A Bootstrapped Modularised method of Global Sensitivity Analysis applied to Probabilistic Seismic Hazard Assessment

Probabilistic Seismic Hazard Assessment (PSHA) evaluates the probability of exceedance a given earthquake intensity threshold like Peak Ground Acceleration, at target site for exposure time. The stochasticity occurrence seismic events is modelled by stochastic processes and propagation wave in soil typically evaluated empirical relationships called Motion Prediction Equations. large uncertainty affecting PSHA quantified defining alternative model settings and/or parametrizations. In this work, we propose novel Bootstrapped Modularised Global Sensitivity Analysis (BMGSA) method identifying parameters most important PSHA, that consists generating artificial datasets bootstrapping an available input-output dataset aggregating individual rankings obtained with modularized from each those. proposed tested on realistic case study Italy. results are compared standard variance-based (GSA) literature. novelty strength BMGSA both fact its application only requires data not use code repeated calculations.