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

Abstract

The classification of fluvial architectural elements is a task typically carried out through detailed examination of core samples, outcrops, or by manual interpretation of well logs. However, these methods are constrained by high costs associated with core sampling, geographic limitations of outcrops, and the time consuming and subjective nature of manual well log interpretation. As an alternative to these methods, this study investigates the use of statistical classification methods to distinguish the fluvial architectural elements of the upper Mesaverde Group of the Piceance and Uinta basins as distinct electrofacies classes. Two statistical classification methods are utilized in this study: 1) a k-nearest neighbor approach (k-NN) and 2) a probabilistic clustering procedure (PCP) in addition to a method involving well-log-indicator flags which incorporates outcrop-based, architectural-element thickness criteria to refine the classifier results. Study data come from 1626 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. Architectural-element thickness criteria is borrowed from previous outcrop based studies of the upper Mesaverde Group. Through an approach which integrated select classifier results with thickness criteria, an overall accuracy (number of correctly predicted samples/total testing samples) of 83.6% was achieved for a simplified four-class architectural element realization. Architectural elements were predicted with user's accuracies (accuracy of an individual class) of 0.89, 0.38, 0.74, and 0.99 for the floodplain, crevasse splay, single-story channel body, and multi-story channel body classes, respectively. Without the additional refinement allowed by the incorporation of thickness criteria, the k-NN and PCP classifiers produced similar results, with the k-NN technique consistently outperforming the PCP technique by a slight margin. In both the k-NN 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 Datapages/Search and Discovery Article #90189 © 2014 AAPG Annual Convention and Exhibition, Houston, Texas, USA, April 6–9, 2014