using envfit with vectors of different magnitudes in R

Question

When using envfit in the vegan package, does the magnitude of each parameter matter?

I have two dataframes:

bwsp: invertebrate abundances

bwenv: five environmental parameters (distance, depth, flow, chlorophyll, phaeophytin)

Three of my parameters have similar magnitudes, and the other two are 2-3 orders of magnitude smaller. The smaller parameters don't appear on the MDS plot.

m <- metaMDS(bwsp)
ef <- envfit(m, bwenv, perm = 1000, na.rm = TRUE)
ef
plot(m)
plot(ef, add = T, p. = 0.05)

score 0 · Answer 1 · answered Feb 23 '17 at 09:58

0

envfit is based on correlation, and the original magnitude does not influence its results. However, you specified P-value limit p. = 0.05 in your plot, and this may have dropped some variables.

This is guessing, because you don't provide a reproducible example.

answered Feb 23 '17 at 09:58

Jari Oksanen

3,287
1
11
15

Got it, thank you - I'm just learning the nuts and bolts of envfit so this helped a lot. – ayesha Feb 23 '17 at 21:59

using envfit with vectors of different magnitudes in R

1 Answers1