We estimated three-dimensional seismic velocity structure in and around the Philippine Sea plate subducting beneath the Kanto district in Japan by applying the double-difference tomography method to arrival time data of earthquakes obtained by the dense nationwide seismic network (Kiban-network). A low S-wave velocity and high Vp/Vs layer with several-km thick, which is shallowly inclined toward the subducting direction of the slab, has been clearly imaged. Comparison with the location of the upper surface of the Philippine Sea slab estimated from seismic refraction surveys, hypocenter distribution of relocated earthquakes, and focal mechanisms shows that this low S-wave velocity and high Vp/Vs layer corresponds to the crust of the Philippine Sea slab. Based on the presently obtained location of the low S-wave velocity and high Vp/Vs layer, hypocenter distribution, and focal mechanisms, we estimated the configuration of the upper surface of the Philippine Sea slab in the Kanto district. Presently estimated configuration of the Philippine Sea slab shows that the slab bends concavely and the depression is located eastward compared with those of the previous studies. Most of the earthquakes associated with the Philippine Sea slab occur along the plate boundary and/ or around the slab Moho. Prominent low S-wave velocity and high Vp/Vs layer was detected at depth of 30 km beneath the region along the latitude of 35.8 degrees, suggesting the serpentinization of the forearc mantle wedge due to dehydration of subducting slab.
The Tokachi-oki earthquake (M=7.9) of May 16, 1968 and the Sanriku-Haruka-oki earthquake (M=7.6) of December 28, 1994 have a partially common source area and the 1968 event consists of two asperities, southern one of which took a role of asperity once again for the 1994 event. Short-period seismic wave radiation zones (SPRZs) have been evaluated for both the events from the inversion analysis of seismic intensity data. We also identified their energy centroids of all the SPRZs. It was concluded that the every centroid of SPRZ was located at the edge of corresponding asperities of the 1968 and the 1994 earthquakes in the forward direction of fault rupture. One SPRZ was located near the terminal slip points in the southern asperity for the 1968 event and was active recurrently for the 1994 event as well as asperity. We calculated magnitude m for both the SPRZs for the 1968 and the 1994 event. m from the southern SPRZ of the 1968 event was larger than that from the SPRZ of the 1994 event. This suggested that the short-period seismic wave excitation from the asperity of the multiple-segment rupture event is larger than that from the same asperity for the single-segment rupture event occurring from the same segment.
We developed a lower-power data logger for earthquake observations, because the lowest-power loggers available now still need heavy lead-acid batteries to operate continuously for longer than two weeks. Two ideas have been made as ways of reducing the power consumption of the data logger presented in this study, keeping its performance high enough for the earthquake activity researches. First, high and low frequency clocks are used for lowering power consumption. Second, a temperature sensing crystal is used for frequency modulation of a crystal clock to keep time. As a result, the data logger can continue 100Hz sampling with 6 alkaline D size 1.5 V batteries for more than 2 weeks.
Maeda and Yokomori (1999) constructed a deterministic model consisting of coupled nonlinear oscillators laid on belt conveyors, where Rayleigh’s friction is to be created between the oscillators and the conveyors. The model reproduces typical statistical properties of earthquakes; their magnitude follow the G-R law, and their occurrence sequences constitute Poisson processes. We show that the model has an approximate recurrence property. Introducing a new index, we investigate systematically relationship between simulated time intervals and quality of recurrence achieved. With model parameters suitably chosen, the model generates sequences matching to observed seismicity of Tohoku and Hokkaido regions in Japan.
Based on a realistic two-dimensional velocity model, we investigated the origins of local sP converted phases observed for earthquakes occurring beneath the Pacific Ocean off Tokachi in Hokkaido, Japan. The sP converted phases appear about 5-10 seconds after initial P waves at stations with epicentral distances of 100-250km. If we assume that the S to P conversion occurs at the sea floor, the observed sP-P times constrain the hypocenters at shallow depths away from the megathrust plate boundary. If we assume that the S to P conversion occurs at the bottom of the unconsolidated sedimentary layer (VP=1.8km/s), the hypocenters relocated using the sP-P times are distributed on or near the plate boundary. The theoretical amplitudes of sP phases converted at the bottom of the sedimentary layer is about 4 times larger than those of the sP phases converted at the sea floor. Therefore the observed sP phases are not likely to be the ones converted at the sea floor. The theoretical amplitude of sP phases converting at the interface between the consolidated sedimentary layer (VP=2.5km/s) and the basement (VP=3.8km/s) is as large as those of the sP phases converting at the bottom of the unconsolidated sedimentary layer. However, if we assume that the observed sP phases are converted at the interface between the consolidated sedimentary layer and the basement, the hypocenters are relocated deep inside the subducted slab. Judging from the distribution of relocated hypocenters, focal mechanisms, and theoretical amplitudes, it is most likely that the observed sP phases are those converted at the bottom of the unconsolidated sedimentary layer.
The major earthquakes of M=7 class, so-called Geiyo earthquake, have occurred in the subducted slab of the Philippine Sea plate under Aki-nada and Iyo-nada sea. The upper surface of the Philippine Sea plate is located in the depth from 40-45km in this region. Five Geiyo earthquakes can be found in 1649, 1686, 1857, 1905, and 2001 since 17th century. Magnitudes M of the older 4 events were larger than 7, while that of the last one is 6.7 in JMA (Japan Meteorological Agency) scale. Magnitudes and locations of focal regions of them were redetermined from seismic intensity data by the method of an attenuation curve fitting and of the seismic intensity inversion. Optimal magnitudes of old events were obtained between 6.7 to 6.9. This results show the old 4 events have almost the same magnitude as the 2001 event. One of the possible reasons why the former values were overestimated is that the empirical relation between magnitude and isoseismal area from inland shallow earthquakes was applied to determine the magnitude of the historical intraslab events such as Geiyo earthquakes.