[S04-2] Historical and macroseismic studies of earthquakes II

[S04-2-02] A New Approach to Comprehend Historical Tsunami Source

^○Ritsuko S. Matsu'ura¹, Yuta Mitsuhashi², Yukitoshi Fukahata³ (1.Earthquake Research Center, ADEP, Tokyo, Japan, 2.Kozo Keikaku Engineering Inc., Tokyo, Japan, 3.Kyoto University, Uji, Japan)

For decades, the initial wave height for a tsunami simulation has been acquired from the crustal deformation of one rectangular fault with a uniform slip in a semi-infinite media, or summation of many small rectangular faults. However, the inevitable singularity around the rectangular edge, especially near corners of a uniform slip rectangular fault is often harmful for tsunami simulations, when the source is close to the target land area. Recently, we made a new easy to use program for calculating surface deformation due to a fault of arbitrary shape with arbitrary slip distributions in a horizontal multi-layered media. We used this program for the forward modeling of some historical tsunamis in Japan, such as 1707 Hoei Earthquake, and 1611 Keicho Sanriku Earthquake. We only know rough tsunami heights for several ten points, and approximative arrival times of tsunami for some places for those historical events. However, some historical materials and paleoseismological surveys allow us to infer inundated area and its status. We show some physically plausible but rather simple models for those examples. We examined only ellipsoidal shape faults with the smoothly tapered slip amount from the center to ellipse. We also show the differences of tsunamis generated by traditional rectangular sources with uniform slip from tsunami generated by our models.
Another fatal problem frequently observed for the tsunami simulation is to confound the crustal deformation with the dynamic solution of the elastic equation, intending to include the rupture process into the tsunami calculation. Since the crustal deformation is nothing but the static solution, it only gives the final deformation after the whole source rupture movement has ceased. If the source is divided into small rectangles and the crustal deformation of each rectangle is used as the initial wave height at a certain time window, that tsunami simulation is invalid.