Hallo!
I used log transformation for daily precipitation interpolation using
universal kriging considering trend of precipitation with altitude and
latitude. If I use antilog transformation for the predictions I get
biased value. I read Cressie's book and found out that the
correction is possible but I need kriging variances and also
Lagrange multipliers for that. Is it possible to get the file with
Lagrange multipliers with Gstat? Have somebody experiences with
such problems?
I have one more question: I would like to get into contact with
person who have experience with cokriging of precipitation, using
altitude or some other geographical variable as kovariable. I have
problems with cross-variogram detection.
Thank's in advance.
With best regards
Damijana Kastelec