Supplementing VS30 with H/V Spectral Ratios for Predicting Site Effects

Abstract

Site amplification models using relatively simple site proxies (e.g., averaged shear-wave velocity and soil depth) do not capture site-specific resonance effects observed at soil sites having a high shear-wave velocity impedance contrast. We present a model to supplement ergodic site amplification equations to consider these effects (i.e., our model is additive to ergodic equations in a natural log sense). The model is conditional on properties of horizontal-to-vertical (H/V) spectral ratios recorded at the site of interest and consists of a frequency-domain pulse function parameterized by site period, pulse amplitude, and pulse width. H/V spectral ratio peaks are observed in 97% of sites in our data set. When no appreciable H/V spectral ratio peak is present, we provide a broadband adjustment to the ergodic model. Because H/V spectral ratios are typically measured under small-strain (effectively viscoelastic) conditions but applied in engineering practice for stronger shaking conditions, our model modifies site period for nonlinear site-response effects. The model adapts in an empirical manner H/V spectral ratio attributes measured from Fourier amplitude spectral ratios to horizontal response spectral amplification using a data set from Japan composed of sites having sufficient recordings to produce stable mean H/V spectral ratios. Compared to an ergodic model that does not consider H/V spectral ratios, application of the proposed model reduces natural log within-event standard deviation by an average of 0.04.