[R-sig-eco] Spatial AND temporal correlation in linear mixed-effects models

[Bob O'Hara]

On 07/26/2012 10:45 AM, Tscheulin Thomas wrote:
> Hi,
>
> I have a question regarding the nesting structure in linear mixed models of data, which is spatially and at the same time temporally correlated.
>
> So far I have tried to get around the problem by averaging out the temporal component but I would really like to keep everything in the model.
>
> Here is the experimental set-up:
> We have measured insect abundance in two different islands, each having 5 sites, each site having 4 plots, in which each we have 5 collection points (traps). The sampling was repeated 5 times. Two times in 2011 and 3 times in 2012.
>
> After averaging the temporal correlation out I composed the following model for the abundance:
>
> model<-lme(abundance~continous_explanatory_variable,random=~1|island/site/plot/trap,method="REML")
>
> This works fine but I have problems making a model that allows the temporal component to stay in.
> How can I do this using the function lme?
>
> I know with lmer I could do this:
>
> model2<-lmer(abundance~continous_variable+(1|island/site/distance/triplet)+(1|year/round))
> but I really want to use the function lme. How can I insert multiple levels of grouping in lme?
>
> Any help is very much appreciated!
> Best,
A bit of googling brought up this, from 10 years ago:
<http://tolstoy.newcastle.edu.au/R/help/02b/2068.html>
I don't think lme has changed that much.

Bob

-- 
Bob O'Hara

Biodiversity and Climate Research Centre
Senckenberganlage 25
D-60325 Frankfurt am Main,
Germany

Tel: +49 69 798 40216
Mobile: +49 1515 888 5440
WWW:   http://www.bik-f.de/root/index.php?page_id=219
Blog: http://blogs.nature.com/boboh
Journal of Negative Results - EEB: www.jnr-eeb.org