Model outputs of surface PM2.5 concentration

Exposure to ambient PM2.5 (fine particulate matter with aerodynamic diameter less than 2.5 µm) can lead to adverse health effects. Air quality forecasting of PM2.5 is critical for informing the general public and decision makers to take preventive cautions. Air quality forecasting models of PM2.5 are subject to large uncertainties due to factors such as the incomplete representation of the physical and chemical processes. Here we develop a computationally efficient bias-correction framework to improve surface PM2.5 forecasts in the United States. We developed an ensemble-based Kalman filter (KF) technique focusing on the non-rural areas in the United States and apply the KF technique to outputs of three chemical transport models (GEOS-Chem, WRF-Chem and CMAQ) for the whole month of June 2012. All three models underestimate surface measured PM2.5 concentration by 20-50%, the KF technique is effective in improving the model forecasts by reducing the model bias.