GNGTS 2015 - Atti del 34° Convegno Nazionale

amplification level of SSR is about 2. For vertical incidence of planar waves, the half- circle and the triangle show obviously a symmetric behavior, even if differences in the occurrence of maxima and minima of SSR can be observed between the half-circle and the triangle. As also shown by the time series, the slope-like shape displays larger SSR amplification in the right part of the model settled at higher elevation, whereas deamplified SSR values (blue color in Fig. 2) are evident at the midpoint of the x-axis which corresponds to the spatial position of the slope. The synthetic signals have been also used to compute the horizontal-to-vertical spectra ratio (HVSR; Fig. 2). Because I kept separated the P- and S- polarization of the seismic input, the HVSR was computed as double ratio (H/Href)/(V/Vref); the term H/Href is the SSR in case of x component and S- input polarization, and the term V/Vref is the SSR computed on z component for P- input polarization. Fig. 2 (right panel) indicates a different pattern for the three geometries. The HVSR of the half-circle shows, in proximity of the uppermost vertex, strong resonance at about 2.5 and 6 Hz with amplitudes around 4. The HVSR of the triangle shows deamplification at the vertex position. Narrow patterns of amplification and deamplification start from the flank of the triangular ridge up to the flat part of the model. The HVSR of the slope still shows larger amplification in the right part of the model located at higher elevation. The maximum levels of HVSR value are 4.5 for the half-circle and the triangle, and 2.2 for the slope (Fig. 2 right panel). The output synthetics are also used to provide insights on the topographic aggravation factor (TAF). According to the previous definition, the comparison between peak ground displacement (PGD) measured at a target receiver and PGD measured at an “ideal” site not affected by topography can be related to TAF. The distribution of PGD computed from synthetics along the topographic profile is illustrated in Fig. 3, where the PGD minima of the three geometries were normalized to 1. The receiver number #1 (placed at a progressive x of 1000 m) is accepted as “ideal” free-field site. The distributions of peak ground velocity and acceleration (PGV and PGA) were also computed after single (or double) differential of the displacement time-series, obtaining very similar trend to the PGD one. Fig. 3 shows that the increase (or decrease) of PGD is strongly depending by the geometrical shapes and source polarization. For x component and S- source polarization (top panel of Fig. 3), the PGD reaches the same level of minimum (with respect to receiver number #1) for all the Fig. 3 – Normalized PGD along the surface of the three geometries. Red, green and black curves show the slope, the triangle and the half-circle (bottom panel). Top panel shows the PGD trend for x component and S input polarization; middle panel shows the PGD trend for z component and P input polarization. 90 GNGTS 2015 S essione 2.2