I missed the point that you actually did a stratified kriging -- this is
indeed including a trend, by ways of a mean value for each stratum (z
layer).
To demonstrate that correlation in the z-direction is present and does
play a role, you'd have to set up the cross validation more clever, and
Hi Edzer, hi Sebas,
Thanks a lot for the comments on my problem. Cross validation was done by
comparing MSEs and the visual inspection of scatter plots. I have indeed a high
signal-to-noise ratio ratio but of course I'm still in search of the method
that gives the best possible results. While i
