Uniting statistical and individual-based approaches for animal movement modelling

Area of Science:

• Ecological modeling
• Animal behavior
• Spatial ecology

Background:

• Animal spatial behavior is shaped by dynamic internal states and environments.
• Integrating these dynamics into habitat selection models is challenging due to data limitations and model constraints.

Purpose of the Study:

• To develop a robust method combining statistical and individual-based modeling (IBM) for habitat selection.
• To overcome limitations of current models in producing dynamic outputs and accessing internal states.

Main Methods:

• Developed a hybrid statistical and individual-based modeling (IBM) approach.
• Utilized GPS data from caribou in Québec to model movement based on dynamic variables.
• Employed Step Selection Functions for empirical parameterization and validated the IBM with k-fold cross-validation.

Main Results:

• The developed method successfully captured the complexity of caribou habitat selection.
• Demonstrated a functional response in habitat selection, indicating model efficacy.
• The model provided accurate projections for future habitat states under varying environmental scenarios.

Conclusions:

• The novel modeling approach effectively integrates dynamic animal states and environmental factors for habitat selection studies.
• This method offers robust projections for testing long-term impacts of environmental changes and informing management plans.