Fault identification by diffraction separation from seismic reflection data using time slice SSA-based algorithm

Abstract

We propose Time Slice Singular Spectrum Analysis (TSSSA) - a filtering algorithm to distinguish diffraction energy from seismic reflection data. The underlying method uses the Eigen Mode separation to decompose a time slice of seismic amplitudes, resulting from linear sum of different interpretable wave fields and uninterpretable noise, into independent wave fields (e.g., reflections, diffraction and noise). The method is found to be effective in identification of sharp boundaries of fault/fracture zones. We present here two examples of filtering on synthetic data: (i) seismic data without noise and (ii) seismic data immersed in Gaussian noise. Our results suggest that TSSSA based algorithm is robust for filtering diffraction energy in both cases. Locations of fault mapped on a 2D migrated section of diffraction energy match well with the faults on the real seismic section even when the fault has very low vertical displacement. Our proposed technique will be useful as a direct method for mapping faults and identification of large fractures by migrating diffraction energy. Presentation Date: Wednesday, October 19, 2016 Start Time: 3:10:00 PM Location: 161 Presentation Type: ORAL