Tectonophysics 483 (2010) 305–309

Contents lists available at ScienceDirect

Tectonophysics
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / t e c t o

Linear poroelasticity of groundwater levels from observational records at wells in

Hokkaido, Japan
Tomo Shibata a,⁎, Norio Matsumoto b, Fujio Akita a, Noritoshi Okazaki a, Hiroaki Takahashi c, Ryuji Ikeda d
a
Geological Survey of Hokkaido, N19 W12, Kita-ku, Sapporo 060-0819, Japan
b
Tectono-Hydrological Research Group, Geological Survey of Japan, AIST, Tsukuba Central 7, 1-1-1 Higashi, Tsukuba 305-8567, Japan
c
Institute of Seismology and Volcanology, Graduate School of Science, Hokkaido University, N10 W8, Kita-ku, Sapporo 060-0810, Japan
d
Department of Natural History Science, Faculty of Science, Hokkaido University, N10 W8, Kita-ku, Sapporo 060-0810, Japan

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

Article history: The natural fluctuation of groundwater level depends not only on the stress applied to an aquifer but also on
Received 20 April 2009 the type of rock that comprises it. Poroelastic parameters, such as strain sensitivity, loading efficiency, and
Received in revised form 23 October 2009 poroelastic bulk moduli, are used to analyze the hydrologic attributes of aquifers. In this study, we use
Accepted 23 October 2009
records of groundwater levels from 19 wells to calculate coefficients of poroelastic characteristics. The
Available online 1 November 2009
records allow a comprehensive description of how groundwater levels vary with fluctuations of volumetric
Keywords:
strain in the crust's interior. We estimate strain sensitivities in two ways: (a) using the wells' response to
Groundwater level tidal strain and (b) using coseismic changes in groundwater level based on historical data. The strain
Linear poroelasticity sensitivities estimated by the two different methods have good linear correlation. We also estimate loading
Strain sensitivity efficiency using observational records and calculate bulk moduli from the loading efficiencies and strain
Loading efficiency sensitivities. The bulk moduli obtained in this manner are roughly consistent with laboratory values.
Bulk modulus © 2009 Elsevier B.V. All rights reserved.

1. Introduction 2. Observations of groundwater levels

Linear poroelasticity provides a compelling relationship between We used data on groundwater levels at nineteen wells spanning the
fluid pressure and deformation of fluid-saturated rock, and thus is an period from 1993 to 2004 as shown in Table 1. Groundwater levels were
excellent parameter for analyzing hydraulic phenomena. One linear in a static condition in all wells, and groundwater depths ranged
poroelastic response, coseismic change in groundwater level, has between 68 and 1500 m. The aquifers tap into volcanic rocks and/or
been studied for many years (e.g., Wakita, 1975; Roeloffs, 1996; Akita sedimentary rocks and are confined. Groundwater level was recorded
and Matsumoto, 2001; Igarashi and Wakita, 1991; Quilty and Roeloffs, with an accuracy of 5–10 mm at 10-minute intervals at each
1997; Akita and Matsumoto, 2004). Poroelasticity, however, cannot observation site (Fig. 1). Observed coseismic changes in groundwater
explain all coseismic changes (Roeloffs, 1998; Matsumoto et al., 2003; levels and volumetric strains are shown in Table 2. A portion of the data
Brodsky et al., 2003). for these changes was reported by Akita and Matsumoto (2001, 2004).
We have long-term data on groundwater levels measured at 19
wells in Hokkaido, Japan, an area where large earthquakes occur 3. Estimation of strain sensitivity
frequently (National Astronomical Observatory of Japan, 2004). In
addition, we have data on the coseismic changes due to six earthquakes Strain sensitivity is a poroelastic characteristic of aquifers and has
with magnitude seven (M 7) or greater in the Hokkaido region from been calculated from coseismic and/or tidal changes in groundwater
1993 to 2004: the 1993 Kushiro-oki (M 7.6), the 1993 Hokkaido-Nansei- levels (e.g., Igarashi and Wakita, 1991; Akita and Matsumoto, 2001,
oki (M 7.8), the 1994 Hokkaido-Toho-oki (M 8.1), the 1994 Sanriku- 2004). If the aquifer is confined, the relationship between the change
Haruka-oki (M 7.5), the 2003 Tokachi-oki (M 8.0), and the 2004 in groundwater level, Δh, and the change in volumetric strain, Δε, is
Kushiro-oki (M 7.1) (Akita and Matsumoto, 2001; Japan Meteorological represented by
Agency, 2005). In this paper, we report changes in groundwater levels
associated with the earthquakes, explore the response of ground water Δh = −ðB Ku = ρw gÞ⋅Δε = −Wε ⋅Δε ð1Þ
levels to tidal strain at the 19 wells, and analyze the linear poroelasticity
of the aquifers by using groundwater level fluctuations. where B is Skempton's coefficient, Ku is the undrained bulk modulus,
ρw is the density of water and g is the acceleration of gravity (Jacob,
⁎ Corresponding author. Tel.: +81 11 747 2473; fax: +81 11 737 9071. 1940; Wang, 2000). In this paper, the coefficient Wε denotes the strain
E-mail address: tomo@gsh.pref.hokkaido.jp (T. Shibata). sensitivity of the groundwater level.

