Benchmark tests for strong ground motion prediction methods: Case for theoretical methods (part 2)

Abstract

Following the 2009 benchmark test, we conducted a subsequent test for the benchmark a the strong motion simulation methods in more complex media, using various theoretical methods (the wavenumber integration method, the discrete wavenumber method, and the thin layer method). All the results show generally good agreements, but the results for damping media show discrepancies, especially in the fling step at very close stations to a surface fault. This is because the frequency-dependent quality factors violate severely the causality condition in the waveforms, especially at lower frequencies.