An AI-supported tool that enables cities to identify major sources of air pollution in near real time and design more targeted, evidence-based mitigation policies.
Date:
-
Location:
France
Partners:
Helmholtz Zentrum München; University of Twente; UMIT TIROL; Healthy Cities; European Public Health Alliance; UNOVA; Tree Technology; Lappeenranta-Lahti University of Technology; University of Augsburg; Substitute ApS; Kintsugi-LowCarbon; University of Vigo; Euroquality; Fundación de la Comunidad Valenciana para la Promoción Estratégica, el Desarrollo y la Innovación Urbana; Nantes Métropole; Air Pays de la Loire; Acoucité; Municipality of Rybnik
Research Area:
Climate-Responsive Societies & Just Transitions
Sustainable Development Goals:
03 - Good Health and Well-Being, 09 - Industry, Innovation and Infrastructure, 11 - Sustainable Cities an Communities, 13 - Climate Action
AI-Powered Near Real-Time Pollution Source Apportionment
Within the Horizon Europe PUREPOLIS project, LEITAT Technological Center is developing a near-real-time pollution source apportionment tool that uses artificial intelligence and existing air quality monitoring infrastructure to identify the main sources contributing to urban air pollution. The three-year project brings together 19 partners, with Nantes Métropole and Valencia serving as pilot cities and Berlin and Augsburg as replicator cities.
The tool will connect to Nantes Métropole’s existing air quality monitoring infrastructure and use machine learning trained on historical air pollution datasets to link specific pollutants to their likely sources. By combining historical data with continuously updated measurements, the system will provide faster and more automated source attribution than conventional analytical methods. It will focus on five principal pollution sources: road traffic, agriculture, urban background pollution, domestic heating, long-distance transport, and industry.
Data-Driven Urban Air Quality Management
The key added value of the solution is its dynamic, near-real-time capability. Unlike conventional source attribution methods that require longer periods of data collection and analysis, the tool is designed to identify potential short-term pollution episodes and provide authorities with timely information within the same day. This will allow policymakers to better understand when and how pollution sources change and to respond more effectively to emerging air quality issues.
By 2028, Nantes Métropole is expected to be able to attribute up to 70% of the main urban air pollutants identified under EU Directive 2024/2881 to five principal pollution sources. This information can support more targeted, evidence-based mitigation policies and help cities prioritize interventions that have the greatest potential impact on public health. Fine particulate matter alone is associated with nearly 240,000 deaths each year in the EU, highlighting the importance of more effective urban air quality management.
Pilot Deployment and Scaling Opportunities
The project is currently at the prototype-in-development stage, with the initial tool being developed and validated within the PUREPOLIS project. By the end of the project, the tool is planned to be operational in Nantes Métropole, with an upscaling strategy targeting five additional replicator cities by 2033. Engagement with city networks is expected to support further deployment beyond the initial project partners.
LEITAT is seeking cooperation with additional pilot and replicator cities interested in testing and adopting the tool. Such partnerships would support further validation, adaptation to different urban contexts, and the wider deployment of AI-supported, data-driven approaches to urban air pollution management.
