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

[Edzer Pebesma]

I can think of two reasons, without having looked at anything else than 
your message:
1. Range=162 may mean a different thing for the exponential model; for 
gstat it means the range parameter as in c1(1-exp(h/a)). This is all 
documented. In ArcGIS it might mean something like c1(1-exp(h/(3a)); for 
you the responsibility to verify this.
2. setting nmax does not necessarly lead to uniquely defined 
neighbourhoods; think of data on a grid.

I understand and do recognize your worries. gstat has been in the open 
source for over 10 years now; before you spread fear, uncertainty and 
doubt I would have appreciated it if you had kept your observations and 
questions directed to this list more factual, and reproducable/verifiable.

Also note that gstat allows YOU to double check the underlying 
algorithms; this in contrast to the other software you mention. You're 
suggesting now to 1000+ subscribers that this burden lies with me, and 
worse, that I didn't do this.
--
Edzer

Yong Li wrote:
> Hi Edzer,
>
> I found a problem with gstat "krige" prediction if I am correct this
> time. For example, with OLSENP variable in my dataset, I developed a
> theoretical model for log(OLSENP) variogram: model=exp, c0=0.2019,
> c1=0.2218 and Range=162 meters. By using this model, I predicted the
> OLSENP surface using gstat krige and ArcGIS Spatial Analyst with the
> same nmax and found the results are significantly different. gstat
> significantly underestimated values (here is where I had an idea to use
> IDW), but Spatial Analyst gave very reliable predictions. I would like
> you to double check the algorithms behind gstat. 
>
> By the way, I start to puzzle because I used gstat to calculate other
> variables in my dataset, such as total nitrogen and extractable
> potassium and also I used gstat to run a series regression-kriging to
> predict soil organic matter distribution and submitted to a journal
> already.
>      
> I appreciate you could give me your response as soon as possible.
>
> Regards
>
> Yong Li
>
> -----Original Message-----
> From: r-sig-geo-bounces at stat.math.ethz.ch
> [mailto:r-sig-geo-bounces at stat.math.ethz.ch] On Behalf Of
> r-sig-geo-request at stat.math.ethz.ch
> Sent: Tuesday, 10 February 2009 10:00 PM
> To: r-sig-geo at stat.math.ethz.ch
> Subject: R-sig-Geo Digest, Vol 66, Issue 10
>
> Send R-sig-Geo mailing list submissions to
> 	r-sig-geo at stat.math.ethz.ch
>
> To subscribe or unsubscribe via the World Wide Web, visit
> 	https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> or, via email, send a message with subject or body 'help' to
> 	r-sig-geo-request at stat.math.ethz.ch
>
> You can reach the person managing the list at
> 	r-sig-geo-owner at stat.math.ethz.ch
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of R-sig-Geo digest..."
>
>
> Today's Topics:
>
>    1. error rgdal library if used from within Grass plugin
>       (Windows) (Agustin Lobo)
>    2. Re: error rgdal library if used from within Grass plugin
>       (Windows) (Roger Bivand)
>    3. Re: FW:  Interpolcation option: IDW or OK? (Tomislav Hengl)
>    4. Calculating descriptive stats from many maps (Rainer M Krug)
>    5. Re: Calculating descriptive stats from many maps (Tomislav Hengl)
>    6. Re: Calculating descriptive stats from many maps (Roger Bivand)
>    7. Interpolcation option: IDW or OK? (Yong Li)
>    8. Re: Calculating descriptive stats from many maps (Robert Hijmans)
>    9. Re: FW: Interpolcation option: IDW or OK? (Robert Hijmans)
>   10. Re: Calculating descriptive stats from many maps (Rainer M Krug)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Mon, 09 Feb 2009 15:06:24 +0100
> From: Agustin Lobo <aloboaleu at gmail.com>
> Subject: [R-sig-Geo] error rgdal library if used from within Grass
> 	plugin	(Windows)
> To: r-sig-geo at stat.math.ethz.ch
> Message-ID: <49903860.8060406 at gmail.com>
> Content-Type: text/plain; charset=iso-8859-1; format=flowed
>
> Hi!
>
> The following error only occurs if the R (2.7.2) session is started
> from a GRASS shell opened through the QGIS GRASS plugin in windows.
> The error does not occur If R is started form its own icon
> or by double click in the .RData  object,  but then the spgrass6
> package would not find the GRASS environment.
>
> It also works in linux.
> The involved  commands  are:
>
> 1. Start  QGIS
> 2. Star GRASS plugin and open mapset
> 3.  Open Grass Shell
> 4.  Run R and execute require(spgrass6)
>
>   
>>>>> Loading required package: spgrass6
>>>>> Loading required package: sp
>>>>> Loading required package: rgdal
>>>>> Error in fun(...) :
>>>>>         GDAL Error 1: Can't load requested DLL:
>>>>> C:\OSGeo4W\bin\gdalplugins\gdal_ECW_JP2ECW.dll
>>>>> 126: N?o foi poss?vel encontrar o m?dulo especificado.
>>>>>
>>>>>
>>>>>
>>>>> Error : .onLoad failed in 'loadNamespace' for 'rgdal'
>>>>> Erro: package 'rgdal' could not be loaded
>>>>>           
>
> Any help appreciated.
>
> Agus
>
>   

-- 
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