SERA WP7/NA5 - Deliverable 7.4: Towards improvement of site condition indicators

Abstract

This report summarizes the research undertaken by ETH in the framework of WP7/NA5 – Task 7.4 of SERA project (“Towards improvement of site characterization indicators”), in collaboration with partners AUTH, INGV, CNRS. We have addressed the broad topic of site condition indicators, or proxies, with a comprehensive work: i) firstly, we have reviewed the state of the art and tracked the present research trends in the use of proxies; ii) secondly, to test their applicability at a wide scale, we have compiled an extensive database of site condition parameters, covering more than 1000 instrumented sites in Switzerland and Japan, and paired it with a companion dataset of empirically-derived local amplification functions. In this phase of data collection, particular attention was dedicated to the harmonization of information derived from disparate sources and referring to different geological and geographical contexts; iii) in a third step, we have systematically assessed the sensitivity of site condition indicators towards local seismic amplification, ranking and collating their behaviour, also in different environments; iv) lastly, we have attempted to assess their potential for the prediction of local site response, resorting to a neural-network structure. The results we have obtained from the two latter stages offer an interesting insight on the varied correlations between local amplification and various typologies of site condition indicators. Furthermore, the findings gained by the neural-network analysis allow us to determine a ‘best’ set of site condition parameters appropriate to predict local site amplification.