Mineral spring gas compositions at several sites in central Japan have been continuously monitored by our laboratory for the earthquake prediction study. In connection with it, a new system for the measurement of mineral spring gas flow rate has been installed at Byakko mineral spring, Miaunami, Gifu Prefecture, a momitoring site. The continuous observation with this system has been conducted since January, 1993. The main body of the gas flow rate meter is a vertically submerged plastic pipe of 4.1cm in diameter attached by pressure transducer. The mineral spring gases issuing from the bottom of the spring pool are collected into the pipe, the upper end of which is closed by an electric valve. As the sample gas is accumulated, the water level in the pipe becomes lower and the inner gas pressure becomes higher. After a fixed period, the inner pressure is measured with the pressure transducer and subsequently the accumulated gas is librated by opening the valve and thus the operation proceeds repeatedly for the continuous monitoring. The gas flow rate is calculated from the measured pressure. These operations including gas analyses are controlled by two personal computers. The obtained data are sent to our laboratory through the public communication line. Although observed gas flow rates considerably fluctuate, the flow rate on moving averages is positively correlated with He/Ar ratio in the gas. Thirteen earthquakes felt around the monitoring site occurred during this one year, and, at the events of the largest three seismic intensity, coseismic anomalies emerged in the gas flow rate. The feature described above suggests that the coseismic anormalies in the rate are caused by the bubbling at the earthquake tremor. On the other hand, it seems that the coseismic change of He/Ar ratio is not always caused by the bubbling. Although the cause of this change still remains unsolved, it is likely that the behavior of He/Ar ratio is controlled by the changes of the compressive stress in the crust.

View full abstract

Earthquake recurrence intervals and characterization of fault behavior provide important data for seismic risk evaluation. Historical records of destructive earthquakes are insufficient to characterize paleoseismicity of inland active faults in Japan, because most of the active faults have recurrence times of more than several thousand years. Geological and geomorphological investigation including excavation studies is indispensable for a long-term prediction of inland earthquakes. We present a study on paleoseismicity regarding the Kannonji fault, which is one of the eastern boundary faults of the Shonai plain in the northeast Japan, based on analysis of the borehole data close to the fault. Stratigraphic analysis and radiocarbon ages of the samples collected from the boreholes have revealed the three surface faulting events in the Holocene. The latest faulting event occurred after 2, 500yBP; the prior one occurred between 5, 500yBP and 4, 300yBP, probably between 4, 500yBP and 4, 300yBP. The third latest faulting event occurred around 6, 300-6, 000yBP. These data suggest that recurrence intervals of earthquakes are about 2, 000 years. According to a previous study by Suzuki and others, no surface faulting occurred at the borehole sites in association with the Shonai earthquake of 1894, although the earthquake was generated from this fault. This fact strongly indicates that other unknown large earthquakes on this fault plane possibly occurred in Holocene time causing no surface faulting. Therefore, a recurrence interval of large earthquakes from this fault is eatimated less than 2, 000 years.

View full abstract

An experiment of explosion seismic observations was conducted by Research Group for Explosion Seismology (RGES) along a north-south profile in the Kitakami region, northeastern Japan. We observed the explosions by RGES and quarry blasts along a 60km profile in the east-west direction crossing the RGES's profile in the central part of the Kitakami region. Based on travel time analysis of reflected waves, velocity boundaries in the middle and lower crust were found at the depth of 15 and 27km, and the crlustal thickness was estimated to be about 34km. The Kitakami region is divided into northern and southern terranes by Hayachine Tectonic Belt (HTB), where maficultramafic rocks intruded during the end of the Jurassic to the beginning of the Cretaceous period. The profile of this study was almost located along the southern margin of HTB. Velocity structure beneath the profile, however, is similar to that along the north-south profile obtained by RGES and shows no significant differences. Record sections from the shots far from the profile show that boundary in the lower crust and the Moho are distinct in the northern part but are vague in the southern part of the Kitakami region. The quarry blasts generated clear S-waves. Using P ans S travel times from quarry blasts, V_p/V_s in the uppermost part of the crust was estimated to be 1.74-1.75.

