RE: [ai-geostats] a linear coregionalization model

Expand Messages
• Dear Nur, In the linear model of coregionalization, each of the basic model must have the same range and anisotropy parameters across the different direct and
Message 1 of 2 , May 17, 2005
Dear Nur,

In the linear model of coregionalization, each of the basic model
must have the same range and anisotropy parameters across the
different direct and cross semivariograms.

Pierre

-----Original Message-----
From: M. Nur Heriawan [mailto:mn_heriawan@...]
Sent: Mon 5/16/2005 11:45 PM
To: ai-geostats@...
Cc:
Subject: [ai-geostats] a linear coregionalization model

Dear All,

I am working with a multivariate geostatistics. In
determining whether the semivariogram or
cross-semivariogram follows a linear coregionalization
model or not...must the spatial range be the same?

For instance, I have semivariogram and
cross-semivariogram from 2 different variables as
follow:
gamma1(h) = 11 nug(h) + 39 sph(h/60)
gamma1(h) = 9 nug(h) + 15 sph(h/100)
gamma12(h)= 14.5 sph(h/80)

If I consider only the variance-covariance components,
such models follow the positive-semi definite
condition. But how about the different spatial range I
have in each semivariogram-cross semivariogram? Is it
essential in a linear coregionalization model?