Hybrid Ground Motion Prediction Equations for Southern Italy

Date: 4/25/2019

Time: 11:30 AM

Room: Pine

HYPSTHER (HYbrid ground motion prediction equations for PSha purposes: the study case of souTHERn Italy) is a research project funded by INGV (Istituto Nazionale di Geofisica e Vulcanologia), mainly devoted to the development of ground motion prediction models, based on the integration between recorded and synthetic data (hypsther.mi.ingv.it). We selected the Southern Italy (Calabria and Sicily regions), as study area, since the expected hazard is high, but the seismic activity was scarce in the last decades.

We present some specific project results, concerning the calibration of the Ground Motion Prediction Equations (GMPEs) for the study area. We developed an empirical GMPEs (SI17ref) for shallow active crustal regions in Southern Italy (for PGA and SA at 0.3, 1, and 3s), introducing a site category for the reference rocks in the functional form, in order to predict the ground motion of sites unaffected by soil amplifications. These empirical GMPEs are not well constrained in near fault conditions, since very few recordings are available at distance lower than 10km.

Within the project, we produced a dataset of more than 180.000 synthetic records for hard rock condition according to two different simulation methods to model (1) a set of finite sources in the magnitude range 5.0-7.5 and (2) point-like sources characterized by lower magnitudes. We developed a methodology to derive the hybrid GMPEs (SI17hyb). Model predictions are larger with respect to SI17ref in near-fault conditions for PGA and lower magnitudes. At distances larger than 50km, SI17hyb predictions are controlled by empirical data, because they attenuate faster with distances, with respect to simulated data.

Since the standard deviation (σ) of the total residuals of GMPEs has a strong influence on the results of PSHA, we propose a heteroscedastic model for SI17hyb, including a dependence on magnitude. We also investigated the dependence of the stress parameter, which is an input parameter of the numerical simulations, on the GMPEs predictions and associated standard deviation.

Presenting Author: Giovanni Lanzano

Authors