ESG6 BLIND PREDICTION. 1D MODELING FOR STEPS 2 AND 3 BY INGV AND ENEA TEAM

Abstract

This paper describes the analyses performed in the framework of ESG blind prediction steps 2 and 3, devoted to the simulation of the weak- (step 2) and strong-ground motions (step 3) recorded at the prediction site located on the Kumamoto plain (Japan) during the seismic sequence of 2016. We employed 1D equivalent-linear (EQL) simulations for predictions of step 2 and both EQL and nonlinear (NL) simulations in step 3. Models were built considering the results of step 1 in terms of shear-wave velocity profile and the available information on the site at regional and local scale distributed by the ESG committee. Nonlinear soil behavior was considered based on available laboratory tests on undisturbed soil samples and on the literature. Models were forced at the bedrock level by the signals recorded at the SEVO reference site assuming an outcropping motion boundary condition (elastic half space). The results showed significant horizontal ground motion amplification at the prediction site for both weak- and strong-motion input. The EQL results are characterized by substantial differences obtained using the NS and EW components of the input motions. The comparison between strong ground motion predictions using the EQL and NL approaches suggests a strong influence of nonlinear soil behavior for the shallow soil of the Kumamoto plain during the stronger earthquakes of the sequence.