Conditional effects, taking into account both occupancy and detection parameters

[Justin Biggerstaff]

Hello,

I am currently generating different conditional effects plots and have been able to create some nice looking plots of my occupancy models, using data generated from the predict() function. To my knowledge, the predict function will either let you make an occupancy prediction, or a detection prediction. What I am wanting to know is if there is a way to account for both of these in a conditional effects plot, essentially providing an estimate of true occupancy (psi/p)? My idea is to create a plot similar to the one attached below, but that keeps detection covariates held constant at their mean (which in my context, are temperature and noise).

Does what I am thinking of make sense, and is it possible? 

[Marc Kéry]

Dear Justin,

the two parts of the model don't interact in the sense you are thinking of. That is, occupancy exists, and therefore, can be predicted, independent of the amount of measurement error that we have in assessing occupancy. Hence, what you want is exactly what spOccupancy (and other, related programs such as unmarked) do: predict occupancy or detection separately.

Best regards --- Marc

--
You received this message because you are subscribed to the Google Groups "spOccupancy and spAbundance users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spocc-spabund-u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/spocc-spabund-users/9931ffa5-f413-415d-9727-3b519a1e5a73n%40googlegroups.com.

--

______________________________________________________________
 
Marc Kéry
Tel. ++41 41 462 97 93
marc...@vogelwarte.ch
www.vogelwarte.ch
 
Swiss Ornithological Institute | Seerose 1 | CH-6204 Sempach | Switzerland
______________________________________________________________

*** Hierarchical modeling in ecology ***

https://www.mbr-pwrc.usgs.gov/pubanalysis/roylebook/