which site covariates

[klg-2016]

https://github.com/kaitlyngaynor/gorongosa-mesocarnivores/blob/d43955d197f0e8d86a5a75bd03e9bfab1e57dd6e/occupancy-mmpp-master/application/02_fit_model_meso.R#L487

I just added in the site covariates I ended up using in my thesis (I believe I chose the correct columns). I don't have a great understanding of how their site covariates are being used in their model, so I'm not sure if that's what makes the most sense

[kaitlyngaynor]

I think you may have better luck reading the paper(s) than trying to figure out it from the code. As I understand it, the occupancy model is essentially, at its core, the Rota model, so it should be fairly easy to adapt.

[klg-2016]

taking notes here as I go through the Kellner paper:

• the only covariate that they mention in the paper is the binary hunting one. everything else seems to be comparing species activity to species activity. this is not a continuous variable, the only factor/categorical variable I see with the GNP data is habitat type
• "Deer detection intensity parameters also depended on the latent presence or absence of coyotes at a site" I'm not sure if this is coded into the model or not. if so, need to figure out where because I don't think we have a priori thoughts about that with genet/civet
• when they say that there are two different detection intensities for deer based on the presence or absence of coyotes, do they mean that coyotes have or have not been detected at a given camera? or something about how they're there at the same time as the deer? I think it's the former

[klg-2016]

and same as I go through the Parsons paper:

• they had three environmental covariates: two scales of tree cover, and housing density
• they use two aspects of temporal thinking: detection intensity as a function of time of day, and time since last coyote detection). again, this assumes an a priori relationship between the two potentially interacting spp
• they have a lot more detail/more questions to test than the Kellner model, but I haven't found their code yet. going to take some time to check out the appendices and see if their code is available publicly