Development of a borehole stress meter for studying earthquake predictions and rock mechanics, and stress seismograms of the 2011 Tohoku earthquake (M 9.0)
© Ishii and Asai; licensee Springer. 2015
Received: 21 August 2014
Accepted: 26 January 2015
Published: 24 February 2015
Although precursory signs of an earthquake can occur before the event, it is difficult to observe such signs with precision, especially on earth’s surface where artificial noise and other factors complicate signal detection. One possible solution to this problem is to install monitoring instruments into the deep bedrock where earthquakes are likely to begin. When evaluating earthquake occurrence, it is necessary to elucidate the processes of stress accumulation in a medium and then release as a fault (crack) is generated, and to do so, the stress must be observed continuously. However, continuous observations of stress have not been implemented yet for earthquake monitoring programs. Strain is a secondary physical quantity whose variation varies depending on the elastic coefficient of the medium, and it can yield potentially valuable information as well. This article describes the development of a borehole stress meter that is capable of recording both continuous stress and strain at a depth of about 1 km. Specifically, this paper introduces the design principles of the stress meter as well as its actual structure. It also describes a newly developed calibration procedure and the results obtained to date for stress and strain studies of deep boreholes at three locations in Japan. To show examples of the observations, records of stress seismic waveforms generated by the 2011 Tohoku earthquake (M 9.0) are presented. The results demonstrate that the stress meter data have sufficient precision and reliability.
Observing precursory phenomena of earthquakes is difficult; however, a case in which precursory fluctuations in strain and inclination were actually observed does exist, although the observations were discovered after the earthquake. This case was related to the swarm earthquakes that often occur in the Izu Peninsula, which is located approximately 100 km southwest of Tokyo (Ishii et al. 1998; Okada et al. 2000; Ishii et al. 2002). The eastern part of the Izu Peninsula is being uplifted at an annual rate of about 2 cm/year. One of the borehole meters that we developed is installed in a deep borehole in the eastern Izu Peninsula. This meter recorded precursory inclination and strain fluctuations before the occurrence of the swarm earthquakes. It was discovered that the precursory phenomena were indicative of tectonic effects and that multi-component observations in deep boreholes is an effective technique for evaluating earthquake occurrences.
Several types of strain observation meters are used in boreholes. To date, volumetric strain meters (Sacks et al. 1971) have been installed at more than 30 locations in the Kanto and Tokai regions by the Japan Meteorological Agency. This type of meter can be used to observe changes in volumetric strain only. The principle of the meter is to convert the movement of silicone oil sealed inside a cylindrical vessel into changes in strain. Sakata (1981) developed a three-component strain meter by applying a Sacks-like principle. These meters are sensitive to temperature changes as the silicone oil is sealed inside the container. Gladwin (1984) developed a three-component strain meter based on the principle of measuring changes in the diameter of a cylindrical vessel by using a capacitance conversion sensor.
Ishii et al. (1997a, 2002) developed a strain meter that had improved sensitivity as a result of mechanically expanding the deformation of the cylindrical vessel. We have developed a comprehensive borehole crustal activity observation device (comprehensive observation device) equipped with strain meters and also tiltmeters, seismometers, and other components that are all co-located in one borehole container. We developed this comprehensive device because we think it is important to obtain as much information as possible to further advance the field of earthquake prediction research.
The comprehensive observation device developed by our institute has been adopted for use by the Japan Meteorological Agency, the National Institute of Advanced Industrial Science and Technology, and others. Such devices are installed in many deep boreholes and are being used to accumulate data. Strain was being analyzed from observational records of the strain meter at the Byobusan borehole observation point (depth of 1,020 m), which is managed by the Tono Research Institute of Earthquake Science when the Kii Peninsula earthquake (M 7.1 and M 7.4) occurred in the southeastern offshore on 5 September 2004 (Okubo et al. 2004; Asai et al. 2005) and when the Sumatra-Andaman earthquake (M 9.1) occurred on 26 December 2004 (Okubo et al. 2005). Crustal deformations concurrent with deep low-frequency tremors have also been observed with the strain meters installed in the deep boreholes, and additional strain analysis studies are being conducted (Itaba et al. 2010).
The Tohoku earthquake (M 9.0) occurred on 11 March 2011, and there are a number of excellent studies about the earthquake (e.g., Nishimura et al. 2011; Yoshida et al. 2011). Outcomes regarding strain seismic motions and crustal deformations concurrent with the 2011 Tohoku earthquake (M 9.0) have been delivered from observational data of meters installed at many deep borehole observation points (Itaba et al. 2011). Results have also been reported based on observations from stress meters installed at deep borehole observation points (Ishii and Asai 2011).