View full abstract

Estimation of the geological structure down to a depth of more than 1, 000m was attempted at two sites KS1 and KS2 in the Ashigara Valley in Odawara City assigned as a test field for the international cooperative research on the Effects of Surface Geology on Seismic Motion (ESG). The technique employed involves the measurement of phase velocities of Rayleigh waves in microtremors.
For the measurement, microtremors were observed at each site by seismic arrays consisting of 7 seismometers deployed within spaces from 100m to 600m in diameter. Applying the frequencywavenumber power spectrum analysis to the observed microtremor data, the phase velocities of the fundamental-mode Rayleigh waves in microtremors were successfully measured in the period range from 0.3s to 0.8s at KS1 and from 0.3s to 1.8s at KS2, respectively.
From these phase velocities, the geological structure was estimated as an S-wave velocity structure to a depth of about 1, 000m at KS1 and to a depth of about 2, 000m at KS2, respectively. Except for the upper structure to a depth of about 20m, the estimated S-wave velocity structure was in good agreement with that determined from the P- and S-wave velocity logging applied to a depth of about 100m at each site. The depth of the basement in the estimated S-wave velocity structures coincides with that in a P-wave velocity structure obtained by a seismic refraction survey, a profile of which runs near KS1 and KS2.
The observation of seismic motion had been made at KS2 on the surface of the ground and a place as deep as 100m in a borehole for the ESG symposium. The averaged spectral ratio of the seismic motions observed on the two places for 16 earthquakes have been obtained.
The period of the first and second peaks in the spectral ratio were calculated for the SH-waves vertically incident on the two places with their estimated velocity structures. The period of the first peak calculated is 1.0s, which coincides with that in the averaged spectral ratio, while the period of the second peak calculated is 0.4s, shorter than that in the averaged spectral ratio by 0.1s. This difference, however, could be well explained by slightly modifying the upper 20m part of the estimated S-wave velocity structure.
The present experiment revealed that the exploration technique using microtremors can be a powerful alternative to conventional seismic methods in estimating geological structures in urban areas.

View full abstract

The Nikko-Ashio region, northern Kanto district, Japan is known for its active seismicity of microearthquakes. Although there are some permanent seismic networks around this region, the detectability of those networks isn't enough to clarify detailed characteristics of seismic activity in this region. We have performed temporary seismic observations repeatedly with high-gain short period seismographs in this region to investigate microearthquake activity since 1980s. A continuous seismic observation was started on september 1992 with seismic network of eight telemetered stations.
According to seismic observations for more than ten years, we have obtained general properties of the microearthquake activity in the Nikko-Ashio region as follows.
There are two active seismic areas; one is the area very close to Mt. Sukai, which is located on the border of Tochigi and Gunma prefectures, extending north-south direction and the other is the area beneath Ashio town, about 10km southeast of Mt. Sukai. The seismic activities of both areas are nearly stationary for more than ten years. The lower limit of focal depths of microearthquakes is about 10km beneath Ashio town and 6km near Mt. Sukai. On the whole, the lower limit of seismicity decreases towards north up to Mt. Nikko-Shirane, an active volcano located about 12km north of Mt. Sukai. This feature forms an aseismic zone in the upper crust beneath Mt. Nikko-Shirane.
From November 1992, new microearthquake swarms began to occur by Lake Chuzenji, about 10km northeast of Mt. Sukai, and beneath Mt. Nikko-Shirane. In particular, the focal depths of some events beneath Mt. Nikko-Shirane are less than 2km. These new swarm activities followed several low-frequency microearthquakes near the Moho boundary beneath Ashio town on November 1992. Low-frequency microearthquakes are generally considered to relate to deep magmatic activity. Meanwhile, several studies suggest the existence of a clear seismic reflector which is considered as a partial melting zone in the middle crust.
These phenomena suggest that the microearthquake activity in the Nikko-Ashio region may have some relation to the volcanic or magmatic activities in the lower crust. We need to monitor the microearthquake activity in this region with the special concern about its relation to volcanic activity of Mt. Nikko-Shirane.

