--> Abstract: Geologically Constrained Electrofacies Classification of Fluvial Deposits: an Example from the Cretaceous Mesaverde Group, Uinta and Piceance Basins, by Daniel Allen and Matthew J. Pranter; #90169 (2013)

Datapages, Inc.Print this page

Geologically Constrained Electrofacies Classification of Fluvial Deposits: an Example from the Cretaceous Mesaverde Group, Uinta and Piceance Basins

Daniel Allen and Matthew J. Pranter
1921 Canyon Blvd Apt 303, Boulder, CO

This study establishes a set of distinct, core-defined electrofacies that represent the fluvial architectural-elements of the middle and upper Mesaverde Group of the Piceance and Uinta basins. Evaluation of electrofacies classification techniques in testing wells suggests that the electrofacies can be accurately predicted in non-cored wells. This allows for wire-line log curve interpretations which are timely, reproducible, and objective. Two different electrofacies classification techniques are utilized in this study: a k-nearest neighbor approach (KNN) and a probabilistic clustering procedure (PCP). Study data come from 1668 samples with known architectural-element classifications determined from the analysis of cores (N=9) with each sample having 4 available measured properties (wire-line log curves). The sampling population is divided into two groups: one for training and one for testing the capability of the trained classifier to correctly predict classes. A novel method involving a well-log indicator flag which pairs the results of the electrofacies classifications with thickness criteria, is applied to further refine classification results. The applicability of the geologically constrained electrofacies classification is demonstrated through batch prediction and mapping of electrofacies in wells (N=~200) throughout the study area. Initial testing yielded overall accuracies (number of correctly predicted samples/total testing samples) of 64% and 62% in the KNN approach and the PCP approach respectively. Testing was conducted again using a new facies realization which was created by combining two geologically similar facies. This yielded improved overall accuracies of 75% and 73% in the KNN approach and PCP approach, respectively. The top performing classification outcome was chosen for the application of the well-log indicator flag approach which increased the overall accuracy to 84%. While both classifiers produced similarly reasonable results, the KNN technique outperformed the PCP technique. In both the KNN and PCP techniques, the combination of wire-line log curves GR and RHOB proved to be the most useful assemblage tested. These curves are common in the study area, and their presence presents the opportunity for further geologic investigations utilizing these electrofacies classification methods.

AAPG Search and Discovery Article #90169©2013 AAPG Rocky Mountain Section 62nd Annual Meeting, Salt Lake City, Utah, September 22-24, 2013