Download A 100 m laser strainmeter system installed in a 1 km deep tunnel at

Journal of Geodynamics 38 (2004) 477–488
A 100 m laser strainmeter system installed in a 1 km deep
tunnel at Kamioka, Gifu, Japan
Shuzo Takemotoa,∗ , Akito Arayab , Junpei Akamatsuc , Wataru Moriic , Hideo Momosea ,
Masatake Ohashid , Ichiro Kawasakic , Toshihiro Higashia , Yoichi Fukudaa ,
Shinji Miyokid , Takashi Uchiyamad , Daisuke Tatsumie , Hideo Hanadae , Isao Naitoe ,
Souichi Teladaf , Nobuo Ichikawac , Kensuke Onouec , Yasuo Wadac
a
Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan
Earthquake Research Institute, University of Tokyo, Bukyo-ku, Tokyo 113-0032, Japan
c
Disaster Prevention Research Institute, Kyoto University, Uji, Kyoto 611-0011, Japan
d
Institute for Cosmic Ray Research, University of Tokyo, Kashiwa, Chiba 277-8581, Japan
e
National Astronomical Observatory, Mitaka, Tokyo 181-8588, Japan
National Institute for Advanced Industrial Science and Technology, Tsukuba, Ibaraki 305-8568, Japan
b
f
Received 24 November 2003; received in revised form 8 March 2004; accepted 9 July 2004
Abstract
We have installed a laser strainmeter system in a deep tunnel about 1,000 m below the ground surface at Kamioka,
Gifu, Japan. The system consists of three types of independent interferometers: (1) an EW linear strainmeter of the
Michelson type with unequal arms, (2) an NS-EW differential strainmeter of the Michelson type with equal arms
and (3) a NS absolute strainmeter of the Fabry–Perot type. These are configured in L-shaped vacuum pipes, each
of which has a length of 100 m. (1) and (2) are highly sensitive (order of 10−13 strain) and have wide dynamical
range (10−13 –10−6 ). Observations with strainmeters (1) and (2) started on June 11, 2003. (3) is a new device for
absolute-length measurements of the order of 10−9 of a long-baseline (100 m) Fabry–Perot cavity by the use of
phase-modulated light. This third strainmeter will be ready for operation before the end of 2004. The laser source
of strainmeters (1) and (2) is a frequency-doubled YAG laser with a wavelength of 532 nm. The laser frequency is
locked onto an iodine absorption line and a stability of 2 × 10−13 is attained. The light paths of the laser strainmeter
system are enclosed in SUS304 stainless steel pipes. The inside pressure is kept to be 10−4 Pa. Consequently,
quantitative measurement of crustal strains of the order of 10−13 can be attained by employing the laser strainmeter
system of (1) and (2) at Kamioka. This resolving power corresponds to that of a superconducting gravimeter. Using
∗
Corresponding author. Fax: +81 75 753 3917.
E-mail address: [email protected] (S. Takemoto).
0264-3707/$ – see front matter © 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jog.2004.07.008
478
S. Takemoto et al. / Journal of Geodynamics 38 (2004) 477–488
the laser strainmeter system, we expect to determine parameters related to fluid core resonance, core modes and
core undertone as well as other geodynamic signals such as slow strain changes caused by silent earthquakes or
slow earthquakes.
© 2004 Elsevier Ltd. All rights reserved.
1. Introduction
A successful detection of intergalactic gravitational waves would open up significant fields of study in
physics and astronomy. Investigation of the nature of the waves themselves and, more importantly, observation of astronomical phenomena through this new means would produce intensive research activity. In
order to detect gravitational waves, the TAMA project started in 1995 with a 300 m interferometer in the
Mitaka campus of the National Astronomical Observatory of Japan (Tsubono and TAMA collaborations,
1997). The name “TAMA” relates to the fact that this laser interferometer was installed in the Tama area
in Tokyo. The aim of the project is to develop the advanced techniques needed for the operation of a
future km-sized interferometer with the aim of observing gravitational waves that may occur by chance
within our local group of galaxies.
The large-scale cryogenic gravitational wave telescope (LCGT), which adopts cryogenic mirrors with
a higher power laser, has the aim of constructing a km-scale gravitational wave detector in Japan to
succeed the TAMA project (Kuroda and LCGT Collaboration, 1999). The plan is to build LCGT in the
Kamioka Underground Observatory, Institute for Cosmic Ray Research, University of Tokyo. If its target
sensitivity is attained, we should be able to catch a few wave events per year.
