GNGTS 2019 - Atti del 38° Convegno Nazionale

670 GNGTS 2019 S essione 3.2 PASSIVE SEISMIC SURVEYS TO DETECT LAVA TUBES. APPLICATION TO “SNOW CAVE” – NE FLANK OF MT. ETNA (SICILY, ITALY) S. Grassi, G. Patti, S. Imposa Dipartimento di Scienze Biologiche, Geologiche e Ambientali, Università degli Studi di Catania, Catania, Italy Introduction. A field survey, performing recordings of environmental seismic noise with a MASW survey, was carried out in an area of the Mt. Etna NE flank, where a well-known natural cave (lava tube) is present. The Mt. Etna area is greatly affected by the presence of numerous lava tubes, affecting the volcano flanks up to its lowest altitudes. The main aim of this study is to test the attitude of these geophysical methods in identifying the presence of these cavities and in mapping their extension, through the reconstruction of impedance contrast sections. As is known, the presence of slow layers or large cavities in the subsoil produces, in the HVSR ratio, the phenomenon of seismic inversion (Castellaro and Mulargia, 2009). When a velocity inversion occur the H/V ratio shows an amplitudes less than 1 for wide frequency ranges. This is because the velocity inversion heavily affects the microtremor horizontal spectral components while leaving almost unaltered the vertical component. In particular, the vertical component of the ground motion has a greater amplitude than the horizontal ones. The depth of the layer or cavity determines the frequency range to which the inversion occurs. It was choose to study a well know lava tube named “ Snow Cave” or “ Thieves Cave” . Located in the north-east sector of volcano, at 1540 m above sea level (Fig.1a), the cave develops on a prehistoric lava flow (Branca et al. , 2011) (Fig.1b), with a length of about 70 meters. It is characterized by the presence of two main halls, the Thieves Hall have a triangular shape with the main side 30 m length, and the Hoüel Hall is extended about 15 m. A gallery 20 m length connects the two halls. At the west end of the Hoüel Hall there is a narrow tunnel that extends for about 30 meters (Fig.1c). Field survey and data processing. Were performed 46 seismic ambient noise samplings, both above and inside the lava tube, and one MASW survey, located to the north of the investigated area (Fig.1c). The environmental noise was acquired whit a sampling frequency of 128 Hz for 20 minutes at each measurement point. The samplings were performed at regular interdistance along different alignments. The microtremor measures have been elaborated with the HVSR technique (Nakamura, 1989). Each time series was processed by considering time windows of 20 s, selecting the stationary part of the signal and removing transients. A triangular window with a smoothing of 10% of the central frequency and a FFT algorithm was applied to each time window to obtain spectra. Each peak in the H/V graph corresponds to a possible reflector (seismic-stratigraphic level) that presents an impedance contrast compared to the neighbour levels. In order to detect the presence of directional effects on HVSR peaks, linked to the lava tube, the spectral ratios were calculated along various directions, turning the NS and EW components of the motion with azimuthal intervals of 10°, proceeding from 0° (north) to 180° (south). In order to retrieve information on the shear waves velocity distribution in the subsoil was performed a MASW survey. The seismic signal was acquired through a digital multi-channel array, formed by 25 vertical geophones to natural frequency of 4.5 Hz, with spacing of 2 meters, for a total length of 48 m. The time series was examined in frequency - phase velocity domain (slant-stack, and Fourier transform) so to discriminate the maximum energy associated with Rayleigh waves and then define the dispersion curve trend. The joint fit between the dispersion curve, obtained from the analysis MASW, and one H/V spectrum (H6), has allowed to get a Vs-depth profile. The 1D Vs-depth profile, the shallow velocity value V 0 and the α coefficient were obtained. These two parameters are necessary to convert the frequency values present in the spectra into depth values, through the Ibs-Von Seht and Wohlenberg (1999) formula.