Biodiversity Without Tournaments: Dominance-Network Ensembles, Community Structure, and Eco-Feedback in Multi-Species Cyclic Games

Generalize Cliff (2024)’s finding that biodiversity collapse under edge ablation depends on the specific tournament network by exploring ensembles of dominance graphs (e.g., circulant and non-tournament digraphs) across 5–9 species. Incorporate community structure and eco-feedback and measure resilience (extinction times, coexistence fraction) under targeted and random edge/node perturbations. This addresses the unrealistic assumption of tournament-like dominance in most multi-species cyclic game studies by quantifying how topological features—cycle covers, chordality, feedback locality—govern coexistence. It connects structural sensitivities from interaction networks to dominance networks and integrates eco-evolutionary coupling. Using Sobol sensitivity analysis, it ranks graph features by impact on resilience. The project provides robust design principles for maintaining biodiversity under uncertainty and a structural theory of coexistence resilient to model misspecification and realistic in community structure and environmental feedback.

References:

1. Tournament versus Circulant: On Simulating 7-Species Evolutionary Spatial Cyclic Games with Ablated Predator-Prey Networks as Models of Biodiversity. Dave Cliff (2024). Proceedings of the 36th European Modeling & Simulation Symposium.