[R-sig-Geo] variogram maps / newbie

[Edzer Pebesma]

Facundo,

your explenation is very clear.

When I once tried, In addition I found different numbers of point pairs 
across the two packages; this could be due to classification of point 
pairs with distances exactly on the bin boundary, but I didn't come to a 
conclusive feeling about it, back then.
--
Edzer

Facundo Muñoz wrote:
> Hi Pedro,
>
> perhaps if you could paste some sample code, we could talk about 
> something more concrete...
>
> I understand that you were able to plot empirical variograms using both 
> variog (geoR) and variogram (gstat).
>
> In both of them you control the maximum distance:
>    in geoR: using max.dist, or uvec which also allows to define the bins
>    in gstat : with cutoff
>
> and the binning: uvec (geoR) and width (gstat).
>
> In despite of this, they will not plot the same empirical variograms 
> even if you define the same bins and maximum distance, since variog 
> (geoR) places the estimates exactly at the bins limits, while variogram 
> (gstat) calculates the mean of the distances of the pairs of 
> observations used for each estimate, and places the estimate at that 
> precise distance. (Ajjj, i'm sure it can be said better)
>
> In sum: even if you manage to make them calculate the same estimates 
> with the same pairs of observations each, you will have the same 
> estimates but located at (more or less slightly) different ordinates.
>
> Reagards
>       Facundo.-
>
> Pedro Mardones escribió:
>   
>> Dear all;
>> Does anyone knows how to obtain a variogram map using the package geoR?
>> I was able to obtain these plots using the gstat package but the
>> results form the variog function are different to those from variogram
>> function in gstat (I guess that has something to do with the bin and
>> maximum distance but not sure)
>> Suggestions will be appreciated
>> PM
>>     
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>