Skip to content

Anisotropy for Cokriging

2 messages · Seyed Jalil Alavi, Edzer Pebesma

#
Dear All
?
I am mapping forest site productivity for one of most important species in 
one of the oldest forests in the world, Hyrcanian Forests, Iran.
 
Dear All, I used rose plot function written by Jon Graham 
(http://www.math.umt.edu/graham/stat544/) and found there is an anisotropy 
in the variogram for kriging and I included the anisotropy in variogram 
fitting.

Dear All, I want to investigate the anisotropy for cokriging and if there is 
anisotropy, I want to consider it for cokriging.

I search many textbooks and also in the internet and found NOTHING for 
anisotropy for cokriging.
I would be very grateful if ANYONE can help me how can I study the 
anisotropy for cokriging and include it the model.

 Another question is about argument cressie=T in gstat package. For krige 
function in gstat package there is cressie=T, if there are outliers. For my 
data, it seems there are some outliers and I want to include cressie=T for 
variogram for fitting cokriging. I did not see any argument in gstat 
function.

I would be extremely grateful for any help

Thanks and regards

Jalil

**********************************
Seyed Jalil Alavi, Ph.D
Assistant Professor in Statistical Ecology
Dept. of Forestry
Faculty of Natural Resources
Tarbiat Modares University
P.O.Box 46417-76489
Noor
Mazandaran
Iran
Mobile Phone Number 00989111580097
Tel:0098(122)6253101-3
http://www.modares.ac.ir/en/Schools/nat/Academic_Staff/~j.alavi
1 day later
#
On 08/01/16 19:52, Seyed Jalil Alavi wrote:
There's nothing special to it, except that very few people actually do
it. Under the linear model of coregionalisation, the range ellipse (or
ellipsoid) has to be identical for all direct and cross variograms.
It is only defined (and implemented) for direct variograms, not for
cross variograms.