Before the construction of LCGT, we decided to install a 100 m cryogenic laser interferometer observatory (CLIO) in the Kamioka Observatory as a proto-type of the LCGT with the support of the grant-in-aid
for Scientific Research on Priority Areas from the Ministry of Education, Culture, Sports, Science and
Technology: “New Century of Gravitational Wave Research”. CLIO will improve the performance of the
cryogenic laser interferometer for the LCGT (Ohashi et al., 2003). In parallel with the CLIO equipment,
we have been installing a laser strainmeter system for geophysical purposes.
The Kamioka Underground Observatory, which is located 1000 m underground in the Mozumi Mine
of the Kamioka Mining and Smelting Co., Gifu Prefecture, Japan, was established in 1983 for the purpose
of researching neutrinos from both supernova explosions and solar activity. The detector “KAMIOKA
Nucleon Decay Experiment” (KAMIOKANDE) was a tank which contained 3000 tonnes of pure water
and had about 1000 photomultiplier tubes (PMTs) attached to the inner surface. Based on data obtained
from KAMIOKANDE, Dr. Masatoshi Koshiba, Emeritus professor of University of Tokyo, was awarded
the 2002 Nobel Prize in Physics for “pioneering contributions to astrophysics, in particular for the detection
of cosmic neutrinos”. In 1996, observations using Super-KAMIOKA Nucleon Decay Experiment or
Neutrino Detection Experiment (Super-KAMIOKANDE) which is the next generation detector after
KAMIOKANDE were started. Super-KAMIOKANDE is a 50,000 tonnes water Cerenkov detector with
11,200 PMTs.
The geophysical group involved in this gravitational wave collaboration has for many years carried
out precise crustal strain measurements using laser strainmeters at Amagase (Takemoto, 1979), RokkoTakao (Takemoto et al., 2003a), Inuyaya (Araya et al., 2002) and Kwasan (Takemoto et al., 2003c). Based
on these experiences, we are now constructing a laser strainmeter system consisting of three types of
independent interferometers in the Kamioka Underground Observatory.
S. Takemoto et al. / Journal of Geodynamics 38 (2004) 477–488
479
Among the laser strainmeter system consisting of three laser interferometers, the EW linear strainmeter
of the Michelson type with unequal arms and the NS-EW differential strainmeter with equal arms have
a resolving power of the order of 10−13 in strain measurements. This corresponds to the resolving power
of superconducting gravimeters (SGs) in gravity measurements. Therefore, these strainmeters can share
the same observation targets as SGs.
2. Observation system
The observation site in Kamioka Underground Observatory is about 2 km from the entrance of the
Mozumi Mine of the Kamioka Mining and Smelting Co. The geographic coordinates of the observation
site are 36.42◦ N, 137.30◦ E. It lies 358 m above mean sea level and about 1000 m below the ground surface.
The surrounding geology has been identified as metamorphic Hida gneiss.
Fig. 1 shows the relative locations of the laser strainmeter system and the Super-KAMIOKANDE in
Kamioka Underground Observatory. The laser strainmeter system was installed in two L-type orthogonal
tunnels newly excavated into the metamorphic rock of Hida gneiss. The velocities of P and S waves at the
observation site were determined from in situ blast data to be 5.54 and 3.05 km/s, respectively (Takemoto
et al., 2003b). The ambient temperature in the observation room is 16.2 ± 0.1 ◦ C.
In order to reduce any disturbances caused by the refractive index changes in the optical path, all of
the interferometric optical components are enclosed in SUS304 stainless steel pipes having a diameter
of 190 mm and a thickness of 3 mm, the inner surface of which is treated with electrochemical buffing
to remove impurities producing outgas. The inside of the pipes is evacuated with turbo molecular pumps
and rotary pumps, and the inside pressure is kept to be lower than 10−4 Pa. The vacuum chambers are
softly connected to the main duct with bellows, so as to decouple any vibration and stress introduced
Fig. 1. Arrangement of the laser strainmeter system in Kamioka Underground Observatory.
480
S. Takemoto et al. / Journal of Geodynamics 38 (2004) 477–488
Fig. 2. Rough sketch of the laser strainmeter system.
through the duct. Each vacuum chamber is mounted on a granite block (80 cm × 80 cm × 60 cm), which
is fixed on basement rock separated from the concrete floor.
