S43B-2780
Simulation of Strong Ground Motion in the Kanto Region during the 2011 Tohoku, Japan, Earthquake (Mw9.0) with the Pseudo Point-Source Model

Thursday, 17 December 2015
Poster Hall (Moscone South)
Atsushi Nozu, Port and Airport Research Institute, Kanagawa, Japan and Atsushi Wakai, National Research Institute for Earth Science and Disaster Prevention, Tsukuba, Japan
Abstract:
In a previous study (Nozu, 2012, Zisin; Nozu, 2013, AGU), we developed a new simplified source model (the pseudo point-source model) to explain strong ground motions from a mega-thrust earthquake. In the model, the source spectrum associated with the rupture of a subevent is modeled and it is assumed to follow the omega-square model. By multiplying the source spectrum with the path effect and the site amplification factor, the Fourier amplitude at a target site can be obtained. Then, combining it with the Fourier phase characteristics of a smaller event, the time history of strong ground motions from the subevent can be calculated. Finally, by summing up contributions from the subevents, strong ground motions from the entire rupture can be obtained. The pseudo point-source model developed for the Tohoku earthquake consisted of nine subevents. The model could reproduce the velocity waveforms and the Fourier spectra quite accurately at sites close to the source region along the east coast of Japan from Iwate to Ibaraki. In this study, we investigated the applicability of the model to larger distances. In particular, we simulated strong ground motions in the Kanto Region including Tokyo. In the application of our model to larger distances, careful consideration of the path effect was required. The original path model used in our previous study can be expressed as P(f)=(1/r)*exp(-πfr/Qβ) and Q=114f0.92(Satoh and Tatsumi, 2002). However, there are two factors not included in the original path model that could potentially lead to underestimation at larger distances: (1) the geometrical spreading term inversely proportional to the square root of distance at larger distances, and (2) the lower bound of the Q value at lower frequencies (e.g., Aki, 1980). The simulation based on the original path model resulted in underestimation of Fourier spectra at larger distances as shown in Figure (a). In this case the error was an increasing function with the distance. Then we employed a revised path model with these effects, in which the Fourier spectra decreases with the square root of distance at distances greater than 80 km and the Q value is constant below 1.0 Hz. By using the revised path model, a good agreement was achieved between the observed and synthetic Fourier spectra in the Kanto Region as shown in Figure (b).