Comparison of real and simulated records using ground motion intensity measures

Abstract

Simulated ground motion records have recently become major alternatives to real records. In this study, differences between simulated and real record sets are evaluated in terms of alternative ground motion intensity measures. As case studies, two regions in Turkey (Erzincan and Duzce) are considered. A large set of simulated records involving a wide range of magnitudes, source-to-site distances, and site properties is prepared. The fundamental seismological properties of stochastically-simulated records are compared with those of actual motions. A real ground motion dataset for each region with consistent source, distance, and site properties is formed. Alternative intensity parameters are computed for real and simulated sets and evaluated through statistical analyses involving Kolmogorov–Smirnov test. Varying patterns are observed for different levels of intensity parameters. Correlations of intensity measures with maximum inelastic displacements obtained from single-degree-of-freedom models are investigated. Better correlations are observed for simulated data emphasizing significance of region-specific ground motions.