Abstract

In order to best optimize reservoir performance of Miocene-aged deep-water reservoirs for the Spiderman gas field (De Soto Canyon Blocks 620 and 621) in the eastern Gulf of Mexico, we have developed a rock-typing methodology that calibrates petrophysical properties to individual depositional facies towards the ultimate purpose of predicting reservoir quality in portions of the reservoir lacking core and/or well control.

First, on the basis of detailed sedimentologic descriptions for the cored intervals, individual depositional facies were identified and conceptual models were developed to predict spatial facies distribution. The models were developed based on the vertical succession of facies, well-to-well log correlations, and seismic-based trends. Second, utilizing routine and special core analysis data from whole cores, hydraulic properties were described and integrated with depositional facies interpretations and log characteristics. Initially, rock typing exercises were successful in relating permeability/porosity properties to pore-size distribution. However, individual rock quality classes included samples from a variety of depositional facies and therefore cannot be used as a tool to adequately predict permeability variations in other portions of the field.

We then compared facies-specific petrophysical parameters to reservoir quality indices in the hopes of developing facies-based permeability transforms to predict permeability in uncored wells. After evaluating a wide range of petrophysical attributes [bed thickness, capillary pressure, grainsize, bed type, water saturation] it was determined that porosity, water saturation, shale volume, and mean grain size showed the closest correlation to reservoir quality indices. Consequently, by using multiple linear regression to compute permeability as a function of their relative volume (as determined from quantitative log evaluation) for each key depositional facies, the relationships between reservoir quality and key petrophysical properties are more clearly observed. The facies-based permeability curves can then be used to map predicted facies within a stochastically-derived geobody volume that predicts reservoir quality and degree of compartmentalization over the entire field.