GNGTS 2015 - Atti del 34° Convegno Nazionale

Results. Boreholes data and Vs values. Following L’Aquila earthquake, within the scope of buildingsreconstruction and seismic microzonation, many boreholes, among which several ones encountered the geological bedrock, and in-hole Vs tests (cross hole, down-hole and seismic dilatometer tests) were carried out in the more damaged area of WAP (Fig. 1) (Del Monaco et al. , 2013; GEMINA, 1963; Gruppo di Lavoro MS–AQ, 2010). These huge quantity of boreholes, along with others previously executed in the same area and geophysical investigations (seismic reflection profiles or geo-resistivity surveying), were used to constrain the z vs f 0 correlation. The Vs values were calculated from borehole tests (Fig. 3) (Del Monaco et al. , 2013; Di Giulio et al. , 2014; Gaudiosi et al. , 2013). The Vs value of unit D comes from a single SDMT test (pers. comm. G. Totani). The Vp values are calculated from Vs values with a Poisson’s modulus of 0.35. Density and Qp and Qs, which are the estimated attenuation quality factor of Vp and Vs respectively, are estimated from literature (Gruppo di Lavoro MS–AQ, 2010). Seismic noise measurements. The f 0 spatial distribution in the studied area has been obtained through seismic noise single station measurements and the application of HVNSR method.The measuring sites were located on an irregular grid with a step of about 100-250 m. About 50 HVNSR measurements were acquired with the Lennartz MARSlite instrumentation with the sensor Lennartz 3D/5s and about 740 ones with Tromino instrumentation (Micromed S.p.A). The sampling frequency was of 250 Hz for the Lennartz MARSlite and 256 Hz for the Tromino instrument. The acquisition time lapse was about 30-45 minutes. The two different instrumentations showed the same results for resonance frequency greater than 1-1.5 Hz (Lanzo et al ., 2011). The acquired seismic noise data were analysed in the 0.1-20 Hz range with the softwares Geopsy (http://www.geopsy.org) and Grilla (Micromed S.p.A) for the Lennartz MARSlite and the Tromino instrument, respectively. Horizontal to Vertical ratio (H/V) curves were obtained by averaging the spectral ratios calculated from non-overlapping windows of 30 s. The spectra of each window were smoothed with a Konno and Ohmachi (1998) smoothing function with a bandwidth coefficient b = 40, and the HVNSR was computed at each frequency and for each window. The quality of HVNSR results were checked by using the criteria of SESAME (2004) to validate the statistical significance of the spectral ratio peaks. The resonance frequency contouring wereplotted by using the Natural Neighbor method. The resonance frequency ( f 0 ) ranges from 0.5 to approximately 15 Hz. The f 0 lowest values (about 0.5-2 Hz) are located along the WAP central axis (downtown L’Aquila, S. Antonio, Hospital areas). In the Coppito area f 0 is higher than 6 Hz, while in downtown L’Aquila is lower than 1 Hz, and a sharp N-S trending boundary between the Coppito and Hospital areas is recognizable. In the Pettino area, the f 0 decreases abruptly from NE to SW. The seismic noise measurements in WAP evidenced a remarkable f 1 only in the Hospital and S. Antonio areas, considering f 1 the resonance frequency peak with a frequency value higher than f 0 in the H/V vs frequency curves. In these areas, the measured f 1 shows an amplitude comparable with the f 0 . It ranges from 6 to 10 Hz and its contouring shows an NW-SE trending elongated shape following the WAP morphology (Fig. 1). Numerical simulation. A “trial and error” iterative procedure on the inversion of HVNSR data by using the software Surface Waves (Lunedei and Albarello, 2009) has been carried out to increase points in the depth ( z ) vs. f 0 diagram. This inversion procedure permitted to test numerically the geological cause of the observed peaks. It is based on: i) modelling surface waves; ii) assuming a flat layered viscoelastic Earth model; iii) arranging random independent point sources at the Earth surface; iv) assuming a uniform probability distribution for source location all around the single seismic station and v) using the same spectral power value of the random independent point sources for all the frequencies (e.g. 0.1-500 Hz). Sedimentary cover (z) vs f 0 : empirical correlation and mapping. The f 0 values of seismic noise measurement sites located nearby the boreholes intercepting the Meso-Cenozoic bedrock were compared to z , as observed in the stratigraphic logs of these boreholes. This procedure GNGTS 2015 S essione 2.2 83