GNGTS 2023 - Atti del 41° Convegno Nazionale

Session 3.2 ___ GNGTS 2023 Extraction of dispersion image from traffic noise data: A case study from Kefalonia M. Karimpour 1 , C. Colombero 1 , F. Khosro Anjom 1 , L. V. Socco 1 , A. Malehmir 2 , G. Apostolopoulos 3 1 Department of Environment, Land and Infrastructure Engineering, Politecnico di Torino, Turin, Italy 2 Department of Earth Sciences, Uppsala University, Uppsala, Sweden 3 School of Mining and Metallurgical Engineering, National Technical University of Athens, Athens, Greece Summary Seismic surface waves data are commonly used to build subsurface velocity models. Surface waves are present in the traffic-induced seismic data. Compared to active sources, the traffic noise can provide a cheap, sustainable, and available source of seismic data. We present a promising case study where the noise generated by individual vehicles is used to generate virtual seismic records through interferometry and then processed to obtain dispersion images. Introduction Surface waves (SW) are usually present in the seismic data. SW are dispersive, i.e., in a heterogeneous medium the phase velocity is a function of the frequency. The seismic data can be processed to produce the dispersion images and the dispersion curves (DCs), which are the phase velocity as the function of frequency, can be picked as the maxima in the dispersion image. The estimated DCs can then be inverted to re-construct VS distribution of the subsurface. The SW are usually dominant in the passive seismic data. One of the sources of the passive seismic data is the traffic-induced noise. Application of traffic-induced noise data has attracted the attention of researchers (Curtis et al., 2006; Behm et al., 2014) due to its cheapness, sustainability, and availability. The passive seismic data are usually processed using seismic interferometry method (Wapenaar, 2004). Seismic interferometry in time domain consists of the cross-correlation between two receivers to retrieve the Green’s functions. As a result of this operation, a virtual source and a virtual receiver is reconstructed. The position of the vehicles that have produced the noise data are usually not considered in the data processing. In this work, we focus on the processing of traffic-induced seismic data considering the different position of the vehicle at different times.