An Assessment Of Some Methods For Correcting Distortions In Seismic Data Over Near Surface Reefs - A Case Study Of The Krill Reef, Zocz 91-01, Timor Sea.
Brigg, Sandy J.
Honours Degree, 1995
University of Adelaide
Many near surface reefs are present in the Northern Bonaparte region of the Timor Sea. These structures have steep flanks rising of the order of 250 metres and significant internal lateral velocity variations. These factors cause events on seismic data acquired over a reef to be "pulled up" in time and their quality to be degraded by conventional common mid-point (CMP) stacking. It is critical to correct these distortions since, in many instances, reefs have grown on structural highs which may represent prospective hydrocarbon traps, requiring accurate mapping.
Conventional CMP processing fails to correct reef distortions since it assumes hyperbolic moveout of reflections. Processing techniques that can deal with non- hyperbolic travel times are required, such as layer replacement statics, non-surface consistent dynamic corrections (eg. Digicon's Reveal process), wave equation datuming, and pre-stack depth migration. There is a large variation in the sophistication and cost of running these processes. The main aim of this project was to compare the ability of such processes to improve sub-reef imaging, and the effort required to implement them. Krill Reef was selected as a typical problem area, and tests were performed on synthetic data generated from models of this reef, and on a 2D seismic line which crosses it.
A simple model of Krill Reef was built with horizontal sediment layers and constant reef velocity. Synthetic shot gathers generated by ray tracing through this model demonstrate the nature of the distortions caused by the reef topography, and verified that water replacement statics and pre-stack wave equation datuming corrects these distortions if the correct velocity model is used.
A second model simulated lateral velocity variations within the reef. Processing with a constant average reef velocity still resulted in major improvement of the data, since most of the strong velocity variation between the water and reef was compensated. Residual distortions were still present due to the deviation of the model velocity from the average reef velocity used in processing. For both models, wave equation datuming produced the more accurate output, but at markedly greater cost.
A seismic line over Krill Reef was processed assuming an average reef velocity using water replacement statics, the Digicon Reveal process, and pre-stack wave equation datuming. All these processes showed major improvement over conventional CMP stacking, however there are significant differences in the resulting sub-reef time structure. Wave equation datuming has enhanced the sub-reef imaging in comparison with the other two processes, and the modelling results suggest that it is likely to provide the. Most accurate correction.
The preferred processing approach to the reef problem depends on the location of the exploration target - if a critical interpretation of a structural high below the reef is required then wave equation datuming may be necessary. Elsewhere, replacement statistics are probably adequate.