Offshore Abu Dhabi is characterized by ultra-shallow waters and thinly layered carbonate formations with high velocity and anisotropy contrasts. This produces complex recorded seismic wave-fields, riddled with noise and multiple energy, which are notoriously difficult to process. In a recent reprocessing project over a producing field offshore Abu Dhabi, we applied innovative processing techniques and obtained an unprecedented uplift in the final seismic image. 3D seismic data was acquired over the study area deploying airgun sources and 4-component receivers with ocean bottom cable (OBC) in an orthogonal cross-spread design. In vintage processing PZ summation was used to obtain the upgoing wavefield with the receiver ghost removed. 1D deconvolution was then applied to attenuate receiver-side pegleg multiples. Strong residual multiple energy nonetheless remained in the overburden and caused challenges in interpretation of the data. For the reprocessing, we applied a method known as up-down deconvolution, which is commonly used in deep water with a smooth water bottom. The implementation of up-down deconvolution in this ultrashallow water environment was challenging due to poor trace sampling and insufficient direct arrival energy in the shallow part of the recorded wavefield in the receiver domain. The mostly flat seafloor and horizontally stratified geology allowed us to implement the method in ultra-shallow water in the crossspread domain, with enhanced trace density. Further, we implemented a three-step method to model velocity and anisotropy parameters. We first used long offset seismic data to estimate effective velocity and anisotropy using a bi-spectral analysis method to pick non-hyperbolic moveout of the reflection data. Secondly, we independently measured Thomsen anisotropic parameters at seismic scale using available well logs and walk-away VSP data. These two independent methods provided an opportunity to build a simple layered anisotropy model. In the third step, we used reflection tomography to derive the final velocity and anisotropy models. The final model shows a wide variation of anisotropy within different geological layers of the survey area, where delta varies between -5% to 10% and epsilon varies between 0% to 40%. The shallow section exhibits high epsilon (30 to 40%) and negative delta values. The new seismic images provide a reliable interpretation of overburden structures. The Vertical Transverse isotropy (VTI) model shows consistency with sonic logs. The seismic images display high correlation with well synthetics at reservoir level. The reliability of the anisotropy model is confirmed by comparing a seismic structural depth map with well tops. The incorporation of anisotropy in incidence angle computation allows accurate angle stacks to be produced, which improves the reliability of subsequent AVO analysis. The workflow showcases the first successful implementation of up-down deconvolution in ultra-shallow waters to remove multiples, and reliable estimation of VTI parameters combining seismic and well data offshore Abu Dhabi.

The 3D ocean bottom cable technique allows for acquiring long offset and wide azimuth seismic data. The use of simultaneous sources reduces the acquisition turn-around and HSE exposure. In shallow water environments, simultaneous source data are highly contaminated by surface waves and interference noise. Poor signal to noise ratio (S/N) affects velocity estimation, wavelet stability and overall image quality. This paper demonstrates the successful implementation of different processing and interpretation tools to deal with these challenges. The initial velocity model was built by extrapolating checkshot corrected sonic velocities along the interpreted key horizons and was subsequently updated to achieve final PSTM velocity. Several passes of noise attenuation were applied. Volumetric curvature analysis was used to monitor and protect fault planes from smearing during the denoising process. Seismic to well ties were continuously monitored to quantify the improvement after each key process was applied and to QC the seismic wavelet through different processing steps. A key factor to achieve a stable wavelet, at the end of the processing in the shallow water environment offshore Abu Dhabi, was the well driven horizon consistent velocity modeling. High seismic to well synthetic cross-correlation was observed on the final processed data due to the high S/N achieved by several passes of denoising, plus attenuation of strong multiple energy by velocity discrimination. High S/N, pickable geological events, and high resolution fault images are some of the key features of the final stacked image. In pre-stack data, long offset information is available to facilitate AVO and AVAz studies. Incorporating geological knowledge in the interpretation of horizons and faults and using well data during the course of seismic processing proved to be effective in obtaining a high quality seismic dataset.

Seismic processing is expected to deliver a reliable product for seismic structure interpretation and reservoir characterization to support subsequent reservoir modeling and drilling operation. 3D OBC wide-azimuth (WAZ) seismic data has a significant advantage in this respect, since it contains large offset and wide azimuth information, which can be utilized for amplitude versus offset (AVO) and amplitude versus azimuth (AVAz) analysis respectively. However, these advanced studies require area-specific data processing efforts in this region for strong linear noise, multiple noise, designature uncertainty etc. Seismic data re-processing was newly conducted in pre-stack time migration (PrSTM) sequence to overcome such technical challenges. Surface-wave noise attenuation was achieved by 2D and 3D model-based dispersive energy attenuation and multi-pass structure oriented filtering in pre-stack and post-stack stages. Source-side deghosting highlighted a low frequency residual bubble and was followed by another pass of de-bubbling based on a smooth spectrum assumption. Shallow water reverberation was minimized by 3D deconvolution and high-resolution radon demultiple. 5D regularization and interpolation with trace densification provided evenly populated traces for each angle and azimuth sectors. Intensive signal preservation protected diffraction energy and resulted in better fault imaging than the vintage processing outcome. Both the vintage and the re-processed seismic data were evaluated in quantitative manners to compare with each other. Angle stacks from the re-processing showed more reasonable response for AVO analysis. Azimuth stacks showed stable S/N ratio leading to higher confidence in azimuthal anisotropy analysis. All these seismic deliverables enable accurate elastic property estimation and detail fault interpreation at target reservoir intervals. The novelty of the proposed workflow is to overcome common technical problems with targeted noise and multiple reduction focusing on angle / azimuth stacks. It unlocks advanced geophysical study for carbonate reservoir in shallow water areas. Processing deliverables will be fully utilized in fault interpretation, AVO and AVAz analysis and will support the forthcoming field development activity.

