Application of WRF/Chem-MADRID for real-time air quality forecasting over the Southeastern United States

A Real-Time Air Quality Forecast (RT-AQF) system that is based on a three-dimensional air quality model provides a powerful tool to forecast air quality and advise the public with proper preventive actions. In this work, a new RT-AQF system is developed based on the online-coupled Weather Research and Forecasting model with Chemistry (WRF/Chem) with the Model of Aerosol Dynamics, Reaction, Ionization, and Dissolution (MADRID) (referred to as WRF/Chem-MADRID) and deployed in the southeastern U.S. during May–September, 2009. Max 1-h and 8-h average ozone (O3) and 24-h average fine particulate matter (PM2.5) are evaluated against surface observations from the AIRNow database in terms of spatial distribution, temporal variation, and domain-wide and region-specific discrete and categorical performance statistics. WRF/Chem-MADRID demonstrates good forecasting skill that is consistent with current RT-AQF models. The overpredictions of O3 and underprediction of PM2.5 are likely due to uncertainties in emissions such as those of biogenic volatile organic compounds (BVOCs) and ammonia, inaccuracies in simulated meteorological variables such as 2-m temperature, 10-m wind speed, and precipitation, and uncertainties in the boundary conditions. Sensitivity simulations show that the use of the online BVOC emissions can improve PM2.5 forecast in areas with high BVOC emissions and adjusting lateral boundaries can improve domain-wide O3 and PM2.5 predictions. Several limitations and uncertainties are identified to further improve the model’s forecasting skill.

► A New RT-AQF Model based on online-coupled WRF/Chem with an advanced aerosol module.
► A good forecasting skill for O3 and PM2.5 that is consistent with other RT-AQF models.
► The use of the online BVOC emissions can potentially improve PM2.5 forecast.
► The model is sensitive to boundary conditions.