Assessment of ground motion in Palermo, Italy, during the 6 September 2002 Mw 5.9 earthquake using source scaling law

The 6 September 2002 M-w 5.9 earthquake caused damage typical of intensity VI on the European Macroseismic Scale to the city of Palermo, 40-50 km southwest of the epicenter. In distinct contrast with the ground accelerations expected for that intensity (30-60 Gal, approximately), the analog accelerographs around Palermo and in the northern coast of Sicily were not triggered, indicating a level of shaking lower than the 10-Gal threshold, at least on the vertical component. To shed light on the apparent contradiction, we estimated ground accelerations of the mainshock in Palermo by using the aftershock seismograms recorded at eight broadband stations in the city on different geological formations. Rock stations in the region were used to compute seismic moment, local magnitude, corner frequency, and stress drop for the events of the sequence. The observed ground motions of mainshock and aftershocks result in a substantially constant (100 bars) stress drop scaling over the seismic-moment interval 10(19) < M-0 < 10(25) dyne cm. To assess the mainshock ground motion in Palermo, we applied this source scaling law to the aftershock seismograms in the city by using three different techniques: random series generation, random summation of empirical Green's functions, and amplitude scaling through spectral ratios between stations. For each of the stations that recorded aftershocks in Palermo, the three methods yield consistent results. Simulated horizontal accelerations of the mainshock vary from 14 +/- 2 Gal on the stiffest formation (Oligo-Miocene Numidian Flysch) up to 51 +/- 14 Gal on Holocene sea and lacustrine soft deposits. Accelerations assessed for the vertical component show a smaller variation, ranging from 11 +/- 4 to 21 +/- 4 Gal, respectively. The minimum average value of the vertical components and the maximum of the horizontal components agree fairly well with the threshold of not triggered accelerographs and ground accelerations statistically corresponding to the maximum felt intensity, respectively. This study indicates that, during the 6 September 2002 earthquake, a large (a factor of 3-4) variability of horizontal ground motion did really occur in Palermo as an effect of local geology.