I'm trying to analyse mark-recapture data in butterflies, mainly to estimate abundance.
I'm struggling to include the "age" information in the model, let's see if someone can push me in the right direction to set it up.
The adults of my species fly for less than a couple of weeks, we did 4 sampling occasions with two days between each (3 in some occasions with bad weather). Every time that we captured (and recaptured) a butterfly we recorded the sex and "how old" was the butterfly.
I wrote "how old" between quotation marks, because is more a subjective measure of how "wasted" the butterflies are when we capture them. We are sure this affects their life span and it's affected by the time, but also by the weather and individual story.
In the preliminary analysis I added this information (just for the first capture of each individual, to make it easier) as a covariate ("span") with 4 levels, from 0 to 3 (Fresh, Good, Old, Worn).
I have around 100-200 individuals and around 30-50 recaptures.
The data looks like this:
- Code: Select all
/ *ch Fem Mal span */
1000 0 1 0
1010 0 1 0
1000 0 1 1
1000 1 0 0
1000 0 1 2
1001 0 0 1
0100 1 1 1
0101 0 1 3
I have been mainly doing the preliminary analysis in R.mark and some in Mark too.
I used "sex" as a grouping variable, and my first attempt was to use "span" as a covariate to set the initial ages and run a POPAN model, but thanks to Jeff Laake I understand that is wrong, but as he suggested this is more a general question about POPAN than a question about an specific model in r.mark.
In chapter 12, (JS models) Schwarz and Arnason clearly say that animals aren't allowed to change from one group to other, and that's why the cohort approach suggested by Jeff made sense. The problem is that I don't think is right to use "span" to make cohorts; because when we classify a individual as "fresh", I'm sure it has just a few hours, and when I classify an individual as "worn", it means it has been flying for long, the wings are broken, and it will probably die soon. But this doesn't mean that a butterfly caught in the second day and classified as "good" (1) is from the same cohort that one caught the first day and classified as "fresh" (0). With the recaptures we have checked that some butterflies take a couple of days to pass from fresh to good, while others may take till four days.
So any suggestions for this?
My last idea is to run the models with "span" in CJS (so I can even add 4 covariates with the "span-age" for every occasion), just to check if that works, and then estimate abundance in POPAN separately without "span". But may be there is a more integrating approach.
Any ideas?
Thanks for taking your time to read this.