GNGTS 2015 - Atti del 34° Convegno Nazionale

in the same geological context and geographic area. Only for depths exceeding 40 m, dynamic curves by EPRI (1993) have been used, in order to not overestimating damping values (usually derived from analysis in shallow samples). Variability ranges in S-wave velocity profiles have been used to establish minimum and maximum values for the bedrock depth, as well as the statistical properties of V S in each strata. Variability of modulus reduction and damping curves have been reproduced by following Darendeli (2001). For each station where the deconvolution has been made, the ground motion to be deconvolved has been provided by other research units participating in the project, and consists in a set of selected earthquake events. Statistics about peak ground acceleration (PGA) and acceleration spectral amplitude (SA) at 0.15, 1 and 2 s have been extracted for each site, based on an ensemble of 100 deconvolution runs for each earthquake record, to account for uncertainty affecting the local seismic stratigraphy. For each earthquake, the two horizontal components have been considered separately, taking into account both the geometric mean and the maximum of PGA and SA of each run. Then the median and the 75 th percentile of the ensemble have been computed. An example: the station of Peglio. The accelerometric station “PGL” is located within Peglio, a village in the Marche region (central Italy). The station is near the top of the hill where the village is situated (468 m a.s.l.), on a not very steep slope (< 15°) above the Metauro River valley. The station lies on syn/post-evaporitic deposits characterized by a siliciclastic succession consist of marl, clay-marl, gypsum, clay, sandstone and silt. The station is situated on a unit ( Formazione a Colombacci ; FCO) formed by clay and marl, dipping with an inclination of about 15° northward and situated on a anticline flank. Due to the nature of these lithologies, this area is characterized by the presence of several landslides, active too (MUSa1q, MUSa). The FCO formation lies on Gessoso-Solfifera Formation (GES) with unconformity relationship; this latter unit, mainly composed by gypseous sandstone and marl, lies on Miocenic torbiditic sandstone and marl ( Marnoso-Arenacea Marchigiana Formation; FAM) with the same stratigraphic relationship. The village historical centre is situated on a Gessoso-Solfifera Formation facies mainly composed by alabastrine gypsum (GESa). Peglio is located in the north of the major north Apennine bending. The main structures are anticlines, synclines and thrusts trending mainly NW-SE (Apennine direction), sometimes cut-off by normal fault with N-S trend. However, the tectonic setting is dominated by Mio-Pliocenic compressive structures. Due to the position of the station “PGL”, it was not possible to deploy a seismic array near the station, so it was decided to realize two arrays in the neighbourhood: one in the same geological unit and the other one on a different unit. However, the first one did not give any readable dispersion curve, so just the curve obtained by the second array was available for the inversion procedure. In the area including the station and the two seismic arrays, 21 single- station ambient-vibration measurements were also performed. Aset of joint inversions of the dispersion curve and a HVSR curve obtained by a single-station measurement located near the relative array was performed, by using the above-mentioned Genetic Algorithm procedure. Each result of the inversion set is the best profile produced by an independent run. The best resulting group, which contains the profile with minimum global misfit value (in red) and the ones whose misfit values are no greater than its double (in green), is shown in Fig. 1. The best fit profile obtained by the inversion (characterized by the minimum misfit value and represented by red lines in Fig. 1) has been chosen as reference to perform a geological interpretation of the subsoil configuration in the place where the seismic array was deployed: this interpretation is exposed in Tab. 2. The whole ensemble of best results (represented by red and green lines in Fig. 1) is considered as representative of the result variability. Consequently, the depth of the seismic bedrock top is estimated to be between 23 and 68 m. GNGTS 2015 S essione 2.2 151

RkJQdWJsaXNoZXIy MjQ4NzI=