Site characterization of the national seismic network of Italy: results at five case studies

Abstract

The seismic characterization of monitoring sites is a fundamental step in any study dealing with the estimation of site effects. The correct assessment of local amplification is also important in the definition of hazard maps, in order to taking into account the possible role of site effects in modifying the ground motion recording with respect to an ideal bedrock site. In the framework of the activity between Department of Civil Protection (DPC) and Istituto Nazionale di Geofisica e Vulcanologia (INGV) (DPC-INGV 2012-202, Allegato B2, Obiettivo 1, Task B), a campaign of site characterization started in 2016 for the estimation of the seismic response at some stations belonging to the National Accelerometric Network (RAN) and to National Seismic Network (RSN). The accelerometric data of these stations are collected in the ITACA database (Pacor et al. 2011) and in a structured archive managed by INGV (please refer to Bordoni et al. in session 2.1 of this conference). In this work, we focus on five seismic stations (CMP0, CDCA, ROM9, SANR, LAV9) installed in a different geological context. We show the strategy adopted for assessing the geological setting and velocity profile below the site and in the estimation of the soil class category. CMPO, SANR and CDCA are situated in alluvial environment (Reno Alluvial Plain, Veneto-Friuli Plain, Alto-Tiber plain, respectively) where the soft deposits show significant thickness (> 100 m), whereas ROM9 and LAV9 are characterized by the presence of volcanic deposits belonging to the Colli Albani hills. At ROM9 the thickness of the volcanic deposits is the order of 50 meters, at LAV9 the thickness is larger (> 100 m). As first step, a conceptual model has been derived by geological field surveys and collecting the available geological information (scientific agreement between ISPRA and INGV). The results of this step are basically 2D geological models and a lithostratigraphic and lithotechnical classification of the outcropping units. Further, a geophysical survey at each site was carried out using surface-wave methods. We deployed 2D arrays of seismic three-components stations recording ambient vibration (or ambient seismic noise) in proximity of the target site to measure the dispersion curve following the recent guidelines (Foti et al, 2017). Passive 2D arrays recorded ambient noise for a total duration of some hours at each site. The array geometry was defined according to the logistic, and when possible two geometries with a progressive larger aperture were used at a same site (Fig. 1). The maximum aperture of the 2D arrays varies approximately from 100 to 400 m using a number of seismic stations from 8 to 14 depending on the site. At LAV9 site, we combined passive 2D array with a linear array of geophones equally spaced and using an active source (a weight body of 50 kg falling down from a height of about 2 m). Frequency-wavenumber and spatial auto-correlation methods derive a dispersion curve (Fig. 2). The inversion of the dispersion curves jointly with the horizontal-to-vertical spectral ratio (H/V curve) provides the local shear-wave velocity (Vs) profile. The soil class was finally assigned computing the mean value of the shear-wave best velocity models in the uppermost 30 m (as prescribed by the national seismic design code). As example, Fig. 3 displays the comparison at ROM9 between the local Vs profile derived from surface-wave analysis and the lithostratigraphical log obtained from the geological analysis. The Vs profile of ROM9 is able to individuate the contact at a depth of about 50 m between the volcanic deposits and the underlying clay (Monte Vaticano Unit, Pliocene). In detail from top to bottom in the velocity model of Fig. 3: after few meters of very soft soil (Vs < 200 m/s), the volcanic deposits show Vs values of 400-500 m/s, whereas the consolidated clay of the Monte Vaticano Formation shows Vs values larger than 600 m/s. The Vs30 at ROM9 resulting from this model is 410 m/s, being B the corresponding soil class category following the national code. As general comment resulting from this experience, a correct use of the surface-wave methods integrated with geological data is able to provide a reliable Vs profile that can be used to include the local effects in the seismic response of the site. However, it is important to highlight that we do not obtain always a perfect match between dispersion curves derived at a same site when we used different array geometry and source. The reasons of these discrepancies are not clear and need deeper investigation. A final consideration is that a suitable site classification is possible only with accurate geological and geophysical surveys. In presence of a reliable estimation of the local velocity profiles at the sites where seismic stations are installed, strong-motion data can be properly used for seismic hazard and site response studies.