GNGTS 2014 - Atti del 33° Convegno Nazionale

74 GNGTS 2014 S essione 2.1 period exceeds the time length of the earthquake catalogue (1000 years). For some SZs, it may be impossible to define a maximum magnitude according to this method in the case that the maximum magnitude computed for them refers to a return period shorter than 1000 years and the events should be, then, contained in the catalogue. The hazard estimates refer to the return periods of a 475 and 2475 years, corresponding, respectively, to the 10% and 2% exceedance probabilities in 50 years, standard references in seismic design. Results. With the characteristic earthquake associated to the DISS sources (DISS Working Group, 2010) and low seismicity uniformly distributed according to the G-R law inside the FRI zones, the first elaboration relates to the applicability of the Cauzzi and Faccioli (2008) GMPE to the near field (no other GMPEs have been used in this test): 1) extrapolating the functional form also outside its distance range of definition or 2) introducing an asymptotic upper bound for ground motion in the near field. As expected, the simple extrapolation of the functional form (Fig. 1b) produces higher values then the asymptotic upper bound option (Fig. 1c) in whole central Friuli for the return period of 475 years. The obtained difference by applying the extrapolation is of less than 0.1 g higher: as the second solution sounds more reasonable, the Cauzzi and Faccioli (2008) GMPE truncated in the near field has been used in the following elaborations. Specific tests on the three GMPEs have pointed out that the Bindi et al. (2011) GMPE produces average ground motions between the low ones calculated with the Ambraseys and Simpson (1996) GMPE and the high ones produced by the Cauzzi and Faccioli (2008) GMPE. The high vertical PGA obtained by the application of the Cauzzi and Faccioli (2008) GMPE is motivated by the fact that this model does not scale the horizontal shaking to obtain the vertical one for high frequencies and, consequently, it produces remarkably larger vertical motions (i.e., equal to the horizontal ones) than the other GMPEs. The computation performed with the characteristic earthquake associated to the DISS sources and low seismicity uniformly distributed according to the G-R law inside the FRI zones, and considering all three GMPEs evenly weighted (0.33 each) in the logic tree, ��� ����� ���� ��� has shown that the highest PGA for a return period of 475 years (Fig. 2a) is expected in central Friuli, because of the presence of 3 individual sources. ��� ���������� �� ���� ��� ����� ��� ���� ���� �� ��� The comparison of this map (Fig. 2a) with that of the horizontal PGA, obtained by the same logic tree simply replacing the vertical parameters of the 3 GMPEs respectively with the horizontal ones, is extremely interesting. This comparison shows that the expected vertical PGAfor the 475-year return period (Fig. 2a) is about 0.1 g lower than the horizontal one (Fig. 2b) and this feature is evident in whole central Friuli. The largest horizontal values of Fig. 2b (between 0.3 and 0.4 g) are in excellent agreement also with those of the recent European seismic hazard map computed in the framework of the SHARE project (SHARE, 2014). It is worth noting that, in addition to the individual seismogenic sources, DISS considers also the composite seismogenic sources. In the case of the study region, the latter ones simply encompass the individual sources in almost all cases. At this point it was interesting to explore the different contribution to hazard given by the introduction of individual seismic sources modelled by a characteristic earthquake seismicity with respect to the wide SZs with a G-R model for seismicity, as it was done in the past for this region (Slejko et al. , 2011). When comparing the map with fault and zones (Fig. 2a) with that obtained considering only the original FRI zonation (Fig. 2c) the contribution of the fault sources results marginal, simply increasing slightly the area of largest PGA. To compare our vertical estimates with those obtained according to the Eurocode 8 (CEN, 2004), where the vertical PGA is equal to 90% of the horizontal PGA, we have computed the vertical ground motion along an E-W oriented transect passing through the most dangerous area of the study region (see its location in Fig. 1a) . The vertical PGA estimated according to the Eurocode 8 (CEN, 2004) standards results conservative with respect to the different results obtained in the present study and mimics quite well the ground shaking computed by applying only the Cauzzi and Faccioli (2008) GMPE (Fig. 2d).