tide-factor of well water level for the prediction of
dp = (2)
Ss earthquakes. There are no studies of stress state change using
where migration of the tide-factor anomalies in space. This work is an
 (1 n) n
 attempt to address these gaps.
Ss = g + (3) (2) Certain factors, such as rainfall and articial exploitation, can
Em Ew
produce considerable variations in the well water level. How-
and Ss is the specic storage of the aquifer,  is water density, and ever, these interferences have no signicant effect on the
g is gravity acceleration. tide-factor of the well water level. Furthermore, the inuence of
Thus, the relationship between the water level change and the meteorological factors such as barometric pressure is mainly on
aquifer dilation can be written as: the tidal constitutes of K1 and S2 waves (Hsieh et al., 1987; Liao
 et al., 2011). After the barometric pressure effects are removed,
dh = (4) the anomalies of the tide-factor, derived from the well water
Ss
level data, can be utilized to detect possible precursory signals
where dh is the water level change in the aquifer. According to the associated with major quakes.
denition of tide-factor, the expression equation of the tide-factor (3) The results of this study indicate that the tide-factor of the pri-
can be written as: mary component, the M2 wave, did not exhibit any evident
dh 1 1 anomalies before the Wenchuan M8.0 event of 2008. However,
= = =
 Ss g[((1 n)/Em ) + (n/Ew )] the tide-factor anomalies of the O1 wave seemed to become
relatively concentrated as time approached the occurrence of
Em Ew
= (5) the event. The mechanistic basis of this event requires further
g [Ew + n(Em Ew )] study.
According to Eq. (5), when the water density and gravity accelera-
tion are constant, the tide-factor is a function of the bulk modulus 8. Conclusions
Em of solid grains and bulk modulus Ew of uids. For a specic
aquifer, Em and Ew are considered to be constant. Thus, the tide- This study attempts to use the tide-factor anomalies derived
factor reects the change in the porosity n of the aquifer. Variations from observations of well water level. This effort is based on the
of n are the result of volumetric strains, which represent the effects following theory: when an aquifer is stressed, its ssures will be
 C. Liu et al. / Journal of Geodynamics 63 (2013) 5461
closed or opened, i.e., its porosity will change. Such a change must
affect the response of a well-aquifer system to the earths tidal force.
Apparently, abnormal variations in the tide-factor are related to a
changing stress state of the aquifer. Thus, the study of tide-factor
anomalies may provide a new approach to detect precursory signals
before major quakes.
We use the observational data of the groundwater level from
16 wells in the Sichuan Province in a two-year period prior to the
2008 Wenchuan M8.0 quake in order to extract tide-factor anoma-
lies and analyze their distribution features in time and space. Our
conclusions are as follows:
(1) Before the Wenchuan event, anomalies of both M2 and O1 wave
tide-factors existed in the well water level.
(2) These anomalies were discontinuous in space, perhaps because
some wells were relatively sensitive, behaving as sensitive
points, and others may not have been so sensitive. Furthermore,
the anomalies were also discontinuous in time.
(3) The tide-factor anomalies of the O1 wave exhibited a distinct
varying rule with time. As time approached the occurrence
of the quake, the number of wells with O1 wave tide-factor
anomalies tended to increase, especially 5 months before the
event. This signal might be helpful for the estimation of the
occurrence time of a future quake.
(4) The spatial distribution of the tide-factor anomalies of the O1
wave was relatively complicated. More than 1 year prior to
the Wenchuan event (before March 2007), anomalies of the
O1 wave appeared on the Longmenshan, Aninhe and Huaying-
shan faults in a disordered and scattered manner. Two or three
months prior to the quake, anomalies seemed to be concen-
trating on these faults. Although the epicenter of future shocks
cannot be predicted with certain accuracy, based on these fea-
tures, it is possible to demarcate the tectonic zones with seismic
hazard.
Acknowledgements
This work was supported by the National Natural Science Foun-
dation of China (40930637), Special Project of the Institute of
Geology, China Earthquake Administration (IGCEA-1205) and Spe-
cial Project of the Seismological Community (200808079). We
gratefully thank Dr. Heiko Woith and Dr. Randell Stephenson for
their very helpful comments and suggestions, which considerably
improved the manuscript.
