Joint full waveform inversion of early arrivals and reflections: A real obc case study with gas cloud

Joint full waveform inversion (JFWI) combines reflection (RWI) and early-arrival (EWI) waveform inversions to build a large-scale velocity model of the subsurface. The misfit function of JFWI requires an explicit separation between the short-spread reflections and the early arrivals, the feasibility of which is illustrated with a real long-offset dataset. Each JFWI iteration is alternated with a waveform inversion/migration of short-spread reflections (IpWI). IpWI provides a short-scale impedance model that is used as an input to build the sensitivity kernel of RWI along the two-way reflection paths. Once a velocity macromodel has been built by JFWI, it can be used as a starting model for classical FWI applied on the full dataset to enrich the high wavenumber content of the subsurface model. We present an application of the above-mentioned workflow to a real 2D OBC profile cross-cutting a gas cloud in the North Sea to review its main promises and pitfalls. Several initial models and offset-driven strategies are assessed with the aim to manage the cycle-skipping issue while producing subsurface models of sufficient vertical and horizontal resolution. Our workflow produces an acceptable FWI velocity model that fits the main early arrivals and reflections when JFWI starts from a smoothed version of an existing traveltime tomography model. Improved scattering-angle illumination provided by the 3D acquisitions should allow us to start the inversion from a crude starting model and would overcome the planar propagation we assume in the 2D geometry.