Skip to content

Compensating smoothing effect by ordinary kriging

2 messages · Swagath, Edzer Pebesma

#
Dear all,

I'm a relatively new with kriging methods using gstat. I am interested 
in performing an ordinary kriging to interpolate heavy metal 
concentrations across Europe  with 1000 data points.  However, I 
realized that when i performed cross-validation, OK tends to 
under-estimate my higher values and over-estimate some lower values. Is 
there anything i can do to correct for this smoothening effect by OK. I 
was doing a web search and came across a paper 'Compensating for 
Estimation Smoothing in Kriging' by Ricardo A. Olea and Vera Pawlowsky 
where they discuss about this smoothening effect. Can anybody tell me  
how to compensate for this smoothening effect by OK and show the 
hot-spots on map.

Thanks a lot for your time and help.

Thanks,
Navin
#
Navin, I believe constrained kriging may do this; I'm Cc:ing its
authors, who may not be on the list.

http://cran.r-project.org/web/packages/constrainedKriging/
On 12/04/2012 03:33 PM, Swagath wrote: