Some pertinent references are:
Conn P.B., Kendall W.L. and Samuel M.D. (2004) A General Model for the Analysis of Mark-Resight, Mark-Recapture, and Band-Recovery Data under Tag Loss. Biometrics, 60, 900-909.
McDonald T.L., Amstrup S.C. and Manly B.F.J. (2003) Tag loss can bias Jolly-Seber capture-recapture estimates. Wildlife Society Bulletin, 31, 814-822.
Arnason A.N. and Mills K.H. (1981) Bias and loss of precision due to tag loss in Jolly-Seber estimates for mark-recapture experiments. Canadian Journal of Fisheries and Aquatic Sciences, 38, 1077-1095.
Seber, G. A. F. and Felton, R. (1981). Tag loss and the Petersen mark-recapture
experiment. Biometrika 68, 211-219.
Wetherall, J. A. (1982). Analysis of double-tagging experiments. Fishery Bulletin 80,
687-701.
I don't know your exact experimental set up, but if tag loss in each period is independent of previous periods (e.g. 5% tag loss in period 1, 7% tag loss in period 2 etc) then tag loss is completely confounded with survival (see Arnason and Mills). In these cases, I'd suggest modelling as straight forward CJS to get estimates of death+tagloss and then adjusting each survival estimate with the estimated tagloss (see Seber and Felton). If tag loss varies by length of time at large and animals are tagged at different time, then this won't work.
Are you interesting in other parameters such as population growth rates? See Rotella J.J. and Hines J.E. (2005) Effects of tag loss on direct estimates of population growth rate. Ecology, 86, 821-827.
Too bad you don't have identifiable fish. Then some recent work by Laura Cowen would be helpful. She has a general methodology to deal with tag loss with individually numbered tags in CJS and JS models. See her thesis at
http://www.stat.sfu.ca/people/alumni/Th ... n-2005.pdf.
YOu can contact Laura Cowen at University of Victoria (
lcowen@math.uvic.ca)
--------------------------------------------------------
Author : Justice
E.mail :
cj15@humboldt.edu
URL :
http://www.phidot.org/forum/viewtopic.php?p=576#576
--------------------------------------------------------
My name is Casey Justice. I am a graduate student at Humboldt State
University in the Fisheries Biology program. I work with the California
Cooperative Fish Research unit under professor Walt Duffy. An aspect of
my study involves survival estimation based on capture-recapture data. I
am fairly well versed in program MARK, however, I've recently encountered
an issue that I thought you might be able to help with. As you know well,
one of the primary assumptions associated with MARK is that of 100 % tag
retention. Because I had clipped adipose fins on all tagged fish, I have
data indicating when PIT tags were shed and I have an estimate of tag
retention. It is obviously not 100%.
My study includes four encounter occasions and fish were only tagged
during the first encounter occasion. Of the entire set of tagged
fish from occasion one, several individuals were recaptured on occasions
2, 3, and 4 with lost tags. Although I am able to identify if a fish had
been tagged previously, I don't know exactly which tag it was originally
given. That is, of the individuals that were never recaptured, I don't
know which of those were lost tags and which were simply not recaptured
due to death, movement, or failure to capture.
My question is this: Is there a way to incorporate estimates of tag
retention into my analysis using program MARK? One idea I had was to use
my estimates of mean growth rates, mean distance moved, and habitat location
data to set up a conditional formula which would select candidate fish from
the population of tagged fish that were never recaptured. Then I would randomly select
individuals from that candidate set and include them as recaptured individuals in
my data set. Although this method might initially sound dangerous in terms of
introducing an element of bias into the data, it seems more realistic than a dataset which
excludes all shed tag individuals. I also thought about randomly selecting the same number
of lost tag individuals from the population of tagged fish that were never recaptured and
removing them from the analysis entirely. This seems like it might also help to account for
my lost tag problem. What do you think?
Thanks so much for your help. :?
--------------------------------------------------------
--
----------------------------------------
Message generated by forum on analysis of data from marked individuals.
