Residual statics correction without NMO - A rank-based approach

Ali M. Alfaraj*, Eric Verschuur, Felix J. Herrmann

*Corresponding author for this work

Research output: Contribution to journalConference articleScientificpeer-review

3 Citations (Scopus)


Surface-consistent residual statics correction for land seismic data does not account for the source - receiver offset. Consequently, it requires normal moveout (NMO) corrected gathers to bring raypaths close to the normal incidence. When the NMO velocity is inaccurate or unavailable, the estimated statics suffer. Therefore, multiple passes of NMO velocity picking and residual statics estimation become essential, which are efforts and time consuming. To avoid this, we utilize a rank-based solution that is capable of estimating non-surface-consistent residual statics. The method is based on the rank property of frequency slices in the midpoint-offset domain, where ideal seismic data is of low-rank nature, while data with residual statics exhibits higher rank. Accordingly, we estimate the statics that lead to the desired low-rank signal via means of low-rank approximation and cross-correlation in an iterative and multi-rank-scale approach. Since we estimate non-surface-consistent statics by accounting for the offset of each trace, it is no longer required to have NMO corrected gathers. Consequently, the method does not require windowing over a noise-free section containing primaries or windowing to avoid the NMO stretch effect, which are required by conventional residual statics correction. Numerical results on simulated and field data suggest that the method has the potential of replacing existing residual statics correction techniques.

Original languageEnglish
Pages (from-to)2565-2569
Number of pages5
JournalSEG Technical Program Expanded Abstracts
Publication statusPublished - 2021
Event1st International Meeting for Applied Geoscience and Energy - Denver, United States
Duration: 26 Sept 20211 Oct 2021


Dive into the research topics of 'Residual statics correction without NMO - A rank-based approach'. Together they form a unique fingerprint.

Cite this