Accurately modeling reservoir quality in Miocene deepwater reservoirs in the K2/ K2-North Field, Green Canyon, Gulf of Mexico, presents many challenges for planning primary and secondary oil recovery. An overlying thick salt canopy prevents adequate seismic imaging at reservoir levels, structural complexities make correlating sand packages difficult, and well spacing is sufficiently large that correlating intra-reservoir surfaces is problematic. Borrowing from successful methodologies developed for seismically better-imaged deepwater reservoirs in the eastern Gulf of Mexico, we utilize welllog, core, and pressure data to calibrate petrophysical properties to individual depositional facies towards the ultimate purpose of predicting reservoir quality in areas which lack core and/or well control.

First, based on detailed sedimentologic descriptions for the cored intervals, grainsize analyses, whole core x-ray scans, well-log patterns, and correlations, we identified four main reservoir facies within two principal Miocene intervals: channel-bypass, thick-bedded amalgamated sheets (axial and marginal), and thin-bedded layered sheets. Second, to distribute facies within time-correlative packages, we used depositional models based on Gulf of Mexico shallow-seismic analogs of distributary channel complexes. While the lower Miocene interval characteristics are more akin to confined, proximal portions of a frontal splay, the middle Miocene interval contains characteristics of the more unconfined medial portions.

Third, using whole core data, we examined a wide range of petrophysical attributes to recognize unique combinations of petrophysical properties for each key reservoir facies. This provided valuable insight for choosing log-based curves used for identifying facies within uncored wells.

To date, pressure transient analysis of production data from the two main intervals are consistent with our facies interpretations. The more sheet-like interval shows few barriers while the more channelized interval suggests a higher degree of complexity.

The resulting improved facies classification scheme has provided a more useful basis for routine reservoir modeling and field management.