A comparative analysis of site-specific response spectral amplification models

In the framework of the Seismic Hazard Harmonization in Europe (SHARE) project, the Swiss Seismological Service (SED) has performed an evaluation of two procedures developed to produce soil amplification models for 5% damped pseudo-spectral acceleration response spectra, each using different parameters to describe the soil properties. The goal of the work presented here is to evaluate the statistical consistency of the methods, with particular regard to their applicability to engineering practice. Additionally, we compare the results with those from a methodology internally developed by the SED, which is based on spectral modeling of ground motion using the quarter-wavelength approximation to parameterize soil conditions. Soil amplification is computed with respect to reference rock condition as defined for the probabilistic seismic hazard assessment performed during the SHARE project. For the comparison, a residual analysis was performed between the computed soil-amplification functions from the three different methodologies, over a number of selected sites spanning different soil classes and ground motion levels. The analysis of the average residuals of these functions is useful to highlight the main differences between the proposed approaches, with special regard to the impact of soil resonances and anelastic attenuation within different frequency bands. The assessment was performed on a group of 88 selected stations of the Japanese KiKNet strong motion network, for which complete logs of the shear-wave velocity profiles are available, in addition to a significant number of earthquake recordings. In a first step, average residuals were computed. Subsequently, amplification variability related to soil classes was investigated. The target of this second step was to perform the comparison by separately analyzing the impact of different soil and velocity classes, according to a soil-classification scheme proposed by Aristotle University of Thessaloniki (AUTH). In this paper the main results of these investigations are summarized and, when applicable, an interpretation of our findings is given. (C) 2016 Elsevier Ltd. All rights reserved.