Given the structural similarities between classic Kendall-extended robust design (with gamma' and gamma" estimated as relating to temporary movement in and out of the sampling region), and multi-season occupancy models, I was wondering if there were equivalent issues for the latter in terms of confounding of epsilon and gamma for fully time-dependent models, , especially over the last interval, analogous to the confounding that arises for the RD with gamma' and gamma". Bill Kendall did a lot of work on that particular problem for the RD -- is there something equivalent for RD occupancy models?
I suspect that the gamma' and gamma" issue doesn't have an equivalent for RD occupancy (and a few cursory simulations suggest this is the case), in part because there are equal numbers of parameters for both epsilon and gamma in the RD occupancy model.