Demographic parameters in multi-season occupancy models

[arianna vicari]

Dear Jeff, Marc and all,

I would like to ask for some thoughts on the methodological approach for my current research. I am interested in landscape connectivity, and after some primary assessment of the connectivity network within the study area, I plan to validate it through different ways. 

One of the approaches I am considering is the assessment of the demographic parameters of the populations, through multi-season occupancy modelling (stPGOcc()), with the aim of evaluating colonization and extinction dynamics in the biologically important areas, and testing if the connecting corridors influence these dynamics.

What I am currently unsure about is how reliable this approach is for validating connectivity, and what its main limitations might be. In particular, I am wondering whether multi-season occupancy models remain informative with only two sampling seasons (e.g. 2020 and 2025), or whether a higher number of consecutive seasons (≥3) is generally required to obtain robust estimates of colonization and extinction.

Related to this, I would also appreciate your opinion on the most appropriate spatial scale for this analysis: would it be preferable to model occupancy dynamics across the entire study area, or to restrict the analysis to predefined habitat nodes and evaluate dynamics among those?  

Any feedbacks or suggestions would be very helpful.

Best regards,

Arianna Vicari

[Jeffrey Doser]

Hi Arianna, 

Thanks for the note. If your interest is in estimating dynamics (e.g., colonization and extinction), spOccupancy is probably not the best package to use for that, as the multi-season occupancy models provided by stPGOcc and related functions do not explicitly estimate colonization and/or extinction like dynamic occupancy models do. I would suggest you look into using the unmarked or ubms packages to fit a dynamic occupancy model, which would then let you better address the questions you're interested in. 

As far as your questions go, it would be possible to estimate effects of colonization/extinction with only two time periods, but the reliability of such estimates would highly depend on the number of changes in occupancy present in your data set. In other words, you would need a sufficient number of sites that were colonized and a sufficient number of sites where the species was originally present but was not present in the second time period. The specific area that you model would depend on exactly the differences between what you're proposing. If the study area includes a large number of areas that you know the species will never exist and hasn't existed in the past, then perhaps you may not want to include them. I'm not quite sure what you mean by habitat node, but if you were interested in colonization/extinction, you would need to make sure you include sites where the species was not originally present, otherwise you wouldn't be able to estimate colonization. 

Kind regards, 

Jeff

--
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/9373e51d-d233-4983-81a7-e1ceb4d771b8n%40googlegroups.com.

--

Jeffrey W. Doser, Ph.D.

Assistant Professor

Department of Forestry and Environmental Resources

North Carolina State University

Statistical Ecology and Forest Science Lab

Pronouns: he/him/his