Assessment of Spatial Variability of Site Response in Japan

Abstract

AbstractLocal soil conditions influence earthquake-induced ground shaking and deformation; a phenomenon known as site effects. The latter correlate with the concentration of damages to the built environment in areas prone to ground motion (GM) amplification. However, the characterization of geotechnical parameters affecting spatially varying earthquake GMs is often oversimplified. Spatial variability of GMs stems from source heterogeneities, varying ray propagation paths, varying local soil conditions, wave scattering, and directivity effects. Regionalization of site effects may also be necessary when specific geologic structures found within a given region control systematic amplifications in the area (e.g., deep sedimentary columns in the US Atlantic and Gulf coastal plains). Evaluating spatially variable GMs is also essential when investigating large, distributed infrastructure, such as water distribution systems. Because soil properties can be spatially correlated at nearby locations, the expected site response will also be spatially correlated. This study focuses on quantifying spatial correlations in site parameters from the Japanese databases, Kyoshin Network (K-Net) and Kiban-Kyoshin Network (KiK-net). Current spatial correlation models for intensity measures are based on correlation length, which neglects for instance, the effects of depositional processes in the spatial distribution of local soil conditions. In this work, we use Kriging to evaluate the significance of the spatial correlation for different site parameters and evaluate potential causes for the observed differences.KeywordsSpatial variabilitySeismic hazardRegional assessmentSite response