SPATIAL CORRELATION OF THE SYSTEMATIC SITE- AND PATHSPECIFIC CORRECTIONS OF A GMPE CALIBRATED IN NORTHERN ITALY

Abstract

The probabilistic assessment of the seismic hazard (PSHA) at an individual site is a standard practice, but an ergodic assumption is commonly made: the ground-motion uncertainty computed by a Ground Motion Prediction Equation (GMPE) from a global dataset is assumed to be the same as the variability at a single site. In this paper, the ergodic assumption is relaxed by means of a residual analysis, accounting for the impacts on both the median and aleatory standard deviation of a GMPE. The aleatory variability is separated from the systematic source, path and site effects using a strong motion data set from Northern Italy with multiple recordings at each site and multiple earthquakes within small regions. A local model, specifically tailored for the area, is used as the reference GMPE, which predicts the geometric mean of horizontal response spectral accelerations in the period range 0.04-4s. The spatial covariance of such repeatable effects is modeled, in order to generate spatially correlated fields of path, source and site corrections and their associated variabilities. The results can be used to prepare fully non-ergodic hazard maps of parameters of engineering interest.