Abstract
Hydrocarbon extraction from shale requires specialized shale play characterizations and analysis for effective, economical, and ecological implementation of oil and natural gas production. In this work, we present a statistical approach that may be used as a preliminary investigation into the hydrocarbon resource potential of shale rocks, particularly to investigate the homogeneity, variability, and maturity of shales within a play. Statistical algorithms for Principal Component Analysis (PCA) and Partial Least Squares Regression (PLS) were used to determine if depositional environments and lithographic boundary characteristics of different plays allow prediction of specific production parameters. This project characterizes the Eagle Ford and Utica formations—two high-producing shale plays in the United States—and the Banff/Exshaw and Colorado formations—two recently assessed shale plays in Alberta, Canada. PLS models were unable to model gas production parameters from predictor variables, highlighting the complexity of gas formation and a need for microscale petrophysical characteristics to model these parameters. In contrast, oil production parameters were better predicted, perhaps due to use of bulk variables such as mineral composition that may correlate to the oil’s location in mineral interfaces. As expected, PLS model predictive capabilities increased with specificity of data sets to particular regions of a shale play. When used by geologists, PCA and PLS modeling could assist in determining optimal sites for hydrocarbon extraction.