Fig. 2 shows the schematic view of the laser strainmeter system in Kamioka consists of three types of
laser interferometers. (1) A simple Michelson type interferometer of unequal arms installed in the EW
direction. (2) An equal arm laser interferometer detecting difference of linear strains in the NS and EW
directions. These two strainmeters (1) and (2) were installed in 2002–2003, and recording started on 11
June 2003. The third type (3) is a new device for absolute-length measurements of the order of 10−9 of a
long-baseline (100 m) Fabry–Perot cavity by the use of phase-modulated light. It will be installed in the
same tunnel before the end of 2004.
The light source of (1) and (2) is provided from a frequency-doubled Nd:YAG laser (INNOLIGHT:
PROMETHEUS 50, λ = 532 nm). The laser frequency is locked onto an iodine absorption line in saturated
absorption signals of an external iodine cell by means of two feedback loops with a piezo-electric actuator
and a thermal actuator. A stability of 2 × 10−13 is attained for the periods of 20–2000 s (Araya et al.,
2002). Laser light is introduced through a polarization-maintaining optical fiber into the Michelson
interferometers (1) and (2). In the case of (1), a short arm between the beam-splitter and the reference
corner-cube prism mounted on the same invar board is used as a reference, while a long arm is used to
measure the distance between the beam-splitter and the end corner-cube prism with a 100 m separation.
Because the incident light involves both horizontal and vertical polarizations, a quarter waveplate in the
reference arm produces a 90◦ phase lag between the polarizations, resulting in bi-directional detection of
the distance variation to be measured. A part of the beam incident on the interferometer is separated off
in order to prevent laser power fluctuations from producing errors in the signal. These optical signals are
detected by photo-detectors, and are supplied to the A/D converter with a sampling rate of 0.005 s. The
second type of strainmeter (2) uses a similar optical arrangement as (1).
Fig. 3 shows the data acquisition system consisting of a successive approximation type A/D converter
(16 bits) for taking in the outputs of interferometers and environmental observation equipment including a
barometer F4711 (Yokogawa Denshikiki Co. Ltd., accuracy: ±0.15 hPa in the range of 500–1300 hPa) and
a quartz thermometer DMT-610B (TOKYO DENPA Co. Ltd., accuracy: ±0.05 ◦ C in the range of −50 to
S. Takemoto et al. / Journal of Geodynamics 38 (2004) 477–488
481
Fig. 3. Block diagram of data acquisition system.
+150 ◦ C). Digital signals are controlled by the UNIX Work Station (Sun Blade 150) trough the IEEE488
bus interface. This machine is robust to multitasking, and is stable to long-term operation, however, it is
not usually suitable for a real-time data acquisition system. In order to solve this problem, we introduced
a time-base module which generates various kinds of clocks synchronized with the external proofreading
pulse (1 PPS), and controls the timing of sampling of the A/D converter. Thereby, the data acquisition
system attained high time accuracy. Data obtained are directly stored in a hard disk (50 GB) equipped
in the EWS with the sampling time of 0.005 s (200 Hz). A release relay adopts the role of releasing the
computer system from the power supply and resets the system when it hangs up.
The third strainmeter (3) to be installed before the end of 2004 is based on the determination of
a free spectral range (FSR) of the cavity from the frequency difference between a carrier and phasemodulation sidebands, both of which resonate in the cavity. Sensitive response of the Fabry–Perot cavity
near resonant frequencies ensures accurate determination of the FSR and thus of the absolute length of the
cavity (Araya et al., 1999). With a modulation frequency of 12 MHz of the frequency-doubled Nd:YAG
laser, the absolute cavity length can be determined with a resolving power of 0.1 ␮m in a distance of
100 m. This device will be effective to monitor the creep movements related to nearby active faults as
well as to detect “strain steps” associated with local earthquakes.