References
Akita, F., Matsumoto, N., 2004. Hydrological responses induced by the Tokachi-oki
earthquake in 2003 at hot spring wells in Hokkaido, Japan. Geophysical Research
Letters 31, L16603.
Allen, C.R., Luo, Z., Qian, H., 1991. Field study of a high active fault zone: the Xian-
shuihe fault of southwestern China. Geological Society of America Bulletin 103,
11781190.
Bodvarsson, G., 1970. Conned uids as strain meters. Journal of Geophysical
Research 75 (14), 27112718.
Burchel, B.C., Chen, Z., Liu, Y., Royden, L.H., 1995. Tectonics of Longmen Shan and
adjacent regions, central China. International Geology Review 37, 661736.
Burchel, B.C., Royden, R.D., Hilst, V.D., Hager, B.H., 2008. A geological and geophys-
ical context for the Wenchuan earthquake of 12 May 2008, Sichuan, Peoples
Republic of China. GSA Today 18 (7), http://dx.doi.org/10.1130/GSATG18A.1.
Cheng, W., Guan, Z., Su, Q., Ruan, X., Zhang, Z., 2011. Precursory anomalies in
Sichuan region before 2008 Wenchuan Ms8.0 earthquake and their statisti-
cal analysis. Acta Seismologica Sinica 33 (3), 304318 (in Chinese with English
abstract).
Chia, Y., Chiu, J.J., Chang, Y., Lee, T., Liu, C., 2008. Spatial and temporal changes of
groundwater level induced by thrust faulting. Pure and Applied Geophysics 165,
516.
Cutillo, P.A., Ge, S., 2006. Analysis of strain-induced ground-water uctuations at
Devils Hole, Nevada. Geouids 6, 319333.
Elkhoury, J.E., Brodsk, E.E., Agnew, D.C., 2006. Seismic waves increase permeability.
Nature 441, 11351138.
61
Hsieh, A.P., Bredehoeft, J.D., Farr, J.M., 1987. Determination of aquifer transmissivity
from earth tide analysis. Water Resource Research 23, 18241832.
Igarashi, G., Wakita, H., 1991. Tidal responses and earthquake-related changes in the
water level of deep wells. Journal of Geophysical Research 96, 42694278.
Kan, R., Zhang, S., Yan, F., Yu, L., 1977. Present tectonic stress eld and its relation
to the characteristics of recent tectonic activity in southwestern China. Acta
Geophysica Sinica 20 (2), 96109 (in Chinese).
King, C., Zhang, W., Zhang, Z., 2006. Earthquake-induced groundwater and gas
changes. Pure and Applied Geophysics 163, 633645.
Kmpel, H.J., 1991. Pore-pressure variation as a precursory phenomenon: the need
for and a list of supplementary data. Tectonophysics 193, 377383.
Lan, S., Chi, B., 2010. Research on abnormal response of groundwater levels in deep
wells in the near-eld to the Wenchun Earthquake. Hydrogeology and Engineer-
ing Geology 37 (2), 15 (in Chinese with English abstract).
Larson, K.M., Burgman, R., Bilham, R., 1999. Kinematics of the IndiaEurasia collision
zone from GPS measurements. Journal of Geophysical Research 104, 10771093.
Li, H., Jiao, J.J., 2001. Tide-induced groundwater uctuation in a coastal leaky con-
ned aquifer system extending under the sea. Water Resources Research 37 (5),
11651171.
Li, H.Y., Su, W., Wang, C.Y., Huang, Z.X., 2009. Ambient noise Rayleigh wave tomogra-
phy in western Sichuan and eastern Tibet. Earth and Planetary Science Letters
282, 201211.
Liao, x, Liu, C.P., Yang, X.H., Tian, Y.P., Shi, Y., Tang, Y.D., 2011. Undrained examination
of tidal response of water-level in conned wells. Acta Seismologica Sinica 33,
234242 (in Chinese with English abstract).
Lin, A., Ren, Z., Jia, D., Miyairi, Y., 2010. Evidence for a Tang-Song Dynasty
great earthquake along the Longmen Shan Thrust Belt prior to the
2008 Mw 7.9 Wenchuan earthquake, China. Journal of Seismology,
http://dx.doi.org/10.1007/s10950-010-9186-7.
Liu, C., Che, Y., Xu, Y., Shi, Z., Yang, X., 2011. Abnormal interruption of water ow in
the Chuan 22 well in Qionglai city, Sichuan Province prior to the 2008 Wenchuan
earthquake, China. Geodesy and Geodynamics 2 (2), 5357.
