Robust analysis of diel activity patterns

Area of Science:

• Ecology
• Animal Behavior
• Statistical Modeling

Background:

• Diel activity patterns are crucial for understanding organismal responses to environmental changes.
• Advances in time-stamped data collection necessitate improved analytical methods.
• Existing kernel density estimators for activity patterns have limitations, including data aggregation and inability to incorporate covariates.

Purpose of the Study:

• To provide a practical guide for estimating animal diel activity using hierarchical models.
• To introduce trigonometric generalized linear mixed-effect models and cyclic cubic spline generalized additive models as superior alternatives.
• To demonstrate the advantages of these models over traditional kernel density estimators.

Main Methods:

• Utilizing hierarchical models, specifically trigonometric generalized linear mixed-effect models and cyclic cubic spline generalized additive models.
• Developing a comprehensive tutorial for data formatting, model fitting, and prediction interpretation.
• Comparing the efficacy of these advanced models against established kernel density estimators.

Main Results:

• Hierarchical models offer a more robust framework for analyzing diel activity patterns.
• These models overcome limitations of kernel density estimators by allowing for covariate incorporation and better data utilization.
• The provided tutorial facilitates the application of these advanced statistical techniques.

Conclusions:

• Hierarchical models are poised to become essential tools for research on animal activity timing.
• The methods presented offer significant advancements for analyzing behavioral responses to global change.
• Further extensions and applications of these hierarchical modeling approaches are anticipated.