These data provide daily estimates of ground-level air pollution concentrations from available predictor data for PM2.5 (fine particulate matter), NO2 (nitrogen dioxide), O3 (ozone), and CO (carbon monoxide) for Functional Urban Areas (FUAs) worldwide, as delineated by the Global Human Settlement Layer (GHSL/JRC). Estimates are produced at a 0.1-degree spatial resolution and aggregated to the FUA level, covering the global terrestrial surface. This release spans April through November 2025.
Estimates are generated using separate random forest models for each pollutant, trained on a combination of satellite observations, meteorological reanalysis data, emissions inventories, and geographic features. Input data sources include: IQAir ground station observations; atmospheric pollutant retrievals from ESA's TROPOMI instrument aboard Sentinel-5P; meteorological variables from the Copernicus ERA5 Land reanalysis database; gridded emissions data from the EDGAR global emissions inventory; and Köppen-Geiger climate zone classifications and topographic/geographic features (elevation, latitude, coastal proximity).
See forthcoming working paper, "Tracking Air Pollution from Space: High-Resolution Daily Estimates for PM2.5, O3, NO2 and CO" by Blankespoor, Dasgupta and Wheeler.