[R-sig-Geo] maximum likelihood variogram parameter estimation, non-Euclidian distances

[Edzer Pebesma]

On 07/15/2010 01:24 AM, Timothy W. Hilton wrote:
> Hello R-sig-Geo list,
> 
> I want to calculate maximum likelihood estimations for variogram
> parameters (sill, range).  My data locations span North America, thus I
> believe that UTM coordinates (or any other projection) will distort
> distances between sites too much for me to use Euclidian coordinates.
> 
> As far as I can tell, geoR will not accept non-Euclidian coordinates or
> allow the user to specify a distance function.  gstat accepts latitude
> and longitude to define locations, but does not include maximum
> likelihood estimation for the range (fit.variogram.reml estimates sill
> only).

True, but you could write your own brute-force loop around it, or
optimize() around it, trying range values and selecting the one with the
maximum likelihood, or write out the likelihood directly and use optim.

Another problem you might be facing is that not all of the usual
variogram models are suitable (in the sense of positive definite) for
spherical distance measures. I'm still searching for a good reference on
this issue.

> Is anyone aware of an R geostats package that handles my situation?
> 
> Many thanks!
> Tim
> 
> --
> 
> Timothy W. Hilton
> PhD Candidate, Department of Meteorology
> The Pennsylvania State University
> 503 Walker Building, University Park, PA   16802
> hilton at meteo.psu.edu
> 
> _______________________________________________
> 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