Although I have never solved this issue myself, I have been interested in it for a while, and as far as I could figure out, I would agree that one of the biggest challenges might be the determination of life history traits.

To measure migrational direction, besides colour markings of putative source populations followed by capture/recapture methods it might perhaps be possible to use isotopic labelling (with all the logistic problems that entails).

After assessing all those hypothesis, and because the animals I was interested in using them in are rather rare, I believe that the best approach would be to perform a populational phylogenetic analysis and/or measure populations genetic diversity (the same method used to produce population size estimates).

But that being said, corridors can be formed by sink populations which existence, even if for a single biologic cycle, can bring connectivity benefits between source populations, which if proven to be the case, would interfere with this type of analysis, as it might affect sink populations genetic diversity.

Wishing I may have revealed more reliable paths than added questions.

S.

You could study insect reproduction (number of eggs deposited/ larvae present and "native" adults) in each vegetation type. In some cases is easy (e.g., dragonflies reproduction in some saline ponds is zero -> are a sink populations) but analysing this fact in other situations is more difficult.