Exploring the dynamics of Lotka-Volterra systems: Efficiency, extinction order, and predictive machine learning

Area of Science:

• Ecological dynamics and food web theory.
• Mathematical ecology and systems biology.

Background:

• Understanding species interactions within food webs is crucial for ecological research.
• The cascade model is a widely used synthetic food web for studying ecological dynamics.

Purpose of the Study:

• To investigate the behavior of Lotka-Volterra ecological systems using the cascade model's connectance properties.
• To determine how trophic efficiency influences species extinction.
• To develop predictive models for species extinction based on ecological parameters.

Main Methods:

• Utilized clustering analysis on summed values of birth, death, self-regulation, and interaction strengths.
• Developed random forest and neural network models to predict extinctions.
• Analyzed Lotka-Volterra ecological systems within a synthetic food web framework.

Main Results:

• Established inequalities based on summed ecological parameters can predict food web stability and endurance.
• Machine learning models (random forest, neural network) accurately predict species extinctions without dynamic simulations.
• Identified the death rate as the most significant variable influencing the order of species extinction.

Conclusions:

• Simplified ecological parameters offer insights into food web stability and extinction dynamics.
• Machine learning models provide efficient tools for predicting ecological outcomes.
• Death rate is a critical determinant of species extinction order in food webs.