GNGTS 2019 - Atti del 38° Convegno Nazionale

268 GNGTS 2019 S essione 2.1 Following modern advances in PSHA, the so-called partially non-ergodic approach (e.g., Rodriguez-Marek et al. , 2014; Faccioli et al. , 2015; Mascandola et al. , 2017; Barani et al. , 2019) was applied to assess the hazard in the study area. In practice, it consists of replacing the ergodic standard deviation of the (logarithmic) ground motion associated with the attenuation equation selected for the hazard assessment with the so-called single-station standard deviation. Compared to the ergodic counterpart, this latter term is deprived of the systematic contribution related to site behavior, which is modeled via logic tree. In the present study, a simple logic tree made of three branches, each corresponding to an alternative value of F S , is used. These alternative values are used to adjust the median rock ground motion predicted by three different attenuation equations: Bindi et al. (2014) and Felicetta et al. (2018) for generic rock conditions (i.e. type A soils), and Felicetta et al. (2018) for reference rock sites (i.e., sites with V S ≥ 800 m/s unaffected by any type of site amplification). This latter model was specifically developed for site-specific seismic hazard assessments that couple the ground motion at the base-rock level (which, at least in principle, should be unaffected by amplification effects) with the response of the overlaying soil. However, it is unsuitable to assess the hazard at the rock outcrop, where the actual site response may not be flat (e.g., due to weathering effects). Hence, it was not used to assess the hazard at those sites characterized by outcropping rock. The calculations were carried out for all nodes of a homogeneous grid with 500 m spacing in longitude and latitude. The hazard computations were carried out by adopting the source model developed by Santulin et al. (2017) for the upcoming seismic hazard map of Italy along with its associated parameterization. Fig. 2 - Soil-specific PGA hazard map for a mean return period of 475 years. As an example, Figure 2 shows the PGA hazard map for a mean return period of 475 years for all nodes falling in the areas subject to microzonation analysis. The maps clearly reflect the significant site-to-site variability of ground response, highlighting areas where the ground