3. Background noise
With regard to the noise level observed with laser strainmeters, Berger and Levine (1974) reported
measurements of earth strains in the frequency range between 10−8 and 10−2 Hz, monitored by laser
482
S. Takemoto et al. / Journal of Geodynamics 38 (2004) 477–488
Fig. 4. Power spectral densities of background noise for Kamioka laser strainmeters (1) and (2), together with those of Gran
Sasso (GS), Pinon Flat Observatory (PFO), Poorman Mine (PM), Queensbury Tunnel (QT) (after Crescentini et al., 1997).
strainmeters at Pinon Flat Observatory (PFO) and Poorman mine (PM), both in the Western United
States. These data sets agree among each other and show a generally smooth spectrum h2 (f) ≈ f−1.8 to the
background noise for frequencies lower than 10−4 Hz.
Crescentini et al. (1997) compared the noise level of the laser strainmeter of the Gran Sasso Observatory
(GS) in Italy with those obtained from laser strainmeters at PFO and PM in USA and the Queensbury
Tunnel (QT) in the UK, and obtained the similar result to these of Berger and Levine (1974). In Fig. 4, we
added our results obtained from Kamioka laser strainmeters (1) and (2) to the figure given by Crescentini
et al. (1997).
At frequencies lower than 10−4 Hz all spectra shown in Fig. 4 are comparable. While the noise level
recorded at Kamioka is lowest in the range of 10−3 to 10−1 Hz, it is two orders of magnitude larger than
other observatories in the range of 10−1 to 100 Hz. This may be caused by the ambient microseismic
ocean noise from the Japan Sea about 40 km distant from Kamioka Observatory.
4. Seismic frequency band
In order to check the response of laser strainmeters of (1) and (2) in the seismic wave band of 10−2 to
10 Hz, we compared the seismographs obtained from laser strainmeters and the CMG-3T seismometer
which is installed in the same tunnel. While the WE component of strain seismographs can be directly
obtained from the laser strainmeter (1), the NS strain component is obtained by adding the strain records
of laser strainmeters of (1) and (2), i.e. (EW) + (NS − EW).
As an example, Figs. 5 and 6 show comparison of seismic signals observed with the laser strainmeter
system and CMG-3T seismometer during the transit of seismic waves generated by the M = 7.0 earthquake
which occurred near the Aleutian Islands on 23 June 2003. In these figures, laser strain seismographs
and pendulum seismographs show almost similar waveforms, though the directional responses of these
instruments differ from each other according to the incident angles of longitudinal or transverse waves
(Benioff, 1935). The maximum peak-to-peak amplitude of the strain signal was about 1 × 10−8 . Owing
to bidirectional signal detection, a strain amplitude of 1 × 10−8 , which corresponds to 1 ␮m in relative
displacement between 100 m distance, beyond a half-wavelength of the laser light, can be fully recorded
0
S. Takemoto et al. / Journal of Geodynamics 38 (2004) 477–488
483
Fig. 5. Comparison of seismic signals observed with laser strainmeter system and CMG-3T seismometer during the transit of
seismic waves generated by the M = 7.0 earthquake which occurred near Aleutian Islands at 21h 12min 34.2s (Japan Standard
Time, JST) on 23 June 2003 (EW component).
Fig. 6. Comparison of seismic signals observed with laser strainmeter and CMG-3T seismometer during the transit of seismic
waves generated by the M = 7.0 earthquake which occurred near the Aleutian Islands at 21h 12min 34.2s (JST: Japan Standard
Time) on June 23, 2003 (NS component).
484
S. Takemoto et al. / Journal of Geodynamics 38 (2004) 477–488
Fig. 7. Comparisons of spectra for 27 min from the P-wave arrival of the earthquake occurred near the Aleutian Islands on 23
June 2003.
without any saturation, hysteresis or steps. There is practically no limit to the range unless the strain
changes too fast to allow detection of the fringes. The P-wave arrival with 1 × 10−9 in strain amplitude
can also be clearly detected.
Comparisons of spectra of two seismographs for 27 min from the P-wave arrival are shown in Fig. 7.
We can recognize that the two seismographs show similar responses. Consequently, strain seismographs
are good quality in the seismic wave band with broad dynamic range.
5. Tidal frequency band
Figs. 8 and 9 show the observed earth strains for a period of 13 days from 12 June 2003, together with
expected tidal strain changes which are obtained from the GOTIC2 program: a program for computation
of the oceanic tidal loading effect (Matsumoto et al., 2001). In these figures, we can recognize that the
observed data are of good quality in the tidal frequency band too.
