--> ABSTRACT: Quantitative Outcrop Analysis of Coastal-Plain Sequences from Eastern Kentucky: Example of High-Quality Data Input to Three-Dimensional Reservoir Modeling, by Stephen S. Flint, Steven J. Van Rossem, and Huw Williams; #91022 (1989)

Datapages, Inc.Print this page

Quantitative Outcrop Analysis of Coastal-Plain Sequences from Eastern Kentucky: Example of High-Quality Data Input to Three-Dimensional Reservoir Modeling

Stephen S. Flint, Steven J. Van Rossem, Huw Williams

An integral aspect of inter-well scale reservoir geological modeling research is the collection of high-quality data on sand-body geometries and facies architectures. Such studies are aided by exceptionally large outcrops in structurally undisturbed areas.

The Pocahontas foreland basin of eastern Kentucky contains well-exposed fluvial-deltaic sequences with negligible tectonic deformation in the basin fill. Coastal-plain sequences are characterized by major mouth bars; stacked, multilateral, low and high-sinuosity distributary channel systems; and a series of minor distributary channel sandstones/coarsening-upward crevasse delta sequences.

Large (up to 500 m × 4 km) roadcut exposures covering a total of some 250 km of highway, coupled with exact coal seam correlation, have allowed three-dimensional mapping of different genetic sandbody geometries, creating a reliable data base for reservoir geological modeling in similar depositional systems. Outcrop correlation panels published by the University of Kentucky have been interpreted in terms of reservoir elements and digitized into a pixel-based data base. The digitized panels were analyzed in terms of horizontal and vertical transitions between reservoir elements. These data allow probabilities to be assigned to both vertical and lateral organization of the various reservoir elements.

Lateral transition matrices have been used to quantify, for example, the probability of an active channel-fill sandstone being in connection with other reservoir sand bodies. Inclusion of such probabilities in probabilistic reservoir models will maximize the accuracy of reservoir configurations predicted from well data alone.

AAPG Search and Discovery Article #91022©1989 AAPG Annual Convention, April 23-26, 1989, San Antonio, Texas.