openCR age modelling

[stefano palmero]

Hello everybody,

I have been trying to estimate phi using age as a factor (phi~age), also creating a hash array with initial ages to be used instead of default 0.

When it is know that openCR doesn't allow for NAs within the sex, I'm not sure wheter it can create problems within the age.matrix that is used by the model.

Errors didn't occur when I run the model but still not sure the method is correct.

Would be grateful if anybody could help.

Thanks for your time,

Stefano

[Murray Efford]

Not clear what is your problem. Animals with unknown initial age?

[stefano palmero]

Hi Murray,

My problem is indeed animals with unknown initialage. Should I remove them from the age.matrix?

Thanks again,

Stefano

[Murray Efford]

Or assign an arbitrary initialage. Need to think about bias that might be caused by removing animals - may be OK if small number and these are unknown at random and only care about rates (phi, lambda etc.) not absolute abundance.

[stefano palmero]

Will test different sets of data, removing NA or assigning arbirtrary ages and see what happens.

One last question. Do I have to use join(CH) when testing age on a multisession capthist or does openCR.fit do it automatically when age is present? I have seen results change.

Regards,

Stefano

[Murray Efford]

openCR.fit() automatically applies join() if the capthist input has a multi-session structure. It doesn't work any differently for age models, and you should get the same result by applying join() yourself. I would be interested if you can provide a simple working example where it has an effect. Perhaps there is some subtle difference in the specification of intervals, or the dots argument of your call to openCR.fit() is passing something unintended to join().

Murray