Retrieving Lateral Variations from Surface Wave Data.

The task of this work is the evaluation of the possibility to identify lateral variation through Surface Wave dispersion analysis even if these techniques are mainly used to characterise 1D subsoil models. This is done exploiting the data redundancy of the ground roll contained in seismic reflection or refraction data through a fully automatic processing procedure that allows to stack dispersion curves obtained from different records and retrieve experimental uncertainties. Hence the dataset to be inverted will be an ensemble of dispersion curves associated to a series of spatial coordinates along the seismic line. In this contest the use of Laterally Constrained Inversion (LCI) algorithm allows to manage such 2D effects in spite of the 1D model assumed for the forward problem solution. Different test have been conducted on different dataset for two synthetic models to evaluate the effects of the processing parameters, of the presence of noise and of lacks of information on the inversion results. All these effects have been observed applying lateral constraints of different strength during the inversion process.