Beyond the jam: field theory modeling of congestion’s ripple effects on traffic dynamics

Urban traffic congestion generates complex ripple effects (i.e. the progressive spread of congestion from a localized point to a broader network), yet existing models lack the capacity to quantitatively capture these spatial dynamics due to the oversimplified representations of traffic behavior, failing to model the continuity and heterogeneity of congestion propagation. This study proposes a network-constrained field-based framework that integrates field theory with a path flow model for fine-grained, heterogeneity-aware quantification of congestion’s ripple effects. First, we develop a traffic field model that captures the continuous influence of congestion across all related road segments. Second, path flow is incorporated to quantify driving subjectivity and propagation heterogeneity. Third, a flow-assisted discretization method is introduced to enable the joint modeling of human mobility patterns and spatiotemporal traffic dynamics, making the model computationally tractable and adaptable to real-world data. Our model achieves 0.88–0.89 overlap accuracy compared to the benchmark model, demonstrating its strong performance in identifying affected road segments while reflecting an 11–12% difference attributed to heterogeneity consideration. Additionally, our model captures a more comprehensive range of congestion effects, identifying 3.64–4.74 more high-order affected segments that the benchmark model may overlook. Results reveal that the stability of congestion characteristics at the macro-level overshadows the variations caused by daily life-driven crowd movement. Turn-level analysis further uncovers directional asymmetries and distinct patterns for vehicles entering and exiting congested segments. This framework provides a practicable tool for quantifying congestion impacts, supporting precise analysis of propagation dynamics and data-driven traffic management strategies.