Item

Abstract

While probabilistic seismic hazard assessment (PSHA) is well-established as the principal methodology for characterizing earthquake shaking at site or across region, quantitative validation (or even simply comparison) of hazard models against recorded ground motions remains challenging. Focusing on recent models from France and Germany, alongside the 2020 European Seismic Hazard Model, we exploit the vast database of weak and strong motion observations provided by the European Integrated Data Archive (EIDA) to compare predicted and observed exceedance of shaking in regions of low-to-moderate seismicity. EIDA vastly expands the number of records and stations available for model-to-data comparison, especially in regions of low seismicity. However, the database of usable accelerations from these stations is not a complete archive of the shaking that has occurred, even at sites in longer operation. We overcome incompleteness using observed seismicity to predict expected distributions of shaking at times and locations where gaps exist, calibrated according to inferences from the distributions of recorded motions. With this joint observed-inferred archive we explore the opportunities for quantitative model-to-data comparison using stations from France, Germany and Italy. The analysis begins with an overview of established methods for quantitative comparison before delving further to assess the influence of non-ergodic PSHA. We also aim to relax the constraint of independence between observation sites by introducing spatial correlation into both the hazard calculation and testing metrics. This effort highlights how systematic seismic data archiving can be vital to help demonstrate the relative suitability of hazard models for application in a region.