Censoring after tag expiry

questions concerning analysis/theory using program MARK

Censoring after tag expiry

Postby aswea » Mon Mar 14, 2022 1:11 am

We are running a multi-year study to look at seasonal survival and harvest probabilities for a population of lake trout. The fish are implanted with acoustic tags and with high-reward external tags to encourage anglers to report captures. The reward is high enough that we are assuming that all harvested fish are reported. We have acoustic receivers in the lake and we class a fish as “alive” each season if it was detected moving between receivers. New fish are tagged each year.

I have been thinking I would use a Burnham model to analyze this dataset with fidelity (F) fixed to 1, and interpreting the probability of being dead and reported (r) as the probability of harvest (i.e. assuming all harvested fish are reported). BUT I’ve just been confronted with the problem that some fish survive past the expiry dates for their acoustic tags. Is it possible to censor fish with expired tags from the study after tag expiry? So far, I’ve found that it is possible to censor by assigned -1 to the frequency field in the CH sequence, but this process will remove all CH digits after the last recapture. For this study, there may be legitimate non-detection between the last recapture and tag expiry.

There is a post in the forum with a similar issue where it was suggested that the digits after expiry be replaced by “.”, but this doesn’t work for the Burnham model. http://www.phidot.org/forum/viewtopic.php?f=21&t=3583&p=11661&hilit=censor&sid=d277f9fa6d234f43aaf4470674da3fef#p11661

Does anyone know how I might handle tag expiry in the middle of a study?

As an alternative, I have been testing using a multistate (states: “tag expired”, “harvested”, and “regular”), but this requires the assumption that the fish survive the season and then transition state (e.g. they are all harvested in a bunch at the end). Clearly this assumption is violated for our study…and it certainly makes the model difficult to explain! The probability of harvest in the multistate is the product of the survival probability for that season and the probability of transitioning to the harvested state (S * Psi).

Thank you!

Aswea
aswea
 
Posts: 27
Joined: Sat Oct 17, 2009 3:32 pm
Location: Gander NL

Return to analysis help

Who is online

Users browsing this forum: Google [Bot] and 14 guests