What is the full model that you are trying to fit?
A repeatability model?
If so, what is the difference between a year by animal interaction and a
permanent environmental effect?
I would hazard a guess and say that the former is related to the variability
and the latter is the similarity between records of an animal.
My understanding is that you have one observation per year per animal.
With your data structure I would think that these are totally confounded.
If you are not happy about the assumptions of the repeatability model you
should try an alternative longitudinal approach such as nonlinear mixed effect
models and covariance functions.
Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml