Subsurface structure identification at the blind prediction site of ESG in Kumamoto under the diffuse field assumption

Abstract

A representative velocity profile for Kumamoto city is presented in this study, derived from ambient seismic noise data collected by three sizes sensor arrays SM, M, and LL. Phase and group velocity dispersion curves for both Rayleigh and Love waves were estimated using techniques such as SPAC and cross-correlations. Additionally, Horizontal-to-Vertical Spectral Ratio (HVSR) analysis was performed individually for each array, and the resulting data were averaged to generate a single representative HVSR curve. By employing joint inversion under the diffuse field assumption, dispersion curves and HVSR data were concurrently fitted, allowing for the estimation of a 1D velocity profile. This approach effectively addresses the non-uniqueness inherent in inversion problems. Furthermore, upper modes were identified, and a comparative analysis of dispersion and HVSR curves was conducted against those of an established model, revealing notable discrepancies requiring thorough examination. Subsequently, seismic energy partitioning analysis was carried out, focusing on the vertical component. The findings were correlated with dispersion diagrams to confirm consistency in mode identification.