Search

Summary In the last decade, Full Waveform Inversion has gained popularity as a very powerful tool to estimate velocity models of the subsurface with unrivalled resolution. Interpreters can now use high-resolution models obtained by FWI directly, in addition to migrated images. However, the FWI velocity models do not necessarily ensure optimal flattening of common image gathers. A fundamental reason for this is that mid to high wavenumbers are brought into the model by the reflection amplitudes, which are only partly explained by changes in velocity. A residual low wavenumber update of the model is then frequently done with the consequence that the positioning of small-scale events in the model may no longer be consistent with the migrated image. In this paper, we first illustrate the above-mentioned pitfalls, and propose an innovative workflow based on FWI-guided tomography. It ensures the optimal focus of the migrated image and it also ensures a better match between the fine-scale velocity structures and the migrated image. The approach is validated on a 2D marine broadband dataset using 32 Hz High-Resolution FWI.