On the influence of multiphysics joint inversion for uncertainty quantification in full waveform inversion

Tianze Zhang, Kristopher A. Innanen

The combination of the elastic model inversion results and the corresponding uncertainty quantification provides a comprehensive solution for full waveform inversion (FWI). Developing methods to reduce the uncertainties associated with inversion results is critical for FWI under non-ideal conditions, as this enhances the robustness of the inversion process and leads to more reliable interpolation and decision making. Each geophysical parameter represents a different aspect of the subsurface, and thus multiphysics joint inversion naturally constrains various geophysical inversions, which can reduce the uncertainties associated with the results. In this study, we investigate how the incorporation of gravity data helps reduce the uncertainties associated with the density parameter in elastic FWI. We demonstrate that including gravity data decreases the eigenvalues relevant to the UQ process in FWI, thereby reducing the uncertainties in the inversion results. We demonstrate this by looking at the eigenvalues, the corresponding Gaussian random field (GRF), and the statistical analysis, proving that the utilization of gravity data can reduce the uncertainties associated with density FWI results.