ESTIMATION OF MODEL UNCERTAINTY OF S-WAVE VELOCITY PROFILE FROM INVERSION OF SURFACE-WAVE PHASE VELOCITY USING MCMC METHOD

Abstract

Applicability of Markov Chain Monte Carlo (MCMC) method in Rayleigh wave phase velocity inversion in shallow soil is examined. Numerical examination was conducted using synthetic phase velocity. Frequency distributions of S-wave velocities and thicknesses of the sampled models were used to estimate optimal model and its uncertainty. The shallow S-wave velocity model was well reconstructed from the inversion. The uncertainties of the parameters are linearly related with the observational errors. The sampled models were used to estimate uncertainty of 1D amplification factors. We furthermore applied the inversion method to actual data from microtremor explorations in damaged area during the 2007 Noto-Hanto earthquake. It is concluded that the MCMC inversion method is effective to estimate optimal model together with their uncertainty.