Modelling directivity effects of the October 21, 2002 (Mw = 5.7), Molise, Southern Italy, earthquake

Abstract

Acceleration time series recorded by the Italian Strong Motion Network (RAN) during the October 31, 2002 (Mw=5.8), Molise earthquake, are employed in order to investigate source effects on the ground motion in the epicentral area. We consider two different seismogenic sources: a fault model inferred from inversion of teleseismic, regional and local seismic signals [Vallée and Di Luccio, 2005], and a fault model based on seismotectonic data [Basili and Vannoli, 2005]. Both source studies suggest a deep location of the earthquake fault plane (ranging from 6.0 to 20.1 km and from 12.0 to 19.9 km, respectively), however, with considerably different fault lengths (5.2 and 10.5 km, respectively), and widths (14.2 and 8 km, respectively). Due to these differences, only the second model allows for effective horizontal unilateral rupture propagation. Finite fault effects are modelled by the Deterministic-Stochastic-Method (DSM) [Pacor et al., 2005], and the Hybrid Integral-Composite source model (HIC) [Gallovic and Brokesova, 2006]. In both methods k-square slip distributions on the faults are considered. We simulate the October 31, 2002 earthquake considering: 1) Vallée and Di Luccio [2005] faultwith a bilateral rupture propagation, and 2) Basili and Vannoli [2005] fault with unilateral directions of the rupture propagation. The spectral attenuation is modelled using a regional estimate of the quality factor [Castro et al., 2004] and k values estimated from acceleration records. Comparison between synthetic and recorded data at nearby stations (hypocentral distances < 60 km) performed in terms of frequency content and peak ground motion, favours the model with unilateral propagation of the rupture. Assuming the source model with unilateral rupture propagation, we utilize both asymptotic and full wave field methods in order to simulate ground shaking scenarios for an area extending up to 150 km epicentral distance. These results are then subjected to comparison with peak ground accelerations recorded in the far field.