GNGTS 2024 - Atti del 42° Convegno Nazionale

Session 2.1 GNGTS 2024 Performance of the Ground Motion Models to predict the significant durations in Italy B. Shoaib 1 , F. Ramadan 2 , G. Lanzano 2 , M. Sadek 3 , H. EL Ghoche 3 1 University of Camerino, Italy 2 Istituto Nazionale di Geofisica e Vulcanologia, Italy 3 Lebanese University – Faculty of Technology, Lebanon Within the framework of seismic probabilistic hazard analysis and other engineering-oriented applications, the common approach to predict the ground motion is to use ground motion prediction models (GMMs). Those models are empirically developed as a function of a few explanatory variables such as magnitude, source-to-site distance, site condition, and focal mechanism. Traditionally, GMMs are developed for peak ground acceleration (PGA), peak ground velocity (PGV), and the ordinates of the acceleration response spectra (SA); however, to fully describe the dynamic response, other intensity measurements should be considered, such as the Arias intensity and ground motion duration. The main aim was to study the ground motion durations caused by moderate-to-high earthquakes in Italy and to contribute to the development of a new duration GMM within the Italian context, by analysing the performance of existing GMMs against recordings of recent earthquakes in Italy. Among several definitions of waveform duration, we considered as a reference intensity measure the relative significant durations (DSR), that are related to Arias Intensity, according to two definitions (Bommer et al., 2009): i) the DSR(5-75) measures the time interval from the 5%up to 75% of the total AI; ii) the DSR(5-95) extends to 95% of the Arias intensity. An Italian shallow crustal dataset (ITA18) (Lanzano et al., 2022) has been considered, already employed for the GMMs calibration of other strong motion parameters (Lanzano et al. 2019; Ramadan et al. 2021); the dataset presents records for 154 events taken from 1637 stations. The dataset is dominant with events of Mw<5, NF focal mechanism, and site type A and B (According to the EC8 site classification). Based on the above mentioned definitions, DSR(5-75) and DSR(5-95) were computed for all records found in the dataset ITA18: Husid function (Arias intensity vs time) was estimated for each acceleration waveform to obtain the time frames corresponding at each specific percentage of the total arias intensity (5%, 75%, and 95%). Figure 1 reports the correlogram of the intensity measures and explanatory variables: it shows the measures of the “strength” of the linear relationship between the DSRs and the variables: a strong positive correlation with RJB and a moderate positive correlation with Mw were observed, as expected. The correlations with the other variables are weaker, probably also due to the uncertainty associated with their estimation.