[R] pcnm - scaling

[robert.ptacnik at niva.no]

Dear listers,
I am using the pcnm function (spacemakeR) to obtain eigenvectors for a
spatial grid of sampling sites. These pcnm eigenvectors are then used in
multivariate ordination to test where community composition follows local
environment or rather shows spatial autocorrelation, and get support for
apatial autocorrelation.
In an RDA analysis, I can see which eigenvectors correlate most with my
data, which can be interpreted that the spatial scale is rather 'fine' or
'broad'. However, is there a way to read the spatial scale on which the
data correlates best, i.e. is it possible to extract a spatial scale
describing the apatial autocorrelation (e.g. from the eigenvalues)?
My data

- irregularly spaced sampling sites, total scale ca. 2000 km, most sites
few tens of km apart
- i use the dfunction rdist.earth to obtain a distance-matrix of all sites
prior to using pcnm
Thanks !
Robert


 Robert Ptacnik
 Norwegian Institute for Water Research (NIVA)
 Gaustadalléen 21
 NO-0349 Oslo
 FON +47 982 277 81
 FAX +47 221 852 00