Modeling Propagation of 2011 Honshu Tsunami

Abstract

In this study, the Honshu tsunami on 11 March 2011 was simulated using the Regional Ocean Model System (ROMS) governed by two - dimensional nonlinear shallow - water equations and driven by three earthquake deformation fields as initial condition. The purpose of this study is to assess the model’s sensitivity to the tsunami source field, especially the case in which the horizontal deformation of earthquake was taken into account. The comparison between the model results and observed tsunami waves at 14 Deep-ocean Assessment and Reporting of Tsunami (DART) buoy stations showed that the average errors of modeled maximum wave amplitude were 23.2%, 26.9% and 37.4% for the three initial deformation fields, respectively. In the best performance case, when the observed wave amplitude was below 0.3 m, the average error of the modeled maximum wave amplitude was 0.034 m and relative error was 23.7%. On the other hand, the average errors of modeled arrival times of the first tsunami - wave peak at 13 DART stations were 2.82, 2.28 and 3.48 min earlier for the three cases of different initial fields, but the error is as big as ~3 hrs at Station 46412 located near the west coast of the USA. The model did not resolve the observed wave at Station 46412, which might be caused by the model’s coarse grid that cannot represent the bottom topography well, especially that near the shoreline. The two - dimensional (2D) model employing shallow - water equations could also be a limiting factor. In Case 4 based on the best performance case with horizontally generated wave, the average relative error of maximum wave amplitude was reduced from 23.2% to 20.6% and the root mean square error (RMSE) was reduced from 0.029 m to 0.014 m. The results suggest that the consideration of horizontal deformation of earthquake, which plays an important role in generating tsunami, can help to improve the performance of simulating tsunami wave.