However, if we want to study earthquake occurrences comprehensively, it would be best to examine the processes of stress accumulation and subsequent release as a fault is generated. It is assumed that earthquakes occur when the stress applied at the fault surface exceeds frictional strength. Thus, the final objective in earthquake prediction research is to monitor the stress distribution and fault strength in earthquake-prone areas over time and space (Ishii et al. 1997b; Iio 2009).
At present, stress measurements are conducted using hydraulic fracturing, stress release methods, and so forth. However, these can only be used to determine the stress value at the point of measurement and cannot be used for continuous observations (Sano 2005). Consequently, we developed a meter capable of continuously observing the stress change. This stress meter can observe strain as well as stress. This article describes the stress meter system and reports case examples of observational results.
Principle and design of the stress meter
At present, the borehole-type strain meters developed by Ishii et al. (2002) are installed for earthquake prediction research at many deep borehole observation sites in Japan. Borehole-type strain meters detect fluctuations in the strain of bedrock (extension and contraction of the bedrock) continuously over long periods of time after they are set within a borehole and integrated with the bedrock by expansible grout with an expansion coefficient of about 0.15%. With conventional technology, strain in the bedrock is detected based on the deformation in the cylindrical case.
The most important factor to be measured when observing the elastoplastic deformation of bedrock for earthquake prediction research is stress, which is directly related to the occurrence of earthquakes. However, conventional technology detects strain in bedrock, and the only way to obtain the continuous changes in stress is to calculate these based on continuous observations of bedrock strain through complex operations using the pre-measured elastic coefficient for the bedrock. That is, stress in bedrock can be obtained only indirectly using conventional technology. To date, no equipment that can directly and continuously detect both stress and strain in bedrock independently over a long period of time has been developed.
Although the case deforms in response to stress or strain from the bedrock, the displacement of the pressure-receiving surface is not affected by the deformation of the case because the pressure-receiving surface material is not mechanically connected to the case, which makes it possible to detect both stress and strain in the bedrock accurately. The two pressure-receiving surfaces are connected to the ring-shaped connection component, which undergoes elastic deformation when subjected to stress. When the stress applied to the pressure-receiving surfaces is removed, the connection component is restored to its original shape and the displacement between the two pressure-receiving surfaces also returns to the value that existed before the surfaces were subjected to stress from the bedrock. The pressure-receiving surface material is installed at several points in different directions against the case, so it is possible to detect stress from the bedrock or strain in the bedrock applied in any axial direction independently.
The stress applied to the two pressure-receiving surfaces from the bedrock is magnified by the stress magnification mechanism, which successively magnifies using the principle of leverage, and the displacement at the final end of the lever is detected by the displacement detection sensor (Ishii et al. 2002). This magnification mechanism is usually used with a magnification rate of approximately 40 times. Thus, even small displacements between the two pressure-receiving surfaces are magnified by this mechanism, and low levels of stress applied to the pressure-receiving surfaces from the bedrock or strain in the bedrock can be detected.
In normal strain meters, the overall deformation of the case becomes correlated and it may be difficult to accurately and independently detect strain in the bedrock applied against the case from a specified direction.
The developed observation instrument
For each component of the stress meter, a mechanism capable of mechanical zero adjustment is installed as necessary. This mechanism adjusts by manually operating a built-in motor, and it can be used to reset the meter in case of a large secular variation.
Examination of stress meter sensitivity
Calibration table for both stress and strain of the meter
Stress (MPa) = −A × output (V) + B
Strain = A × output (V) × 10−5 + B × 10−7
Examination of strain meter sensitivity
Results and discussion
Stress seismograms of the 11 March 2011 Tohoku earthquake (M 9.0) and stress tidal records
Stress and strain waveforms
Comparison of stress seismic motion with water pressure seismic motion
Stress changes caused by spring water
Case examples of a comparison between vertical stress seismic motion and water pressure seismic motion as well as concurrent changes with spring water were shown above. The structure of the meter is identical for both the horizontal stress component and the vertical stress component. Based on these results, it can be surmised that the stress meter properly records changes in stress. The reliability of the horizontal stress component can be examined by calculating the invariant for stress. This matter will be discussed in a different report.
Considerations on the reliability of the developed stress meter
Here, we briefly describe the installation method of the borehole stress meter before discussing its reliability. The instrument is fixed inside a deep borehole with expansible grout. A rock sample of predetermined depth is extracted by overcoring before installing the instrument. Then, the borehole is examined to see if there are any cracks or water paths. If there are any, a deeper borehole is dug. Finally, the instrument is installed in uniform bedrock.
Check if the observed stress data satisfy the elastic invariants.
