GNGTS 2022 - Atti del 40° Convegno Nazionale

GNGTS 2022 Sessione 2.2 375 SEISMIC SITE AMPLIFICATION AT PERMANENT STATIONS IN NORTHEASTERN ITALY P. Klin, G. Laurenzano, C. Barnaba, E. Priolo, S. Parolai Istituto Nazionale di Oceanografia e di Geofisica Sperimentale - OGS, Trieste, Italy The awareness of the seismic risk in northeastern Italy has fostered the deployment of an ever-growing number of permanent stations devoted to seismic monitoring and emergency management purposes (Bragato et al. , 2021). The evaluation of seismic site amplification at the seismological stations represents a prerequisite for the correct usage of the earthquake recordings in many applications ranging from the rapid quantification of the earthquake impact on the territory (e.g. Poggi et al. , 2021), to earthquake source characterization studies (e. g. Franceschina et al. , 2013). In the present study, we exploit the considerable number of earthquake recordings collected in northeastern Italy, from 1996 to 2017, to estimate the site response at the recording network stations in terms of the following quantities 1) frequency-dependent amplification function A ( f ) with respect to a reference site and 2) median amplification factors for peak ground acceleration (PGA) and velocity (PGV) with respect to a well-established ground-motion prediction equation (GMPE). The analysis is based on the recordings of 368 earthquakes in the magnitude range of M 3.2–5.8 collected by 67 stations in a distance range of 10–300 km. The stations are located in an area that spans the territories of Friuli-Venezia Giulia, Veneto and Trentino-Alto Adige/ Südtirol regions and the bordering parts of Lombardia and Emilia-Romagna. We evaluate A ( f ) in the frequency band 0.5–20 Hz through the decomposition of the S-wave amplitude spectra in terms of source, propagation, and site response. We solve the decomposition with a nonparametric, single step generalized inversion technique – GIT (Oth et al. , 2011). We select the station Forcella Aurine (FAU), in the central part of the map in Fig. 2, as the reference station for the GIT analysis, in virtue of the absence of peaks in the microtremor horizontal to vertical spectral ratio (HVSR). On the basis of the minimum ( A min ), maximum ( A max ) and average ( A avg ) values of A ( f ), we group the retrieved amplification functions into the following five categories of stations: a. Neutral ( A min > 0.5 and A max <2), b. Deamplifiying (non-neutral stations with A avg < 1.0), c. Narrowband amplification (stations with A max >2 and A max A min > ( A avg ) 2 ) d. Broadband low amplification (non-neutral and non-narrowband stations with 1.0 < A avg < 2.0), e. Broadband high amplification (non-narrowband stations with A avg > 2.0 ) The stations considered in the present work are mapped in Fig. 1 with the color of the group they result to belong. We can observe that only 11 out of 66 stations belong to the neutral category (i.e., amplification function with values bounded in the interval 0.5–2 and that we can assume flat and unitary with respect to station FAU). A minor percentage (7%) exhibits narrowband amplification, whereas most stations (44) belong to the categories of broadband amplification. 23 of these (34% of network stations) exhibit a mean amplification value larger than 2. The deamplification in a number of stations could be due to their installation on a medium that is stiffer than the one at the reference station and thus presents a relatively lower rock site amplification. We evaluate the amplification factor for PGA and PGV with respect to the GMPE ITA10 for Eurocode 8 site class A (Bindi et al., 2011) following the approach presented in Laurenzano et al . (2019) and find that only a limited fraction of stations presents an amplification factor near unity. Most stations exhibit either positive or negative bias, consistently with the amplification functions obtained with GIT. We found good correlation between the amplification factors and the frequency-averaged amplification functions, which suggests the possibility of predicting time-domain peak ground- motion values from amplification functions estimated by generalized inversion.