I suggest starting with simpler models (a few parameters, not 16) and then adding a parameter at a time to find out exactly which additional parameter is causing the failure.
16 parameters Is a lot, so unless you have an enormous data set it may not be possible to fit such a model. In any case, stepwise building up of the model and using the estimates of each previous model as starting values will often solve the problem.
The order of the starting values will be exactly the order that the parameters are printed, which you can diagnose by running a simpler occupancy model and inspecting the output.
regards
andy