Strong pulse-like ground motion has been often observed during the earthquake near the fault, and considered to cause severe damage. In the present paper, the physical mechanism for strong pulse near the fault is investigated. Ground velocity waveform near the fault is calculated using the nearfield displacement waveform expressions generated from slip on the fault [AKI and RICHARDS (1980)] assuming slip rate time function to be a slightly modified Kostrov's type time function that has a linearly increasing initial part with duration of Td simulating the effect of slip weakening frictional model. We found that the nearfield and intermediate terms are canceled out, and the farfield term is predominant in ground velocity waveform near the fault. The strong pulse in velocity waveform is generated mainly by the rupture propagation of an equivalent line source and slip acceleration on the fault. If Td is short enough and the slip velocity function is impulsive, the strong pulse near the fault is generated mainly by the rupture propagation effect and the pulse width is expressed as L(1/ν-1/β), where L is the length of the equivalent line source, i. e., the line segment inside the fault which connects from the initial point of rupture to the observer, and ν is the rupture velocity, and β is the shear wave velocity. When Td becomes longer corresponding to the faulting process for large critical displacement of slip weakening friction law, the pulse width becomes longer than L(1/ν-1/β).
The Euler vector of the Philippine Sea plate relative to the Eurasian stable craton was estimated using the Global Positioning System observations. For this purpose, we used repeated observation data at Okino Torishima (Parece Vera) and velocity vectors at some sites on the Philippine Sea plate in the nationwide continuous monitoring network by the Geographical Survey Institute. First, GPS data at Okino Torishima for the period of 1995-1997 were reanalysed taking its reference at Tsukuba IGS site (TSKB). Then, the obtained velocity was converted taking local velocity of Tsukuba relative to the stable Eurasian plate estimated by HEKI (1996) into account. Obtained velocity vector was then combined with Minami Daito-jima, Chichijima and Hachijo-jima velocity data to estimate the Euler vector using the linear least squares method. The estimated location of the Euler pole was (41.55±0.42°N, 152.46±0.57°E) and the anguler velocity was -1.50±0.04deg/my taking counterclockwise rotation as positive. Based on this estimated Euler vector, we discussed the tectonics along the Suruga trough and Ryukyu trench region as the followings: (1) Motions of northern Izu-Bonin islands such as Izu Oshima and Miyake-jima that resembles the plate motion are significantly affected by local motions which presumably be volcanic origin, (2) Directions of slip vectors for the interplate large earthquakes [SENO et al. (1993)] and the azimuth of subducting plate motion show significant offsets of about 10 degrees at the Nankai trough region. This offset might be due to the eastward motion of the Eurasian continent. (3) At the southwestern islands region, rifting of the Okinawa trough is predominant and the Okinawa ridge moves almost toward the opposite direction of the subduction. (4) Two small Taiwanese islands east of Taiwan mainland show significantly slow displacement rates as estimated from the rigid motion model, which might be due to compressional force due to continental collision at Taiwan island.
The northern peripheral area of the Osaka basin is the most complicated in geological structure in the basin. It is featured by three geological structures; Arima-Takatsuki tectonic line sharply dividing the Hokusetsu mountains and the Osaka basin, the Senri hills located 2 or 3km south of the line, and the graben structure between them. To incorporate the subsurface structure in this area into a 3 dimensional model of the Osaka basin for the simulation of strong ground motions, we conducted a survey of the subsurface structure below this area using the reflection method and the microtremor array experiment. As a result, we revealed a shape of the basin-bedrock interface and S wave velocity structure in the sediments. Main results are summarized as follows. (1) Arima-Takatsuki tectonic line has a normal-fault component. (2) The depth of the graben structure is more than 1km. (3) The sediments of the Senri hills gradually increase their thickness toward the south. (4) The basement rock in the east of Butsunenjiyama fault raises up at least 0.5km relative to the basement rock in the other side.
