species-specific occ.covs and pred.covs in msPGOcc?

[Jeanne Käser]

Hi all,

I have a question regarding msPGOcc models. I am trying to fit a model on spider data: 78 species distributed in 50 plots, each visited 5 times.

I have some data on species-specific functional traits which would be relevant to add to the model, e.g. the hunting strategies of each species, which impacts their trapping rate.

I haven’t found documentation on species specific traits in msPGOcc, and the few tries I made have failed. I am now looking to confirm that this feature indeed does not exist in spOccupancy? And, do you have any alternative strategies which you could recommend?

Thanks a lot in advance :)

Jeanne

[Jeffrey Doser]

Hi Jeanne, 

Thanks for the note. You are correct that it is not possible to directly include species-specific traits/covariates into msPGOcc or any multi-species functions in spOccupancy. There are two alternative strategies I would suggest: 

1. You could use the postHocLM() function to quantify relationships with species-traits after fitting the model. Basically, you would extract species-specific estimates from msPGOcc, then use those in a subsequent linear model that would relate those estimates to species-specific traits, while fully taking into account uncertainty. This is sort of a work around for not being able to fit species-specific traits directly in the spOccupancy model. 
2. If you aren't planning to explore any of the spatial functionality in spOccupancy and you don't mind leaving the Bayesian framework, you can fit a multi-species occupancy model with species traits in unmarked using the relatively new occuComm function in unmarked. 

Kind regards,

Jeff