Accounting for animal movement in estimation of resource selection functions: sampling and data analysis

Area of Science:

• Ecology
• Movement Ecology
• Wildlife Biology

Background:

• Population-level resource selection patterns arise from individual behaviors.
• Statistical models often overlook individual-environment interactions in habitat use.
• Incorporating individual movement into resource selection models is challenging.

Purpose of the Study:

• To propose and evaluate a novel step selection function (SSF) model.
• To assess the impact of different sampling schemes on movement model bias.
• To investigate methods for reducing bias in resource selection estimates.

Main Methods:

• Developed an SSF model combining a resource-independent movement kernel and a resource selection function (RSF).
• Analyzed simulated movement data using three distinct control point sampling schemes.
• Applied logistic regression, accommodating sampling schemes and incorporating a distance function.

Main Results:

• Ignoring sampling schemes and the movement kernel led to biased selection estimates.
• Bias magnitude depended on control point generation, selection strength, and resource map scale.
• The addition of a distance function significantly reduced bias across sampling methods.

Conclusions:

• Resource selection analyses must account for sampling schemes and movement kernels.
• Including distance as a covariate in SSF analyses is crucial for accurate estimates.
• This integrated approach enhances understanding of animal movement ecology and habitat selection.