Spatial replication and occupancy covariates

[bmitchel]

Hello Phidot,

I'm curious about how Presence handles occupancy covariates. Do these covariates always need to be site level, or can they vary by occasion?

I'm working with a data set that has 5 point counts along routes at different elevations, and using the point counts as spatial replicates. This species tends to occur at high elevations, so I'd like to be able to use an elevation covariate for occupancy that varies with each survey occasion. I've entered the covariate into Presence as a sampling covariate, and the model seems to run but I'm not sure that it is doing what I want (adjusting psi based on covariate values for each spatially replicated survey). From the descriptions I have found, it looks like there is an assumption that psi is constant across occasions within a season.

If Presence is not using the sampling covariate to adjust psi for each occasion, how does it handle the use of a sampling covariate for psi? I've tried to find posts on Phidot about this, and have RTFM...

Thanks,

Brian Mitchell

[jhines]

Since occupancy is assumed constant within a season, it cannot change from one survey to the next. If you model psi as a function of a survey covariate, PRESENCE will just use the first column (survey) of the covariate to compute initial occupancy. If the other 4 surveys are near the same elevation, then using the first one is probably OK. If there are huge differences in elevation among the 5 surveys, then you might want to break the site into two sites.

Jim