Category: Ecology

Summary: Finding the juvenile-adult coupling where modular food webs lose their stability advantage.

Ecological communities with modular structure (groups of tightly interacting species) are generally more stable than random networks. But many species have stage structure (juvenile vs adult stages with different interactions). This experiment asks: at what juvenile-adult coupling strength does modularity stop providing a stability advantage over a randomly shuffled network?

The experiment constructs paired ecological Jacobians — one with modular block structure, one with the same degree distribution but shuffled connections — and measures the stability difference across coupling strengths. It maps the threshold where modularity no longer helps.

This bridges the modularity-stability and stage-structure literatures in a way not previously mapped at finite sizes.

Method: GPU dense Jacobian eigensolve with iterative deepening. Paired modular vs shuffled Jacobians, bisection on juvenile-adult coupling.

What is measured: Stability difference (leading eigenvalue), modularity advantage threshold, transition sharpness.