Unequal time intervals in CJS, Pradel, and POPAN models

[stshroye]

I am interested in detecting and accounting for closure violation in capture-recapture samples of Largemouth Bass collected by shoreline electrofishing. I am using CJS models to detect losses from the sampled population and Pradel reverse-time models to detect gains (to avoid having to make assumptions about unmarked fish). Sampling consisted of 3-8 electrofishing occasions spanning 8-37 days, with 3-13 days between occasions. Estimates of daily survival and recruitment might be near the boundaries of 1 and 0, but cumulative losses and gains between occasions could indicate substantial closure violation. I would normally account for unequal time intervals in MARK, but in this situation doesn't it make sense to leave the intervals all 1 rather than having MARK estimate the daily rates?

To evaluate the potential effects of closure violation, I am also comparing abundance estimates from closed-population models with model-averaged gross population estimates (N*) from POPAN models. Then, shouldn't I also leave the time intervals all 1 in the POPAN models?

[jlaake]

I'm not entirely sure what you are doing but if your concern is having rates close to 1 for survival then you can always use a decimal value to represent the time intervals as a proportion of a year or month. For example, if daily survival is .9981 and you use the interval of 1/365, the estimates survival will be an annual rate of 0.5 well away from any boundary. What you set for intervals is under your control but you should account for unequal intervals.

Now that said, pent in popan models and other non survival parameters are not adjusted for unequal time intervals. However you can use a time model for those parameters to account for different interval lengths or in some cases I believe there may be a way to use a log link with an interval length covariate.

If I understand correctly, I would be concerned that the length of time to conduct a survey occasion was longer than gap between occasions.