Loop-Scale Seismic Facies Analysis: Challenges in Use of Automated Versus Manual Picking Techniques in Delivering Geologically Constrained Interpretations to Reservoir Modelers

O’Byrne, Ciaran J., Carlos Pirmez, Vaughan Cutten, Ed Kruijs, Mike Stovall, Shell International E & P, Houston, TX

In the last decade, there have been impressive advances in technologies which automate interpretation and rapid assessment of the prospectivity of seismic volumes. Analytical tools have greatly enhanced the interpreter’s ability to rapidly extract meaningful planform and voxel realizations of complex sub-surface stratigraphy in deepwater slope reservoirs. However these realizations in themselves are often insufficient in terms of rendering geolog­ically constrained geometric bodies and or surfaces for reservoir simulation. While voxel or attribute based auto-picking tools can optimize and greatly accelerate our understanding of gross sub-surface geology there is not yet a replacement for manually picking complex geo­logical surfaces (e.g. channels) with highly variable juxtaposed lithofacies or rock proper­ties. This critical limitation has real implications for accurately representing sub-surface geology in static reservoir.models. However, a combination of manually picking grids guid­ed by horizon slices/voxel attributes generated by these tools allow us to produce far more accurate interpretations, leading to more accurate depositional model scenarios and in turn provide geologically constrained input to reservoir modeling packages.

An example of such an interpretation workflow is presented here for an ongoing project in offshore, deepwater Nigeria, with a demonstration of the critical geological insights gained from mapping a slope reservoir in this manner. This dataset defied successful appli­cation of traditional voxel based tool technology where the geology is sufficiently complex and poorly resolved that extracted bodies had little to no meaningful geometries. Sculpting the data based on the higher resolution manual picks yielded more successful results.