0040-1951/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.tecto.2009.10.025
306 T. Shibata et al. / Tectonophysics 483 (2010) 305–309

Table 1
Description of observation wells.

Site Lat. Lon. Alt. Depth of well Screened depth Geology Distance from the coast
N(°) E(°) (m) (m) (m) (km)

AB 42.533 140.796 83.0 120 59.5–109 Qs, Pv 1.8

AK1 43.435 144.100 430.0 118.5 72–115.7 Pv, Mv 52
AK3 43.430 144.085 420.0 91.8 24–91.8 Pv, Mv 50
JS 41.980 140.638 169.0 1103 570–1078 – 16
KN 43.739 144.925 157.0 792 500–792 Mv 11
KT 42.022 140.626 165.0 974 638–947 – 9.3
NW 42.468 140.877 15.0 1010 580–1000 Ms 1.0
OB1 42.919 143.131 58.5 1328 950–1060 Ps, Ms, PTs 50
OB4 42.909 143.202 47.0 1400 1235–1400 Ps, Ms, PTs 45
OB6 42.919 143.131 47.0 220 164.5–192.5 Ps 47
SK 41.778 140.783 3.8 85 45.2–80 Ms, Mv 0.38
SP1 43.081 141.316 10.4 657 287.9–309.9, 354.1–376.2 Qs 12
SP2 43.081 141.316 10.4 657 539–597 Ps 12
SP3 43.056 141.341 20.0 1001 800–1001 Ps, Ms 15
SR2 42.455 141.204 21.0 508 – Ps, Pv, Ms, Mv 0.15
TS 43.485 144.464 102.3 416 – Mv, Ms 58
TS2 43.505 144.452 138.0 1500 – – 57
YC 41.753 140.715 6.0 200 150–200 Mv 0.35
YN 41.774 140.790 4.5 68 48–68 Ms, Mv 0.08

Descriptions of some wells quoted from Akita and Matsumoto (2001, 2004). Qs: Quaternary sedimentary rocks, Ps: Pliocene sedimentary rocks, Pv: Pliocene volcanic rocks,
Ms: Miocene sedimentary rocks, Mv: Miocene volcanic rocks, Pts: Pre-Tertiary sedimentary rocks.

