American Geophysical Union Fall Meeting, San Francisco, California, USA, 5-9 December 2005
F. Pacor, G. Cultrera, A. Emolo, F. Gallovic, A. Cirella, I. Hunstad, A. Piatanesi, E. Tinti, G. Ameri and G. Franceschina
This work is developed within the framework of the Italian project S3 "Expected shaking and damage scenarios in strategic and/or priority areas" (Italian Dipartimento della Protezione Civile in the frame of the 2004-2006 Agreement with Istituto Nazionale di Geofisica e Vulcanologia). One of the goals is to perform a sensitivity analysis using various simulation techniques to test the ground motion variability for different source parameters. We applied deterministic (Compsyn, Spudich and Xu, 2002; Disp, Okada, 1985; 1992) and hybrid simulation techniques (DSM, Pacor et al., 2005; HIC, Gallovic and Brokesova, 2006) to simulate an earthquake with complex source characteristic of the 1980 Irpinia earthquake, M6.9, Southern Italy. The large wave-length characteristics of the adopted source model were first validated, for the 0s sub-event, comparing the observed and synthetic acceleration envelopes. Then synthetic time series (acceleration, velocity and displacement) generated with the simulation techniques were compared with recorded and synthetic data both in time and in frequency domains. The deterministic model fits very well the low frequency content (f<1 Hz) both considering waveforms and spectra. The DSM method is able to reproduce the observed peak ground values and spectral content (f >1 Hz) although the simulated waveforms are more simple than the recorded ones. The HIC model provides broadband seismograms in relatively good agreement with the observed ones. Starting from the validated model, we fixed the source geometries and varied the kinematic parameters of the rupture scenarios (rupture velocity, nucleation points and asperities). For each model we generate ground acceleration and velocity at a dense regional grid of receivers. Peaks ground motions increase as the rupture velocity increases. The acceleration and velocity distributions around the fault strongly depend on the position of the nucleation point. On the contrary peak ground displacement are more sensible to the position of the asperity. Strong peak values are simulated at directivity sites when the rupture starts near the fault edges. When we compare the simulation results with the empirical attenuation relationships we observe that the mean values are similar but the variability obtained from the shaking scenarios exceeds the empirical standard deviation.