[R-sig-Geo] Interpolcation option: IDW or OK?

[Edzer Pebesma]

Yong Li wrote:
> Hi ALL,
>
> I have been with the attached dataset and R code to use OK to
> interpolate soil OLSEN P spatial distribution for a couple of weeks. So
> far I have not found a satisfactory solution using OK or local OK or
> block OK, compared to IDW method. However theoretically OK is always
> better than IDW as also a journal editor advised me in my submitted
> manuscript.
>   
"always better" is quite a strong statement. It does have a minimum 
variance property, but only under a number of assumptions that need to 
hold. IDW has not a naturally quantified variance, but has e.g. the nice 
property that the interpolated values do stay within the data range, 
which is not true for OK.
> Normally if we do not find a significant spatial structure for a soil
> variable, we may choose IDW or other methods. How is your guys' opinion
> on this or may you help me to find a better solution using OK with my
> dataset?
>   
Why would IDW be useful when no significant spatial structure is 
present? Why not use the global mean as predictor?

Other questions are whether variables exist out there that have no 
spatial structure, and also what significance means in your comment. 
Should we conclude that spatial correlation is zero when it is not 
significant? I would say no.
> I appreciate any help.
>   
And I'm looking for more opinions--anything!
> Regards
>
> Yong Li
> ------------------------------------------------------------------------
>
> _______________________________________________
> 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.springer.com/978-0-387-78170-9 e.pebesma at wwu.de