Matsumoto, N., Kitagawa, G., Roeloffs, E.A., 2003. Hydrological response to earth-
quakes in the Haibara well, central Japan: I, Groundwater level changes revealed
using state space decomposition of atmospheric pressure, rainfall and tidal
responses. Geophysical Journal International 155, 885898.
Molnar, P., Deng, Q., 1984. Faulting associated with large earthquakes and the aver-
age rate of deformation in central and eastern Asia. Journal of Geophysical
Research 89, 62036227.
Nathanial, J.T., Todd, C.R., 2007. Removal of barometric press effects and earth tides
from observed water levels. Ground Water 45, 101105.
Pawlowicz, R., Beardsley, B., Lentz, S., 2002. Classical tidal harmonic analysis includ-
ing error estimates in MATLAB using T TIDE. Computers and Geosciences 28,
929937.
Quilty, E.G., Roeloffs, E.A., 1991. Removal of barometric pressure response from
water level data. Journal of Geophysical Research 96 (B6), 1020910218.
Ran, Y., Chen, L., Chen, J., Wang, H., Chen, G., Yin, J., Shi, X., Li, C., Xu, X., 2010. Pale-
oseismic evidence and repeat time of large earthquakes at three sites along the
Longmenshan fault zone. Tectonophysics 491, 141153.
Reoloffs, E., Sneed, M., Galloway, D.L., Sorey, M.L., Farrar, C.D., Howle, J.F., Hugher,
J., 2003. Water-level changes induced by local and distant earthquakes at Long
Valley caldera, California. Journal of Volcanology and Geothermal Research 127,
269303.
Rhoads, G.H., Robinson, E.S., 1979. Determination of aquifer parameters from well
tides. Journal of Geophysical Research 84, 60716082.
Riedel, T., Lettmann, K., Beck, M., Brumsack, H.J., 2010. Tidal variations in groundwa-
ter storage and associated discharge from an intertidal coastal aquifer. Journal
of Geophysical Research 115 (4), 40134022.
Taniguchi, M., 2002. Tidal effects on submarine groundwater discharge into the
ocean. Geophysical Research Letters 29 (12), 15611563.
Woith, H., Barbosa, S., Gajewski, C., Steinitz, G., Piatibratov, O., Malik, U., Zschau, J.,
2011. Periodic and transient radon variations at the Tiberias hot spring, Israel
during 20002005. Geochemical Journal 45, 473482.
Xu, X., Wen, X., Yu, G., Chen, G., Klinger, Y., Hubbard, J., Shaw, J., 2009. Coseis-
mic reverse- and oblique-slip surface faulting generated by the 2008 Mw 7.9
Whenchuan earthquake, China. Geology, 515518.
Yan, R., 2009a. Characteristics analysis of coseismic response to well water level
caused by Wenchuan 8.0 earthquake in China. Recent Developments in World
Seismology 4, 32.
Yan, R., 2009b. Analysis of Earthquake Precursor Data. In: Report of Scientic
Research on Wenchuan Ms8. 0 Earthquake, Edited by monitoring and predic-
tion division. China Earthquake Administration, Seismology Publishing House,
Beijing, pp. 175184 (in Chinese).
Yang, Z., Deng, Z., Liu, C., 2008. Coseismic changes of water level and water temper-
ature caused by Ms8.0 Wenchuan earthquake. Seismology and Geology 30 (4),
895905 (in Chinese with English abstract).
Zhang, P., Shen, Z., Wang, M., Gan, W., 2004. Continuous deformation of the Tibetan
Plateau from global positioning system data. Geology 32, 809812.
Zhang, P., Xu, X., Wen, X., Ran, Y., 2008. Slip rates and recurrence intervals of the
Longmen Shan active fault zone, and tectonic implications for the mechanism
of the May 12 Wenchuan earthquake, 2008, Sichuan, China. Chinese Journal of
Geophysics 51 (4), 10661073.
Zhu, Q., Tian, Z., Zhang, Y., Gu, Y., Gao, S., 1997. Distortion of earth tide form water
level in the Taipingzhuang well in Beijing and its relation to earthquakes. Seis-
mology and Geology 19 (1), 5052 (in Chinese with English abstract).