GNGTS 2016 - Atti del 35° Convegno Nazionale

576 GNGTS 2016 S essione 3.2 is characterized by pyroclastic deposits enriched in yellow or grey pumices, interspersed with ashy levels, down to a depth of 5±1 m; carbonate deposits follow in the succession. Data acquisition and method. We used three geophysical methods, which include passive method (microtremor HVSR) and active methods (Ground Penetrating Radar - GPR, and MASW analyses) to estimate the depth and reconstruct the internal architecture of pyroclastic deposit mantling the northern slope of the Vallatrone Mount in southern Italy. Horizontal-to-Vertical Spectral Ratio (HVSR) method, applied to microtremor, provides an estimate of the resonance frequency of cover sediments, in the hypothesis of a 1D medium. In particular, a 1D medium, consisting of a sedimentary layer resting on bedrock, generates a peak in the HVSR curve. This peak, due to a site effect, is defined stratigraphic peak (���������� Haghshenas et al. , 2008; Castellaro and Mulargia, 2009)�. GPR is an active prospecting method, which is based on the analysis of the reflections of electromagnetic waves transmitted into the ground. It allows the detection of electromagnetic discontinuities in the soil generated by layers or isolated bodies, having different dielectric properties (Davis et al. , 1989). The instrumentation consists of a control unit, which generates an electromagnetic pulse of a few nanoseconds; this is sent to an antenna, which, in turn, transforms it into a higher-amplitude pulse, and radiates it into the ground. The frequency of the antenna can vary from a few MHz up to GHz and is chosen according to the depth of investigation to be reached and to the desired resolution. These two parameters are inversely proportional to each other (Annan, 1976; Coco e Corrao, 2009). The acquisition technique used is that of the continuous profile, that is performed with the displacement of the pair of antennas moved in a continuous manner along the predetermined profile, with low and constant speed drag. We used a GPR antenna provided by IDS, with a frequency of 80 MHz; this means that the source of radiation has a frequency range between 80 and 120 MHz. The acquisition was carried out in monostatic mode. The multichannel analysis of surface waves (MASW) method deals with high-frequency Rayleigh waves, to estimate the near-surface S-wave profile. The MASW technique provides the Rayleigh wave dispersive function from which one can derive the 1D Vs profile, with the assumption that the subsurface is a layered half-space in horizontal and parallel layers (Dal Moro, 2008). Instrumentation is generally composed of a multi-channel seismograph, vertical geophones with natural frequency of 4.5 to 10 Hz, and a seismic source. Using three different geophysical methods, as described above, we produced satisfying results that highlight the physical properties of the soil, useful to characterize the pyroclastic blanket. Finally, geophysical results were put in relation with the stratigraphic data. Data processing and results. The HVSR microtremor technique has proved an effective tool for assessment of site resonance frequencies (Haghshenas et al. , 2009). Seismic signals were recorded using a Tromino (Micromed) triassial station, 24 dB, 0.1-256 Hz. We collected the seismic signal for thirty minutes, at 128 Hz sample rate. Data analysis was performed using the Grilla software provided by Micromed. Signals relative to each component were subdivided in non-overlapping time windows of 16 s, and tapered with a 5% cosine function. Spectra were smoothed using the Konno & Ohmachi window (b=40). The resulting spectral ordinates relative to NS and EW components were geometrically averaged. HVSR curves over the 16- s-long time windows were then averaged to compute the final HVSR curve, as shown in Fig. 2. The HVSR curve shows a clear amplitude peak at 6.44 ±1,62 Hz. This peak corresponds to a local maximum in the horizontal components (green and blue lines in Fig. 2c) and to a local minimum in the vertical component (purple line), and this pattern proves the stratigraphic origin of this peak (Castellaro and Mulargia, 2009). Clearly, the pyroclastic layer resting on the carbonate bedrock generates this significant peak in the HVSR curve (stratigraphic peak). We also acquired two 2DGPR profiles: Profilo01 for a total length of 23 m, and Profile02 for a length of 21 m. GPR data displayed directly after the acquisition, generally also contain "noise", making interpretation difficult. The identification of the optimal processing sequence depends