Inversion Analysis on Surface Wave Dispersion Curves and H/V Spectra by Neighbourhood Algorithm

1. Abstract In order to predict strong earthquake ground motion, it is inevitable to construct a relevant subsurface structure model, especially S-wave velocity structure of the site of consideration. Inversion analysis with micro-tremor observation is frequently used to acquire S-wave velocity structure due to its convenience and inexpensiveness. Usually inversion analysis is carried out for either dispersion curves or H/V spectra (Spectral ratio of horizontal component to vertical component on the ground surface); however, simultaneous inversion might be advantageous to cover deeper depth and expanded frequency range. In this paper, we have studied applicability of the Neighbourhood algorithm for the simultaneous inversion of dispersion curves and H/V spectra with numerically synthesized data as well as the ones obtained by micro-tremor observation. First, we compared single inversion for either synthesized Rayleigh wave dispersion curves or synthesized Rayleigh wave H/V ratio to simultaneous inversion for both of them to find the advantage of the simultaneous inversion in terms of precisely evaluated S-wave velocity structure with layer thickness fixed. Then we carried out micro tremor array observation at water front district of Tokyo to evaluate surface wave dispersion curves by SPAC (Space autocorrelation coefficient) method and H/V spectra. From these examples, we can empirically conclude effectiveness of Neighbourhood algorithm for the simultaneous inversion.