--> Outcrop-based Reservoir Modeling of the Sappington Formation (Montana); Developing Strategies for Reservoir Characterization and Modeling of Middle Bakken Reservoirs (North Dakota)

AAPG Annual Convention and Exhibition

Datapages, Inc.Print this page

Outcrop-based Reservoir Modeling of the Sappington Formation (Montana); Developing Strategies for Reservoir Characterization and Modeling of Middle Bakken Reservoirs (North Dakota)

Abstract

A key challenge facing us is how to best describe and model the geologic characteristics of Bakken reservoirs in the context of limited subsurface spatial information and practical limits to numerical model size. How do we create geologically meaningful reservoir models that are also computationally efficient? To determine the most appropriate strategies for reservoir modeling of the Bakken Formation, Williston Basin, we have completed a suite of geo-cellular models using outcrop information from the Sappington Formation, SW Montana. The Bakken and Sappington formations are time-equivalent (Mississippian-Devonian) and share many geologic attributes owing to similarities in depositional histories and environments. Hence, the outcrops offer clues about subsurface reservoir heterogeneities that may impact production characteristics. Specifically, some outcrops reveal stratigraphic and facies variability over a scale that is similar to a Bakken Drill Spacing Unit (DSU). We present study results from one such outcrop and highlight the impact of geologic variability at a range of model resolutions. The first step was to create a high-resolution, DSU-scale model to approximate the vertical and lateral geologic variability observed in the outcrop. Bakken reservoir petrophysical properties were assigned to the model. Next, a suite of progressively coarser models were created and compared to the high-resolution model. This process entailed subjecting the original model to a “scale-up experiment” and documenting successive changes to the model design characteristics. Finally, we analyzed the models in terms of their petrophysical properties (connected pore-volume) as a method of screening for expected dynamic (simulation) behavior. Together, these results reveal the scale of resolution at which the geologic characteristics of the outcrop were lost in the scale-up process and simulated fluid flow characteristics are compromised. In ongoing and future phases of work, we will incorporate results of a behind-the-outcrop coring program, create a suite of simulation models to analyze their dynamic behavior and analyze the impact of inclusion of natural and induced fracture networks.