Joint FWI for velocity model building : a real case study in the viscoacoustic approximation

W. Zhou,R. Brossier,S. Operto,J. Virieux,P. Yang
2017-01-01
Abstract:Joint full waveform inversion (JFWI) combines reflection (RWI) and early-arrival (EWI) waveform inversions to build a large-scale velocity model of the subsurface. It is alternated with a waveform inversion/migration of near-offset reflections to build a short-scale impedance model that is used as an input to build the sensitivity kernel of RWI along the two-way reflection paths. The velocity macromodel built by JFWI can be used as the initial model of standard FWI to enrich the high wavenumber content of the model. We present an application of this workflow to a real 2D OBC profile across a gas cloud in the North Sea. First, we highlight the footprint of attenuation by comparing recorded seismograms with the synthetics computed in a viscoacoustic velocity model previously developed by 3D FWI. Then, the main promises and pitfalls of JFWI are highlighted using two initial models of increasing accuracy. When starting from a crude 1D initial model, JFWI is influenced by cycle-skipping artifacts and fails to update the lowwavenumber content of the subsurface model. When a more accurate initial model is used, the procedure of JFWI followed by standard FWI (with resulting JFWI model as initial model) succeeds in building a velocity model which is more accurate than the one built directly by standard FWI. This study suggests that JFWI is more efficient than FWI to update the low horizontal wavenumbers along the reflection wavepaths during the velocity macromodel building task, hence leading to a more accurate reconstruction of the gas cloud. However, it remains prone to cycle skipping when a conventional differencebased misfit function is considered. Therefore, more robust misfit function must be used in the future to reduce the demand on an accurate initial model.
What problem does this paper attempt to address?