redundancy analysis to explain mortality causes

Hei! I am analyzing some data on mortality of roe deer and I was wondering
if it would be correct to use a multivariate method like redundancy analysis
(or canonical correspondence analysis). I have two sets of variables, one
including the number of animals that die each year for each age class and
municipality and another with the age class and the roe deer density in that
year / municipality.
I have run first a DCA to get the gradient length and then a RDA since the
gradient was below 3. I get a very nice multidimensional graph where the
position of the response variables with respect to the vectors representing
the explanatory variables makes sense, the overall model is significant
(according to the permutation test) and explains 42% of common variance.
But I have never seen this type of analysis used for mortality causes, I
have seen it used in community ecology and in a diet study (use versus
availability). I would like to use it because the graph is neat and easy to
understand and because it is a multivariate analysis and I think that the
mortality causes are not independent from each other.

What do you think?

Thanks a lot for any help and sorry if I have not been clear!

Claudia

claudia.melis@xxxxxxxxxxx
http://www.bio.ntnu.no/users/melis/


.