We estimate the strain sensitivity of the groundwater level by constituents, and the M2 tidal constituent has a period of over 12 h,
comparing the observed response of groundwater level to the during which there are large peaks of temperature and barometric
volumetric strain of theoretical tides. The amplitude and phase shift effects (Roeloffs, 1996). The amplitudes and phase shifts of the
of the tidal response of the groundwater level were determined using theoretical tidal strain and the groundwater-level response to the M2
the BAYTAP-G software (Tamura et al., 1991). The amplitude and tidal constituent at the 19 wells are shown in Table 3, where positive
phase shift of the theoretical volumetric strain, taking into account values denote a delay of the phase shift.
both the earth and the ocean tide loadings, were estimated using the The calculated strain sensitivities W(M2)
ε for eleven wells (AK1,
GOTIC2 software (Matsumoto et al., 2001). AK3, JS, KN, KT, OB1, OB4, SP1, SP2, SP3, and TS2) are reasonable
We calculated the strain sensitivity of the groundwater level with because the phase shifts of the tidal response of groundwater level at
the M2 tidal constituent (period: 12.4206 h), W(M2)
ε , those wells are almost the same as those of the theoretical tidal strain.
The sensitivities of OB1, OB4, SP1, and SP2 also agree well with those
ðM2Þ
Wε = TW = TT ; ð2Þ reported by Akita and Matsumoto (2004). However, the phase shifts
of the tidal response in the other wells (AB, NW, OB6, SK, SR2, TS, YC
where TW is the groundwater level amplitude of the M2 tidal and YN) differ from the theoretical tidal strains. Except for OB6 and TS,
constituent and TT is the theoretical amplitude of M2 tidal volumetric these wells are situated near the coast (within 2 km), and the
strain. The reason we chose the M2 tidal constituent is that the estimated amplitude of the M2 tidal volumetric strain contributed by
amplitude of the M2 tidal strain is the largest among the tidal the ocean tide loading, To, is more than 80% of the estimated
amplitude contributed by the earth tide, TE (Table 3).
There are two possible explanations for the disagreement between
phase shift of the groundwater levels at wells AB, NW, SK, TS, and YC
and theoretical tidal strain. First, estimates of the M2 tidal strain
contributed by the ocean tide loading could be incorrect. Second, the
groundwater levels in these aquifers may depend not only on ocean
tide loading but also on other phenomena not considered in Eq. (1),
such as water flow from the ocean. These explanations, however,
cannot account for the phase shift discrepancies in wells OB6 and TS.
If several of these coseismic changes in groundwater level, Δheq,
are proportional to the coseismic changes in volumetric strain, Δεeq,
we can estimate the coefficient, W(E)ε ,

ðEÞ
Δheq = −Wε ⋅Δεeq ð3Þ

where W(E)ε can be thought of as the strain sensitivity of the

groundwater level estimated by earthquake response if Δheq
correlates well with the volumetric strain and the coefficient has a
plausible value. The coseismic change in groundwater level has the
opposite sign from volumetric earthquake strain. Table 2 presents the
calculated strain sensitivities for the 19 wells, where coseismic
changes in volumetric strain were estimated by Okada's computer
code (Okada, 1992).
The coseismic changes in groundwater level, for the most part,
Fig. 1. Location of observation sites. have good linear correlations with the change in volumetric strain
 T. Shibata et al. / Tectonophysics 483 (2010) 305–309 307

Table 2
Observed and compiled data of groundwater change and volumetric strain change induced by earthquakes at each well.

Site Kushiro-oki Hokkaido- Hokkaido-Toho- Sanriku-Haruka- Tokachi-oki Kushiro-oki Wε(E) R2

Nansei-oki oki oki
Eq. (1993) Eq. (1993) Eq. (1994) Eq. (1994) Eq (2003) Eq. (2004)

Resp. Strain Resp. Strain Resp. Strain Resp. Strain Resp. Strain Resp. Strain
(cm) (10− 8) (cm) (10− 8) (cm) (10− 8) (cm) (10− 8) (cm) (10− 8) (cm) (10− 8) (mm/10− 8)

AB – – – – – – – – − 59 31.4 − 35 0.2 18.9 –

AK1 – – – – – – – – −4 12.0 0 6.1 2.6 0.589
AK3 – – – – – – – – −7 13.8 −3 6.0 5.1 0.999
JS – – – – – – – – − 22 18.6 – – 11.8 –
KN – – – – – – – – – – −9 5.8 15.5 –
KT – – – – – – – – −5 19.1 0 0.1 2.6 0.999
NW – – – – – – – – −55 33.4 – – 16.5 –
OB1 27 − 24.2 −4 7.0 − 12 25.0 0 − 0.1 −130 191.4 0 1.6 6.8 0.991
OB4 – – – – – – – – 430 170.3 30 1.7 − 25.3 0.992
OB6 – – – – – – – – − 65 190.3 – – 3.4 –
SK – – – – – – – – 27 16.7 0 0.1 − 16.2 0.999
SP1 2 4.0 −9 32.9 −3 10.3 −1 1.1 −6 53.2 0 0.3 1.6 0.641
SP2 – – −6 32.9 −2 10.3 −2 1.1 − 16 53.2 0 0.3 2.7 0.912
SP3 – – – – – – – – − 40 54.2 – – 7.4 –
SR2 – – – – – – – – − 13 48.1 0 0.2 2.7 0.999
TS – – – – – – – – 4 − 21.8 – – 1.8 –
TS2 – – – – – – – – – – −3 8.2 3.7 –
YC – – – – – – – – 10 15.2 0 0.1 − 6.6 0.999
YN 9 0.1 − 73 44.6 0 3.4 − 12 7.3 − 30 16.7 0 0.1 16.5 0.973

