[R] vegan cca: syntax

fpgraz fgraz at unam.na
Mon Sep 26 14:36:16 CEST 2011

Dear all,
I am a new member to the list - and to the analysis that I am attempting. 

I have the following case

A group of us have been monitoring (over a period of a few years) a number
of paired plots that were flooded and / or burnt. 

The plots are located in two topographical settings, some were burnt, some
were flooded, some were burnt & flooded and some were not affected at all.
At each location one plot was fenced to exclude grazing while another was
left unfenced to include the effect of grazing. Effectively, I guess, we
have a split plot design in a longitudinal study.

The plots were established opportunistically, so there were some practical
aspects that might not be ideal... but then when is this ever the case... 

I now want to determine importance of various effects of these … effects...
and used vegan's cca in the following context:

1. I read in the count data

2. transformed the counts using a 4th root transformation into count.dbf2 –
based on a suggestion from a colleague and following up in “Quinn, G. P.,
and M. J. Keough. 2002. Experimental design and data analysis for
biologists, 1st edition”.

3. I then used the following two commands to generate an ANOVA table:
> out.cca2 = cca(count.df2 ~ Flood*(Fence+Fire+age)+topo, predictor.df)
> anova(out.cca2, by="term", step=2000)

4. Since the order in which the factors are entered seems to matter, I tried
a number of iterations obtaining similar results to:

Model: cca(formula = response.df2 ~ Flood * (Fence + Fire + age) + topo,
data = predictor.df) 
                Df   Chisq       F     N.Perm Pr(>F)    
Flood           1  0.4643 13.3495   1999  5e-04 *** 
Fence          1  0.1507  4.3342   1999  5e-04 *** 
Fire             1  0.1049  3.0152   1999  5e-04 *** 
age             1  0.1756  5.0492   1999  5e-04 *** 
topo            1  0.1520  4.3710   1999  5e-04 *** 
Flood:Fence  1  0.0868  2.4949   1999  5e-04 *** 
Flood:Fire     1  0.0952  2.7378   1999  5e-04 *** 
Flood:age     1  0.1235  3.5510   1999  5e-04 *** 
Residual    171  5.9470                          

I am somewhat isolated here and have no-one to cross check the syntax. I
could not include the 'split plot' aspect in the equation. Should I worry
about this?. 

Any assistance would be very welcome.



View this message in context: http://r.789695.n4.nabble.com/vegan-cca-syntax-tp3843617p3843617.html
Sent from the R help mailing list archive at Nabble.com.

More information about the R-help mailing list