Power grids to epidemics: study shows small patterns trigger systemic failures

Why do some systems collapse suddenly after what seems like a minor disturbance? A single transmission line failure can cascade into widespread blackouts. A delayed shipment can ripple through a global supply chain, emptying store shelves far from the original disruption. A rumor spreading in a small online network can spark nationwide panic. In nature, a slight environmental shift can throw an ecosystem into chaos, and a local disease outbreak can quickly escalate into an epidemic.

New research suggests that in many of these cases, the key isn’t the entire system – but its smallest building blocks. Tiny clusters of interacting components, called network motifs, can act as amplifiers, triggering outsized reactions that ripple through the entire system. By identifying these critical patterns, scientists hope to better predict – and possibly prevent – cascading failures across the complex, interconnected systems that shape our world.

The study, published in the Proceedings of the National Academy of Sciences, was led by researchers from Florida Atlantic University, the Carl von Ossietzky University of Oldenburg, and the University of California, Merced.

Complex systems

Scientists often study complex systems such as food webs, social networks or infrastructure systems by mapping them as networks of connected parts. Within these large networks are smaller recurring patterns of interaction.

In ecology, one classic example involves two species competing for the same limited resource. This simple pattern helps explain the “competitive exclusion principle,” which says that species competing for exactly the same resources cannot stably coexist. What makes this pattern powerful is that its consequences hold true no matter how complicated the rest of the ecosystem is.

The researchers wanted to know whether other small patterns have similar system-wide effects. Using mathematical models and computer simulations, they tested thousands of small interaction patterns embedded in larger networks. They examined when the presence of a small cluster of species could determine how the entire system behaves – and when the larger network could override it.

Small clusters

Their analysis showed that small patterns rarely determine whether a system will ultimately remain stable or collapse. However, the team discovered that these small clusters often control something just as important: how strongly a system reacts immediately after it is disturbed.

This property, known as “reactivity,” describes whether a system temporarily surges, swings or destabilizes in response to a shock – even if it later settles down. A system can be stable in the long run but still experience sharp, potentially dangerous spikes after small disruptions. The researchers found that even motifs involving just two or three species can account for a large share of a network’s overall reactivity. In other words, small groups of interacting components can act as amplifiers, intensifying disturbances in ways the rest of the network cannot fully counterbalance.

Ecological food webs - and more

Although the study focused on ecological food webs, the findings apply much more broadly. The same mathematical principles hold for other systems that can be represented as networks, including supply chains, power grids and social networks that spread information or disease. In each case, small clusters of tightly connected parts may be responsible for triggering outsized responses to disruptions.

The results suggest a practical new direction for research. Instead of trying to measure and predict the behavior of an entire complex system at once, scientists may be able to identify specific small patterns that are especially prone to amplifying disturbances.

In ecological networks, for example, pinpointing these high-reactivity clusters could help researchers identify groups of species most likely to drive sudden ecosystem changes. Similar approaches could help engineers locate vulnerable sections of power grids or help public health officials identify risky clusters in disease transmission networks.

System properties

“We hope future research, perhaps inspired by this work, will deepen our understanding of which system properties arise from parts of a network and which emerge from the network as a whole,” said Ashkaan K. Fahimipour, Ph.D., co-author, an assistant professor in the Department of Biological Sciences, and a member of the Center for Complex Systems within FAU’s Charles E. Schmidt College of Science. “If we can figure out when small interaction patterns are responsible for big responses, we can focus attention on the most critical parts of complex systems and better anticipate how they might react to change.”

READ MORE: Food systems, climate change, and air pollution: Unveiling the interactions and solutions

READ MORE: New review reveals how microbial communities accelerate the global spread of antibiotic resistance

Ultimately, the study highlights a powerful idea: in complex systems, small patterns can have big consequences. By learning to spot the clusters that amplify disturbances, scientists may move closer to predicting – and possibly preventing – cascading failures across the interconnected systems that shape our world.

Study co-authors are first author Melanie Habermann, a Ph.D. candidate at the Carl von Ossietzky University of Oldenburg; Justin D. Yeakel, Ph.D., associate professor of life and environmental sciences at the University of California, Merced; and Thilo Gross, Ph.D., a professor and network and data scientist at the Carl von Ossietzky University of Oldenburg.