With measures such as low-emission zones, modern buses, and the expansion of both public transport and cycling infrastructure, Berlin has succeeded in improving air quality. Nonetheless, further reductions in air pollution remain necessary. Concentrations of certain pollutants still exceed the air quality guideline levels set by the World Health Organization (WHO) across the city. These limits are designed especially to protect vulnerable population groups. Particulate matter (PM₁₀, PM₂,₅) and nitrogen dioxide (NO₂) are considered particularly harmful to human health. They can cause cardiovascular and respiratory diseases, especially in children, older adults, and individuals with preexisting conditions. Diesel soot has been classified as carcinogenic by the WHO. NO₂ is primarily emitted by traffic. It contributes to the formation of particulate matter and ozone, and places additional strain on soils and ecosystems.
The ‘FAirQ’ model, developed by INWT Statistics GmbH, was designed to provide a citywide overview of air quality in Berlin. It supports both short-term predictions for the coming four days and the calculation of average air pollution levels for an entire calendar year. The model is built on AI and big data methods, drawing on a range of input data including air monitoring values, weather forecasts, traffic data, large-scale pollution predictions, and urban structure information. Using decision trees, it analyses relationships among these variables to predict concentrations of NO₂, PM₁₀, and PM₂,₅ across a 50 × 50 metre grid. The model’s outputs largely align with actual measurements, although deviations may occur in certain spots, such as near busy roads.
The model provides a comprehensive and nuanced picture of air pollution levels across Berlin and is regularly updated.