(Table 2), suggesting that the W(E)

ε coefficients are plausible. However, indicates a transitory change of hydraulic conductivity induced by
two coseismic changes, at wells AK1 and SP1, which have coefficients waves of an earthquake. Such a transitory effect could cause the
of determination (R2) less than 0.9, do not have good linear unexplained negative changes in groundwater levels. Poroelasticity,
correlations with the change in volumetric strain, and three however, cannot explain some of the observed coseismic changes,
coefficients (OB4, SK, and YC) have negative signs. These groundwater such as those in wells AK1, SK, SP1 and YC.
changes cannot be explained by poroelastic changes induced by Fig. 2 shows the comparison of the coefficients W(M2) ε and W(E)
ε at
earthquakes. The negative sign of the coefficient at OB4 may be caused 14 of the 19 wells. Wells AK1, OB4, SK, SP1, and YC are omitted
by water flow from a shallow aquifer to a screened aquifer or leakage because the W(E) ε of those five wells have negative signs and thus are
from a shallow aquifer to the well because the water temperature not plausible as strain sensitivities. Open circles in Fig. 2 indicate the
decreased after the Tokachi-oki earthquake in 2003 (Akita and wells in which there are discrepancies in the phase shifts between the
Matsumoto, 2004). observed and theoretical values of W(M2)ε . In Fig. 2b, the ratio of W(E)
ε /
Other possible explanations for changes in poroelastic response W(M2)
ε is classified by aquifer rock type. The W(E)ε coefficients are
have been proposed (e.g., Roeloffs, 1998; Matsumoto et al., 2003; proportional to those of W(M2)
ε within a range from 0.1 to 10 at the
Masterlark, 2003). The barrier-clearing model of Brodsky et al. (2003) wells, despite having different aquifer rock types. These linear

Table 3
Calculated amplitude and phase shifts of M2 tidal constituent, responses of water level to tidal constituent, strain sensitivity using M2 tidal constituent, barometric efficiency (B.E.)
and loading efficiency (γ).

Site TE TO TT TW Wε(M2) B.E. γ

amplitude (10− 8) water level (mm) strain sensitivity (=1 − B.E.)
[phase shift (deg.)] [phase shift (deg.)] (mm/10− 8)

