Capturing and translating the dynamics of traffic emissions using a congestion-based framework

Traffic congestion remains one of the biggest environmental and social issues in urban cities. Insights from traffic reports, modelling results, and real-world measurements show that traffic congestion would exacerbate vehicular emissions of up to 55%, compared to optimal driving conditions in highly congested urban areas.

To capture the dynamics of traffic patterns, we built our geospatial framework by utilising multiple sources of traffic data: traffic counts and speeds by local in-situ traffic counters, open-access aggregated floating car data (TomTom and Google Traffic), and a standardised functional road classification. The framework also incorporates meteorological parameters that affect the traffic capacity of urban road network to calculate the traffic density. Together with a projected fleet composition and its corresponding speed-dependent traffic emission factors, we computed the resulting dynamic traffic emissions of greenhouse gases (e.g. carbon dioxide CO₂) and air pollutants (e.g. carbon monoxide CO and nitrogen oxides NO_x) in gridded format. These can then be deployed in existing urban climate models to quantify climatic effects and air pollutant exposure induced by road transportation, and in particular congestion.

We applied the framework in two cities in Europe with distinct traffic behaviour: Helsinki and Munich. The preliminary results show relatively good performance in capturing the dynamics of traffic density in both cities (R² = 0.78–0.88). The framework was further evaluated against their local emission inventory. However, this gave varying results for different emittants for different road classes in both cities. Beyond local applicability, we also explored the scalabilty of the framework. Applying the calibration coefficients trained in one city and testing in another, we found that road classes such as local connecting roads behaved similarly in both places (r = 0.70–0.96 ) while some others did not.

This initiative sheds light on the feasibility of translating the framework to a larger scale beyond a few cities in Europe. Our future step is to improve the scalability of the framework by including existing large-scale multi-city traffic datasets on urban roads worldwide to better model the heterogeneity of the traffic patterns and emissions in the world.