# MODELAIR Uses AI-Powered 3D Digital Twins for Street-Level Air Quality Insights > Source: > Published: 2026-09-18 10:00:30+00:00 The EU-funded **MODELAIR** project is developing [AI-enabled 3D digital twins](https://scalevise.com/ai-visibility-geo-checker) designed to predict urban air pollution at street level in real time. Rather than treating air quality as a city-wide average, the project combines city geometry, simulations and observed data to give authorities a more granular view of where pollution is likely to be concentrated and where interventions may have the greatest effect. The project is part of Horizon Europe under grant agreement No. 101072559. According to [MODELAIR's official project website](https://modelair.eu/), its goal is to build AI-based tools that support decisions on urban air-quality control through simulated 3D city representations and real-time insights. Pilot work is taking place in **Brussels, Madrid and Bristol**. That matters because conditions can vary sharply between streets. Traffic, the shape and height of surrounding buildings, and local airflow can all affect how pollutants move through an urban area. MODELAIR's approach is intended to make those local patterns visible and predictive, giving city authorities a basis for traffic planning, monitoring and more targeted pollution controls. MODELAIR is not simply a sensor dashboard. Its central technical proposition is the integration of AI with detailed modelling methods and 3D city data. The project uses computational fluid dynamics, often shortened to CFD, to model airflow and pollutant dispersion. It also uses reduced-order models, or ROMs, which are intended to support faster calculations than high-fidelity simulation alone. Project documentation and related CORDIS materials describe work on high-fidelity CFD and ROM approaches, data assimilation, and drone-based LiDAR geometry. Data assimilation is important in this context because it connects modelled conditions with observed data, helping the digital twin use available measurements rather than relying on a static simulation. | MODELAIR component | Role in the digital twin | Why it matters for street-level insight | |---|---|---| | 3D city models and drone-based LiDAR geometry | Represents the physical urban environment | Street and building geometry can influence local airflow and pollution dispersion. | | High-fidelity CFD | Models airflow and pollutant behaviour | Provides detailed simulation capability for urban conditions. | | Reduced-order models and machine learning | Supports faster predictive modelling | Helps make near real-time insight a practical project objective. | | Sensor data and data assimilation | Connects observed information with the model | Allows the digital twin to incorporate real-world measurements. | The distinction is significant. A conventional monitoring approach can show conditions at the location of a sensor. MODELAIR is aiming to combine those observations with a detailed representation of the surrounding city and predictive models. That can help authorities examine conditions across streets where sensors may not be installed, although the usefulness of any deployment will still depend on the quality of the underlying data and local implementation. The project's documented use cases centre on public decision-making. Horizon Magazine describes the work as supporting traffic planning, air-quality monitoring and targeted pollution controls. A city could use street-level forecasts to identify areas that need closer attention, instead of applying the same response across an entire district. MODELAIR's technical work is ongoing. Its wider deployment beyond pilot locations is not established by the available project information and would depend on factors including local sensor networks, usable city data and adoption by public authorities. The project should therefore be understood as an active effort to develop and validate this capability, not as a broadly available commercial platform. MODELAIR is primarily aimed at cities, but better local environmental intelligence could become relevant to organizations operating in urban areas. The immediate value is not a guaranteed return on investment for every company. It is the prospect of making operational decisions with more location-specific information when cities make such insight available or incorporate it into local actions. For example, urban service providers and logistics operators may need to understand where traffic planning or pollution controls affect routes and schedules. Building managers may have an interest in localized external conditions around a property. Businesses that work with municipalities, sensors, mapping, mobility or building data may also see opportunities to support the data and operational layers around similar digital-twin initiatives. The practical lesson is that **[data quality and integration](https://scalevise.com/services/api-system-integrations) are as important as the AI model**. A useful urban digital twin needs accurate geometry, relevant observations, reliable modelling and a clear decision process for acting on the output. Without those elements, a visually impressive 3D model may not translate into better operations. Urban data projects create value only when predictions lead to workable decisions and connected operations. Scalevise helps businesses assess AI use cases, map data and system dependencies, and design practical implementation paths that [reduce manual work](https://scalevise.com/services/ai-automation) instead of adding another dashboard. If local environmental intelligence could affect your fleet, facilities, customer service or operations, [Scalevise AI consultancy](https://scalevise.com/services/ai-consultancy) can help turn that opportunity into a defined plan. **Request a consultation.** **What is MODELAIR?** MODELAIR is a Horizon Europe project developing AI-enabled tools, including 3D city digital twins, to support urban air-quality decision-making. **How does MODELAIR predict street-level air pollution?** MODELAIR combines 3D city geometry, CFD and reduced-order modelling, machine learning, sensor data and data assimilation to model and predict local air-quality conditions. **Which cities are testing MODELAIR?** The project has pilot testing activities in Brussels, Madrid and Bristol. **Is MODELAIR available for broad city deployment?** The verified information describes an ongoing project and pilot work. Broader deployment would depend on local data, sensor networks and city adoption. **Why could street-level air-quality data matter to businesses?** Where local authorities use or share this information, it could help businesses understand location-specific conditions that affect urban operations, such as routes, facilities or services. MODELAIR shows how AI, detailed urban models and environmental observations can be combined to move air-quality analysis closer to the street where decisions are made. Its pilots in Brussels, Madrid and Bristol will be important tests of whether this technically ambitious approach can deliver useful, timely insight for urban planning and pollution control. For businesses, the nearer-term opportunity is to watch how this kind of local data becomes part of city operations and services.