Abstract: Predrill Prediction and Evaluation of Porosity and Permeability in Sandstone

S. Bloch, K. P. Helmold

Mean porosity and permeability of many sandstones can be predicted accurately prior to drilling. The predictive approach depends mainly on the availability of empirical data in the area of interest. In unexplored or sparsely explored areas, general relationships between reservoir quality and simple geologic parameters (e.g., Scherer's regression equation, the relationship between porosity and vitrinite reflectance quantified by Schmoker and coworkers) can be useful. This approach requires reasonable estimates of the parameters used in the calculation of porosity. Approximate values of these parameters can be obtained from seismic data combined with a geological analysis of the area of interest.

In basins with sufficient data to generate calibration data sets, the predictive technique utilizes multivariate regression analysis. The effectiveness of this approach has been proven by numerous case studies (e.g., North Slope of Alaska; Haltenbanken area, offshore Norway; Yacheng Field, South China Sea; Taranaki Basin, New Zealand; southern San Joaquin basin, California). Accurate predictions can be obtained even when the target is 20 miles away from the nearest control point.

The critical variables controlling porosity are detrital composition, sorting, and burial history. Permeability can be predicted independently of porosity using the similar independent variables. Importantly, these variables can be accurately estimated from seismic data and facies models. The predictive applicability of the empirical approach is constrained by the limits imposed by the calibration data set and is generally limited to samples containing less than 10% cement. A satisfactory predictive model for samples with a wide range in cement content can be obtained by dividing the calibration data set into two or more subsets and developing a predictive model for each. For example, one subset can be limited to samples with less than 10% cement, whereas the second subset can consist of samples with more than 10% cement. Porosity and permeability in the first subset are then expressed by multivariate regression equations, whereas reservoir quality in the more heavily cemented samples can be estimated prior to drilling if the geologic factors controlling the spatial distribution of the cement or cements are predictable. This approach is particularly effective in sandstone subjected to "closed-system" diagenesis.

AAPG Search and Discovery Article #90986©1994 AAPG Annual Convention, Denver, Colorado, June 12-15, 1994