Burnham model - lack of banding/recapture effort

[cnicolai]

I'm running a Burnham model and have one year with no banding/recapture effort. Obviously recoveries still occur.

Trying to figure out what to do with the survivals around this period. p's are allowed to be estimated and are so, at basically zero.

So, I could arbitrarily make the missing survival equal to the prior or following survival. But, I would like to try to make it a bit more robust than that, maybe the average of the prior and following, or even as fancy as a mean of all the others. I've been exploring some of the functions for the DM, but haven't had any luck.

Any suggestions?

[Jochen]

In the "run" window you have the option to fix parameters. You may set the recapture probability p to zero for a year where no capturing occurred at all. This usually works - good luck
Jochen

[cnicolai]

Thanks Jochan, but I think its beyond that. Since there were no releases or encounters, it can't estimate survival. I am letting the p for that occasion to be estimated rather than setting it to zero and it seems to be basically estimating it at zero.

I know I need to constrain the survival to something. In making the general structure, I don't want to go to a trend model right away.

So, here is an example of my options for the DM. Let's imagine a 6 year data setand the year without banding or encounters is year 3.

OPTION 1. Let it estimte with no constraints (obvious problems)
1 1 0 0 0 0
1 0 1 0 0 0
1 0 0 1 0 0
1 0 0 0 1 0
1 0 0 0 0 1
1 0 0 0 0 0

OPTION 2. Constrain it to equal the year prior. This allows the estimation of one less Beta. Seems like a reasonable approach. Others have done it.
1 1 0 0 0
1 0 1 0 0
1 0 1 0 0
1 0 0 1 0
1 0 0 0 1
1 0 0 0 0

OPTION 3. Constrain it equal to the following year. Again, one less beta to estimate. Again, others have done this.
1 1 0 0 0
1 0 1 0 0
1 0 0 1 0
1 0 0 1 0
1 0 0 0 1
1 0 0 0 0

Option 4. Constrain to a value that uses both the prior and following period. But, this causes problems as it is adding both betas and actually produces a higher estimate than either prior or following. Kind of looking for more of an average.
1 1 0 0 0
1 0 1 0 0
1 0 1 1 0
1 0 0 1 0
1 0 0 0 1
1 0 0 0 0

OPTION 5. I think this would be the way to do it, but I don't know how. Basically some type of function would be benficial (fx). Some how to average the betas from the prior and the following. This approach would estimate the original number of betas, of which the one of interest would be a mean of the prior and following.
1 1 0 0 0 0
1 0 1 0 0 0
1 0 1 fx1 0
1 0 0 0 1 0
1 0 0 0 0 1
1 0 0 0 0 0

Does this make sense? Anyone understand the functions and if this approach may be valid?

Thanks in advance,
Chris