Cross correlation

Cross-correlation (Davis 1986) is carried out on two column(s) of evenly sampled temporal/stratigraphic data. The x axis shows the displacement of the second column with respect to the first, the y axis the correlation between the two time series for a given displacement. The "p values" option will draw the significance of the correlation, after Davis (1986).

For two time series x and y, the cross-correlation value at each lag time m is computed.

For positive lags, x is compared with a y that has been delayed by m samples. A high correlation value at positive lags thus means that features in y are leading, while x lags behind. For negative lags, features in x are leading. A reminder of this is given by the program.

The p value for a given m is given by a t test with n-2 degrees of freedom, with n the number of samples that overlap.

It is important to note that this test concerns one particular m. Plotting p as a function of all m raises the issue of multiple testing – p values smaller than 0.05 are expected for 5% of lag times even for completely random (uncorrelated) data sets.

Missing data supported.

Reference

Davis, J.C. 1986. Statistics and Data Analysis in Geology. John Wiley & Sons.

Published Aug. 31, 2020 7:59 PM - Last modified Aug. 31, 2020 7:59 PM