Toward Predicting the Influence of Facies on Well Performance in the Marcellus Formation: Development and Integration of a Core-Derived Lithofacies Model

Abstract

In the Marcellus Formation of northern PA, widely variable well performance over a short geographic distance challenges the common notion of homogeneous and predictable shale gas plays. Our research tests the idea that lithofacies variability and stacking patterns, controlled by the formation's depositional system, could also vary over short distances and be partly responsible for differences in well production. Since shale lithofacies provide the first order of control on reservoir quality, the ability to model and more accurately predict the distribution of lithofacies in a reservoir is essential in maximizing hydrocarbon production. This research represents the initial step in creating a comprehensive facies model for the largest unconventional shale reservoir in the United States, with a specific focus on the Union Springs Member (lower Marcellus). Over 7000 feet of core from 25 Marcellus wells spanning a 75 x 25 mile area of northern PA were described in sub-decimeter-scale detail. Twenty one different lithofacies were identified in these cores, with 5 facies commonly comprising the majority of the Union Springs Member; some of the criteria for facies designations included carbonate vs. silicilastic content, fauna, grain size, rock fabric and sedimentary structures, and shale color variations. Interpretations of depositional environment included consideration of relative water depth, position in the basin relative to foredeep, and possible chemocline fluctuations. Even over short distances (less than 10 miles), significant lateral changes in facies and their cyclicity occur, suggesting a complex basin with potentially important and predictable paleotopographic variations, some of which may have been influenced by the movement of underlying salt deposits. A preliminary comparison of the core lithofacies data with petrophysical log and well production data sets implied some correlations between facies and rock quality indicators (core-derived TOC, porosity, and permeability measurements) but hinted at subtle “unseen” variability within the facies identified in core. This problem prompted subsequent statistical analyses aimed at determining the influence of facies variation and other factors on well performance, and it also highlighted the need for a petrophysical model that could detect facies variations and potentially “see” facies that we were unable to observe in core descriptions (see abstracts by Mintz and Syrek).

AAPG Datapages/Search and Discovery Article #90189 © 2014 AAPG Annual Convention and Exhibition, Houston, Texas, USA, April 6–9, 2014