"occuMulti" conceptual doubts

[Mateus Zimmer]

Is it possible to obtain an 'average model' (as commonly done for single-species models) including the multi-species models that obtain a delta AIC lower than 2? Those models would have to use the same covariates for the 'average model' make sense?

Also, what should my question be for me to use a covariate in the occupancy interaction between two or more species? And how to interpret the beta values obtained in such case?

[Ken Kellner]

I believe that the AICcmodavg package supports model averaging of occuMulti models with the modavg() function. I have not tried this myself.

If you have a positive beta value associated with an interaction term, that means that as that covariate increases in value, it is increasingly likely that the (two or more) species will occupy the same sites.

If it is negative, then as the corresponding covariate increases in value, it is less likely that the species will occupy the same sites.

[Mateus Zimmer]

Thanks Ken!  

I understand... But I mean using a site covariate, for example vegetation cover, within an interaction term

[Ken Kellner]

My previous explanation applies in that situation. If you include vegetation in the formula for the interaction term, and the result is a positive beta for vegetation, then increasing vegetation cover means the two species are increasingly likely to both occur at a site. Maybe I'm misunderstanding.

[Mateus Zimmer]

No, that's it. I was just wondering if it was as simple as that... Thank you!