Optimal transportation theory for species interaction networks.

Area of Science:

• Ecology
• Theoretical Ecology
• Network Theory

Background:

• Biotic interactions like pollination and predation depend on population densities, trait matching, phenology, and random events.
• Existing models often struggle to integrate these diverse factors for a comprehensive understanding of interaction networks.

Purpose of the Study:

• To propose optimal transportation theory as a unified framework for modeling species interaction networks with varying interaction intensities.
• To develop a method that simultaneously considers population densities, functional traits, phenology, and stochasticity in ecological interactions.

Main Methods:

• Formulating the coupling of species distributions as a constrained optimization problem, maximizing system utility and entropy (randomness).
• Applying the maximum entropy (MaxEnt) principle to derive a model from this optimization problem.
• Developing a framework for estimating pairwise species utilities from empirical data.

Main Results:

• The proposed model can simulate changes in species relative densities.
• It effectively disentangles the impacts of trait matching and neutral (stochastic) forces on interactions.
• The framework successfully predicts couplings in real-world pollination and host-parasite networks.

Conclusions:

• Optimal transportation theory provides a powerful, unified approach to studying species interaction networks.
• This method offers new insights into the drivers of ecological interactions, separating deterministic (trait-based) and stochastic influences.
• The framework is experimentally validated and applicable to diverse ecological networks.