Lithology classification from seismic tomography: Additional constraints from surface waves

An efficient way of interpreting a seismic profile cross-section is a joint interpretation of velocity models of different types of seismic waves. This study performs tomographic inversion of surface wave travel times observed during the seismic profile carried out in Namibia in the framework of the SIMBA project. The thus obtained surface wave velocity model is used to complement the previously computed P- and S-wave models. Profile sections characterised by similar seismic velocities are identified as lithological classes and remapped in model space. Two methods are used to identify such classes: a manual identification of high probability zones in a probability density function, and an automatic neural network approach. The results of these two methods are consistent with each other. The availability of the surface wave velocity model as additional independent physical parameter increases the correlation between the remapped lithological classes and the geological map, leading to the conclusion that the identified classes correspond to real geological formations.