Closed-population models in Mark vs CAPTURE

[matobler]

I am using capture-recapture analysis to analize camera trap data of cats. So far I have done all of my analysis using the program CAPTURE. However CAPTURE has some limitations as far as model sellection goes. I am now looking into using Mark but have some questions:

As far as I understand there are differences in the way CAPTURE and Mark estimate parameters for some of the models. Does Mark use a MLE for M(h). Is the AIC available for models with heterogeneity? Which program performes better with poor data? Has anyone used Mark to analize camera trap data?

I was also wondering if there was more detailed information available on how to build different models with heterogeneity in Mark. Most documentation of Mark seemst to focuse on open-population models.

Thanks,

Mathias

[Paul Lukacs]

Mathias,

If you are referring to the jackknife M(h) in MARK, it is not fit with maximum likelihood and thus AIC is not available for the jackknife. MARK simply calls CAPTURE to fit that model. Other forms of heterogeneity models are available in MARK such as finite mixtures and through the use of individual covariates. Those models produce MLEs.

There is now a chapter in the MARK book on closed models.

Paul Lukacs