View full abstract

Observation of crustal deformation using Global Positioning System (GPS) has been conducted in Shikoku region, Southwest Japan, to monitor the interaction between the Eurasian and the Philippine Sea plates along the Nankai trough. Four repeated surveys conducted in 1990 to 1994 have revealed that the stations located at the coast line are displaced to the northwest to west-northwest direction relative to the inland station. This direction is parallel to the moving direction of the Philippine Sea plate and opposite to the coseismic displacement at the time of the 1946 Nankai earthquake (M=8.0). During the period of four years, the maximum rate of 35mm/yr is obtained at Muroto, the station nearest to the Nankai trough and located in the source region of the 1946 earthquake. This displacement rate amounts to three quarters of the relative velocity between two plates estimated from seismological data. This implies that the present interplate coupling along the Nankai trough off Shikoku region is strong and causes considerable deformation of the plate-convergence region. As a whole the southeastern part of Shikoku is contracted with a constant rate of about 3×10^-7/yr. Strain pattern obtained in this study is very similar to the long-term triangulation and trilateration results.

View full abstract

In 1992, the eastern part of the Nankai Trough off Tokai area, Japan, is seismically surveyed by the dense refraction profiling using an ocean bottom seismograph (OBS) array and airguns. In this area, the Philippine Sea plate is subducting under the Japan Arc along the Nankai Trough, and a remarkable topographic high, the Zenisu Ridge, exists along the southern side of the trough. The aim of this experiment was to obtain detailed seismic velocity structures of the crust at the plate boundary and to reveal the subducting pattern of the Philippine Sea plate at its north margin. We could obtain detailed velocity structure models because of the dense profiling, the good data quality and precise analysis by a ray-tracing method using both the travel-time data and the amplitude data. Results of this study are summarized as follows:
1) The eastern Nankai Trough, similar to the southwestern Nankai Trough off Shikoku, has a typical oceanic crust with the thickness of 8-10km, which varies along this profile. Igneous layers are composed of the oceanic layer 2A (P-wave velocity of 3.6-4.3km/s), 2B (4.7-5.6km/s), 3A (6.3-6.4km/s) and 3B (6.7-6.9km/s). And the velocity of the uppermost mantle is 7.8-8.0km/s.
2) The Zenisu Ridge has a similar velocity structure to that of the Nankai Trough. It shows a difference with the crustal structure of the Izu-Ogasawara Arc which is characterized by existence of the 6km/s layer. This is clear evidence that the Zenisu Ridge is a topographic high originated from the oceanic crust. Beneath the ridge, the sediment thickness is thin (0.5-2km) and the igneous basement, the oceanic layer 2, shows severe undulations as compared with the structure of the Nankai Trough.
3) Existence of a subducting oceanic crust beneath the continental slope was imaged by clearly observed wide-angle reflection phases. Our data support that the subducting oceanic crust lacks the layer 2A and the upper part of the layer 2B. It indicates a remarkable contrast to crustal structure models obtained at the southernmost Kuril Trench and the Ryukyu Trench where the oceanic crust subducts with the whole of the oceanic layer 2.
4) The velocity of the upper crust on the subducting oceanic crust was determined as 5.3-5.8km/s, which is relatively low as compared with that of the Japan Arc. The continental slope is characterized by an accretional prism composed of a thick sediment layer of 3.0-3.5km/s and a wedge shaped layer of 3.8-4.2km/s. The wedge seems to contain material of the upper part of the oceanic layer 2 which was scraped off during the subduction because of the similarity of velocities between them.
5) Beneath the south margin of the Zenisu Ridge, the oceanic layer 3 was found to have a discontinuity in layering. Furthermore, the layer and the Moho show a sudden dip to the north in the south side of the discontinuity. Although the feature of the structure indicates a possible subduction of the oceanic crust beneath the ridge, the extent of the dipping layer just reaches beneath the south margin of the ridge. This result may suggest that a new subduction of the Philippine Sea plate has started there.

View full abstract

In order to deepen understanding of the characteristics of near-field ground motion, case histories of upthrow of objects during 21 earthquakes are reviewed. The fault distance to the site where the upthrow was observed becomes larger with increase of the earthquake magnitude. The area where the upthrow was observed roughly corresponds to that of the J. M. A. intensity VII which is X or greater in the M. M. scale. The upthrown objects are boulders, stone mouments, human bodies, small structures and wooden houses, with limitation in size. In case of the boulder, one on a shallow socket of surface soil and with the diameter of approximately 50cm tends to have a large displacement. These facts may suggest the soil-object interaction system with a limited vibration period causes a large response of the object and the resulting upthrow under strong shaking of the M. M intensity X or greater.

View full abstract