Individual differences in habitat selection mediate landscape level predictions of a functional response

Area of Science:

• Ecology
• Wildlife Management
• Spatial Ecology

Background:

• Predicting animal space use is crucial for conservation and management.
• Animal habitat selection is influenced by both habitat availability and individual variation.
• Existing models like Generalized Functional Response (GFR) and resource selection functions (RSFs) with random effects attempt to capture these dynamics.

Purpose of the Study:

• To compare the predictive performance of GFR models and RSFs with random effects in predicting elk (Cervus canadensis) space use.
• To investigate how individual variation in habitat selection affects model performance.
• To determine which modeling approach is superior for future space use predictions.

Main Methods:

• Utilized elk selection data for mixed forest and avoidance of roads in relation to habitat availability.
• Compared predictive performance of GFR models (fixed effects) and RSFs with random effects.
• Assessed model performance under varying degrees of individual variation in habitat selection.

Main Results:

• Individual variation in road avoidance negatively impacted the performance of both GFR and random effects models.
• Neither model's performance was significantly affected by changes in individual road and forest availability.
• Both models performed similarly in predicting mixed forest selection and road avoidance, with GFR showing slightly better estimates but higher variance.

Conclusions:

• Neither GFR nor random effects models demonstrated superior predictive performance for future space use predictions.
• The choice between models involves a bias-variance trade-off, with GFR offering potentially better estimates but increased variance.
• Accounting for individual variation in habitat selection is critical for accurate ecological modeling.