AB 0.889 [0] 1.095 [− 102] 1.255 [− 58] 5.03 ± 0.23 [34.6 ± 2.6] 4.0 0.75 ± 0.22 0.25 ± 0.22
AK1 0.863 [0] 0.234 [−96] 0.871 [− 16] 3.32 ± 0.30 [− 21.7 ± 5.2] 3.8 0.77 ± 0.20 0.23 ± 0.20
AK3 0.863 [0] 0.234 [− 96] 0.871 [− 15] 0.32 ± 0.21 [− 52.1 ± 37.3] 0.4 0.82 ± 0.20 0.18 ± 0.20
JS 0.905 [0] 0.049 [− 51] 0.936 [− 2] 5.00 ± 0.12 [−22.4 ± 1.4] 5.3 0.78 ± 0.20 0.22 ± 0.20
KN 0.854 [0] 0.200 [−97] 0.854 [−13] 25.16 ± 1.57 [− 19.0 ± 3.6] 29.5 0.59 ± 0.25 0.41 ± 0.25
KT 0.904 [0] 0.207 [−96] 0.905 [− 13] 3.17 ± 0.14 [− 22.2 ± 2.5] 3.5 0.83 ± 0.16 0.17 ± 0.16
NW 0.891 [0] 0.999 [− 102] 1.193 [−55] 23.18 ± 0.18 [34.2 ± 0.4] 19.4 0.54 ± 0.28 0.46 ± 0.28
OB1 0.878 [0] 0.217 [− 96] 0.882 [− 14] 8.29 ± 0.12 [− 6.4 ± 0.8] 9.4 9.4* 0.72 ± 0.23 0.28 ± 0.23
OB4 0.878 [0] 0.215 [−96] 0.882 [− 14] 6.95 ± 0.78 [− 10.5 ± 6.4] 7.9 8.7* 0.65 ± 0.22 0.35 ± 0.22
OB6 0.878 [0] 0.219 [− 96] 0.882 [−14] 1.83 ± 1.01 [58.2 ± 19.4] 2.1 2.1* 0.71 ± 0.24 0.29 ± 0.24
SK 0.911 [0] 0.743 [− 106] 1.003 [− 45] 45.42 ± 0.83 [52.7 ± 1.0] 45.3 0.55 ± 0.25 0.45 ± 0.25
SP1 0.873 [0] 0.164 [−92] 0.882 [− 11] 3.36 ± 0.15 [−9.5 ± 2.6] 3.8 3.9* 0.81 ± 0.17 0.19 ± 0.17
SP2 0.873 [0] 0.164 [− 92] 0.882 [−11] 3.48 ± 0.29 [− 11.2 ± 4.9] 4.0 3.0* 0.84 ± 0.16 0.16 ± 0.16
SP3 0.874 [0] 0.157 [− 91] 0.885 [− 10] 5.20 ± 2.35 [−5.0 ± 25.9] 5.9 0.56 ± 0.26 0.44 ± 0.26
SR2 0.891 [0] 1.126 [−103] 1.267 [−60] 80.69 ± 0.16 [62.1 ± 0.1] 63.7 0.38 ± 0.25 0.62 ± 0.25
TS 0.862 [0] 0.243 [−96] 0.869 [− 16] 8.67 ± 1.64 [130.4 ± 10.8] 10.0 0.64 ± 0.25 0.36 ± 0.25
TS2 0.861 [0] 0.242 [−96] 0.868 [−16] 4.67 ± 0.44 [− 10.7 ± 5.3] 5.4 0.82 ± 0.20 0.18 ± 0.20
YC 0.911 [0] 1.075 [− 107] 1.187 [− 60] 24.56 ± 0.11 [62.5 ± 6.5] 20.7 0.44 ± 0.26 0.56 ± 0.26
YN 0.911 [0] 0.775 [− 106] 1.018 [− 47] 49.13 ± 0.36 [51.1 ± 0.4] 48.3 45.3* 0.53 ± 0.25 0.47 ± 0.25

TE: volumetric strain by M2 earth tide, TO: volumetric strain by M2 oceanic tidal loading, TT: volumetric strain by earth and oceanic tide, TW: M2 amplitude of water level. *Strain
sensitivities from Akita and Matsumoto (2001, 2004).
308 T. Shibata et al. / Tectonophysics 483 (2010) 305–309

shown in Table 3. The loading efficiencies of the wells near the coast
(NW, SK, SR2, YC, and YN) are more than 0.4.
The loading efficiency can be rewritten as

ðuÞ
γ = B⋅Ku = Kv ; ð5Þ

where K(u)
v is the uniaxial undrained bulk modulus (Jacob, 1940).
Combining Eqs. (1) and (5), we see that the strain sensitivity, Wε, is
proportional to the loading efficiency, γ:

ðuÞ
We = γ⋅Kv = ρw g: ð6Þ

Fig. 3 shows the correlation between the loading efficiency and the
strain sensitivity classified by aquifer rock type. As mentioned above,
the strain sensitivities of some wells near the coast may be inaccurate.
In particular, as wells SK, SR2, YC, and YN are within 500 m of the
coast, the strain sensitivities, W(M2)
ε , are larger than those of the other
wells. We infer that the large response of the groundwater level to
tidal strain and inaccurate estimation of tidal strain at SK, SR2, YC, and
YN yield large strain sensitivities W(M2)
ε . At SR2 and YN, we substituted
W(E)
ε as the strain sensitivity response for the strain sensitivity by tidal
response, W(M2)
ε . The substitute strain sensitivity and bulk modulus at
SR2 and YN are shown in Fig. 3. However, we do not use W(E) ε as the
strain sensitivity at SK and YC because the sensitivities of W(E) ε at SK

