Closed captures model Mth in Mark

[gpauley]

I'm puzzling over how to create model Mth (N, p(t), c(t)) in Mark. The obvious solution is to allow c and t to vary fully with time, but the obvious problem is that pt (last capture probability) is not conditioned on anything, so pt=1 and N=Mt+1. Help!

[cooch]

I'm puzzling over how to create model Mth (N, p(t), c(t)) in Mark. The obvious solution is to allow c and t to vary fully with time, but the obvious problem is that pt (last capture probability) is not conditioned on anything, so pt=1 and N=Mt+1. Help!

Easiest approach is probably with a design matrix for the full closed with heterogeneity - where column 1 (first column in design matrix) is the pi estimate, column 2 would be the intercept for the p and c parameters, columns 3 -> n are time effects, column n+1 is the additive mixture effect, column n+2 is the behavioral effect, and column n+3 is the population estimate. This would give you model M(tbh). As per usual, you get sub-models by using selected elements of thi more general design matrix - in this case, model M(th) is constructed by using only columns 1, 2, 3 -> n, n+1, and n+3.

See pp. 11 -> end in Chapter 14 (note: Paul Lukacs has introduced a new model naming system for closed models, so it might not be immediately obvious at first reading that this section of chapter 14 has to do with closed models with heterogeneity.