[R-sig-Geo] specifying correlation structure in (n)lme

[Roger Bivand]

On Fri, 14 Jun 2013, Gregoire, Timothy wrote:

> Gabriel,
>
> Look at the Pinheiro & Bates book on Mixed Models, it is part of the 
> Springer series and is prob'ly an online resource thru the Harvard 
> library system.
>
> An alternative, outside the mixed models construct, is to use the 
> likfit() function of the geoR package by Diggle & Ribeiro.

But note that the coordinates are said to be given in "decimal long/lat", 
which if this means decimal degrees in geographical coordinates, and the 
study area is not on the Equator, may lead to trouble. If the study area 
is not global, you should project the coordinates to a planar 
representation, otherwise the distances used may not be what you think, 
and there may be instability numerically.

Roger

>
> Tim
>
> Timothy G. Gregoire
> J. P. Weyerhaeuser Professor of Forest Management
> School of Forestry & Environmental Studies, Yale University
> 360 Prospect Street, New Haven, CT  06511-2104  U.S.A.
>
> office: 1.203.432.9398  mobile: 1.203.508.4014, fax: 1.203.432.3809
> timothy.gregoire at yale.edu
> G&V sampling text: http://crcpress.com/product/isbn/9781584883708
>
>
> -----Original Message-----
> From: r-sig-geo-bounces at r-project.org [mailto:r-sig-geo-bounces at r-project.org] On Behalf Of Gabriel Gartner
> Sent: Thursday, June 13, 2013 8:43 PM
> To: r-sig-geo at r-project.org
> Subject: [R-sig-Geo] specifying correlation structure in (n)lme
>
> Hi All,
>
> I have what I imagine is a fairly basic question, but I am a relative 
> beginner at using spatial data in my analyses.  I am trying to address 
> spatial autocorrelation in my data which examines lizard morphology 
> across a broad geographic range.  My spatial data is in the form of 
> decimal long/lat and I am trying to select which variogram model to use.
>
> My question is, what is the form of the correlation structure within 
> each of the corClasses for decimal lat/long data.  In other words, what 
> goes after "form" in the following example script:
>
> exp.cybotes<-update(null.model, correlation =corExp(1, form = ???), method= "ML")
>
> I apologize for the basic question but I am stuck!
>
> Thanks,
> Gabriel
>
>
> Gabriel E.A. Gartner
> Postdoctoral Fellow
> Museum of Comparative Zoology
> Harvard University
> 26 Oxford St.
> Cambridge, MA 02138
> Office: 617-384-8437
> http://www.oeb.harvard.edu/faculty/losos/Gartner/index_2.html
>
>
>
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> 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
>

-- 
Roger Bivand
Department of Economics, NHH Norwegian School of Economics,
Helleveien 30, N-5045 Bergen, Norway.
voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no