A New Approach for Multimodal Inversion of Rayleigh and Scholte Waves

The knowledge of the near surface velocity can be of paramount importance for assessing near surface effects on exploration seismic data both on land and offshore surveying. Ground roll (Rayleigh waves) and mud roll (Scholte waves) present in seismic records can be processed and inverted to supply a shear velocity model of the first hundreds meters. In this paper an automatic multimodal inversion for Rayleigh and Scholte waves using a new approach is presented. The proposed algorithm minimizes the L1 norm of the modulus of the Haskell-Thomson matrix determinant of the synthetic profile evaluated on real data. This algorithm is very fast, and it allows considering simultaneously all modes. Tests on both synthetic and real data, relative to Rayleigh and Scholte waves, have shown the importance of accounting for higher modes and the capability of the proposed method to handle their inversion. Improvement on the reliability of the final model has been obtained with a fast and efficient technique.