Smoothing for spatiotemporal models and its application to modeling muskrat-mink interaction

Area of Science:

• Ecology
• Statistics
• Dynamical Systems

Background:

• Analyzing spatially distributed data with short time series presents estimation challenges.
• Existing methods may not adequately address the complexities of temporal dynamics in such datasets.
• Understanding species interactions, like predator-prey relationships, requires robust modeling techniques.

Purpose of the Study:

• To propose a novel method for estimating temporal dynamics parameters in spatially dependent models.
• To enhance parameter estimation accuracy, especially for short time series data with spatial dependencies.
• To apply the developed methodology to a real-world ecological dataset.

Main Methods:

• Development of a spatial smoothing technique for parameter estimation.
• Asymptotic analysis to evaluate the method's performance, particularly concerning nugget effects.
• Application to a 25-year dataset of annual mink and muskrat populations across 81 Canadian locations.

Main Results:

• Spatial smoothing improves parameter estimation, especially with nugget effects, even with large sample sizes per location.
• The method successfully models temporal dynamics in the mink and muskrat population data.
• The modeled dynamics reveal insights into the food chain interaction between mink and muskrats.

Conclusions:

• The proposed spatial smoothing method offers an effective approach for analyzing spatially dependent dynamical models.
• This methodology is particularly valuable for ecological studies involving spatially distributed populations and short time series.
• The analysis provides a quantitative understanding of the predator-prey relationship between mink and muskrats.