GNGTS 2023 - Atti del 41° Convegno Nazionale

Session 3.3 ______ ___ GNGTS 2023 value decreased to 128 m for cluster A that are characterized by lower velocities (shorter wavelengths). In Fig. 2e and 2f, we show the iso-surfaces of the estimated pseudo 3D VS and VP models from the W/D method. The estimated models show a significant contrast between the eastern and the western parts, which is in agreement with the geological information of the site. We applied the LCI and SWT methods to the data set to estimate the VS. Then, we transformed the results to VP model using the a priori Poisson’s ratio from the W/D method. We linearly interpolated the VS and VP results of the three methods to obtain the velocity values at voxels of 1 1 0.1 m 3 corresponding to , and (depth) directions. We computed the average × × differences at each and location as: , (1) where is the velocity value (either VS or VP) at each voxel, and is the number of voxels in (depth) direction. In Fig. 3a and 3b, we show the difference between the estimated VS and VP models from W/D and LCI methods. Similarly, in Fig. 3c and 3d, we compare the results of the W/D and SWT methods. The differences between the estimated models are less than 5% in majority of the area (green colour). Conclusions We showed the application of the W/D method to estimate the VS and VP models, using data from a stiff location. The prior applications of the method to synthetic and real 2D data sets had provided good results; the current application of the method to the 3D data set acquired in scheme of carpet recording shows encouraging results that are consistent with the geological information available for the site. The methodology can be also used to provide the Poisson’s ratios required for other surface-wave techniques, such as LCI and SWT methods and calibrate these methods for the estimation of both VS and VP models.