--> ABSTRACT: Stochastic Modeling of Dakota Sandstone (Rocktown Channel Member) in Central Kansas Using Architectural-Element Analysis, by John Holbrook, Timothy Ray, Yuzhi Cui, and M. D. Raghunath; #91019 (1996)

Datapages, Inc.Print this page

Stochastic Modeling of Dakota Sandstone (Rocktown Channel Member) in Central Kansas Using Architectural-Element Analysis

John Holbrook, Timothy Ray, Yuzhi Cui, and M. D. Raghunath

Architectural-element analysis provides a system for describing depositional units in terms of three dimensional geometric building blocks for simulation of aquifer heterogeneity. We predicted the probable growth pattern(s) of the Rocktown channel section by using stochastic modeling to assess the most likely stacking arrangement for component architectural elements.

Detailed architectural-element analysis of the Rocktown channel member at twenty-nine outcrops in the central-Kansas Saline River valley revealed that these upper Dakota strata are dominated by four elements, each with distinct geometries. These are, in order of abundance, channel-fill (Holbrook ^approx44%, ribbons), overbank-fine (^approx27%, tabular), sandy-bedform (^approx17%, prismatic), and lateral-accretion (^approx12%, lobate) elements. The Rocktown channel member was deposited from a meandering stream system.

Our stochastic model uses these field data to predict elements between data points. The initial model lattice contains the observed outcrop data. The lattice is then partially filled by using Monte Carlo methods to place additional elements according to the abundance percentages listed above. The rest of the lattice is then filled by allowing each existing element to grow according to its shape and size constraints, taking into account surrounding elements, growth likelihood probabilities, and other geological factors.

This architectural approach frees us from facies models, allowing us to define deposits in terms of three-dimensional, genetically significant units. This method has the advantage over geostatistic techniques of being able to incorporate boundaries into the model, while still accounting for the chaotic aspects of natural systems.

AAPG Search and Discover Article #91019©1996 AAPG Convention and Exhibition 19-22 May 1996, San Diego, California