[R] vegan CCA I am Completely new to ordination analyses

[Jari Oksanen]

kitty <kitty.a1000 <at> gmail.com> writes:

> 
> Dear list,
> 
> I am trying to predict species volume from bioclimatic data, I have various
> sites and I have a data frame with species volume and
> the corresponding bioclimatic data for each site.
> 
> I read on a discussion forum that you can use ordination to predict species
> abundance (in my case volume) from 'new' climate data for sites where you do
> not know the abundance.
> 
> Unfortunately I cannot work out how to do this in the r package vegan, in
> fact I cannot even get a simple CCA to work; I have attached a sample of my
> data (volume and bioclimate data for around 100 of my sites) and the code I
> used is below.
> 
> Does the output look right? why am I getting a 0 output? Is
> there something wrong with my data, data format, r code etc?
> 
Ordination, including cca(), is multivariate analysis. If you only have one 
dependent variate, you should use univariate methods (regression etc.). 
Some multivariate methods fall back to univariate methods if you only 
have one dependent variable, but cca() does not do so in a meaningful way. 
In fact cca() does do this, but not meaningfully: your residual variation is 
zero and rank is zero. Switch to regression.

Cheers, Jari Oksanen