Missing covariates when marked individual "not seen"

[blhill]

Hello! I am using CJS models to estimate survival of color-marked Dunlin. I would like to use an individual's nest initiation date as a covariate, which would be different for each bird for each of the 8 years. In terms of the input file, what should I use as a code for years before the bird was banded? Also, many birds were captured at a nest once and then never seen again. How do you code for the following years when we can't be sure whether the bird died or emigrated? Finally, how do you code for a year when a bird was not seen, but we know it was alive because it was seen the following year? I can see the value of using median initiation date for the birds we did not see but know were alive, however, I am hesitant to use median initiation date for birds that might not be born yet, and birds that have most likely died. Is there a way around this? Or will MARK "know" (based on capture history) that since the bird was not banded yet (or not seen) it wouldn't contribute to the estimate?

Thank so much for any help!

[cooch]

Hello! I am using CJS models to estimate survival of color-marked Dunlin. I would like to use an individual's nest initiation date as a covariate, which would be different for each bird for each of the 8 years. In terms of the input file, what should I use as a code for years before the bird was banded?

Nothing. The CJS model conditions on initial capture, and does not model the 'leading zeros' in any way.

Also, many birds were captured at a nest once and then never seen again. How do you code for the following years when we can't be sure whether the bird died or emigrated?

In a CJS model, permanent emigration and mortality are confounded. But, given the level of your question, I'll suggest you work through the MARK book, chapter 1 -> 7 (at least).

Finally, how do you code for a year when a bird was not seen, but we know it was alive because it was seen the following year? I can see the value of using median initiation date for the birds we did not see but know were alive, however, I am hesitant to use median initiation date for birds that might not be born yet, and birds that have most likely died. Is there a way around this? Or will MARK "know" (based on capture history) that since the bird was not banded yet (or not seen) it wouldn't contribute to the estimate?

Thank so much for any help!

Seriously, you need to go through the first 7 chapters of the MARK book. You question indicates your have little -> no background. Meaning, you need to do a *lot* of work before you start working with your own data.

The MARK book can be accessed here:

http://www.phidot.org/software/mark/docs/book/