Geological reservoir layering for simulation and development usually models depositional sequences. Another approach looks for flow units, i.e., rock units with uniform flow properties. The two are not always identical and should be distinguished. Numerical core descriptions show that porosity and permeability in Reeves Field are determined by both depositional and diagenetic parameters. Flow units defined in this way cut across depositional cycles and cannot be traced between wells with 40-acre spacing. High oil cuts are obtained from 20-acre infill wells despite a long-term waterflood. Reservoir behavior is controlled by the geometry of these flow units, not be depositional cycles.
The reservoir at Reeves Field was discovered in 1957 (Fig. 1). It is a Lower San Andres dolostone, with sulfates and some pyrite, silica and clay minerals, originally deposited on a shallow, south-sloping ramp. This study is confined to a square-mile area in the north of the field, in which there appear to be no complications such as faulting or bioherms. The average porosity in this area is 9% (cutoff 6%), the average permeability 1.6 md (cutoff 0.1 md).
Geologists commonly subdivide reservoirs by depositional sequences1. These are packages of rocks that represent sediments resulting from changes in sea level relative to adjacent land, commonly due to uplift or subsidence, or to climatic changes. They are mapped as stacked or interleaving layers, with one depositional sequence (or, in previous terminology, depositional cycle) per layer. Gamma-ray peaks are normally assumed to be both time-markers and the tops of cycles or sequences. Each layer is treated as a flow unit.
The Reeves reservoir was first modeled in terms of depositional cycles2. Since then, sequence stratigraphy3, gamma-ray picks (Model 1)4, and gamma-ray logs plus core-based interpretation of depositional cyclicity (Model 2)5, have all been used.
Another approach to reservoir modeling looks in detail at the reservoir rock, searching for those aspects of the core that are directly relevant to reservoir properties. Attention may be focused on the nature and connectivity of pore spaces, on grain and crystal size, and on depositional texture6. A modification of this method, using additional aspects of core petrography in numerical format, divides the Reeves reservoir into discrete units that tend to vary in K/PHI value, but are not isolated (Model 3)7.
In Model 4, presented here, the sequence of diagenetic events is taken into account as well, refining the unit top depths from Model 3 so that the flow units are better-defined. The flow unit geometry of the Reeves reservoir is not controlled by depositional textures alone; diagenetic textures and paragenetic sequence are needed to locate flow unit tops precisely.
A new method of core description was developed as part of this study, providing numerical data that can be used in statistical analyses7. A total of 1379 feet of four-inch diameter core from eight wells (Fig. 1). Whole core analyses and sidewall neutron porosity logs were available, supported by thin sections in places. Petrographic descriptions were entered directly into spreadsheets on a laptop computer. The cores were all from a square mile area, all from the same stratigraphic interval, and all had similar depositional and diagnetic histories, so the data were regarded as representing a single statistical population; they were loaded into one database table and plotted against porosity and permeability to assess relationships.