[R-sig-Geo] Variogram estimation without spatial reference

[Edzer Pebesma]

What you did sounds to me like an (unbalanced) nested sampling scheme.
One of the first computer programs I ever wrote analysed such data (in
the late 80's, in Pascal) but it got lost somewhere in the migration
from 5.25" 360 Kb floppy disks to 256 Gb SSD drives.

As the underlying model was, IIRC, a mixed effects ANOVA model, I guess
lme() in package nlme would now provide you the means to analyse them;
you'd need to define a factors for each distance class, and have this
nested within the factor meadow.

On 06/12/2012 09:53 PM, Tom Gottfried wrote:
> Am 12.06.2012 17:29, schrieb Gunnar Oehmichen:
>> Dear R-Sig-Geo list,
>>
>> as part of an applied field-laboratory course, I will try to estimate
>> variograms of soil-lipid contents for two meadows differing in use.
>>
>> Samples were taken in the following scheme:
>>
>> Code dist X Y Wat mead Pair
>> 21 B6-1 0.0 443840 5451412 B B6 6
>> 22 B6-1 0.5 443841 5451410 B B6 6
>> 23 B6-1 1.0 443843 5451410 B B6 6
>> 24 B6-1 2.0 443844 5451411 B B6 6
>> 25 B6-1 4.0 443841 5451410 B B6 6
>> 26 B6-1 8.0 443848 5451409 B B6 6
>> ...
>>
>> column dist represents the distance [m] to the preceding point.
>>
>> This scheme was replicated on each meadow 5 times to facilitate
>> replicates of each lag (0.5, 1.0, 2.0, 4.0, 8.0). We picked a random
>> point (dist = 0.0) sampled, moved a distance of 0.5 m in a random
>> direction, sampled and repeated the procedure with increasing distances.
>> Allthough lattitude and longitude were recorded (X and Y), I do not
>> consider these values to be helpful, since the measurement error of the
>> GPS device is ~ 6m, so a much coarser resolution than needed.
>>
>> Results of the analysis are still missing, so no need to worry about
>> that.
>>
>> but to the point, do you know of a way to use functions like variogram()
>> with this kind of data?
> 
> No, but it should be pretty easy to implement a function that calculates
> semivariance between the rows of your data (using diff()). Taking the
> average of semivariances for each value of dist gives you a
> semivariogram. But it gives you not more than 10 point pairs for each
> distance, which is not much.
> 
> HTH,
> Tom
> 
>> Thanks a lot,
>>
>> Gunnar Oehmichen
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> 
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> 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