An inversion method to separate effects of source, propagation path, and local site from S-wave spectra is applied to earthquake observation data obtained in the southwestern Kanto plain, Japan. First, effects of building in which seismometer is installed are evaluated from spectral ratio of microtremor records that were simultaneously obtained in the building and at the free surface in its vicinity for correcting the S-wave spectra. Theoretically calculated amplification factor due to shallow weathered layers and the spectral ratio of microtremors at a rock site are used as a constraint condition in our inversion. Estimated source spectra show a ω-2 decay with increasing frequency. Qs-value along the propagation path in the Earth's crust is in an agreement with results of previous studies. Strong site-dependency was found in the separated site amplification factors. In particular, soft soil site has an amplitude decrease in frequencies higher than 2Hz. The amplification factors were well explained by theoretical amplifications of reverberating S-waves in subsurface structural models from geophysical surveys. These results suggest the appropriateness of the constraint condition used in our inversion.
The generating frequency of the North-West American tsunamis is relatively lower than that of the South American region, but there are historical records of a large tsunami accompaning with the January 1700 earthquake (M 9) in the Cascadia subduction zone (SATAKE et al., 1996). In the present paper, tsunami magnitudes on the Imamura-Iida scale, m, are investigated by using the diagram of wave-height attenuation with distance. The regional characteristics of tsunami magnitudes are discussed in relation to earthquake magnitudes, Ms, during the period from 1899 to 1997. The tsunami magnitudes in the South-East Alaska to Canada region are nearly normal compared to earthquakes with similar size in the other Pacific regions, and the 1899 Yakutat tsunami being m=3 is the largest. The magnitude values in the California region are mostly m=0 or less (amplitude: 50-100cm), but those of a few tsunamis vary by the faulting mechanism. For example, the magnitude value of the 1906 San Francisco tsunami accompaning with a strike-slip earthquake (Ms=8.3) is m=-4. On the contrary, that of the 1927 Lompoc tsunami caused by a high-angle thrust earthquake (Ms=7.0) is m=1, and this tsunami was observed in Hawaii and Japan. According to the epicenter distribution of the earthquakes (Ms≥6.5) since 1812, a seismic gap exists at the segment of 700km off the Washington to Oregon states. It should be considered a region of relatively high tsunami risk.
According to SHIMAZAKI (1986), the large and small Japanese intraplate earthquakes obey the different scaling laws: M0∝L3 for small events but M0∝L2 for large events, where L and M0 are fault length and seismic moment, respectively. This is caused by the fact that the fault widths W for the large events are bounded by the thickness of the seismogenic layer in the crust. We examined the relations among source parameters for 33 Japanese intraplate earthquakes from 1885 to 1995 and confirmed the validity of the results obtained by SHIMAZAKI (1986) for the events of M=5 to 8, where M is the magnitude in the scale of the Japan Meteorological Agency (JMA). Relations between source parameters and JMA magnitude M were also derived from the relations among source parameters, using the M0-M relation by TAKEMURA (1990). SHIMAZAKI (1986) also indicated the offset of the L-M0 relation at the transition between small and large earthquakes, and suggested that the offset appeared to be due to the difference in boundary conditions between buried and surface faults. We found an offset from 6.5 to 6.8 in the JMA magnitude M, as well as the offsets of a factor of about 2 in D and M0, but no offset in L and W, where D is the average slip of the fault. Also we found that almost all events with M≥6.8 accompanied the surface faults, while most of the events with M≤6.5 did not accompany any surface fault. These results strongly supported that the offsets in D, M0, and M were caused by the surface fault breaks for the large earthquakes. Furthermore, we examined the relation between the damages from the Japanese intraplate earthquake and its JMA magnitude M. The damages suddenly increased from M=6.5 to M=6.8. The scaling law obtained above gave the large earthquake a strip fault whose location was very shallow. Because of these conditions, the intraplate earthquakes with M≥6.8 bring about strong ground motions in the wide area.