Dear all,
Im working in a field experiment with 20 traps (pitfall traps to collect ground atrhopods) in a treatment field and 20 traps in a reference field (so potentially spatially autocorrelated).
I performed a nMDS (non param multidimensional scaling) plot to assess multivariate ordination of those samples and I plotted also 95% confidence ellipses to visualize effective discrimination between the treatment and the reference field. Then I would like to have a statistical measure of this discrimination so my idea was to perform a perMANOVA (adonis function in R software) to test dissimilarity between fields. So my question is:
-Can I use perMANOVA with such experimental design? If not, is there a way to deal with such autocorrelation? Suggestion on alternatives?
Thanks a lot
Alessandro
While this is not a complete answer, it may get you started. Why don't you just determine whether you actually have spatial autocorrelation (e.g. variograms and Moran's I). Then if you pass that assumption, just proceed as normal?
Hope that helps!
Cheers,
Remi
Thanks Remi, I will try to make a Mantel test comparing animal abundance distance matrix with the location (of the traps in the field) matrix. I should solve the problem in this way. Thanks a lot
- Badges
- Science topic
- Similar topics
- Biological Science
- Biota
- Eukaryota
- Fungi