0
$\begingroup$

I'm doing partial distance based (db-)RDA analysis of nematode community data. I have got the triplots already. They contain vectors of environmental data.

Is it possible to overlay a second set of vectors, describing the relationship of biodiversity/nematode community Indices with community structure of the sites? Or do i have do perform a second db-RDA with these Indices as explanatory variables, and hence replace the environmental data for these Indices. Or is the use of Biodiversity indices, calculated on the basis of species communities, to explain these communities a circle reasoning?

$\endgroup$
0
$\begingroup$

So it is possible to do a constrained ordination of a species matrix by environmental variables then extract these reduced-space site scores and examine their relationship with vectors of biodiversity/any other variables. But the question it answers is only related to the proportion of variation in the primary (community) matrix explained by those secondary (environmental) variables. I.e. is the structure of the primary matrix constrained by environmental variables in any way related to vectors of biodiversity, if that’s what the question is.

Alternatively, you can constrain the species matrix by vectors of biodiversity, i.e. using them as the secondary ‘environmental’ matrix, which is what you are probably referring to in ‘replace the environmental matrix’. This is an entirely different analysis/question of what biodiversity vectors/metrics are most important in explaining/describing the variation in community composition. I think those two are mixed up a bit in the question above or else the goal is not entirely clear.

$\endgroup$
  • $\begingroup$ I'm doing distance based rda, because this enables me to use constraints, by calculating multiple multivariate regression a priori and extract the important environmental variables, wo only these were displayed. As far as i know with NMDS I'm only able to infer hypothesis a posteriors, like with PCA. $\endgroup$ – Pharcyde Aug 19 '15 at 7:35
  • $\begingroup$ If this question is is too applied/field-specific for CV, which forum would you recommend? $\endgroup$ – Pharcyde Aug 19 '15 at 7:47
  • $\begingroup$ @Pharcyde, I edited it to clarify, keep specifying if it still does not answer your question. $\endgroup$ – katya Aug 19 '15 at 18:05
  • $\begingroup$ Thanks a lot for your answer, it elucidated my way of thinking about that problem a lot. Just to make sure i got it right: Constraining species community matrices by biodiversity indices, calculated on the basis of the same species matrix, is not a circular reasoning? I'll edit my question to clearify. But thanks for answering! $\endgroup$ – Pharcyde Aug 27 '15 at 11:33
  • $\begingroup$ It is not entirely circular because site position in the biodiversity indices space is not identical to how those sites are ordinated in a species space; as for reasoning, you'd need to find a good reason why you want to do it, ie what is the question you are trying to answer with this analysis. $\endgroup$ – katya Aug 27 '15 at 19:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.