modeling territory fidelity of marked individuals

[CarrieO]

Hi,

My organization has banded approximately 5,000 Common Loons across North America over the last 25 years. COLOs are highly territorial during the breeding season and territories are highly contested; therefore, much turnover occurs and we'd like to estimate territory fidelity and model potential covariate effects, such as sex, body mass, etc. I'm trying to determine if MARK is the software for this analysis.

The data types we have include: capture, resights, recaptures, and recoveries and so it seemed like the Barker model would be suitable. However, survival is well established in COLOs and so I'm most interested in fidelity and movements between territories. And so I wonder if the multi state model could work where state A = territory where originally captured and banded and B = some other territory. Or do the states need to be specific locations where all the individuals may occur, such as the island example in the MARK book? If a loon is booted from its original territory, it could turn up in another nearby territory and could potentially remain in that second territory for many consecutive breeding seasons. It could potentially regain occupancy of the original territory after some years of occupying another territory or could go unsighted for several years before turning up in its original or new territory. As you can see there are a number of scenarios and it seems as though at least one would be violating some key assumption.

Lastly, since the banding of individuals occurs over a period of 25 years, many encounter histories will not overlap for many of the individuals. Is that ok? If so, would all individuals have a 25 column encounter history and 0s will occur until the year the individual was banded?

I realize there are a lot of questions here and any insight or guidance would be greatly appreciated.
Thank you for your time.
Carrie