Computer-Assisted Association of Well Log Signatures with Depositional Environments

James D. Clippard, Marianne K. Clippard

Depositional environments are usually interpreted from well log data by identifying a set of log signatures and evaluating shape, absolute value, and the spatial interrelationships of these signatures. This task may be complex and time consuming when dealing with log signatures defined by an aggregate of several log types and many wells; as a consequence, many hypotheses may go unformulated or untested. The efficiency and thoroughness with which this process is performed may be greatly enhanced through the use of a computer-assisted algorithm consisting of the following five steps: (1) delineation of strata boundaries in all wells being studied, (2) principal component analysis to determine the minimum number of parameters required to describe the observed set of log data, (3) fuzzy c uster analysis to group the wells into signature groups, (4) comparison of log signature with known lithofacies indicators, and (5) spatial mapping of cluster probabilities to enhance perception of geometric relationships.

This algorithm requires that several wells represent each discrete lithofacies and that the number of transitional facies sampled is small relative to the total number of wells. When these conditions are satisfied, the cluster probabilities may be viewed as a generalized lithofacies similarity indicator. Through consideration of the mapped cluster probability geometries and the lithofacies indicated by the signatures associated with each cluster, direct initial associations between log data and depositional environment may be made in both clastic and carbonate settings.

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