[R-sig-Geo] semivariogram calculation - discrepancy between gls and gstat

[Edzer Pebesma]

Guido, I suspect that nlme works with models for spatial correlation and
not for spatial covariance, meaning that the variance (variog) gets
standardized - see ?corClasses in nlme. Do the variog values approach 1
for longer distances?

On 07/05/2010 11:02 PM, Guido Lorenz wrote:
> Dear R-sig-Geo members,
> 
> a semivariogram for spatial soil data, calculated by the gls (nlme library) function gives the following estimates:
> 
>> lmm.dap2b <- gls(Dap.sa ~ 1 + ID.sitio, Pd2006mm, correlation=corGaus(form=~(easting.m+northing.m)|ID.sitio, nugget=TRUE, metric="euclidean"), na.action=na.omit, method="REML")
>> Variogram(lmm.dap2b)
>       variog      dist n.pairs
> 1  0.3059585  2.828427     122
> 2  0.4040062  4.269282     125
> 3  0.5744688 18.110770     123
> 4  0.5266091 20.000000     125
> ....
> 
> whereas the variogram function of the gstat library gives, for similar distances (although different number of sample pairs), very different gamma values: 
> 
>  > variogram(Dap.sa ~ 1, locations = ~ easting.m + northing.m, data=Pd2006mm, cutoff=80)
>     np      dist       gamma dir.hor dir.ver   id
> 1  219  3.225693 0.005578128       0       0 var1
> 2   20  6.478671 0.004254656       0       0 var1
> 3    3 13.513045 0.009896324       0       0 var1
> 4  307 19.201322 0.009708390       0       0 var1
> .....  
> 
> Can anyone explain what is happening?
> 
> Thanks for any advice,
> Guido Lorenz
> 
> Can anyone explain
> 
> _______________________________________________
> 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