GNGTS 2015 - Atti del 34° Convegno Nazionale

GNGTS 2015 S essione 2.1 deviation σ, and the coefficient of aperiodicity α . Finally, these values are then used to calculate the hazard rates, for a given exposure time, following a BPT probability density function (time- dependent) and a Poissonian distribution. In our application the FISH code has been customized to take into account empirical scaling relationships in volcanic worldwide domains, including the one obtained in this study for the Etna region. In particular, the calculated α values (Tab. 1) suggest a “less periodic” behavior of faults with respect to ones obtained from the intertimes analyses of historical earthquakes (~0.4, see Azzaro et al. , 2012b), and comparable to those derived from instrumental data (0.64-0.72 obtained by GR, see Azzaro et al. , 2013). Tab. 1 -AR-FiSH output and comparison with estimations based on historical and instrumental earthquake data sets. M min : minimum magnitude for which is calculated the probability of occurrence (M max -sd M max ); BPT prob.: time-dependent probability to have an earthquake ≥ M min in the next 5 years. From source to site. The second important element of seismic hazard assessment is related to the propagation of the seismic energy from the source to the recording site. In this respect, the volcanic areas exhibit specific seismic propagation properties: the attenuation of seismic energy is very high, especially for shallow (H < 5 km) earthquakes, and GMPEs commonly used for tectonic areas are not suitable to attenuate this kind of local seismic sources. Thus, a new GMPE has been calibrated for Etna by using data recorded by the seismic network of Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Etneo (INGV-OE). The data set consists of 91 earthquakes with local magnitude M L ranging 3.0 to 4.8, and hypocentral distances between 0.5 km and 100 km. The shallow events on Mt. Etna occur in peculiar geological conditions, with foci that fall into a thick sedimentary substratum with strong lateral heterogeneities. On this basis, the data set were divided into two groups: Shallow Events (SE, focal depth < 5 km), and Deep Events (DE, focal depth > 5 km). In order to compare our data to those recorded in Italy and Europe, we adopted the formulation proposed by Boore and Atkinson (2008), which is also used in the Italian standard equation “ITA10” (Bindi et al. , 2011). In addition to the “standard” peak ground motion parameters such as acceleration (PGA) and velocity (PGV), we also calculated empirical relations for spectral amplitude (PSA) referred to 0.1 s, 0.5 s, 1.0 s and 2.0 s periods (Tusa and Langer, 2015). For a shallow event of M = 4, at low frequencies, the corresponding spectral values are higher than those predicted by ITA10; on the other hand, PSA for higher frequencies are well below the values obtained by ITA10. The new GMPEs for the Etna region have been finally implemented for the two softwares (C risis , Ordaz et al. , 2013; O pen Q uake , Pagani et al. , 2014) used in our analysis. Other relevant implementations we have done concern the capability of calculation to take into account both the topography (elevation) and local site effects. The former is particularly relevant in a volcanic edifice as Mt. Etna, where earthquakes are very shallow and elevation increases sharply moving from the coast (sea level) to the Central Craters (more than 3000 meters a.s.l.), just in 20 km. At first, thanks to the cooperation with the software developer team of C risis (by M. Ordaz and coauthors, University of Mexico City), a new release now accomplishes the general modeling of a 3D surface (effect of topography) in order to compute