Removing Individuals from Mark Resight Data

[arlitt]

(Please note: I am posting this question for a graduate student at my institution. Thank you, Andrea)

I have several years of mark-resight data for white-tailed deer collected using trail cameras. I am using the poisson log-normal option under the mark-resight data type in program MARK. One of the parameters of interest is apparent survival. During the course of the study, we have removed tagged animals to adjust population densities. Is there a way to separate these removed animals from those that have died or permanently emigrated naturally? I know with other data types the group code can be set
negative to indicate that the animal was removed. Can something similar to this be done with mark resight data? When I change the sign of the group code, the number of known marks tabulated by the program decreases by that number as though the entire capture history for that individual had been removed.

-Robin
--
Robin Donohue
Graduate Research Assistant, M.S. Candidate
Caesar Kleberg Wildlife Research Institute
Texas A&M University-Kingsville
MSC 218, 700 University Blvd.
Kingsville, TX 78363
(203) 644-4607

[bmcclintock]

Hi Robin (via Andrea),

Because of the different data type and encounter history entries (e.g., '+0' and '-0'), the Poisson log-normal mark-resight model does not allow use of '-1' in the encounter history frequencies. To censor an individual that is no longer in the population of interest, simply use '..' entries in the encounter history for each primary occasion after removal. Once removed, these individuals will have no influence on parameter estimation.

For example, if there are four primary occasions and an individual that was resighted 2 times during the first primary and 3 times during the second primary was then removed after the second primary occasion, the encounter history entry for this individual would be:

0203.... 1;

To help avoid potential mistakes in the future, I believe Gary has now put in an error message if one attempts to use '-1' frequencies when using this model. Hope this helps!

Cheers,
Brett