Abstract: Data from multiple surveys of an area of the earth are integrated for use in oil and gas exploration and production. The data of each survey is first gridded and grid node values determined. A map index is assigned to each grid node and grid node values are associated with each map index to form a grid node suite. The set of grid node suites is processed to identify and determine covariant clusters of variables in multivariate space which includes various survey variables but which does not include location as a variable. Then, after determining to which of the clusters each grid node most probably contributes, a cluster identifier is associated with a respective grid node index for each grid node. Then a covariant cluster map of cluster identifiers as a function of grid node indexes is generated. Cluster locations on the map are correlated with other information about the area to assign earth features to the covariant cluster map.

Abstract: A set of representative subpopulations of a dataset such as a set of depth related formation log responses is selected by (a) classifying the dataset into n=a disjoint clusters; (b) determining discriminant function for the thus produced n=a disjoint clusters; (c) generating a measure respresentative of agreement/disagreement betweeen the classification generated by cluster analysis in step (a) and the classification generated by discriminant analysis in step (b); (d) repeating steps a-c for other numbers n=a, b, c, . . . ; and (e) selecting the set n=k where k is a, b, c, . . . , where the measure shows best agreement.