Application of acoustic global fwi on a shallow seismic reflection data set acquired in Luni (Tuscany)

The archaeological site of Luni in Tuscany (Italy) is located on an alluvial plain where geological surveys carried out in the field recognized different units such as alluvial fans of the Magra river, sedimentary swamps and sand bar, all indicating a flat layer stratification of the shallow subsurface. The archaeological context of the site makes Luni a very interesting location for the application of various geophysical techniques, such as GPR, electrical tomography and magnetic surveys. Furthermore, the good seismic response observed in previous seismic reflection and refraction tests, makes also Luni attractive for a seismic reflection survey aimed at investigating the near-surface layers for a more detailed geological study. All these motivations lead us to the acquisition of a reflection seismic data set that can also be used for the application of different processing and inversion algorithms especially tailored to the near-surface investigations. In this work, we describe the processing and the results of a preliminary 2D acoustic Full Waveform Inversion (FWI) of a seismic reflection data set acquired in Luni. The processing is carried out to obtain the best image in time and also to estimate a preliminary interval velocity field (by means of the Dix equation) that is later used as input to the pre-stack depth migration. Successively, a global 2D acoustic FWI is applied to produce a more accurate velocity field in depth. We make use of a parallel implementation of a genetic algorithm (GA) optimization (Sajeva et al., 2016), where the data misfit is evaluated as the L2 distance between the envelope of the predicted and observed data on the first arrivals to reduce the non-linearity of the misfit function. Pre-stack depth migrated common image gathers (CIGs) computed with the obtained velocity field show a good horizontal alignment of the observed reflections.