Selection of Ground Motion Prediction Models for Subduction Zones in Costa Rica

Abstract

The selection of the Ground Motion Prediction Models (GMPM) is a key task for any seismic hazard analysis. The Interface and Intraslab subduction zones of Costa Rica are considered in this research. For the first time, a database with more than 1700 strong ground motion records, exclusively from Costa Rica, is used to check the model-data fitting. A set of GMPM is proposed for each tectonic region, also weights for the logic tree analysis are proposed. The selection process and the assignation of weights use the distribution and trends of the residuals, the likelihood, and the log-likelihood approach.