Fig. 2. (a) The coefficient Wε(E) relating coseismic changes in groundwater level and
coseismic strain step versus the strain sensitivity by the M2 tidal constituent (Wε(M2)),
and (b) Wε(E) normalized by Wε(M2) in each well given by aquifer rock type. Data for five
wells AK1, OB4, SK, SP1 and YC are not plotted in (a) and (b), because Wε(E) in three of
the wells have negative signs and are not plausible as strain sensitivities. The other
two do not have good linear correlations with the change in volumetric strain. Open
circles indicate the wells with phase shifts that differ between observational and
theoretical values in Wε(M2). Solid and dotted lines denote factors of 0.1, 1, and 10 in the
correlation.

proportions are known to vary by orders of magnitude (Koizumi et al.,

2002). Furthermore, the ratios of W(E) ε /Wε
(M2)
are close to 1 when
observational phase shifts in W(M2)ε are similar to theoretical ones,
suggesting the reliability of the coefficients W(E)
ε and Wε
(M2)
. Therefore,
both coefficients can serve as poroelastic parameters in determining
groundwater fluctuations.

4. Poroelastic compression coefficients and aquifer rock type

We can estimate another poroelastic aquifer parameter, the

loading efficiency, which is the well response to barometric or
ocean tide loading (Jacob, 1940; Wang, 2000) using observations of
groundwater level and barometric pressure. The loading efficiency, γ,
is related to the barometric efficiency (B.E.) as

Fig. 3. (a) Correlation between the loading efficiency, γ, and the strain sensitivities,
γ = 1−B:E:; ð4Þ
Wε(M2) and (b) estimated uniaxial undrained bulk modulus, Kv(u), in each well given for
the aquifer rock type. The wells are classified by the aquifer rock types. Circles, squares,
and triangles indicate sedimentary rocks, sedimentary and volcanic rocks, and volcanic
where B.E. can be calculated using the BAYTAP-G software (Tamura rocks, respectively. The strain sensitivities of Wε(M2) at SK, SR2, YC, and YN may be
et al., 1991) with observed groundwater levels and barometric inaccurate for several reasons. The Kv(u) moduli at YN and SR2, estimated using Wε(E) are
pressures. The estimated barometric and loading efficiencies are shown in the open squares.
 T. Shibata et al. / Tectonophysics 483 (2010) 305–309 309

