| R codes & Models - |
|
|
1. R code for calculating measures of Functional Diversity and Trait Dispersion in a set of samples.
This code will calculate Petchey and Gaston's FD, FD based on the Rao Index, Mason et al.'s FDvar,
MST -- the minimum spanning tree distance (an overlooked measure of FD, which is about the same as the Petchey and Gaston Index -- it's older than all these),
mean Nearest-Neighbor Distance in trait space, as well as H, effective richness, and evenness.
Here's a link to the code used by Villéger, Mason, and Mouillot (Ecology 89: 2290-2301)
where they calculate Functional Richness as the convex hull volume, Functional Evenness, and Functional Divergence.
I some misgivings about the convex hull because you must have one more species than trait dimensions
in order to have a non-zero volume. I think this is problematic.
If trait dimensionality is high, then species-poor communities have zero Functional Richness.
Villéger says that in such a case, FR is greater than zero, but it cannot be measured.
This answer is not particularly satisfying to me.
Never the less, their methods are helpful and should prove to be quite useful.
2. shortly: R code for null modeling some of these.
3. shortly: polishing this page
4. Some favorite links for ecological computing:
Jim Grace's Structural Equation Modeling site with helpful tutorials on the
use of AMOS and more.
Michael Palmer's Ordination site
Owen Petchey's group - and while we're at it, |