Figs. 10 and 11 show comparisons of “observed” and “theoretically expected” amplitudes of eight
major constituents of tidal strains. The “observed” amplitudes were determined by applying the tidalanalysis program: BAYTAP-G (Tamura et al., 1991) to the data of 66 days from 12 June. On the other
hand, “expected” amplitudes were obtained from the GOTIC2 program. The “observed” amplitudes are
about 10% smaller than the “expected” amplitudes. This discrepancy may be explained by effects of
local inhomogeneities around the observation site, such as topographic and cavity effects. We are now
examining this possibility by means of finite element calculations.
S. Takemoto et al. / Journal of Geodynamics 38 (2004) 477–488
485
Fig. 8. Example of record obtained from the laser strainmeter system in Kamioka together with theoretically expected tidal
strains calculated by the GOTEC2 program (EW).
Fig. 9. Example of record obtained from the laser strainmeter system in Kamioka together with theoretically expected tidal
strains calculated by the GOTEC2 program (NS).
486
S. Takemoto et al. / Journal of Geodynamics 38 (2004) 477–488
Fig. 10. Comparison of observed and theoretically expected tidal constituents (EW).
Fig. 11. Comparison of observed and theoretically expected tidal constituents (NS).
6. Discussion and concluding remarks
We have installed a highly sensitive EW linear strainmeter and a NS-EW differential strainmeter with
resolving powers of 10−13 in the Kamioka Underground Observatory. Results from preliminary analysis
show that the qualities of these data are high in the seismic wave band and the tidal band.
With regard to the data in the seismic wave band, the response of the strain seismograph differs considerably from that of the pendulum seismograph and that in consequence, a comparison of seismographs
written with the two instruments should provide information on wave characteristics which cannot be
S. Takemoto et al. / Journal of Geodynamics 38 (2004) 477–488
487
derived from observations with either instrument alone (Benioff and Gutenberg, 1951). While the pendulum type of seismograph records the velocity of seismic waves, i.e., the time derivative of ground
displacements, the strain seismograph shows the spatial derivative of ground displacements. Therefore,
it is possible to determine the phase velocity of the seismic wave passing through the observation site
by taking ratio of both seismographs (Mikumo and Aki, 1964). Although this approach is attractive,
the sensitivities of the previous instruments have not been high enough to determine phase velocities.
Now, our highly sensitive equipment with broad dynamic range could meet the expectation to repeat and
improve on this approach. Moreover, it is expected that seismic core modes will be detected using our
sensitive and stable laser strainmeters.
In the tidal band, we expect to determine parameters related to fluid core resonance and core undertone.
Also, it is likely that we will be able to detect other geodynamic signals such as slow strain changes caused
by silent earthquakes or slow earthquakes.
Another target of our laser strainmeter system is the investigation of phenomenon associated with the
Earth’s background free oscillations. It has long been thought that Earth’s free oscillations are induced
only by large earthquakes. However, based on data obtained from a superconducting gravimeter at Showa
station, Antarctica, Suda et al. (1998) showed that there is a possibility that the Earth has been oscillating
persistently without triggering of earthquakes. Atmospheric motions and wind-driven ocean waves are
considered to be the potential causes of the oscillations. We are planning to install a superconducting
gravimeter in the same tunnel in 2004. The combined use of highly sensitive laser strainmeters and the
superconducting gravimeter will be an effective tool to investigate this phenomenon. We should be able
to separate the “spheroidal modes” and “toroidal modes” of the Earth’s background free oscillations. The
amplitudes of the toroidal modes, which involve only horizontal ground motion, would impose significant
constraints on the possible source of the background free oscillations.
Acknowledgements
We wish to express our sincere thanks to Professor Yoichiro Suzuki and the staff of Kamioka Observatory, Institute for Cosmic Ray Research, University of Tokyo for their help in maintenance of the laser
strainmeter at Kamioka Underground Observatory. We are also grateful to John P. Matthews, Carla Braitenberg and an anonymous reviewer for their critical reading of the manuscript and useful comments. This
work was mainly supported by the grant-in-aid for Scientific Research on Priority Areas from the Ministry
of Education, Culture, Sports, Science and Technology (grant 14047213), and partially supported by the
grant-in-aid for Scientific Research (A) (14253004), Japan Society for Promotion of Sciences (JSPS),
and the grant-in-aid for the 21st Century COE Program (Kyoto University, G3).
