We have a sparse primary data set and a densely sampled secondary data.
When calculating the cross-semivariogram, one needs a secondary and
primary datum collocated for all pairs. However, the cross-covariance
can be calculated with primary at the head and secondary at the tail.
SGems appears to only calculate collocated paired data for the
cross-covariance, which is problematic because it restricts the number
of possible pairs unnecessarily. Is there any way around this?