Published 1994 as part of Computer Applications 3
Copyright © 1994 The American Association of Petroleum Geologists.  All Rights Reserved.
 

A stochastic model comprising a truncated Gaussian field with a linear expectation trend has been developed for modeling facies tract distribution in shoreface reservoirs. The expectation trend is described in terms of a progradation direction f, which controls facies tract orientation, and an aggradation angle q, which governs the distances over which the tracts can prograde. Interfingering between facies tracts is regulated by the spatial covariance function of the underlying Gaussian field. The estimation of f and q uses a statistical method known as Bayesian inference, which uses general geologic information and analog data, in addition to well observations.

A controlled test has been done based on two Cretaceous parasequences exposed in the Book Cliffs, near Helper, Utah. The test involved a synthetic extrapolation exercise where four measured sections were treated as appraisal wells and a fifth (blind test) well was used to evaluate the results of the simulations. Fifty realizations of facies tract distribution in each parasequence were simulated. The stochastic model generated geologically realistic facies tract distributions and the thicknesses of shoreface facies in the blind test well were predicted satisfactorily. The main limitations of our model are that it cannot capture abrupt updip pinch-outs or systematic distal thickening of facies tracts. Modifications are necessary to include such effects.

Despite the limitations, our modeling procedure has potential to assist in the management of shoreface reservoirs where the facies tract distribution is uncertain. The procedure also has potential to help identify and estimate risk for exploration prospects.