RSF population level availability

[Marina Garrido]

Hi Chris,

I have a question about accurately estimating RSF at the population level using CTMM, particularly regarding the definition of availability.

I am working with multiple migratory individuals across several years. I am focusing on the breeding season, and my goal is to assess how habitat selection differs at the population level. Then I would also love to check the differences between age classes and migratory strategies, as some migrate large distances and others short distances.

I see two possible approaches:

1. Fit individual RSFs and derive population-level estimates using mean().
2. Define a population-level availability (e.g. merged seasonal home ranges or joint UD) and fit a pooled RSF including individual ID as a random effect. I saw it in some papers that do it for habitat selection Type II.

My main questions are:

• Is mean() the appropriate way to estimate population-level selection in ctmm, given that availability differs among individuals? And if so, would you use the mean across individuals and across years to get the population-level selection?
• Alternatively, is it valid to define availability at the population level (e.g. merged home ranges) for a pooled RSF?
• Should availability be defined separately by year, or can it be combined across years?

My main goal is to understand selection within the area during this season at the population level, not just within individual home ranges. This way, in the future, I want to compare it with the wintering areas.

Any guidance would be greatly appreciated. Thank you in advance for your time!

Best,

Marina

[Christen Fleming]

Hi Marina,

mean() tells you about population variation in selection, but for the population-level selection itself, the mean of selection is not the selection of the mean, because of the logarithmic relation between intensity and selection. I have a graduate student working up this with camera-trap integration this summer.

Best,

Chris