The Arabian Gulf near-surface geology is complex, with extremely shallow waters and a hard water bottom generating high amplitude short period multiples and thinly bedded high and low velocity layers creating high apparent anisotropy in the bandwidth of seismic surveys. Obtaining an accurate description of the velocity variations in the near-surface and at intermediate depths is a necessity for reliable imaging and positioning of the reservoir layers located underneath. We propose a two-step full-waveform inversion (FWI) of ocean-bottom node (OBN) seismic data from offshore Abu Dhabi. We update a velocity model, using both diving and reflected waves, to reach the required depth of penetration. FWI has become an industry standard for velocity model building. However, due to the oscillatory nature of seismic data, FWI is known to be subject to cycle-skipping, where the inversion process falls into a local minimum. This risk is mitigated by using an accurate enough initial model and the use of low frequencies. In the shallow waters of offshore Abu Dhabi, near-offset data suffer from strong mud-roll and guided-wave energy that are not properly modeled with acoustic FWI. We exclude these offsets from the input data and use diving waves, starting at 3.5Hz, to update the near surface. The diving waves penetration is limited to approximately one kilometer in this area and corresponds to the base of a shallow high velocity layer. To reconcile the kinematics of reflected waves, travelling mostly vertically and used for imaging, and diving waves, travelling mostly horizontally, and used in the velocity update, we need an accurate estimation of the anisotropy. This is obtained using Backus averaging from available well logs. For deeper updates, the data are processed to remove the mud-roll and guided wave energy. This allows for the inclusion of reflections and near offsets. The FWI update is performed to 10Hz and penetrates about 3km into the sub-surface. We applied this FWI workflow to a recent node survey acquired offshore Abu Dhabi. The velocity model obtained follows the main geological structures and accurately describes the velocity variations in the shallow sub-surface. The estimation of anisotropy is important to ensure good convergence of the FWI and for imaging and vertical positioning of the migrated events. The reverse-time migration (RTM) image obtained with the updated model shows improved focusing and simplified depth structures compared to the RTM image obtained with the smooth initial model. To the best of our knowledge, this is the first successful implementation of FWI, here combining diving and reflected waves, on a dataset from offshore Abu Dhabi. It is a step towards resolving buried anomalies such as karst features, that cause imaging distortions at deeper reservoir levels.

Low fold poorly sampled vintage seismic data often suffers from poor fault imaging. This can have a critical impact on reserve estimation and well planning. Acquiring high density seismic data over producing fields requires overcoming logistic challenges along with additional costs and increased acquisition time. However, advances in seismic processing technology could improve the fault resolution of vintage seismic data in a cost effective manner. This has been proven in a case study from offshore Abu Dhabi. The presence of strong surface wave energy, resulting from the shallow water environment and near surface heterogeneity, masked events in the deeper part of the section. Poor and irregular spatial sampling caused aliasing of the surface wave. In the vintage processing, strong de-noising was applied to tackle the aliasing issue, which smeared the fault definitions. During the re-processing, a joint low-rank and sparse inversion was applied to regularize and densify the input data to obtain a de-aliased surface wave noise model. Subsequent adaptive subtraction of the noise from the input removed strong surface waves without damaging the body waves. The stack quality was improved by application of cascaded surface wave attenuation algorithms. Additional five dimensional Fourier reconstructions of the data improved the signal quality. A carefully designed fault-preserving residual noise attenuation workflow further reduced the residual noise content. Automatic picking of key stratigraphic horizons was carried out in order to evaluate the spatial resolution of the re-processing outcome. Sharper discontinuities along fault planes observed compared to the interpretation of the vintage seismic data. Increased confidence in fault interpretation is of value for structural restoration study and further reservoir understanding. In addition, several new, previously not-visible, small fault features were highlighted as evident from volumetric curvature and semblance analysis. They have been effectively utilized in a forthcoming drilling campaign to de-risk well operation. Multi-dimensional data densification to de-alias surface waves and five dimensional re-construction of the signal proved to be beneficial to enhance the fault features on the poorly sampled seismic data.