[R-sig-eco] ADONIS and pseudoreplications

[syrovat at sci.muni.cz]

Dear listers,

I intended to send a separate question but it fell under comments on
Eduard Szöcs's question. Sorry for that.

The problem:
I collected aquatic insect larvae from 17 sites three times (spring,
summer, autumn), giving a total of 51 samples. During each sampling
campaign several chemical and substratum variables were measured, while
some others were measured only once, as they are not expected to change
over time.
Now I want to relate the assemblage taxonomic structure to environmental
variables and find those variables which explain considerable portion of
variance in the assemblage data-set.

Suggested analysis by Zuur et al.:
In chapter 26 of Zuur et al. 2007 (Analysing Ecological Data) the authors
analyze similar kind of data (12 time seris) and suggest using direct
ordination (RDA, CCA), or Mantel test if there are many zeros (because of
the possibility to select a measure of dissimilarity), to relate the
communities to the environmental variables and to identify the important
ones.

My thoughts:
As there are many zeros in the data, I decided to use adonis() (as a
better alternative to Mantel test).
But it seems that the samples are not independent - I have only 17 sites
with 3 replications for each. Am I correct?
I thought about setting strata=season in the adonis model as I am
interested in the differences among sites, but don't know if it solves the
problem? Moreover, the effect of environmental variables might differ
among seasons, therefore interaction with season should probably be
included.

I ended up with this model:
adonis(log(spe+1) ~ env1 %in% season + env2 %in% season + ..., distance=
'bray', permut= 999, data= env)
(where season is a factor with 3 levels)

I'm not convinced the model is correct and would greatly appreciate any
comments.

And, please, let me know if I am completely wrong.

Thank you very much and sorry for the long text,

Vit

--
Vit Syrovatka
Department of Botany and Zoology
Masaryk University
Kotlarska 2
CZ-611 37 Brno, Czech Republic

Tel.: +420-532146342
E-mail: syrovat at sci.muni.cz