GNGTS 2014 - Atti del 33° Convegno Nazionale

242 GNGTS 2014 S essione 2.2 Multiple station measurements (seismic arrays). To define the seismostratigraphic reference model, 34 seismic arrays were deployed, uniformly distributed in the whole survey area (Fig. 1). Ambient vibrations were recorded for 20 minutes at a 128 Hz sampling rate by using vertical geophones (4.5 Hz) and a digital acquisition system (BrainSpy 16 channel acquisition system by Micromed for UNISI and RER and by a Geode 24-channel modular acquisition system by Geometrics for UNIBAS). Sensors were irregularly distributed (with inter-geophonic distances in the range 0.5-40 m) on the ground surface along two crossing perpendicular branches (UNISI) or along L and C shape branches (UNIBAS), with maximum dimensions between 100-250 m. Effective Rayleigh waves phase velocities ( V R ) as a function of frequency (the dispersion curve) were obtained from cross-spectral matrixes by the Extended Spatial AutoCorrelation (ESAC) technique (Ohori et al. , 2002; Okada, 2003). All the dispersion curves show very similar trends, with velocity values that monotonically decrease when frequency increases. In particular minimum V R values are delineated between 100 and 150 m/s and maximum values​ between 300 and 400 m/s. Unlike as was done for the HVSR curves, in this case it has not been possible to identify some representative patterns of dispersion curves and associate them with distinct geographical areas; moreover, no systematic pattern of dispersion curves has been found for the four zones described above. This features shows that, at least for the shallower 50-100 m of subsoil (the maximum investigation depth of seismic arrays), the entire area (about 2500 km 2 ) can be considered as homogeneous from the seismostratigraphic point of view and that outcropping materials are characterized by rather low velocity values. This hypothesis is also confirmed by the values ​of shear wave velocity obtained from 13 borehole seismic tests (10 down-hole and 3 cross-hole) carried out on behalf of the Emilia- Romagna regional administration. In the first 50 m of depth, these V S profiles show differences lower than 150 m/s with values fully comparable with the ones obtained from ambient vibration array measurements. Shear wave velocity profile and seismic substrate’s depth. To assess the depth of the transition among sedimentary cover and the seismic substrate, the dispersion curvewere analyzed by a simplified approach (Albarello et al. , 2011). This method is based on the assumption that Rayleigh wave phase velocity V R , roughly corresponds to the average S-wave velocity up to a depth h that decreases with the corresponding frequency f . In particular, it is assumed that: (1) With the aim to obtain S V values compatible with experimental data, c value is chosen in order to maximize the compatibility of the dispersion curves with the average S-wave velocity profiles as a function of depth derived from available borehole data. The depth profiles of the average Vs values determined in this way, were then modelized by considering a power law pattern in the form V S (z) = V 0 ·z x , where V S is the average S-wave velocity up to the depth z. The parameters V 0 and x obtained in this way were used to attribute to the frequency f 0 of the HVSR maxima a depth H of the relevant resonant interface following the formula: (2) (Ibs-Von Seht and Wohlemberg, 1999). Initially, to test and calibrate the entire procedure, data available in the Mirandola-Medolla area were considered as a benchmark. Only in this area, in fact, borehole seismic tests intercept the seismic substrate. In particular, the two cross-hole tests present in this area show a significant shear wave velocity increase in correspondence of the contact between the alluvial Quaternary