References
Araya, A., Telada, S., Tochikubo, K., Taniguchi, S., Takahashi, R., Kawabe, K., Tatsumi, D., Yamazaki, T., Kawamura, S.,
Miyoki, S., Moriwaki, S., Musha, M., Nagano, S., Fujimoto, M., Horikoshi, K., Mio, N., Naito, Y., Takamori, A., Yamamoto,
K., 1999. Absolute-length determination of a long-baseline Fabry–Perot cavity by means of resonating modulation sidebands.
Appl. Opt.-OT 38 (13), 2848–2856.
Araya, A., Kunugi, T., Fukao, Y., Yamada, I., Suda, N., Maruyama, S., Mio, N., Moriwaki, S., 2002. Iodine-stabilized Nd: YAG
laser applied to a long-baseline interferometer for wideband earth strain observations. Rev. Sci. Instrum. 73 (6), 2434–2439.
488
S. Takemoto et al. / Journal of Geodynamics 38 (2004) 477–488
Berger, J., Levine, J., 1974. The spectrum of earth strain from 10−8 to 10−2 Hz. J. Geophys. Res. 79, 1210–1214.
Benioff, H., 1935. A linear strain seismograph. Bull. Seis. Soc. Am. 25, 283–309.
Benioff, H., Gutenberg, B., 1951. The response of strain and pendulum seismographs to surface waves. Bull. Seis. Soc. Am. 41,
229–237.
Crescentini, L., Amoruso, A., Fiocco, G., Visconti, G., 1997. Installation of a high-sensitivity laser strainmeter in a tunnel in
central Italy. Rev. Sci. Instrum. 68 (8), 3206–3210.
Kuroda, K. and LCGT Collaboration 1999. Large-scale cryogenic gravitational wave telescope, Int. J. Mod. Phys. D 8(5),
557–579.
Matsumoto, K., Sato, T., Takanezawa, T., Ooe, M., 2001. GOTIC2: a program for computation of oceanic tidal loading effect.
J. Geod. Soc. Jpn. 47, 243–248.
Mikumo, T., Aki, K., 1964. Determination of local phase velocity by intercomparison of seismograms from strain and pendulum
instruments. J. Geophys. Res. 69, 721–731.
Ohashi, M., Kuroda, K., Miyoki, S., Uchiyama, T., Yamamoto, K., Kasahara, K., Shintomi, T., Yamamoto, A., Haruyama, T.,
Saito, Y., Higashi, T., Suzuki, Y., Sato, N., Tomaru, T., Tatsumi, D., Telada, S., Ando, M., Araya, A., Takemoto, S., Higashi,
T., Momose, H., Akamatsu, J., Morii, W., 2003. Design and construction status of CLIO. Class Quantum Grav. 20 (17),
S599–S608.
Suda, N., Nawa, K., Fukao, Y., 1998. Earth’s background free oscillations. Science 279, 2089–2091.
Tamura, Y., Sato, T., Ooe, M., Ishiguro, M., 1991. A procedure for tidal analysis with a Bayesian information criterion. Geophys.
J. Int. 104, 507–516.
Takemoto, S., 1979. Laser interferometer systems for precise measurements of ground-strains. Bull. Disas. Prev. Res. Inst.,
Kyoto Univ. 29 (2), 65–81.
Takemoto, S., Yamamoto, T., Mukai, A., Otsuka, S., Fujimori, K., 2003a. Crustal strain observation for nine years with a laser
strainmeter in Kobe, Japan. J. Geodyn. 35, 483–498.
Takemoto, S., Araya, A., Akamatsu, J., Morii, W., Higashi, T., Fukuda, Y., Onoue, K., Ichikawa, N., Kawasaki, I., Ohashi, M.,
Telada, S., Momose, H., 2003b. Installation of 100m laser strainmeters in the Kamioka Mine. Ann. Disas. Prev. Res. Inst.,
Kyoto Univ. 46B, 749–755 (in Japanese with English abstract).
Takemoto, S., Momose, H., Fujimori, K., Higashi, T., 2003c. Crustal strain observation with a laser extensometer in a shallow
tunnel. J. Geod. Soc. Jpn. 49, 215–225 (in Japanese with English abstract).
Tsubono, K. and TAMA collaborations, 1997. TAMA Project, Proceedings of the TAMA Workshop on Gravitational Wave
Detection, Saitama. Universal Academy Press, 1996, pp. 249–257.