an open population that becomes closed

[cassowary43]

Hi everyone,

I wonder if anyone has a creative solution to deal with a winter bird population that is open as birds arrive on wintering grounds (due to searching for "optimal" habitat within the wintering area), but quickly becomes closed as birds settle into small (<1ha) home ranges?

We captured birds by systematically walking the 2.25-plot with a team of workers and chasing flushed birds into mist-nets. We sampled 10 plots 5 times (one month intervals) during the winter. The last 4 sampling periods are during the site faithful period and the 1st is during the movement period (banded birds disappear).

I would like to know if capture probabilities differ between the "open" and "closed" period. I suspect not given the temporal scale of sampling a plot (1-3 hrs in the day) vs the temporal scale at which birds are moving around (we do not witness extra-plot movements, which may take place at night).

In other words I believe the population is closed within a trapping occasion, but open between the 1st and 2nd trapping occasions. Afterwards, the population can be assumed to be closed within and between trapping occasions.

I realize that a Robust design is the optimal way of estimating survival and capture probability in an open/closed situation, but we did not design the study to run "sessions" of sampling, partially because this is an unexpected find and hasn't been previously described in this species. Should I simply use a CJS model (recognizing that phi reflects movements and not survival during the movement period)? If so, will capture probabilities be meaningful? Or is there a better alternative like an open JS model?

Many thanks for any ideas you might have,
Erik
ejohn33 AT lsu.edu