GNGTS 2019 - Atti del 38° Convegno Nazionale

GNGTS 2019 S essione 2.2 409 3D MODELLING OF THE SEISMIC BEDROCK MORPHOLOGY IN THE TUSCAN BASIN (CENTRAL ITALY) L.M. Giannini 1 , C. Varone 2 , C. Esposito 1 , G. Scarascia Mugnozza 1 , L. Schilirò 3 1 Department of Earth Sciences of «Sapienza» University of Rome and Research Center for the Geological Risks (CERI), Italy 2 NHAZCA S.r.l., Spin-off «Sapienza» University of Rome, Italy 3 Department of Pure and Applied Sciences, University of Urbino “Carlo Bo”, Italy Earthquakes strike each day wide areas of the Planet determining, in several cases, high level of structural damage and loss of human life. The amplitude and energy intensity of the earthquake results from the combined effect of several physical processes affecting the propagation of the motion from the seismic source to the soil surface. These effects are widely known as site effects and they can be grouped in two main typologies: stratigraphic and 2D/3D effects. The thickness and geometry of soil deposits, especially those lying on the seismic bedrock, assume a fundamental role in case of local seismic response analysis and seismic risk assessment. The definition of the geological model, including the position and geometry of the seismic bedrock, then represents the starting point for the assessment of seismic site effects. Thus, in this work we proposed a methodology to perform a 3D reconstruction of the seismic bedrock morphology by integrating geognostic and geophysical in situ measurements, such as borehole data and ambient noise measurements. Specifically, these data have been analyzed and extrapolated through spatial statistical analysis to obtain a homogeneous spatial distribution of fundamental resonance frequency within the basin. The latter information was then elaborated to obtain a 3D reconstruction of the seismic bedrock position and geometry. The abovementioned approach has been applied in the northern central part of Firenze – Prato – Pistoia basin (Tuscany, Central Italy), which is a densely urbanized area, also characterized by the presence of important industrial and infrastructural assets. The basin is located in the northern part of the Apennine mountain chain and extends in NW-SE direction over a length of 45 km and a width of 10 km. It is a structural basin characterized by an asymmetric semi-graben geometry controlled by the Fiesole fault on its northeastern border and by two minor structures on the northern and southern portion. The basin filling is mainly represented by clays, silts and sands composing Pleistocene fluvio-lacustrine deposits and recent alluvial deposits from the Arno river (Capecchi and Pranzini, 1986; Coli et al. , 2004), which overlay the local geological bedrock made up of External Ligurian Units in NE and SW-SE part and Cervarole tectonic Unit in NW sector (Puccinelli et al. , 2015). Over two hundred noise measurements homogeneously distributed within the valley have been collected from previous microzonation studies and then geo-localized and stored in a database. The noise measurements were analyzed according to HVSR (Horizontal to Vertical Spectral Ratio) technique proposed by Nakamura (1989) in order to assess the main resonance frequency (f 0 ) of the soil. Amachine learning approach was then used for analyzing the data and a relation between f 0 values and distance of the point from the outcropping bedrock has been defined. The resulting f 0 distribution as a function of the distance from the outcropping bedrock is characterized by a negative trend with several outliers. Afterwards, for exploiting in detail the resulting data, a spatial cluster analysis has been performed with the aim of defining local portion of the basin characterized by a specific relation between f 0 and distance which, in turn, reflects specific geo-structural conditions. Several natural clusters have been identified through the abovementioned analysis: in this sense, it is worth noticing that the points previously considered as outliers are instead ascribable to two well-defined clusters in the northern portion of the basin, just where two main geo-structural conditions are located according to preceding studies (Capecchi et al., 1975). Other two clusters have been detected in the northeastern and southwestern portion of the basin. The first one is characterized by an evident decrease of f 0 with the distance, while the second one shows a slight decrease of f 0 moving away from the outcropping bedrock. These results are then consistent with the asymmetric geometry of the basin related to the presence of the Fiesole fault.