Thermodynamic phase transitions reveal the resilience structure of urban traffic congestion

Abstract

Understanding how cities transition from free-flowing to congested traffic remains a central open problem in urban science. Here we show that city-scale congestion undergoes a reproducible nonlinear transition analogous to an order-disorder phase transition in statistical mechanics, in which aggregate mobility acts as a control parameter and jam extent as a collective order parameter. Crucially, this analogy is not merely formal: we derive and empirically identify an effective thermodynamic temperature with concrete physical meaning, quantifying infrastructural heterogeneity and how broadly a city explores congestion configurations as demand increases. Low-temperature cities are congestion-fragile: small mobility increases trigger sharp, system-wide jam transitions. This framework further reveals that the macroscopic fundamental diagram is an incomplete description of the traffic state: it emerges as a projection of a richer free-energy landscape governed by entropy-capacity trade-offs. Validated across 46 cities in Latin America and the Caribbean and independently confirmed with loop-detector data from 8 cities on three continents, these results establish a physics-based foundation for comparing urban traffic resilience and anticipating congestion regime shifts under changing mobility demand.