A geospatial approach for dynamic on-road emission through open-access floating car data

This paper presents a geospatial approach for quantifying street-level on-road emissions of carbon dioxide (CO ₂), nitrogen oxides (NO _x), and carbon monoxide (CO). By leveraging an existing open-access database of real-time congestion information derived from floating car data, we tested three methods to map high-resolution dynamic traffic emissions. To demonstrate the robustness and accuracy of the methods, we showcased results for summer workdays and winter weekends in the Helsinki Metropolitan Area (HMA). The three methods employed include (1) a physics-based relation known as the macroscopic fundamental diagram, (2) a data-driven input-adaptive generalized linear model (GLM), and (3) their ensemble (ENS). These methods estimated traffic density with satisfactory accuracy (R ² = 0.60-0.88, sMAPE = 31%-68%). Utilizing speed-dependent emission factors retrieved from a European database, the results compared favorably against the downscaled national emission inventory, particularly for CO ₂ (R ² = 0.70-0.77). Among the three methods, GLM exhibited the best overall performance in the HMA, while ENS provided a robust upscaling solution. The modeled emissions exhibited dynamic diurnal and spatial behavior, influenced by different functional road classes, fleet compositions and congestion patterns. Congestion-induced emissions were calculated to account for up to 10% of the total vehicular emissions. Furthermore, to anticipate the forthcoming transportation transformation, we calculated emission changes under scenarios with various penetration rates of connected and autonomous vehicles (CAVs) using this geospatial approach. The introduction of CAVs could result in emission reductions of 3%-14% owing to congestion improvements.