Technical Paper: Time- and Frequency-Domain Metrics for the Evaluation of Full-Aperture, Time-lapse Data from Multi-Sensor Streamers

Society: EAGE
Paper Number: Th G102 01
Presentation Date: 2014


Recently introduced multi-sensor streamers measure particle motion as well as acoustic pressure. The generalized matching pursuit (GMP) algorithm applied to independent, coincident wavefield measurements allows dealiased signal reconstruction and is well suited to finite crossline sampling in marine data. GMP also allows receiver deghosting which broadens signal bandwidth. The performance of wavefield reconstruction and deghosting can be evaluated using time-lapse repeatability metrics if the reference signal is known. However, these metrics are usually calculated in the time domain making it difficult to evaluate reconstruction performance near spectral notches and at high frequency. We analyze time-lapse data acquired over the North Sea Forties Field by a commercial multi-sensor streamer. Using frequency-dependent normalized root mean square (NRMS) attributes we evaluate algorithm performance across the spectral bandwidth. We find that wavefield reconstruction can significantly improve data repeatability early in processing while later processes have a smaller impact, in contrast to repeatability of single measurement data, leading to a simpler and more time-lapse friendly workflow. A reduction in NRMS following GMP is retained throughout subsequent processing resulting in a final migration with improved repeatability over a broader bandwidth. This emphasizes the importance of multi-sensor acquisition for high quality wavefield reconstruction in time-lapse seismic.

Related services and products

Request More Information

Determine Changes in the Reservoir Over Time

Visit 4D Seismic page
We offer a complete 4D solution, from feasibility studies to inversion and history matching, to the latest acquisition technology and global expertise. Visit 4D Seismic page