SITE CHARACTERIZATION DATABASE OF INGV ITALIAN SEISMIC NETWORK

Abstract

A critical issue in the performance of a seismic network is the characterization of site response where stations are located. This information is essential to improve some aspects related to seismic surveillance and the publication of products in near-real time following an earthquake. A proper evaluation of the site effect is also necessary to improve the quality of recordings databases, facilitating their use for research purposes. The Italian National Seismic Network of the INGV (Rete Sismica Nazionale, RSN) consists of about 400 seismic stations equipped with a velocimeter and, for one-third of the sites, an accelerometer. They are connected in real time to the INGV data center in order to locate earthquakes for civil defense purposes and their records are distributed through the EIDA node (eida.rm.ingv.it/). Recently INGV has addressed the site characterization of RSN with an internal project (funded within the INGV research line T3 “Seismic hazard and contributions to the definition of risk”), as well as within the INGV-DPC Agreements (INGV-DPC Agreement 2016-17-18, Annex B2 Objective 1 - Task B “Characterization of accelerometric sites”, funded by the Civil Protection Department), with the purpose of characterizing the seismic response of all the stations acquired in real time by its data center. The basic goal is building a geographic relational database, integrated with the other INGV infrastructures, designed to archive homogeneous parameters through the seismic network useful for a complete site characterization, including housing, geological, seismological and geotechnical features as well as site and topographic class according to the European and Italian building codes. The system resides on a dedicated server and the data are organized in an internal storage based on PostgreSQL DBMS (acronym CRISP). It will be directly related to SeisNet, the INGV database used for the network management, but it is still possible to insert new sites not belonging to the RSN. The backend of the system includes several procedures that allow the information updating through web services created ad-hoc, such as those of the Institute for Environmental Protection and Research (ISPRA) for geological and lithological attributes and for visualization of geological maps and related legends. On the other hand, specific programming interface services – API- expose the shared information to allow the transfer to other strong-motion data providers (e.g. ITACA, http://itaca.mi.ingv.it, and ESM, http://esm.mi.ingv.it) in semi-automatic way. The collection of geological, morphological and seismological data followed a nationwide approach, aimed at obtaining homogeneous data for the RSN sites. We started from the revision of all available geological and geophysical data and the analysis of noise waveforms, storing the analysis results as images and searchable data. Thanks to the collaboration with the Geological Survey of Italy (ISPRA-SGI), a review of the geological map of Italy (at a scale of 1:100,000 and 1:50,000) and their relative explanatory notes, including also many other available published data (borehole logs, local geographical portal, etc.), allowed to develop a stratigraphic conceptual model under each site. As for the attribution to each site of a topographic class according to the Italian building code, a morphometric analysis using an automatic procedure has been carried out on two DEM datasets with resolution at 30 m and 10 m. Regarding the seismological parameters, noise velocimetric records at all the stations were homogeneously analyzed by using mostly continuous data, as follow: 1) estimation of data quality with annual and seasonal noise analysis; 2) selection of noise traces (day/night and seasonal), horizontal-to-vertical spectral ratio computation and determination of directionality of the amplification peaks; 3) in case of directionality, we proceeded with the polarization analysis of the signal to identify the preferred direction of the movement, slope and straightness. A preliminary statistical analysis highlights that only 26% of the RSN accelerometric stations do not have amplification peaks, while 29% show a polarization of the signal in a preferential direction. Finally, we are collecting all the available information about the station housing, to account for possible soil-structure interaction. The database includes also 15 sites that have been fully characterized by performing a geological survey followed by the 1:5,000 geological and lithotechnical maps, a geological cross section and report, the S-wave velocity profile inferred through seismic noise arrays and, for one site, downhole measurements. With the contribution of the Site-Characterization Team: S. Amoroso, R. Azzaro, R. Bianconi, M. Cattaneo, R. Cogliano, D. Di Naccio, C. Felicetta, A. Fodarella, S. Lovati, A. Mandiello, C. Marcocci, C. Mascandola, M. Massa, A. Mercuri, G. Milana, S. Pucillo, G. Riccio, G. Tusa, M. Vassallo, et al. (INGV); M. Amanti, G. Conte, C. Cipolloni, G. M. Monti, C. D’Ambrogi, M. D’Orefice, P. Di Manna, D. Fiorenza, R. M. Gafà, B. Roberto, M. Roma, L. Vita (ISPRA)