[R-sig-Geo] geoR interpolation with log transform

[Paulo Justiniano Ribeiro Jr]

Dear Kapo

I've replied your private message with some details
Let me add the "honouring the points" will dependes on which 
interpretation is given to the nugget (micro-scale variation measurement 
error or a combination of these)

In geoR there are arguments micro.scale and signal in n
krige.control() and output.control() respectivelly, to control such 
behaviour

Paulo Justiniano Ribeiro Jr
LEG (Laboratorio de Estatistica e Geoinformacao)
Universidade Federal do Parana
Caixa Postal 19.081
CEP 81.531-990
Curitiba, PR  -  Brasil
Tel: (+55) 41 3361 3573
VOIP: (+55) (41) (3361 3600) 1053 1066
Fax: (+55) 41 3361 3141
e-mail: paulojus AT  ufpr  br
http://www.leg.ufpr.br/~paulojus

On Sat, 13 Nov 2010, kapo coulibaly wrote:

> I?m using geoR to interpolate hydraulic conductivities. I constructed the
> variogram using the log transform as follows:
>
> dummy.var<-variog(dummy.geo,max.dist=40000,uvec=25,lambda=0)   # Then I
> fitted a spherical model using eyefit parameters as starting parameters
>
> dummy.param<-eyefit(dummy.var)
>
> param1<-variofit(vario=dummy.var, ini.cov.pars=dummy.param[[1]]$cov.pars,
>
> cov.model = dummy.param[[1]]$cov.model,nugget=dummy.param[[1]]$nugget)
>
> spatPara<-krige.control(type.krige="ok",obj.model=param1)
>
> # I then interpolate across the map
>
> InterMap<-krige.conv(dummy.geo,loc=loc,krige=spatPara)
>
>
> Ma question is why are my interpolated values higher than the actual
> measurements. I expected the interpolation to honour measurement points at
> least. Am I missing something?
>
>
>
> Thanks
>
> 	[[alternative HTML version deleted]]
>
>