A multi-model operational air pollution forecasting system for China

In this study we used an ensemble of nine chemistry transport models (CTMs) to provide operational forecasts of air quality in China for 3-5 days ahead (http://www.marcopolo-panda.eu/forecast/). The performance of the multi-model prediction system is evaluated using available datasets in China including ozone and particulate matter. The ensemble approach enhances the accuracy of the forecasts and shows improved statistical skills compared to individual model predictions. We discuss the different uncertainties affecting the performance of current air quality forecasting systems and the different approaches that can be used to reduce such uncertainties.