Paper submitted to GESGB Conference (G-CON) 2026, by Stuart Greenwood, Marisa Domingos, Rachel Collings, Øystein Korsmo, Thomas Hallett, James Wright (TGS), Nick Benfield (Shell), Tom Cafferkey and Matthew Dack (NEO Next)
Introduction
In recent years, elastic Full Waveform Inversion (FWI) has become standard on most seismic imaging projects. This method is considered favorable due to the ability to utilize the full recorded wavefield to resolve the pressure velocity model (Vp-model) that can reconstruct the sub-surface image correctly. Despite the early conceptional deployment of FWI (Tarantola, A., 1986), significant research continues within this field, resulting in a fast-moving technology. The research has primary focused on; improving physics in modeling (Liu et al., 2024), improving robustness through advanced objective-functions (Mao et al., 2020), parameter-decoupling, alternative imaging conditions (Ramos-Martinez et al., 2016) and multi-parameter inversion algorithms (Yang et al. 2021). All of this has contributed significantly to large evolutionary steps for the FWI results. Regardless of this, far more important is the quality of the seismic data used by the algorithm; where long offsets, low frequencies and azimuthal rich data enable full illumination of the subsurface and reduce the risk of ending in a local minima.
In this case study we make use of elastic FWI on vintage narrow azimuth (NAZ) short offset streamer datasets, with max offset of 2500 meters, in 20 meters water depth combined with large lateral geological complexity, involving chalk out-crops and salt structures. The methodology is based on elastic modeling, dynamic matching (DM) objective-function and inverse scattering imaging condition (ISIC). Despite the non-favorable input to the inversion, we demonstrate how modern elastic FWI can resolve valuable velocity estimations that improve the structural imaging. Furthermore, we discuss the limitations and alternative approaches required due to the limited available offset within the data.