[R-sig-eco] pDCA

[Jari Oksanen]

On 7/06/11 12:45 PM, "Klara Dolos" <klara.dolos at uni-bayreuth.de> wrote:

> Hallo all,
> 
> I ´d like to calculate a partial detrended corresponcence analysis to get the
> gradient length (and the axis coordinates) in sd for one paricular
> environmental variable. I tried to use decorana(vegan), but the function
> cannot perform partial analysis (?). Is there an R function that does this
> analysis or is there a reason why it should not be done?
> 
> As an alternative I could employ an environmental fit (line enfit(vegan))
> after calculating a dca and take the coordinates on the vector?
> 
> I would be happy if anybody could give me hints on literature or an R package
> I could use.
> 
Kalra,

As far as I know, vegan has the only DCA implementation in R similar to the
original DECORANA. You cannot use that for partial or constrained detrended
analysis. This would require changing the original Fortran code which in
turn requires a good knowledge of the original design of the function and
cannot be done casually. I have no plans to make the changes needed for
partial or constrained DCA.

I abstain from commenting the use of DCA in estimating gradient lengths (but
don't recommend anybody to do so).

Cheers, Jari Oksanen