Deriving priors from the low-frequency component of seismic data significantly improves the anti-aliasing capabilities of the matching pursuit Fourier interpolation at high frequencies. However, for coarsely sampled data, the low-frequency region gets contaminated with aliasing, thus making it unsuitable to compute priors to de-alias high frequencies. One such acquisition scenario is the ocean-bottom node (OBN) survey where we sample sources at a much denser grid compared to the receivers. Thus, any standard interpolation framework can reconstruct the common-receiver gathers, but struggles to produce reliable reconstruction results for common-shot gathers, which is necessary for successful multidimensional deconvolution in an OBN scenario. To stabilize the interpolation for common-shot gathers, we propose to exploit the principle of reciprocity in combination with wavefield extrapolation to build priors from well-sampled interpolated common-receiver gathers to reconstruct common-shot gathers across the full frequency spectrum. Using real data examples from the Valhall field, we demonstrate that reciprocity-based priors bring substantial uplift in the interpolation and enhance the quality of reconstruction significantly.
Presentation Date: Wednesday, October 14, 2020
Session Start Time: 8:30 AM
Presentation Time: 9:20 AM
Presentation Type: Oral