PROBABILISTIC TSUNAMI HAZARD ANALYSIS USING RADIAL BASIS FUNCTION NETWORK AND QUASI-MONTE CARLO SIMULATION AND ITS APPLICATION TO RISK ESTIMATION OF DISASTER WASTE CAUSED BY THE ANTICIPATED NANKAI TROUGH EARTHQUAKE

Abstract

 This paper presents a methodology for probabilistic tsunami hazard analysis (PTHA) using radial basis function (RBF) network and quasi-Monte Carlo simulation (MCS). RBF network associated with tsunami intensity is obtained from the results of tsunami numerical simulations based on an experimental design, which can reduce the computational cost drastically. Quasi-MCS can be performed with fewer simulations than ordinary MCS even when estimating tsunami intensity corresponding to low exceedance probabilities. The proposed method is validated using the tsunami propagation simulation results. The contribution of the proposed methodology in terms of the computational cost associated with PTHA is discussed. In an illustrative example, the proposed method is applied to the risk estimation of tsunami disaster waste caused by the anticipated Nankai Trough earthquake in Mie Prefecture, Japan.