Spatially correlated multiscale Vs30 mapping and a case study of the Suzhou site

Abstract

The average shear-wave velocity in the first 30m of subsoil, Vs30, is a key indicator of site response affecting the ground-motion amplification for many earthquake engineering applications. Mapping of Vs30 over a large region is commonly done through proxy-based models correlating Vs30 with geological or topographic information. In this paper, a multiscale random field-based framework is presented and applied to mapping Vs30 over extended areas. This framework accounts for spatial variations of Vs30 values across different length scales and is able to adaptively refine around areas of high interest while maintaining consistent description of spatial dependence. In the case study site, Suzhou City, a total of 309 shear-wave velocity measurements are compiled and used to calculate Vs30 values, from which the statistical and spatial parameters for the random field model are inferred. USGS topography-based Vs30 data are also collected and used as secondary information to improve the accuracy of predictions. The random field models are coupled with Monte Carlo simulations to obtain a multiscale Vs30 map and its associated uncertainties at the Suzhou site. The new Vs30 map is then applied to site classification and amplification factor characterization in the studied region to demonstrate its applications.