Abstract

Improvements in computer processing power are facilitating the development of more detailed environmental models with greater geographical coverage. We developed a national-scale model of outdoor air pollution (Hybrid Air Dispersion Exposure System − HADES) for rapid production of concentration maps of nitrogen dioxide (NO2) and ozone (O3) at very high spatial resolution (10m). The model combines dispersion modelling with satellite-derived estimates of background concentrations, land cover, and a 3-D representation of buildings, in a statistical calibration framework. We developed an emissions inventory covering England and Wales to implement the model and tested its performance using concentration data for the years 2018–2019 from fixed-site monitoring locations. In 10,000 Monte Carlo cross-validation iterations, hourly-annual average R2 values for NO2 were 0.77–0.79 (RMSE: root mean squared error of 5.3–5.7 µg/m3), and 0.87–0.89 for O3 (RMSE = 3.6–3.8 µg/m3) at the 95% confidence interval. The annual average R2 was 0.80 for NO2 (RMSE = 4.9 µg/m3) and 0.86 for O3 (RMSE = 3.2 µg/m3) from aggregating the hourly-annual estimates. The air pollution surfaces are freely available for non-commercial use. In using these surfaces for exposure assessment, all residential locations, and neighbourhoods in urban areas, are unlikely to be below the 2021 World Health Organisation Air Quality Guidelines threshold (10 µg/m3) for annual average NO2 concentrations (10 µg/m3). Rural and suburban areas are likely to exceed the peak-season 8-hour daily maximum O3 threshold (60 µg/m3).