We have developed multi-component borehole instruments for borehole stations deeper than 1000m. The instruments are composed of 7 strainmeter components (4 horizontal, 2 inclined, and 1 vertical), 2 tiltmeters, 3 seismometers, 4 magnetometers, and a high-resolution thermometer. The instruments are also equipped with new systems for data transmission and data monitoring, even during installation, and for determining the instrument direction at the bottom of the borehole. We have also developed an intelligent-type strainmeter based on the Ishii-type strainmeter for measuring in situ stresses. It is equipped with strain sensors, an A/D converter, a CPU, a memory, and a battery, and it has no outside cable. It is cemented into a deep borehole with expansion grout and then taken out by overcoring after it has coupled with the basement rock. By this procedure we can estimate in situ rock stresses. Observation in deep boreholes can avoid the problems of both artificial noise and meteorological disturbance. It enables the performance of high S/N ratio observations for detecting very small signals. The results obtained from deep borehole observations on the Izu Peninsula and in the Tono area of Gifu Prefecture have shown some interesting variations of strains and tilts related to the precursory phenomena of earthquakes. An example of in situ stress determination has also been demonstrated. In the case of deep borehole observation, we first measure the in situ stresses and then install a multi-component borehole instrument to monitor crustal activity. By this process we can continuously monitor stress variation. This kind of stress monitoring is very important in earthquake prediction research.
We evaluated local site effects and Qs-values in the western part of Kanagawa Prefecture, Japan by a spectral inversion method using the ground motion data from nine middle class events that occurred near the target area. The results were as follows: (1) Qs-value was approximated to be Qs=20×f in the frequency range of 1-20Hz by assuming the geometrical spreading factor to be r-1. (2) The seismic moments evaluated in this study were slightly larger than those of previous studies. (3) The stations in the southern part of Ashigara valley showed high amplification at around 1Hz. (4) The relative site factors in reference to KNO coincided with the theoretical amplification factors calculated for one-dimensional subsurface structure models with SH-wave input motion from the layer of Vs=1200m/s.
We conducted a seismic reflection survey along the coast line of the eastern part of the Kanto Plain, central Japan. Four reflectors were clearly recognized in the reflection profile: Reflector A is nearly flat (about 0.1-0.25s in 2-way time); Reflector B is gently tilted to the south at about 0.25-0.5s; Reflector C is waved around the 0.5-1.0s at the northern and southern parts of the profile; Reflector D is hugely waved deeper than 1.5s. With respect to borehole data adjacent to the seismic line, the strata between Reflectors C and D can be correlated with the early middle Miocene half-graben deposits distributed along the Tanagura Tectonic Line. It suggests that the southern trace of the Tanagura Tectonic Line should be concealed at just east of the present seismic line.
Three damaging inland earthquakes have occurred in the northern Miyagi Prefecture of the Tohoku District in Japan since 1900. Magnitudes M of the 1962 and the 2003 events were assigned to be 6.5 and 6.4, respectively, from local seismic records observed by the Japan Meteorological Agency (JMA). That of the 1900 event was 7.0, which was determined mainly from the seismic intensity data, while the rank of damage was very lower than other M=7 class inland shallow earthquakes in Japan. In the present study, damage rate data, seismic intensity data, and old seismograms were re-examined to evaluate the magnitude and the location of the focal region of the 1900 event. Re-evaluated magnitude of this event is almost the same as those of the 1962 and 2003 events. The seismic gap between the 1962 and 2003 events is filled with the focal region of the 1900 event obtained. The possibility of a big shallow earthquake occurrence must be very low in near future in the seismic zone of the northern Miyagi Prefecture.
Peculiar seismic activities are occurring beneath Lake Hamana inside the subducted Philippine Sea slab at depths of about 30km. They consist of three spindle-shaped earthquake clusters with NW-SE axes, in a left-stepping alignment in the EW direction, with several kilometers between them. Illustrations of the focal mechanisms indicate that they act like an open crack under NW-SE compression and NE-SW tension and are caused by a right-lateral shear force acting on the Philippine Sea slab. A stress pattern simulation model suggests the following explanation of the situation. A localized locked zone is positioned on the plate boundary just north of the clusters. Separated from the main locked zone, it is considered to be one of the satellite asperities surrounding the main one. The main locked zone is located eastward from Lake Hamana and is expected to become the seismogenic zone of the forthcoming Tokai earthquake. These clusters have demonstrated a remarkable decrease in activity since the second half of the year 2000. An anomalous tectonic movement detected by GPS measurements occurred almost simultaneously. This indicates that a slow slip event progressed on the plate boundary beneath Lake Hamana; that is, the locking must be released there. Since the current change in seismic activity corresponds with this movement, it can be attributed to tectonic stress change due to the slow slip. We estimate that at least three similar periods of quiescence have occurred during the last quarter century. Tidal gauge findings at Maisaka and the crustal tilt at Mikkabi, both of which were observed near Lake Hamana, have indicated almost simultaneous occurrences of similar anomalies. As a result, three episodes of slow slip were identified: the first occurred before 1980, the second around 1990, and the last has been ongoing since late 2000. This implies that the slow slip repeats quasi-periodically with an interval of about one decade. We consider that the locked zone beneath Lake Hamana is a small asperity with a potential of slowly and intermittently slipping due to a weak coupling condition in an area sandwiched between two seismogenic zones of the Tokai and Tonankai megathrust earthquakes.