GNGTS 2023 - Atti del 41° Convegno Nazionale

Session 3.2 ___ GNGTS 2023 Field data, Methodology, and Results The data were acquired in Kefalonia, Greece. The dataset includes a 2D line of 234 vertical 4.5 Hz receivers with 5 m spacing. The seismic line was deployed along a road (Fig. 1) with regular passage of trucks. The noise induced by these trucks can be analysed and used to retrieve virtual gathers which are here compared with active data acquired with a weight-drop source. The data were recorded continuously for 4 days. We split the recordings into 5 s windows and searched for records containing patterns related to vehicle passage. We show an example of a record in Fig. 2a: where the red arrow points to the pattern associated with the passage of the truck between trace 171 and trace 164. The event has a constant velocity of about 51 km/h and can be automatically recognised and tracked within the record. We consider the truck as a moving source and, at each time, we identify the receiver which is closest to the truck. This receiver is then used as a virtual source by cross-correlating its trace with all the other traces and producing a virtual gather as there was an active source located at the position of the receiver. This can be repeated for all the receivers along the truck passage to create a set of virtual gathers for each vehicle. In Fig. 2b, we show an example of the computed virtual gather where we considered the 171st trace as the virtual source (shown by the red arrow in Fig. 2b). We depict an example of the recordings from a near-by active source in Fig. 2c for comparison. To test the possibility of using the virtual gathers for surface wave analysis, we consider a window of receivers (rectangle in Fig. 2b) and for each virtual gather we compute the velocity spectrum using the phase-shift method (Park et al., 1999). In Fig. 3a we show the dispersion image obtained by one virtual gather. The image is noisy but shows clear dispersive trends. To improve the signal-to-noise ratio of the dispersion image, several virtual gathers were computed, and the dispersion image computed in the same receiver window were stacked (Fig. 3b). We display an example of the obtained dispersion image from an active source in Fig. 3c. We can see that the obtained dispersion images from the virtual and active data are in agreement. The SW with higher frequency (16-30 Hz) are dominant in the active example (Fig. 3c) while the lower frequencies (9-16 Hz) are dominant in the passive example (Fig. 3a and b). We can also see that the fundamental mode is clearer in the dispersion image from the passive data (Fig. 3a and b) while the higher mode is dominant in the active example (Fig. 3c). Conclusions We showed that it is possible to use vehicles as moving seismic sources when a seismic line is deployed along a road. Truck passages can be easily identified and tracked within passive records and through interferometry it is possible to compute virtual seismic gathers using the receiver closest to the source as a virtual source. The obtained seismic gathers can be used for surface wave analysis and stacking improves the dispersion image retrieval.