Analysis of anomaly correlations

Abstract

Recognizing correlations between data sets is the basis for rationalizing geophysical interpretation and theory. Procedures are presented that constitute an effective process for identifying correlative features between two or more digital data sets. The procedures include the development of normalization factors from the mean and variance properties of the data sets. Using these factors, the data sets may be transformed so that they have common amplitude ranges, means, and variances, thereby allowing a common graphical representation of the data sets that facilitates the visualization of feature correlations. Anomaly features that show direct, inverse, or no correlations between data sets may be separated by the application of correlation filters in the frequency domains of the data sets. The correlation filter passes or rejects wavenumbers between coregistered data sets based on the correlation coefficient between common wavenumbers as given by the cosine of their phase difference. Standardizing and summing the filtered outputs where directly correlative features have been enhanced yields local favorability indices that optimize the perception of these features. Differencing the standardized outputs where inversely correlative features have been enhanced, on the other hand, provides favorability indices that improve the perception of the inverse correlations. This study includes a generic example, as well as magnetic and gravity anomaly profile examples that illustrate the usefulness of these procedures for extracting correlative features between digital data sets.