Dear Edzer Pebesma,
Is it possible to cokrige a regionalized factor
("factorial kriging") with gstat.
It seems to me that the cleanest way to do this might be
to do a cokriging of the original N variables toghether
with an additional newly created variable (the factor) with
no data (or with all NA, or with some arbitrary data but setting
appropriately the local neighbourhoods parameters to get no points
form the factor), having specified all variograms and cross-variograms
between the N+1 variables.
This should reduce the usual cokriging system to the right
factor cokriging system.
However doing this with gstat results in a planar (at infinity)
map for the factor variable. Where am I wrong?
Is there any other way to cokrige the factor?
Best Regards
Marco
------------------------------------------------------------------------
Dr. Marco Minozzo
Dipartimento di Scienze Statistiche Tel. +39 075 5855226 or 5855242
Universita' degli Studi di Perugia Fax +39 075 43242
Via A. Pascoli, C.P. 1315 Succ. 1 E-mail: [EMAIL PROTECTED]
06100 Perugia (Italy)
------------------------------------------------------------------------