Perform a comparison between the observed stress tidal variation and theoretical tidal variation.
Perform a comparison between the observed stress waveform for the Tohoku earthquake and the theoretical stress waveform.
According to our preliminary investigation, we have obtained good results, and we will soon submit a more comprehensive paper on the reliability of the meter. Here, we just present data on the reliability of the developed stress meter that are clear at the present moment.
Only the pressure-receiving system of the stress meter detects stress and strain. We will now describe the reason why the cylindrical case does not affect the detection. The system is composed of two pressure-receiving surfaces, a frame, a magnifying device, and a displacement sensor. This pressure-receiving system is on the axis orthogonal to the case axis, and it is exposed at the outer wall. The system is thus movable in the axial direction without a mechanical connection as an O-ring is installed in the space next to the case. The device also incorporates a displacement detection sensor to detect stress and strain in the bedrock based on the displacement between the two pressure-receiving surfaces. Therefore, the cylindrical case does not at all affect the detection of stress and strain. The pressure-receiving system only contributes to the detection of stress and strain. The device detects pressures (stresses) and displacements (strains) applied only to the pressure-receiving surfaces.
In conventional strain meters, the overall deformation of the case caused by cracks, water paths, and so on becomes correlated, and it may be difficult to accurately and independently detect strain in the bedrock applied against the case from a specified direction. These types of problems have been solved by the stress meter.
As examples of normal stress observations, we present Figures 16 and 17; these data were obtained at borehole station STG200. The area where the station is located is considered to be poroelastic. Figure 15 shows a schematic depiction of the arrangement of the stress meter and water pressure meter, taking into account the poroelastic medium. The waveforms recorded by the water pressure meter and stress meter in borehole station STG200 for the 2011 Tohoku earthquake (M 9.0) are compared in Figure 16. The vertical component of stress seismic waveforms (red line) and water pressure seismic waveforms (black line) are plotted. It is apparent that both waveforms are very similar from the time of wave arrival to the later phases of the waves.
We can interpret this observation as follows. The waves generated by the earthquake arrive at the station and cause stress variation in the elastic part of the poroelastic medium. Stress variation in the elastic part produces a pressure change of pore water in the poroelastic medium, so the water pressure meter records the seismic waveforms. Therefore, the water pressure seismic waveforms are delayed by about 0.35 s behind the arrival of the stress seismic waveforms.
Figure 17 shows a case where spring water was generated by drilling at a site approximately 300 m away from the water pressure meter (Figure 15). It is thought that a water vein connects the 300-m point and the observation point and that the changes appeared as a drop in pore water pressure as a result of the spring water runoff. A drop in pore water pressure causes extensional changes in the vertical stress in the elastic part of the poroelastic medium, and the stress meter detects extensional changes in vertical stress in the bedrock. Both amplitudes were almost equal. These examples indicate that the stress meter performed well in recording the stress variation.
There are two deep borehole stations, TOS and HYS, in addition to borehole station STG200 that are equipped with the stress meter. There are also six borehole stations equipped with conventional strain meters. The tidal strain variations observed at these sites by the stress meters are similar to the tidal variations observed by conventional strain meters. These observations also indicate that the stress meters are performing well.
Based on the discussions above, we believe that the stress meters are performing well in recording both stress and strain variations.
Although we believe that precursory signs of an earthquake can occur before the event, it is difficult to observe such evidence with precision. This is especially the case when observations are made on the earth’s surface, where the precursory signals of earthquakes can be disturbed by artificial noise and other factors. One possible solution to this problem is to make observations with meters installed in the deep bedrock where earthquakes occur.
If continuous observational data can be obtained for stress, we gain access to new information that previously could not be obtained for studying earthquake occurrence and mechanisms and further advances in earthquake research can be expected. However, no instrument capable of continuous observations of stress had been developed. We therefore developed a new borehole stress meter capable of continuous observations of stress change. This stress meter can record continuous observational data for both stress and strain.
This report introduced the design principles for the stress meter and showed the results of sensitivity examinations for stress and strain. At present, such meters are installed and accumulating data in deep boreholes at three observation sites. We also studied the reliability of the stress meter that we manufactured.
Favorable correspondence in the data was found during comparisons of stress meter data and water pressure meter data. As a case example, stress seismic waveforms generated by the 2011 Tohoku earthquake were presented. The results showed that the stress meter that we have developed delivers data with sufficient precision and reliability.
The authors thank Mr. Hideo Sugaya of Techno Sugaya Co., Ltd. for discussions and comments in regard to developing the stress meter. We also thank Mr. Shigeo Matsumoto for providing useful suggestions throughout the development and installation process. Lastly, we are grateful to the MACOME Corporation, Sumiko Resources Exploration Development Co. Ltd., and System Giden Co. Ltd. for their cooperation and support.
