[R-sig-Geo] error in spautolm spdep and model selection?

[Michael Denslow]

Thank you Roger and others who responded on and off list,
Roger, I am happy to send the data that generated the error and will do so off list.

> > 
> > Perhaps to larger question is which is the appropriate
> SAR/CAR model to use for my data. I realize that this is not
> an R question but any help would be most appreciated! My
> data has some spatial autocorrelation in the predictor
> variables. The residuals of the lm() model are not
> autocorrelated. Also the elevation variable is strongly
> autocorrelated and directional in my data set. As I
> understand it this would rule out CAR as an option.
> 
> The models you are fitting are trying to harvest spatial
> structure in the residual. If there isn't any spatial
> structure to harvest, lm() is quite good enough. Spatial
> pattern in the X variables is not a problem, especially if
> it matches the spatial patterning in the response. Do you
> have a reference for your claims above, they sound like
> heresay to me? 

I think that the problem is that I miss understood a statement that I read about the CAR model. In Dormann et al. (2008 p. 614, Methods to account for spatial autocorrelation in the analysis of species distributional data: a review) they state:

'CAR is unsuitable when directional processes such as stream flow effects or prevalent wind directions are coded as non-Euclidean distances, resulting in an asymmetric covariance matrix.'

I originally took this to mean that if you have some effect that is strongly directional CAR is not appropriate. After looking at it again, I am frankly not sure what it means. In any case, it appears that I may not need any of these models at this time. This leads to my last question. 
Should I present the results of my Moran's I test of lm() residuals (lm.morantest) to show that they are indeed not autocorrelated and a spatial regression is therefore not needed? 

Thank you again,
Michael

> > 
> > Michael Denslow
> > 
> > I.W. Carpenter Jr. Herbarium [BOON]
> > Appalachian State University
> > Boone, North Carolina U.S.A.
> > 
> > -- AND --
> > 
> > Communications Manager
> > Southeastern Regional Network of Expertise and
> Collections
> > sernec.org
> >