cgres wrote:I am analyzing within-season variation of the components of reproductive success looking at temporal changes and their relationship. I have a problem with pre-fledging survival (the survival probability from hatch to the time of banding in late summer, shortly before fledging). I 've been looking for a model but conventional capture-recapture methods cannot be used to estimate pre-fledging survival because it requires at least two recapture events and after marking young at the nest and we only have one recapture event (one month later during banding).
I have some information available that I can use, here is the idea: estimate the partial brood loss (i.e. broods in which at least one individual survives to banding) and total brood loss (when no individuals survive). Because young remain together with their parents, mortality due to partial brood loss can be estimated based on the recapture of web-tagged young at banding by the difference between the number of young recaptured that belong to a particular family/nest and the original value for this brood. This will allow us to see how many young are missing. Total brood loss can be estimated by combining young:adult ratio in our captures at banding with the average brood size independently estimated during field observation at that time, and assuming that all families have two adults. I can split the data into 2 groups according to hatching date, there will be two groups of pre-fledging survival (before and after the median hatch date).
Any ideas/input? Thanks a lot!
I'd suggest that you (and Gilles) start by reading an older paper (I'd say classic, but I was involved with it, which means it misses classic by several steps):
Williams, T.D., E.G. Cooch, R.L. Jefferies & F. Cooke. (1993) Environmental degradation, food limitation and reproductive output: juvenile survival in lesser snow geese. Journal of Animal Ecology, 62, 766-777.
All goose studies have the same issue -- you have some information from nesting, some information at banding, and generally nothing in between. So, you tell creative stories about what happens over the interval, complicated by the fact that individual goslings are often 'exchanged' between families (yes, even on Bylot), and fates of goslings aren't independent.
The whole issue of partial brood loss, total brood loss and all other such things has been looked at in some detail (lesser snow geese at LPB, brant in Alaska etc). The paper I cite mentions some, but there haven't been a huge number of 'methodological updates' since the hey-dey of the 90's, when everyone was publishing like crazy. I have no doubt there are more clever approaches that you could consider now -- for example, Paul Lukacs and Vicki Dreitz have published on related issues -- paper in The Condor 106:926-931 (and a chapter in the MARK book).
My guess is that there is some amalgamate of this sort of recent work, combined with some ways of handling non-independence, and uncertainty about brood association, which would provide a useful methodological update to what we did in the 90's (which is really just a tweak on a Mayfield-based approach). But, regardless of how 'fancy' things get, the end result (which has been demonstrated in every goose study) is almost inevitably that (i) early hatching goslings grow faster, (ii) and have higher survival rate. Sedinger could never find a brood size effect, whereas every other goose study did. Throw on some interactions with adult age, and propensity of old vs young mom to use habitats deferentially and you have a story (albeit, one that has been told now a few different times).
