uneven encounter occasions

[Th.pol]

Hello,

I'm a french student who is studying spoonbill population. The individuals have been banded in the nest, and I have datas concerning band re-sigthing from 1990 to 2006
I organised my datas in order to get encouter histories file showing observation (or not) for each year of age. The goal is to determine the survival probability in function of the age of the individual.

So I do have encounter histories with various number of encounter occasions, depending on the year of banding. (ie: one individual banded in 1996 have encounter occasion for 10 years of life, whereas one banded in 2000 will only have 6.)

I would like to know what is the best way to deal with this data set.
.As much as I know, MARK requires a specified encounter occasion for each analysis
.I do not want to loose data by using only the firsts occasions of the longest encounter histories

Thanks a lot for your help to a beginner !

[jlaake]

You should review Evan and Gary's electronic book and the help files on the Cormack-Jolly-Seber model. The CJS model is designed with triangular PIMS for exactly the reason you specify in which there is a new release cohort at each occasion (year in this case). The capture histories of the the 1996 start with a 1 to signify their release on the first occasion but the capture histories for the 2000 cohort would start with 00001 to specify that they were not released until 2000 in the 5th cohort. The CJS model structure allows you to use all of the data but since you are marking only the young be careful in how you specify the model. I believe this is discussed in chapter 8 of the electronic book.

--jeff