Modeling utilization distributions in space and time.

Area of Science:

• Ecology
• Spatial Ecology
• Wildlife Biology

Background:

• The traditional utilization distribution (UD) models animal occurrence in two spatial dimensions (x, y).
• Existing methods often struggle to incorporate temporal variations in animal movement and space use.
• There is a need for a more comprehensive approach to understanding animal space-time utilization.

Purpose of the Study:

• To extend the definition of the utilization distribution (UD) to four dimensions, incorporating both space and time.
• To develop and evaluate a novel product kernel model estimation method for four-dimensional UDs.
• To assess the ecological informativeness of the new method using simulations and an empirical case study.

Main Methods:

• Redefined the utilization distribution (UD) to encompass four dimensions: three spatial and one temporal.
• Developed a product kernel model using a wrapped Cauchy distribution for circular temporal covariates (e.g., day of year).
• Utilized Monte Carlo simulations to evaluate the performance of the product kernel estimator.

Main Results:

• The product kernel method produced models highly correlated with true probabilities of occurrence (Pearson's r = 0.975).
• The method successfully captured temporal variations in the density of occurrence.
• Empirical application to bighorn sheep (Ovis canadensis) demonstrated the model's ability to depict seasonal migration patterns.

Conclusions:

• The four-dimensional utilization distribution and product kernel method offer a powerful tool for analyzing animal space-time use.
• This approach provides ecologically informative models that capture temporal dynamics in animal space use.
• The method has significant implications for wildlife research, conservation, and management, particularly for migratory species.