trap cov and/or state space cov

[Gilles Maurer]

Hi again, still wondering about covariates but now is more about the traps covariates.

I do not understand clearly at what stage trapcovs are taken into account in the model and how to 'call' them ? I guess it is linked with the detection process. 
But here are many points that are still unclear for me,  knowing that in my case traps are located at certain  statespace pixels (or cell's centroid) or let say confounding with statespace pixel because traps are indeed selected 'visited cells' within the statespace (same location / same resolution). As a result, covariates values such as habitat, or spatial covariates have the same value at one trap location and corresponding state space pixel. 

As a result, cov traps are labelled the same than state space cov, eg. I do have habitat as a trap cov but also in the state space. 

Question 1 : How to include or not include a trap covariate in the model ? Is that when  adding the cov in the detection process ? eg : 'habitat' is a trapcov stored in the scrframe ( for all session/occasion)that is considered when calling  a model such as D ~ 1,  p0 ~ habitat, sig ~ 1 

Question 3 : When preparing the data, I duplicated all statespace covs  in trapcovs (except the sspace is much bigger that the number of traps). Is that an issue ? I used the same names, eg habitat, elevation, distance_village, etc. both in trapcov and the statespace.? Is that confusing ? Could that increase computation time ?  PS At the moment I have 10 covs both in trap covs and state space covs using the same names. Is it a potential source of mess ? 

Thanks for your help

Gilles 

[Chris Sutherland]

Hi Gilles, 

We cover this in an oSCR workshop lecture - have you seen this:  https://www.youtube.com/watch?v=GzYYnl8n6rI&t=429s. From 49minutes onwards, you will find a section on extending the data objects to include covariate data, and how to then fit models using those variable names.

If this doesnt clarify things, feel free to reach out again :)

Chris