[R-sig-Geo] Bayesian Block Kriging?

Edzer Pebesma edzer.pebesma at uni-muenster.de
Fri Feb 19 10:12:58 CET 2010


Candan, that is a very interesting question.

You should note that by block kriging one usually refers to the 
situation where the observations are (nearly) points, and the prediction 
refer to mean values over larger areas, the blocks. This is the 
assumption made in gstat, anyway. If your data are spatial averages, and 
you want to predict averages over areas of the same size, you could 
still use point kriging as there is no change in support from 
observations to predictions size / area / volume. Is your prediction 
support different from your observation support?

csoykan at mail.sdsu.edu wrote:
> Hello,
>     I'm interested in doing Bayesian kriging using R.  I see that the package geoR has a function that will allow one to do this (krige.bayes).  However, my data are not in the form of points, but rather they are blocks that represent spatial averages (i.e., the number of fishing hooks per month in a given lat x long square).  I am therefore interested in treating the data as spatial blocks, rather than points.  I see that package gstat has a function that will allow one to
> do block kriging (krige).  However, I cannot find a function that will allow me to do both, Bayesian and block kriging at the same time.  My questions are:
>
> 1) Can the function "krige.bayes" be made to accommodate data in areal form (rather than point data)? or
> 2) Does the function "krige" allow for Bayesian inference (i.e., priors, posteriors)? or
>   
No.
> 3) Is there another package that can do both? or
> 4) Is there a method for combining functions to do what I would like to do?
>   
I believe that krige.bayes does everything simulation-based. Of course 
you could use krige() to do simulations for you. You could also simulate 
on a (fine) point scale, and do the upscaling to blocks realisation by 
realisation, yourself. But again, I'm not sure if this is what you need.
> A response to any and/or all of these questions would be greatly appreciated.  Thank you in advance.
>
> Candan Soykan
> csoykan at mail.sdsu.edu
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>   

-- 
Edzer Pebesma
Institute for Geoinformatics (ifgi), University of Münster 
Weseler Straße 253, 48151 Münster, Germany. Phone: +49 251 
8333081, Fax: +49 251 8339763  http://ifgi.uni-muenster.de 
http://www.52north.org/geostatistics      e.pebesma at wwu.de



More information about the R-sig-Geo mailing list