Urban traffic is one of the main sources of greenhouse gas and air pollutant emissions, making accurate estimation a key challenge for cities aiming to develop effective climate strategies. A research team led by Jia Chen, professor of Environmental Sensing and Modeling at Technical University of Munich (TUM), has developed a new data-driven framework that provides detailed insights into urban traffic emissions.

What’s the situation in Munich?
Overall, traffic emissions in Munich have been declining since 2022. Encouragingly, the impact of stricter emission standards and cleaner vehicle fleets is clearly visible in the proportionally stronger decrease of nitrogen oxides (NO₂ and NOₓ) and carbon monoxide compared to carbon dioxide. Despite this, Munich continues to struggle with heavy traffic congestion, and overall traffic volume has increased since 2020, although it has not yet returned to pre-COVID levels by 2024. On the Mittlerer Ring, however, a contrasting trend is emerging between 2023 and 2024: traffic volumes have fallen slightly, and emissions have fallen more sharply than the city average.

DRIVE: Capturing a citywide emissions picture
The model, called DRIVE (Data-driven Road-Transport Inventory for Vehicle Emissions), combines three key data sources that have rarely been integrated at this level of spatial and temporal detail: a traffic model of the City of Munich, real-world data from more than 100 counting stations, and emission factors that reflect the technical characteristics of vehicle fleets. It accounts for greenhouse gases (CO₂, CH₄) and air pollutants (CO, NOₓ, NO₂, PM₁₀).
Applied to Munich, DRIVE provides hourly emission estimates down to individual road segments, enabling a much more precise understanding of how emissions vary across space and time, including realistic uncertainty assessments.

COVID‑19 pandemic causes sharp drop in traffic
The researchers analyzed traffic emissions from 2019 to 2022, capturing major disruptions such as the COVID‑19 lockdowns. Traffic volumes dropped significantly in 2020, leading to a 15.9 percent decrease in total emissions. Although traffic activity increased again in 2021 and 2022, it remained below pre‑pandemic levels. The results also show that air pollutants declined more quickly than CO₂ emissions, reflecting stricter emission standards and the gradual shift toward cleaner vehicle fleets.

Potential applications: From speed limits to infrastructure planning
Beyond retrospective analysis, DRIVE is designed to operate nearly in real time once current traffic data is available. This makes it a valuable tool for evaluating implemented policy measures such as speed limits, traffic restrictions like low‑emission zones and driving bans, or infrastructure changes. The model also provides uncertainty estimates across different spatial and temporal scales, strengthening the reliability of urban emissions inventories.
By integrating multiple data sources into a unified model, the study demonstrates how data science can support more accurate monitoring and more targeted reduction of urban emissions.

The research results were produced in collaboration with the Netherlands Organisation for Applied Scientific Research (TNO), with support from the EU project “ICOS Cities”. The research is being continued as part of the current “MCube” (Munich Cluster for the Future of Mobility in Metropolitan Regions) project and in collaboration with the City of Munich.

Publication:
Kühbacher, D., Chen, J., Aigner, P., Ilic, M., Super, I., and Denier van der Gon, H.: DRIVE v1.0: a data-driven framework to estimate road transport emissions and temporal profiles, Geosci. Model Dev., 18, 9967–9990, https://doi.org/10.5194/gmd-18-9967-2025, 2025.