3D seismic data de-noising and reconstruction using Multichannel Time Slice Singular Spectrum Analysis

• Published in: Journal of applied geophysics Vol. 140; pp. 145 - 153
• Main Authors: Rekapalli, Rajesh, Tiwari, R.K., Sen, Mrinal K., Vedanti, Nimisha
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.05.2017
• Subjects: Data-gap filling; De-noising; MSSA; Rank reduction algorithm; MSSA; Rank reduction algorithm; De-noising; Data-gap filling
• ISSN: 0926-9851; 1879-1859
• DOI: 10.1016/j.jappgeo.2017.04.001

Noises and data gaps complicate the seismic data processing and subsequently cause difficulties in the geological interpretation. We discuss a recent development and application of the Multi-channel Time Slice Singular Spectrum Analysis (MTSSSA) for 3D seismic data de-noising in time domain. In addition, L1 norm based simultaneous data gap filling of 3D seismic data using MTSSSA also discussed. We discriminated the noises from single individual time slices of 3D volumes by analyzing Eigen triplets of the trajectory matrix. We first tested the efficacy of the method on 3D synthetic seismic data contaminated with noise and then applied to the post stack seismic reflection data acquired from the Sleipner CO2 storage site (pre and post CO2 injection) from Norway. Our analysis suggests that the MTSSSA algorithm is efficient to enhance the S/N for better identification of amplitude anomalies along with simultaneous data gap filling. The bright spots identified in the de-noised data indicate upward migration of CO2 towards the top of the Utsira formation. The reflections identified applying MTSSSA to pre and post injection data correlate well with the geology of the Southern Viking Graben (SVG). •A time domain Multichannel SSA based algorithm for 3D seismic data de-nosing and gap-filling is presented.•Data gaps are filled using L1 norm based residual minimization.•The method is compared with frequency domain MSSA on synthetic and real 3D seismic data sets.