GNGTS 2017 - 36° Convegno Nazionale

GNGTS 2017 S essione 2.2 401 DIFFUSE CODA WAVEFIELD AND SEISMIC NOISE TO INVESTIGATE SUBSOIL STRUCTURES: THE CAMPI FLEGRI CASE L. Nardone 1 , R. Esposito 1 , D. Galluzzo 1 , L. Margerin 2 , M. Calvet 2 , F. Bianco 1 1 Istituto Nazionale di Geofisica e Vulcanologia – Sezione Osservatorio Vesuviano, Italy 2 CNRS / Université de Toulouse, France Introduction. In this work we studied the diffuse coda wavefield and the seismic noise by using the H/V spectral ratio to analyse the local site response to examine more accurately the velocity structures in the area of Campi Flegrei (CF). The CF area can suffer the effects of local earthquakes, that are typically shallow (depth less than 4 km) and with low energy (M max = 4, Aster et al., 1992) and regional earthquakes (mostly tectonic) generally coming from the Apennine chain. The CF Caldera is a geologically complex area in terms of shallow and intermediate velocity structures (Santacroce et al., 2003). The study of these crustal features is of great importance for the risk mitigation and for improving the knowledge of the volcano system dynamics. The Caldera has been the scenario of two big eruptions; the Campanian Ignimbrite (IC) occurred 39 ka ago and the Neapolitan Yellow Tuff (NYT), 15 ka ago. After the NYT eruption there have been about 70 eruptions, separated by quiescent periods (Orsi et al., 2009). The caldera structure has been subject to several scientific studies aimed at the investigation of different and complementary parameters: deep drillings (De Natale et al., 2016), tomographic studies, using both local and teleseismic earthquakes (Battaglia et al., 2008), noise based monitoring (Zaccarelli and Bianco, 2017), gravimetric (Capuano et al., 2013) and magnetic surveys (Fedi et al. 1991), deep temperature measurements and heat flux (Afanasyev et al., 2015). Recent studies of seismic attenuation (De Siena et al., 2010, 2017) reveal the presence of melted rocks at 4-5 km depth that are not showed by the seismic tomography. These volumes are in agreement with the evidence of petrological and geochemical data. Chiodini et al. (2010, 2016) explain the trend of the geochemical data suggesting the presence, beneath the Solfatara Volcano, of an hydrothermal system at 2 km depth, which feeds the two surface fumarolic areas. They suppose a 4 km deep zone of magmatic gas accumulation that gives rise to pulsed episodes of magmatic gas emission supplying fluids to a shallower part, where they vaporize liquid of meteoric origin to form a 2 km deep vertical plume of gas. The purpose of this work is to evaluate H/V spectral ratios of the diffuse wavefield and to study, jointly with the application of array technique on seismic noise, the subsoil structure (from shallow to deep layers) of the Campi Flegrei area. Dataset, methodology and inversion. The data set. The data set consists of 24 earthquakes (their location is depicted in Fig. 1), recorded from 2014 to 2017 by 22 seismic stations of the Mobile Seismic Network of the INGV – Osservatorio Vesuviano, installed in the CF area, and 3 hours of seismic noise recorded at the underground seismic Array “ARF” (La Rocca and Galluzzo, 2015) (Fig. 1). The magnitude of the earthquakes is between 3.2 and 6, and the epicentral distances vary from 65 km to 630 km. The local and the regional earthquakes, chosen to have the maximum azimuthal coverage, allowed us to investigate the dependence of the site response from the epicentral azimuth. In order to have the diffusive wavefield conditions, after a selection of the earthquakes with the best signal-to-noise ratio, we selected the coda starting from twice S-wave travel time to 3 times the signal-to-noise ratio. The spectral analysis was performed over a 32 s moving windows with 30% of overlapping. Therefore, for each earthquake, depending both on the magnitude and epicentral distance, we had a different number of windows. For each seismic event, in the frequency range 0.3 – 25 Hz (Fig. 2), we made two spectral analysis: H/V spectral ratio (Nakamura, 1989) and H/V M as defined by Margerin (2009). Spectral analysis. Not every station of the Mobile Seismic Network have shown significant