From Deterministic to Generative Deep Learning for Urban Air Quality Reconstruction from Sparse Observations

2026-07-28

AI Summary

This study addresses the challenge of city-wide air pollution reconstruction, which is hindered by complex pollutant interactions, meteorological uncertainty, and sparse monitoring stations. The authors propose a diffusion-based generative framework for multi-pollutant reconstruction that efficiently generates high-resolution spatial distributions of NO₂, O₃, PM₂.₅, and PM₁₀ from sparse observations. A novel data augmentation strategy, requiring no retraining, is introduced to enable effective transfer from simulated to real-world scenarios. Experiments on real monitoring data from Paris demonstrate that the method produces realistic pollution patterns with high structural similarity, significantly outperforming deterministic deep learning models. The approach exhibits strong generalization capability and promising potential for practical deployment in urban air quality monitoring.

Abstract

Full-field reconstruction of air pollution is essential for evaluating pollution exposure and supporting public health decision-making. However, the complex interactions among pollutants, hard-to-predict weather patterns, and limited monitoring station coverage make this a complex task. We apply deep learning techniques to provide fast and accurate reconstructions from sparse observations of four key pollutants: NO2, O3, PM2.5 and PM10. Models are trained on full-field simulation data and evaluated on real-world observations collected from 9 to 28 monitoring stations in the city of Paris. We introduce a diffusion-based generative framework for multi-pollutant reconstruction and benchmark its performance against deterministic deep learning models. Despite noisy observations and strong spatial variability, the models achieve high structural similarity on simulated validation data and produce realistic spatial patterns on real-world observations, as indicated by power-spectrum analysis. We introduce data augmentation methods that enable transfer to real-world observations without retraining, allowing the models to generalise beyond the training period. These findings highlight the potential of ML models for reliable real-world deployment in air pollution reconstruction tasks.

Problem

• air quality reconstruction
• sparse observations
• multi-pollutant
• urban air pollution
• full-field reconstruction

Innovation

• diffusion model
• generative deep learning
• air quality reconstruction
• data augmentation
• multi-pollutant modeling