Only a small percentage of original oil in place (OOIP) is produced during primary production from most carbonate reservoirs. Secondary recovery programs, such as waterflooding, commonly double the amount of oil recovered, but much of the OOIP remains in the reservoir. To effectively improve recovery, development programs must target the location of remaining oil. The reservoir characterization problem is to build a model that images remaining oil saturation and that can be used to predict the outcome of various development programs.

Building a carbonate reservoir model begins with an understanding of the relationship between pore space and petrophysical properties. This relationship must be linked to depositional and diagenetic models so that the petrophysical properties can be distributed in 3D space. The rock fabric method for making this link has been developed at the Bureau of Economic Geology by an integrated team of geologists, petrophysicists, and reservoir engineers. Pore-size distribution is the key link between petrophysical measurements and rock fabric descriptions, and rock fabric is the key link to sequence stratigraphic models. Rock fabrics are divided into matrix fabrics—which contain interparticle and separate-vug porosity—and nonmatrix fabrics—which contain touching vugs. How a reservoir performs will be related to the volume and distribution of these basic fabrics. For a reservoir with matrix fabrics, sequence stratigraphic framework is crucial. The basic stratigraphic element is the high-frequency cycle, within which basic rock fabrics are systematically distributed. However, the basic petrophysical element is the rock-fabric flow unit, which is defined by facies stacking within a high-frequency cycle. The result is a static 3D model of porosity, permeability, and initial oil saturation suitable for input into a numerical flow simulator. Production history of the field is simulated, and the end result is an image of the location of remaining oil saturation.