Ground-motion model for Housner’s spectrum intensity based on a novel hybrid-scenario approach

Abstract

We use a novel hybrid-scenario approach to develop a ground-motion model (GMM) for Housner’s spectrum intensity (SI) using estimates of pseudo-spectral acceleration (PSA) from the authors existing GMMs based on a scenario approach proposed in the literature and predictor variables from the same database used to develop the PSA GMMs. These estimates of SI are used together with the predictor variables to develop a hybrid-scenario GMM using mixed-effects regression analysis. Because the GMM is based on predicted values of PSA, the aleatory variability from the regression is not indicative of the actual variability of observed values of SI. Instead, a hybrid-scenario model for magnitude-dependent between-event, within-event, and total aleatory standard deviations is derived from the PSA GMMs using the scenario approach. The predicted values of SI and its standard deviations from the hybrid-scenario model are found to be relatively consistent with the residuals between these predictions and estimates of SI from the database (i.e. the observations). However, the values of SI predicted from a purely empirical GMM developed directly from the observations are different than those from the hybrid-scenario model by up to a few tens of percent. These differences are the result of an insufficient number of observations resulting from bandwidth limitations of the database that lead to a bias in the empirical results. The near-source standard deviations from the hybrid-scenario GMM are found to be generally consistent with both those from the empirical model and those from the residuals between the hybrid-scenario model and the observations. However, the far-source standard deviations of the hybrid-scenario model are smaller than those from these other methods.