COVA FUNCTIONS FOR UNEVENLY AND NONCORRESPONDINGLY SPACED PROCESSES

A cova function is a representation of coherency of two time series, which is defined for unevenly and noncorrespondingly spaced series akin to the cross-variogram of geostatistics. Whereas cross-variogram calculation implicitly assumes that data are sampled at the same points in both processes, the cova functions introduced here are defined also for unevenly and noncorrespondingly spaced data processes. Residual cova functions are valid for nonstationary time series. In the program COVA, a FORTRAN 77-code is given for time series. As a demonstration, COVA is applied to simulated processes and thereafter used to compare climatological data with marine-geochemical data in order to investigate common variational patterns underlying the observations through time. © 1990.