- Asai Y, Okubo M, Ishii H, Aoki H, Yamauchi T, Kitagawa Y, Koizumi N (2005) Co-seismic strain-steps associated with the 2004 off the Kii peninsula earthquakes - observed with Ishii-type borehole strainmeters and quartz-tube extensometers. Earth Planets Space 57:309–314View ArticleGoogle Scholar
- Gladwin MT (1984) High-precision multicomponent borehole deformation monitoring. Rev Sci Instrum 55:2011–2016View ArticleGoogle Scholar
- Iio Y (2009) Why inland earthquakes happen? Kinmirai company, Nagoya, Japan, pp 1–174 (in Japanese)Google Scholar
- Ishii H, Asai Y (2011) Development of borehole stress meter and observed stress data, Poster presentation at: European Geosciences Union, General Assembly, Wien, AustriaGoogle Scholar
- Ishii H, Yamauchi T, Kusumoto F (1997a) Development of high sensitivity borehole strain meters and application for rock mechanics and earthquake prediction study. In: Rock Stress. Balkema, The Netherlands, pp 253–258Google Scholar
- Ishii H, Chen G, Ohnishi Y (1997b) Estimation of far-field stresses from borehole strain meter observations. In: Rock Stress. Balkema, The Netherlands, pp 259–264Google Scholar
- Ishii H, Nakao S, Matsumoto S, Hirata Y (1998) Precursory crustal movements observed by multi-component borehole instrument in the case of earthquake swarms off east coast of Izu peninsula. In: Proceedings of Continuous Crustal Movement Observation in Deep Underground., pp 47–51, in JapaneseGoogle Scholar
- Ishii H, Yamauchi T, Matsumoto S, Hirata Y, Nakao S (2002) Development of multi-component borehole instrument for earthquake prediction study: some observed example of precursory and co-seismic phenomena relating to earthquake swarms and application of the instrument for rock mechanics. Seismogenic Process Monitoring, Balkema, The Netherlands, pp 365–377Google Scholar
- Itaba S, Koizumi N, Matsumoto N, Ohtani R (2010) Continuous observation of groundwater and crustal deformation for forecasting Tonankai and Nankai earthquakes in Japan. Pure Appl Geophys 167:1105–1114, Doi:10.1007/s00024-010-0095-zView ArticleGoogle Scholar
- Itaba S, Matsumoto N, Kitagawa Y, Koizumi N (2011) Strain changes after the 2011 off the Pacific coast of Tohoku earthquake (Mw 9.0) observed by a distant borehole strainmeter array. Poster presented at the 2011 European geosciences union, general assembly, Wien, Austria, 8 April 2011Google Scholar
- Nishimura T, Munekane H, Yarai H (2011) The 2011 off the Pacific coast of Tohoku Earthquake and its aftershocks observed by GEONET. Earth Planets Space 63:631–636View ArticleGoogle Scholar
- Okada Y, Yamamoto E, Ohkubo T (2000) Coswarm and preswarm crustal deformation in the eastern Izu Peninsula, central Japan. J Geophys Res 105:681–692View ArticleGoogle Scholar
- Okubo M, Asai Y, Ishii H (2004) The seismological and geodetical roles of strain seismogram suggested from the 2004 off the Kii peninsula earthquakes. Earth Planets Space 57:303–308View ArticleGoogle Scholar
- Okubo M, Asai Y, Ishii H, Aoki H (2005) Thousand kilometers breaking of the 2004 Sumatra earthquake proved by dynamic strain analysis. Rep Tono Res Institute of Earthquake Sci 16:81–83Google Scholar
- Sacks IS, Suyehiro S, Evertson DW, Yamagishi Y (1971) Sacks-Evertson strain meter, its installation in Japan and some preliminary information concerning strain steps. Papers Meteoro Geophys 22:195–208Google Scholar
- Sakata S (1981) On the concept of some newly-invented borehole three-component strainmeters. Rep Nat Res Center Disaster Prev 25:95–126 (in Japanese with English abstract)Google Scholar
- Sano O (2005) Measurement methods of contemporary stress field and rock mechanics. J Geogr 114(6):1003–1021 (in Japanese with English abstract)View ArticleGoogle Scholar
- Yoshida Y, Ueno H, Muto D, Aoki S (2011) Source process of the 2011 off the Pacific coast of Tohoku Earthquake with the combination of teleseismic and strong motion data. Earth Planets Space 63:565–569View ArticleGoogle Scholar
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.