open population - real survival estimate - multistrata

[simone77]

Hi,

I would appreciate a lot some opinion on it:

I have a dataset with yearly (during breeding period) marking (readable rings) of chicks of a species of bird and several resightings occasions all through the year. The resightings were done during sessions with different time gaps among them. Moreover the resighting effort was not constant (different time length of the sessions).

The population is geographically and demographically open: according to a preliminary look at our dataset it seems clear that a part of population stays here all around the year, another part emigrates, someone come back, some individuals immigrates from other regions, someone recruits and so on.
Now I’m interested in estimating the survival rate of the post-fledging age classes (every individuals still alive when they begin to fly) during different months.
I organized my dataset by considering only the individuals seen at least once after their first breeding seasons (talking about the chicks) and I regarded them as “1” (seen) when seen at least once in a month and otherwise as “0” (not seen). I suppose that being the resighting effort not constant all troughs the year the monthly estimates of the parameters (i.e. survival rate) will depend on that and I would trust more in estimates from months with more resighting sessions (more effort), wouldn’t I?

Apart of how to “manage” the different factors (age, cohort, gender, environmental factors, etc.) possibly affecting the estimate of the parameters I’m interested in, in this context what is the best way of performing the analysis in order to better estimate real survival of post fledging individuals?
I am thinking of a multi-strata approach by considering two states (A=seen inside my study area, B=seen outside). In such a way I can account for emigration rates (A -> B) and better estimate real survival, isn’t it?

How should I need the heterogeneity in terms of resighting effort?

I assume that it is very important the sample size and the number of parameters I want to estimate so: I have 13 years, a population growing up almost exponentially, more than 7.000 ringed chicks, a 3.000 individuals seen at least once, a 400 individuals seen outside of my study areas.

Thanks in advance for any suggestion

[Fernando Arce]

Dear Simone, I think you may better explore Mark-resight models (Chapter 18 of the book) as is not exactly a capture-recapture experiment.

Anyway, I can try to comment some of your questions (but I'm not an expert) out of the forum, as maybe are a bit off topic.

Good luck with the moritos

Fernando
farcego@gmail.com

[simone77]

Hi,

I don't believe I should use Mark-Resight models because I'm interested in estimating survival and emigration rate and I think that Mark-resight models are an adaptation of the NOREMARK models implemented to estimate closed population and superpopulation abundance.
I hope my questions are not off topic: I'm asking for statistical and analytical help in a MARK analysis context, don't I?.

Anyway, thank you very much for your answer.
Cheers from moritos

Simone