Benchmark tests for strong ground motion prediction methods: Case for numerical methods (Part 3)

Abstract

We performed a benchmark test for strong motion simulation methods using numerical methods (finite difference method and finite element method). We simulated four observed earthquakes considering a Kanto basin model. All the results calculated by six teams from different institutions generally show good agreement to each other. We found that the choice of grid space makes difference of allocation of soil parameters near the surface. It generates difference of waveforms.