Robust Design with POPAN for the secondary occasions

[simone77]

Hi,

I have Capture-Mark-Resight data from a study of a breeding colony of birds during several years.
The captures (and markings without losses on capture) are from ringing operations during the breeding seasons in the breeding colony.
The resightings are from surveys performed quite randomly and sparsely (I know it is not the best way to do it) all the year along in the neighbour area, most of them are during the breeding season (at the breeding colony).

Question: Is it possible to analyze these data by using simultaneously a (Open?) Robust Design where the secondary occasions are modeled by POPAN parameterization?

Before posting I have been looking at (i) the GMI, (ii) messing around with MARK and also (iii) searching in the bibliography, but I haven't been able to find any answer to this.

In fact:
(i) In the GMI I did not find direct reference to that.
(ii) When selecting Robust Design the window that pops up shows several closed capture data types to be chosen among but there is no reference to POPAN parameterization.
(iii) There is a paper (Loman and Madsen 2010, "Sex ratio of breeding Common toads (Bufo bufo) influence of survival and skipped breeding") where they analyze their data by using both models but separately. They do that for three main reasons: 1) because they are interested in estimate the (existing) pattern of immigration/emigration from and to the breeding site (POPAN allows it), 2) because they can use the POPAN results to pool secondary occasions in order to fulfill the closure assumption of the Closed Robust Design they are going to use and, 3) to achieve a sort of "test of the robustness of the results" about the doubly estimated population size (by both POPAN and RD).

I consider this way to perform those analyses very adequate (and very well explained) but still I would like to know if there exist a way to use the POPAN parameterization "inside" the RD to achieve a unique breeding population size estimate and to make the estimate of the other parameters (survival and Total Population size) more accurate depending on a more realistic model fitting the secondary occasions.

Many thanks for any help you can provide to me,


Simone

[simone77]

Perhaps it is worth adding another question to make the previous post more "answerable".
In case it was not possible to perform such analysis in MARK, do you know an alternative way to do it?

Thanks agains

[Bill Kendall]

If you read far enough into the robust design chapter of Evan's book, you'll see a discussion of the open robust design. In MARK you choose the Open Robust Design Multistate option. With this model, an extension of the POPAN model is used within the season. The only difference is this: whereas with POPAN the probability of remaining at the breeding colony for another period within the season (phi) can be modeled as time-dependent or constant, with the ORD-MS model it can also be modeled as a function of the number of time periods since arrival at the colony.

[simone77]

Many thanks for answering,

I did not catch that it represents an extension of the POPAN.
The fact it allows to model the probability of staying depending on how much time they have previously stayed there, makes biological sense also in my case. I will try to use that approach.