Physics-based ground motion simulations for earthquake engineering: seismological validation across multiple intensity measures

Abstract

This work assesses the seismological reliability of a physics-based ground-motion simulation approach that couples an extended-source representation of the seismic rupture with Discrete Wavenumber (DWN) wave propagation in a horizontally layered anelastic medium. The validation is performed through a multi-intensity-measure procedure in which simulated intensity measures are compared with empirical GMM predictions derived from recorded earthquake datasets by means of standardized residual analysis and residual decomposition. The procedure is applied to a purpose-built database of simulated ground-motion signals, generated by systematically varying the main source, path and site parameters controlling ground-motion generation, including focal mechanism, moment magnitude, rupture variability, crustal structure, local site conditions, source-to-site distance and source-to-receiver azimuth. The results indicate that, within the investigated magnitude–distance–site domain and with respect to the adopted empirical reference models and to the simulation methodology assumptions, the simulated motions capture the main median trends and variability components of the GMMs, with standardized residuals generally within the empirical uncertainty range. The residual decomposition highlights interpretable trends related to VS30-dependent site effects, distance-dependent deviations in less constrained ranges, and offsets associated with the adopted deep crustal models. This work represents a seismological validation step toward the future use of simulated accelerograms in performance-based seismic assessment workflows, subject to dedicated engineering validation.