I am using the R package mvabund to examine how environmental conditions and species traits are correlated with ecological community structure. The traitglm() function is a nice tool for this. However, my data consist of ~40 'sites' repeatedly sampled annually over a 20-year time period. I think an appropriate method would be to include 'year' as a fixed effect and 'site' as a random effect, but I do not see an option in mvabund to include random effects. Note that the environmental conditions are not expected to dramatically vary across space as much as through time. So, incorporating the temporal element is very important, but the sites of course are resampled each year. If there is no such way to include random effects in mvabund, how should I handle time?
Asked
Active
Viewed 32 times
1
Joshua Smith
- 125
- 6