A Robust FWI Method for Model Updating in High Contrast Bodies

We present a robust method for Full Waveform Inversion (FWI), enabling the recovery of long-wavelength features of a velocity model. By using both transmitted and reflected waves the dynamically weighted FWI gradient enables high-resolution model building deeper than those achieved by diving waves alone. This reduces the dependency on long offset data acquisition.

The FWI approach uses a sophisticated regularization scheme to stabilize the inversion space. This methodology, which forms an extra constraint on the objective function, overcomes some of the limitations of the inversion in the presence of high contrast bodies. The implementation uses the split Bregman method, making it efficient and accurate.

We demonstrate the benefits of using the new gradient and regularization scheme by presenting the results on an intra-volcanic reservoir velocity model build from the Faroes-Shetland Basin.