Time since marking

[Ryan Fisher]

Hi everyone,
A bit of a interpretation question here. I'm analyzing a dataset of Northern Flicker (a woodpecker) survival with 665 individuals over 6 years (using the live recapture model). We're looking at basic sex and age differences in apparent survival. We've completed those basic models and it suggests no differences in apparent survival between different sexes or ages. No problem here yet.
However, we noticed that there were a lot of individuals that were marked and then never recaptured again. So we decided to include a time since marking model, with a different survival estimate for birds one year after marking compared to 2+ years after marking. This model then becomes the top model (survival for the interval between first marking and second capture is 36%, versus 56% for all other capture events after, 95% CI don't overlap). Although this seems like a transients problem, we only mark birds that are breeding on the study site (birds are trapped on the nest in late incubation stages) and therefore this removes any problem of birds just migrating through. There doesn't seem to be a problem with birds abanndoning nest sites after we mark them either (or problems with nest success - as nest success for all banded birds is around 80-85%). Also, we have very little mortality until the end of the breeding season when we stop monitoring them (although this doesn't eliminate problems during migration). Our recapture rates are also high (~90%) suggesting that we find most birds on the study area given that they are breeding there. Obviously permanent eimgration could also be contributing to this low estimate.
Can anyone offer some advice or point me toward any literature that may have found similar results. The literature dealing with transients doesn't seem to apply to this particular situation...
Thanks,
Ryan

[Ken]

Not exactly sure what you’re asking. Are you asking about the biological implications or a statistical fix to the problem? I can help with the former but the latter is beyond my capabilities. This is an extremely common situation when banding nestlings. Many species disperse post fledging never to return to their natal area. Some species exhibit natal philopatry, returning to breed near to where they were reared, but more commonly this is not the case. While surviorship is generally lowest for birds in this first, post fledging year, you are also dealing with the confounding fact that many species do not return to breed where they were reared. In the Birds of North America account for the Northern Flicker, William S. Moore states that only 4 of 383 Flickers that he banded as hatchlings were ever resighted. This is rather old data, however, and I’m not sure what the design of his study was. Anyway, hope this wasn’t a complete waste of your time.

[ganghis]

Hi Ryan,

This does look like an issue with permanent emigration.

1) Have you tried conditioning on animals seen more than once (i.e. deleting the first '1' in all encounter histories and throwing out all those with '00000')? This approach may help to eliminate some problems with permanent emigration, although it will greatly reduce your sample size and you may get a slightly biased sample (those that have already survived may have greater fitness than those that died)

2) Take a look at
Cilimburg AB, Lindberg MS, Tewksbury JJ, et al.
Effects of dispersal on survival probability of adult yellow warblers (Dendroica petechia)

-Paul Conn
AUK 119 (3): 778-789 JUL 2002

They use a nested grid approach where they estimate apparent survival over sequentially expanded search areas to get a handle on permanent emigration.

[Ryan Fisher]

Yes I just wondering about potential avenues to examine this problem more thoroughly in a statistical sense and will look at censoring the data for birds that have been seen more than once.

In response to Ken, this data set isn't including birds marked as fledglings since we only get about 3% per year coming back to the study site (similar to Moore in BNA), so it doesn't include that problem.

Thanks for the suggestions, I'll check out that Cilimburg paper

[Ken]

DOH!!!!!! Sorry Ryan - misunderstood your design. : oops: