Maximizing the Value of Sparsely and Irregularly Sampled OBC Seismic Data Offshore Abu Dhabi

OBC seismic data reprocessing is performed to deal with insufficient spatial sampling in the data. Conventional linear noise attenuations using multi-channel filters are not effective with under-sampled coherent events. We hence apply a model-based approach that separates coherent noise and other signatures using a Fast Wavelet Transform. The process then estimates a noise model in the FX domain including aliased components, and adaptively subtracts it from the recorded data. To enhance spatial sampling, 5D data interpolation and regularization based on anti-leakage Fourier transform is implemented. This yields regular directional sampling from near to far offsets. Additionally, 3D algorithms are utilized through the course of reprocessing to preserve offset and azimuth information such as denoise, de-multiple, imaging and velocity analysis.

The reprocessing successfully provides an appropriate data for azimuthal anisotropy analysis. AVOAz inversion is subsequently performed to estimate azimuthal attributes without undesired effects and limitations in amplitude and velocity based approaches. The inversion results provide an insight into stress induced azimuthal anisotropy in the field along with a good consistency with the regional geology and well data. The results enhance the value of sparsely and irregularly sampled OBC seismic data as if it was acquired with optimum spatial sampling.