On the Accuracy of Vs30-Based Site Response Amplifications in California

Abstract:

To provide more accurate models of site response in ShakeMap, earthquake scenarios, and other seismic hazard applications, we have developed a database of site-specific empirical amplification factors (EAFs) based on the Pacific Earthquake Engineering Research Center’s Next Generation of Attenuation ground motion database. These amplification factors are developed for each site with a minimum of five recordings by averaging amplification as a function of frequency (after correcting for event terms where we require each event include a minimum of five recordings). The EAFs employ a non-reference site approach, where the denominator in the site response ratio is defined as the spectral acceleration predicted by the Boore et al. (2014) ground motion prediction equation for rock conditions, i.e. a time averaged shear-wave velocity to 30 m (Vs30) of 760 m/s. Given the large number of sites in the database, the EAFs provide a valuable dataset for addressing the accuracy of site-response models that are based on Vs30. By focusing on a handful of sites where the EAFs are well determined and the Vs30 profile is measured (rather than inferred based on a proxy), we are able to analyze the limitations of Vs30-based site amplifications and discuss strategies for improving site response models. In many cases, the Vs30-based amplifications accurately reproduce the frequency-dependence of the EAFs, though under some conditions the EAFs exhibit distinctive trends with frequency that cannot be reproduced by Vs30-based amplification factors. EAFs at recording sites provide an alternative to Vs30-based factors that allow for a more realistic representation of site response. Yet, an additional challenge that must be addressed entails developing a strategy for proper geospatial interpolation between EAFs for estimating amplifications at locations without multiple recordings.