Synthetic Tsunami Waveform and Character of Tsunami Sources in Japan Sea, a Normal Mode Approach

Abstract:

Normal mode method is widely used in seismology. Although some possibilities of applying normal mode to tsunami researches are already pointed out, the potential power of this method still remains.

One advantage of normal mode method is that, once the mode solution is obtained and stored in the computer, we can synthesize tsunami waveforms as superposition of normal modes (Satake and Shimazaki, 1987). Although this theory is scientifically sound and there are some simple numerical tests, it has never been applied to a real case, due to computational difficulty. Recently, Wu and Satake (2014) developed a faster method to calculate normal mode based on Loomis (1975) method, and then obtained high resolution normal mode solution of Japan Sea, which contains 4000 modes (down to period of about 10 min). This makes it possible to verify the theory using real earthquake source parameter.

With the source parameter of the Mw 7.7 1983 Japan Sea earthquake, we succeeded to synthesize tsunami waveforms by normal mode method for the first time. Comparisons with computed waveform based on finite different method (FDM) show good fit, thus proved the goodness of this idea. Unlike FDM, the calculation time does not depend on the duration for computation.

Another merit of normal mode method is that the weight of excited tsunami due to an earthquake source may give us some direct information about the feature of tsunami generated by that source.

In this study, we also calculated the excitation weight of 60 potential submarine faults in Japan Sea proposed by MLIT (2014). From these obtained weights, we can clearly see, for example, that when the source is located at shallower part or with larger magnitude, the generated weight will usually be larger.