[R-sig-Geo] regression kriging with singular variogram

[Edzer Pebesma]

On 01/25/2015 01:49 PM, Bede-Fazekas Ákos wrote:
> Dear All,
> I have a simple question about regression kriging (gstat:::krige()) and
> variogram fitting (gstat:::fit.variogram()), which I have not found exact
> answer for (neither in the description of package gstat, nor on the
> internet). Does it cause any problem if I do regression kriging with a
> fitted variogram (eg. a spherical or exponential type) that is singular (ie.
> I got a message 'warning: singular model in variogram fit') but shows a
> curve in plot(experimental_variogram, fitted_variogram) that is expected
> (ie. it fits nicely to the points and has a shape as drawn in the books)?

No; the only problem is that with this message printed, you still need
to look at the variogram and see if it make sense at all, as the fit
method could not find some optimum.

> 
> A bonus question: is the message 'warning: singular model in variogram fit'
> a warning or a simple message? I tried to use
> fitted_variogram <- suppressWarnings(fit.variogram(...))
> and
> fitted_variogram <- suppressMessages(fit.variogram(...))
> but none of them hides the message. Please let me know how to hide it.

You can't, because it uses Rprintf(). I've changed this to use the R
warning mechanism, so in future versions you can.

-- 
Edzer Pebesma,  Co-Editor-in-Chief Computers & Geosciences
Institute for Geoinformatics (ifgi), University of Münster
Heisenbergstraße 2, 48149 Münster, Germany. Phone: +49 251
83 33081 http://ifgi.uni-muenster.de GPG key ID 0xAC227795

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 490 bytes
Desc: OpenPGP digital signature
URL: <https://stat.ethz.ch/pipermail/r-sig-geo/attachments/20150126/5ab2cbc3/attachment.bin>