Abstract
Effects of 3D subsurface structural model must be accurately included in estimation of strong ground motion for earthquakes. In this article we examined performance of two 3D models of the Kanto basin, Japan in a finite difference ground motion simulations. The 3D contour maps proposed by Suzuki (1999) and Yamanaka and Yamada (2002) were digitized and used in simulations of ground motions for two moderate earthquakes, which occurred with an intermediate-depth under the Uraga channel [EQ1 (MJMA5.9)] and with a shallow depth near Izu-Oshima island [EQ2 (MJMA6.5)], for aiming at understanding effect of differences in the models on simulated motion. These two events were chosen to know the differences such effect during the events at the difference location in Kanto basin.
The results of the EQ1 ground motion simulation showed a good agreement between observed and synthetic velocities that were mainly characterized by an impulsive S-wave onset. Although, the synthetic waveforms could qualitatively explain some characteristics of the observed motions, which were a long duration and a predominant long-period component, the observed motion in the EQ2 were not fully reconstructed by simulation. We concluded that this difficulty was caused by the uncertainty subsurface structure which existing propagation path for the EQ2 simulation. A quantitative comparison with between the synthetic motions in the two models was tried by using difference of envelope function. The area with major differences of the synthetic was found around center of Tokyo. The model by Suzuki (1999) showed better result than the model by Yamanaka and Yamada (2002) in the area. However, at the other area showed opposite tendency. These differences for the 3D synthetic motions clearly indicated a performance of the two models. These features can be used for constructing a new 3D underground structure model.
