Earthquake Scenario Modelling for the Validation of Risk Model

Destructive earthquakes such as the 2010 M7.0 Haiti and the 2011 M9.1 Tohoku events underline the importance of reliable seismic risk models. Key to this are Probabilistic Seismic Risk Assessment (PSRA) models, crucial in earthquake engineering. Their validation and calibration, however, can be limited by sparse historical loss data. This research aims at testing the European Seismic Risk Model (ESRM20) considering past events. Using the GEM Earthquake Scenario Database, we formulated earthquake scenarios for Europe based on significant past events. With the OpenQuake-engine, we produced ground motion fields (GMFs) integrating a rupture model with seismic station recordings. These fields were combined with the ESRM20's exposure and vulnerability models to estimate economic losses, building damages, and casualties. We compared these estimates to historical impact data, using a "normalisation" process considering inflation, population changes, and construction advancements. By analysing two recent, similar-sized earthquakes, we demonstrated how different assumptions can lead to varying loss predictions. This approach was then extended to other scenarios, further validating the methodology's broader applicability. This research provides recommendations to simulate past events and validating existing risk models, which is fundamental to improve the reliability of existing risk models.