Mechanical-statistical modeling in ecology: from outbreak detections to pest dynamics

Area of Science:

• Ecology
• Population Dynamics
• Mathematical Biology

Background:

• Assessing population control and adaptability requires understanding large-scale, long-term dynamics.
• High-resolution surveys are often infeasible for pest populations with irregular outbreaks.
• Discrepancies between phenomenon and sampling scales complicate population inference.

Purpose of the Study:

• To develop a mechanical-statistical model for analyzing spatio-temporal outbreak occurrence data.
• To infer population dynamics, estimate densities, and identify hidden factors from limited observations.
• To apply the model to understand the dynamics of the European pine sawfly (Neodiprion sertifer).

Main Methods:

• Proposed a spatio-temporal mechanical-statistical model incorporating ecological mechanisms and observation processes.
• Employed Bayesian inference for estimating model parameters, pest densities, and unobserved variables.
• Integrated a dynamical Beverton-Holt model with a hidden regime variable to handle population density variations.

Main Results:

• Supported the view of European pine sawfly dynamics as metapopulations with alternative equilibria.
• Confirmed the significant impact of extreme minimum winter temperatures on sawfly outbreaks.
• Quantified a novel positive correlation between the lake area to total area ratio and outbreak occurrence.

Conclusions:

• The developed model effectively analyzes outbreak data, bridging scale discrepancies.
• European pine sawfly outbreaks are strongly linked to specific environmental factors, including temperature and landscape features.
• Metapopulation dynamics and hidden variables are key to understanding and managing this pest.