Correlation analysis: application of DFA and DCCA in well log profiles

Detrended fluctuation analysis and detrended cross correlation analysis are used in this work to identify and characterize correlated well log data. This is performed by first separating the different fluctuations due to external trends, and evaluating the autocorrelation and cross-correlation exponents to determine whether scale properties persist as the size of the series changes. Two new methodologies were developed to identify optimal values of the cross-correlation coefficients and graphically display them, which we call the automatic search procedure and correlation map . The methodologies were applied to well logs from the Jequitinhonha Basin, Brazil, to verify the existence of scale property in these data. For practical purposes, our goal is to use a local analysis framework to detect all points of high cross-correlation among different physical parameters in the same well, and among one same physical parameter in different wells. The correlated events suggested the continuity of geological features, including the vertical displacements of rock layers. In particular, it was possible to identify layers of calcilutites in a specific depth range. These rocks are of particular importance to the study of stratigraphic correlations due to their great regional extent and regular layering.