and YC have negative signs, so are obviously incorrect. As shown in References

Fig. 3a, where the wells are classified by aquifer rock type, the strain
Akita, F., Matsumoto, N., 2001. Coseismic groundwater level changes in hot spring wells
sensitivity tends to correlate positively with the loading efficiency. in Hokkaido induced by four earthquakes larger than M7.5 between 1993 and 1994
From the strain sensitivity and the loading efficiency, we can estimate (in Japanese with English abstract). Journal of the Seismological Society of Japan 53,
Kv(u) of each well. Although the Kv(u) values are scattered, the estimated 193–204.
Akita, F., Matsumoto, N., 2004. Hydrological responses induces by the Tokachi-oki
moduli roughly correspond to laboratory values for sandstone (13– earthquake in 2003 at hot spring wells in Hokkaido, Japan. Geophysical Research
47 GPa) and granite (61–66 GPa) in Fig. 3b (Wang, 2000). In Letters 31. doi:10.1029/2004GL020433.
particular, Kv(u) estimated by Wε(E) at SR2 and YN are within the Brodsky, E.E., Roeloffs, E.A., Woodcock, D., Gall, I., Manga, M., 2003. A mechanism for
sustained groundwater pressure changes induced by distant earthquakes. Journal
laboratory values, while Kv(u) estimated by Wε(M2) at SR2 and YN are of Geophysical Research 108. doi:10.1029/2002JB002321.
much larger than the laboratory values. This result shows that Wε(E) is Igarashi, G., Wakita, H., 1991. Tidal responses and earthquake-related changes in the
more plausible as strain sensitivity than Wε(M2) at SR2 and YN. water level of deep wells. Journal of Geophysical Research 96, 4269–4278.
Jacob, C.E., 1940. On the flow water in an elastic artesian aquifer. Transactions American
Geophysical Union 22, 783–787.
5. Summary and conclusions Japan Meteorological Agency, 2005. Seismological summary (in Japanese). Newsletter
of the Seismological Society of Japan 16, 76–85.
Coseismic changes in groundwater levels at 19 wells were Koizumi, N., Kitagawa, Y., Takahashi, M., Sato, T., Matsumoto, N., Ito, H., Kuwahara, Y.,
Cho, A., Sato, T., 2002. Changes in groundwater level and crustal strain in and
observed for six earthquakes from 1993 to 2004 in Hokkaido, Japan. around the Kinki district related to the 2001 Geiyo earthquake (in Japanese with
These coseismic changes are related to strains induced by the English abstract). Journal of the Seismological Society of Japan 55, 119–127.
earthquakes and are recognized as poroelastic responses. Poroelastic Masterlark, T., 2003. Finite element model predictions of static deformation from
dislocation sources in a subduction zone: Sensitivities to homogeneous, isotropic,
response depends not only on the stress applied to an aquifer but also Poisson-solid, and half-space assumptions. Journal of Geophysical Research 108.
on its rock type, and is described by the strain sensitivities of the doi:10.1029/2002JB002296.
groundwater level. We estimated two types of strain sensitivities, Matsumoto, K., Sato, T., Takanezawa, T., Ooe, M., 2001. GOTIC2: a program for
computation of oceanic tidal loading effect. Journal of the Geodetic Society of Japan
from groundwater responses to earthquakes and to tidal constituents, 47, 243–248.
respectively, and found that they have a linear correlation. Matsumoto, N., Kitagawa, G., Roeloffs, E.A., 2003. Hydrologic response to earthquakes in
Loading efficiencies of groundwater level at these wells were the Haibara well, central Japan: 1. Groundwater-level changes revealed using state
space decomposition of atmospheric pressure rainfall, and tidal responses.
found to have values of 0.16–0.62. Loading efficiency and strain
Geophysical Journal International 155, 885–898.
sensitivity are both related to the poroelastic bulk modulus, so we can National Astronomical Observatory of Japan, 2004. Rika Nenpyo. Maruzen Co. Ltd.,
obtain correlations between them. From the correlations, we Tokyo. 945 pp.
Okada, Y., 1992. Internal deformation due to shear and tensile faults in a half-space.
estimated poroelastic bulk moduli that correspond roughly to
Bulletin of the Seismological Society of America 82, 1018–1040.
laboratory values, suggesting that the natural fluctuation of ground- Quilty, E., Roeloffs, E.A., 1997. Water level changes in response to the December 20,
water level is constrained by linear poroelasticity. This study is a 1994 M4.7 earthquake near Parkfield, California. Bulletin of the Seismological
preliminary approach to the use of observational records and was Society of America 87, 310–317.
Roeloffs, E.A., 1996. In: Dmowska, R. (Ed.), Poroelastic techniques in the study
accomplished by compiling observed data for many wells. These of earthquake-related hydrologic phenomena. Advances in Geophysics, vol. 34.
results have an additional possible application in hydrological Academic, San Diego, pp. 135–195.
observation, another method for measuring volumetric strain, but Roeloffs, E.A., 1998. Persistent water level changes in a well near Parkfield, California,
due to local and distant earthquakes. Journal of Geophysical Research 103,
this requires verification. 869–889.
Tamura, Y., Sato, T., Ooe, M., Ishiguro, M., 1991. A procedure for tidal analysis with a
Acknowledgments Bayesian information criterion. Geophysical Journal International 104, 507–516.
Wakita, H., 1975. Water wells as possible indicators of tectonics strain. Science 189,
553–555.
We used the software package GMT (Wessel and Smith, 1998) to Wang, H.F., 2000. Theory of linear poroelasticity with applications to geomechanics and
draw some of the figures. The Department of Health and Welfare of hydrogeology. Princeton Univ. Press, Princeton. 287 pp.
Wessel, P., Smith, W.H.F., 1998. New, improved version of the Generic Mapping Tools
the Hokkaido Government and many municipalities in Hokkaido
released. EOS Transactions American Geophysical Union 79, 579.
provided records for analysis. We would like to thank Naoji Koizumi
and an anonymous reviewer for their helpful and constructive
reviews.