Alton A. Brown1
(1) Consultant, Richardson, TX
ABSTRACT: Prediction Of Reservoir Quality (Porosity/Permeability) In Frontier Basin Carbonates
Reservoir quality (RQ) predictions are unreliable in the absence of calibration data. Predictions improve where wells are available for calibration. RQ has been predicted using seismic detection techniques, statistical prediction techniques, petrographic analysis, and numerical modeling.
Seismic detection is most reliable where 3D seismic data are used to image targets at shallow to moderate depth. Porosity may be detected, but porosity estimates are typically qualitative. Depositional facies indicative of favorable RQ development may be interpreted from seismic even where RQ cannot.
Statistical approaches are best used where burial diagenesis dominates or where facies relationships are clearly related to RQ. Most statistical predictions estimate mean porosity, but porosity risk can be approximated from cumulative probability curves. Appropriate calibration data are essential.
Petrographic approaches help characterize diagenetic and depositional patterns which control RQ. Pore fabric data aid prediction of permeability from porosity. A general understanding of porosity loss mechanisms also allows qualitative predictions of settings preserving or enhancing porosity.
Numerical modeling is perhaps the least effective prediction approach because so many variables affect RQ development, and few of these are precisely known during the exploration phase. Some burial diagenetic models do a moderately successful job of predicting average porosity, and some early diagenetic models can construct realistic vertical RQ variations.
In all cases, effects such as regional dolomitization, unexpected facies changes, and late anhydrite cementation/dissolution can completely upset carefully constructed prediction methodologies. For this reason, frontier RQ predictions should always be properly risked and used with caution.
AAPG Search and Discovery Article #90906©2001 AAPG Annual Convention, Denver, Colorado