GNGTS 2013 - Atti del 32° Convegno Nazionale

The area of damage stretched over 4 km and was centred around the villages of Santa Venerina and Guardia (Azzaro et al ., 2006). In the literature, studies concerning the seismic hazard regarding the area of Mt. Etna have been published by Albarello et al . (2000) and Montaldo et al . (2005). However, Albarello et al . (2000) estimated peak ground accelerations using a specific attenuation relation derived from macroseismic data of Italian earthquakes (Margottini et al ., 1987), while Montaldo et al . (2005) used an attenuation relationship specifically for the volcanic areas starting from different predictive relationships. Notwithstanding the high seismic potential of the investigated area, no attenuation relationships of peak ground motion parameters are available in the literature so far. In this study, we used the background seismicity of the area to derive empirical attenuation laws of the peak ground-motion parameters for the shallow events on Mount Etna, considering both small (< 15 km) and intermediate distances (< 50 km) from the source. Data set and record processing. The data set we used in this study consists of 54 shallow selected seismic events (Fig. 1), recorded by the stations of the “ Rete Sismica Permanente della Sicilia Orientale ” (RSPSO), operated by Istituto Nazionale di Geofisica e Vulcanologia (INGV) – Osservatorio Etneo . The starting set of data includes earthquakes recorded from August 2006 to October 2012 having local magnitude (M L ) between 2.5 and 4.3, focal depth and epicentral distance less than 5 km (ca. 86% of events has focal depth less than 2 km) and 50 km, respectively. The seismic network includes about sixty digital stations equipped with Nanometrics Trillium broadband seismometers (natural period of 40 sec), located in the volcanic area of Mt. Etna. The data are digitized by a 24-bit A/D converter with a sampling rate of 100 Hz. Focal parameters and local magnitude values of the events were taken from the “ Catalogo dei terremoti della Sicilia Orientale - Calabria Meridionale (1999-2011)” (Gruppo Analisi Dati Sismici, 2013) of the INGV. Before processing the data, we visually checked the two horizontal components of each event at all stations in order to exclude traces with electronic glitches and phenomena of amplitude saturation. We also looked for good signal- to-noise ratio (higher than about 3) to guarantee the best seismic records. After this selection, a collection of 1980 horizontal-component seismograms (considering both the two horizontal components) was thus available for the analysis. Fig. 2 depicts the characteristics shown by our digital records in terms of magnitude-distance (Fig. 2a), number of data–magnitude (Fig. 2b), and number of data – soil classes (Fig. 2c), distributions. Up to a distance of about 30 km, the data show a fair degree of homogeneity with respect to spatial sampling. In order to avoid the risk of potential bias due to poor distribution of data in some range of distance, the records from distances larger than 30 km were excluded from the analysis. Finally, a total of 962 horizontal PGA was available for the analysis. As an initial step in the processing, the digitized velocity time histories of the horizontal components were first baseline corrected, removing offsets and the linear trends from the data, and instrument corrected to obtain ground velocity in units of m/sec. Then, they were bandpass filtered using corner frequencies of 0.1 and 25 Hz for the high-pass and low-pass filtering, Fig. 2 – Statistical characteristics of our data set. (a) Distribution of epicentral distance vs. local magnitude (ML); (b) number of data as a function of ML; ( c) number of data vs. EC8 soil classification. 154 GNGTS 2013 S essione 2.1