A Journey of FWI Offshore Abu-Dhabi – From Concept to Reality

Abstract

Abstract Commencing in 2018, ADNOC undertook a mega seismic campaign (Cambois et al., 2019), acquiring approximately 30,000km2 of ocean bottom sensor (OBS) data in the challenging imaging environment of offshore Abu Dhabi. The region posed difficulties due to ultra-shallow water depths, a hard water bottom, and a complex near surface. To accurately image deeper reservoirs, a precise velocity model of the complex near surface was crucial. This study discusses how FWI was utilized to overcome these challenges and its impact on velocity model building and imaging. Although FWI is a standard technique in many geological basins, its application in the offshore Abu Dhabi region proved to be challenging. The presence of a hard water bottom resulted in strong elastic effects, while the near surface exhibited significant velocity variations and high anisotropy (with epsilon parameter > 30%). Nonetheless, Hermant et al. (2022) recently achieved successful application of FWI in the offshore Abu Dhabi region. The key to their success lay in careful data selection, conditioning, and the choice of an appropriate initial model. Data was meticulously chosen to avoid interference from mud-roll and guided wave energy, and pre-conditioning was kept to a minimum. Given the high apparent anisotropy in the near surface, the selection of a good initial model was critical. Well data and Backus averaging (Backus, 1962) were utilized to generate initial velocity and anisotropy models. Diving wave updates were limited to the very near surface, while reflection data was employed to extend the model updates and incorporate higher frequency details. Encouraged by the promising initial results, diving wave FWI was subsequently applied to a 14,000km2 area of offshore Abu Dhabi as part of an ongoing depth velocity model building process. The inclusion of the FWI model for the near surface led to a significant enhancement in the final depth velocity model and the resulting migrated image. FWI utilizing reflection data exhibited notable improvements in capturing fine details, such as karst features, within the velocity model. Consequently, FWI with reflection data was implemented in two other surveys to generate high-resolution velocity models, thereby improving structural imaging, and identifying deeper faults. In summary, this study demonstrates that diving wave FWI successfully produced a near surface velocity model over 14,000km2, leading to substantial improvements in imaging and facilitating the delineation of deeper structures. The incorporation of high-frequency updates, including reflections, enhanced the resolution of the existing model and the interpretation of geological structures, such as faults. These findings underscore the transformative potential of FWI in velocity model building within the offshore Abu Dhabi region.