Squared-slowness Gradient Profile from Surface- wave Inversion as a Starting Model for Full- waveform Inversion

SUMMARY Full-waveform inversion (FWI) is widely used for the detailed reconstruction of subsurface velocity models from seismic data collected in different environments. The method requires an adequate starting model. Otherwise, an incorrect model may be obtained, due to the presence of local minima in the least-squares objective function. This is in particular true in case of a complex near-surface velocity structure. Classic surface-wave inversion provides a layered representation of the medium, which can subsequently be smoothed to serve as a starting model for FWI. As an alternative to the layered representation of the medium, we consider a continuous, smoothly varying P-wave velocity profile. The profile has its squared slowness increasing linearly with depth and allows for analytical evaluation of the dispersion curves. We investigate the application of the reconstructed profile as an initial guess for acoustic FWI. A synthetic example shows how we can reconstruct both the background structure as well as a high-velocity inclusion in this way.