Bidirectional yet asymmetric causality between urban systems and traffic dynamics in 30 cities worldwide

Abstract

Understanding how urban systems and traffic dynamics co-evolve is crucial for advancing sustainable and resilient cities. However, their bidirectional causal relationships remain underexplored due to challenges of simultaneously inferring spatial heterogeneity, temporal variation, and feedback mechanisms. To address this gap, we propose a novel spatio-temporal causality framework that bridges correlation and causation by integrating spatio-temporal weighted regression with a newly developed spatio-temporal convergent cross-mapping approach. Characterizing cities through urban structure, form, and function, the framework uncovers bidirectional causal patterns between urban systems and traffic dynamics across 30 cities on six continents. Our findings reveal asymmetric bidirectional causality, with urban systems exerting stronger influences on traffic dynamics than the reverse in most cities. Urban form and function shape mobility more profoundly than structure, even though structure often exhibits higher correlations, as observed in cities such as Singapore, New Delhi, London, Chicago, and Moscow. This does not preclude the reversed causal direction, whereby long-established mobility patterns can also reshape the built environment over time. Finally, we identify three distinct causal archetypes: tightly coupled, pattern-heterogeneous, and workday-attenuated, which map pathways from causal diagnosis to intervention. This typology supports city-to-city learning and lays a foundation for context-sensitive strategies in sustainable urban and transport planning.

Figures (7)

• Figure 1: Spatio-temporal correlation between urban systems and traffic dynamics during rest days across 30 cities. (a) Average $R^2$ values across 30 cities, representing the spatio-temporal correlation estimated by the STWR model. (b) Bubble plot of city-specific $R^2$ values, where the x- and y-axes indicate urban form and function, respectively, and the bubble size reflects urban structure. (c-e) Violin plots showing the distribution of STWR coefficients for features derived from urban structure, form, and function, respectively. (f) Global overview of STWR results across 30 cities, with urban structure (green), form (blue), and function (orange) shown for each city. Cities are visualized as concentric circles: the innermost circle represents the $R^2$ value, the middle ring shows negative STWR coefficients, and the outer ring depicts positive STWR coefficients. Within each feature group, individual features are arranged counterclockwise in the same order as in the violin plots. For example, in urban form, TA (total area) is positioned at the rightmost segment, followed counterclockwise by NP (number of patches) and ending with SHEI (Shannon’s evenness index).
• Figure 1: Spatio-temporal association between urban systems and traffic dynamics during work days across 30 cities, with STWR results for structure (green), form (blue), and function (orange). Each city is visualized as a concentric circle, with the inner circle indicating its $R^2$ value, the middle ring showing negative STWR coefficients, and the outer ring displaying positive STWR coefficients.
• Figure 2: Bidirectional causal patterns between urban systems and traffic dynamics during congestion periods on rest days. (a)$L$-$\rho$ plots from the STCCM model for each city, represented by urban structure (green), form (blue), and function (orange). Increasing $\rho$ values with larger library sizes $L$ indicate the presence of spatial causality. (b-d) Average $\rho$ values at the largest $L$ for all 30 cities, depicting bidirectional causal relationships between traffic dynamics and (b) urban structure, (c) form, and (d) function. Each point represents a city, with the x-axis indicating the $\rho$ value for traffic dynamics $\rightarrow$ urban systems and the y-axis indicating the $\rho$ value for urban systems $\rightarrow$ traffic dynamics. The dotted 45-degree line marks directional differences: cities above the line exhibit stronger causality from urban systems to traffic dynamics, while those below indicate stronger causality in the reverse direction.
• Figure 2: Bidirectional causal patterns between urban systems and traffic dynamics during congestion periods on work days across 30 cities. The $L$-$\rho$ plots illustrate STCCM results for each city, with structure (green), form (blue), and function (orange). Increasing $\rho$ values with larger library sizes ($L$) indicate the presence of spatial causality.
• Figure 3: Categorizing global cities through bidirectional causality between urban systems and traffic dynamics during rest and work days. (a) Hierarchical clustering dendrogram of 30 cities based on the shape and magnitude of their bidirectional causality curves. The x-axis lists cities, and the y-axis shows the dissimilarity between their causal profiles. (b) Heatmap of average $\rho$ values at the largest library size $L$, estimated for each city on rest and work days. Columns follow the city order in the dendrogram. Rows indicate causal directions between urban systems and traffic dynamics, with the first six for rest days and the next six for work days.