[R-sig-eco] envfit on constrained ordination

[Tim Richter-Heitmann]

Dear group,

i was reading about this issue, in which envfit was performed on a dbRDA 
model

https://github.com/vegandevs/vegan/issues/248

I think many people would perform some kind of a variable selection on 
this model, and possibly get a subset of significant predictors by 
anova.cca.

In this example, envfit is used instead.

After

plot(dbrda)
plot(vf, col = 'red')

one can see that the differences are very small between the vector 
arrangements (vf vs. original dbRDA). So, my question is, in which 
scenario its recommended to perform envfit on top of an constrained 
ordination, which already tries to find a visual representation of 
environmental gradients through the ordination.

Thank you!



-- 
Dr. Tim Richter-Heitmann

University of Bremen
Microbial Ecophysiology Group (AG Friedrich)
FB02 - Biologie/Chemie
Leobener Straße (NW2 A2130)
D-28359 Bremen
Tel.: 0049(0)421 218-63062
Fax: 0049(0)421 218-63069