Question about Including Daily Vegetation in Occupancy Models

[arianna vicari]

Hello everyone!

I’m currently working on a single-species occupancy model using survey data collected from 2019 to 2023. I would like to include a dynamic vegetation variable that changes daily. My initial approach was to include it as a detection covariate, but I keep running into formatting errors when trying to do so.

I’m starting to wonder if this is the right approach. Do you have any suggestions on how best to include a daily-changing covariate in the model?

Thank you for your help!

Best regards,

Arianna Vicari

[Marc Kéry]

Dear Arianna,

that sounds like a sensible idea, if you think that vegetation height might affect detection probability. However, when you say daily values for the covariate, then that requires that you also organize your detection/nondetection data by daily occasions --- is this so ? If not (e.g., if you have weekly occasions), then you will have to summarize the covariate by occasion as well.

Can you be more explicit what kinds of errors you encounter ?

Best regards  ---- Marc

--
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/149e9a6a-f932-4e54-ac97-4a0f4351cc62n%40googlegroups.com.

--

______________________________________________________________
 
Marc Kéry
Tel. ++41 41 462 97 93
marc...@vogelwarte.ch
www.vogelwarte.ch
 
Swiss Ornithological Institute | Seerose 1 | CH-6204 Sempach | Switzerland
______________________________________________________________

*** Hierarchical modeling in ecology ***

https://www.mbr-pwrc.usgs.gov/pubanalysis/roylebook/

[arianna vicari]

Hi Marc,

Thank you for the explanation.
My data (from camera traps) is structured with daily records for 91 cameras across 2610 days. The detection covariates I currently include are:

• Effort (number of days each camera was deployed), and

• Dynamic MSAVI, with one value per camera per day when the camera was active (NA when the camera was not active).

I think I have now fixed the formatting issue I mentioned before, but I would appreciate your opinion on whether structuring the data in this way (daily resolution for both detections and MSAVI) is appropriate.

In addition, I am wondering how inference rasters should be handled in this case. If MSAVI is included as a daily covariate, does this imply that I would need to generate a prediction raster for each day, or is there a more standard approach to representing such dynamic covariates in the inference stage?

I thank you in advance for the help.

Kindly,

Arianna Vicari

[Marc Kéry]

Dear Arianna,

thanks for this information. Purely format-wise, this looks OK. But, to be honest, I have some big questions about your analysis. First of all, I have never seen an occupancy model that was fit to data that span 7 years. Clearly, you can't assume closure over such a long period (unless perhaps your study system is some big tree); I'd bet a lot of money that your population is open over this duration. Hence, you must address this in some way. (Apologies if you intend to adopt a model that accommodates an open population, but this is not apparent from your email.)

Then, daily temporal resolution is almost certainly overkill. At least, if your main questions are about occupancy, I'd perhaps envision weekly or monthly occasions. (If you want to model detection probability as a way of inferring something about the behaviour of the animals, then perhaps this resolution might be OK.)

Your decisions about how to discretize time into primary occasions (e.g., years, springs or mating seasons) and into secondary ones (e.g., weeks or months, as just mentioned) will depend on the dynamics of your study system and importantly also on your questions. So, what questions do you want to answer with your occupancy model ?

Best regards  ---- Marc

To view this discussion visit https://groups.google.com/d/msgid/spocc-spabund-users/073ee297-9ac4-4243-bcb7-4775c73eb369n%40googlegroups.com.