5.3      Numerical simulation of an extreme haze pollution event over North China Plain based on initial and boundary condition ensembles.

Liu, Hongbo, Xiaobin Li, JuanJuan Liu, LASG, Institute of Atmospheric Physics, China, and Ziyin Zhang, Institute of Urban Meteorology, China Meteorological Administration, China

North China Plain often suffers heavy haze pollution events in cold season due to the rapid industrial development and urbanization in recent decades. In the winter of 2015, the megacity cluster of Beijing-Tianjin-Hebei (BTH) experienced a 7-day long extreme haze pollution episode with peak PM2.5 [particulate matter (PM) with an aerodynamic diameter ≤ 2.5 μm] concentration of 727 μg/m3. In this study, the sensitivity of PM2.5 concentration variation in this event to meteorological conditions is examined through numerical experiments with WRF-Chem model. A control run (CTRL) and three groups of ensemble experiments (INDE, BDDE, INBDDE) have been carried out based on different initial and lateral boundary conditions derived from ERA5 reanalysis and its 10 ensemble members (EDA). The CTRL run reasonably well reproduces the meteorological conditions and the overall life cycle of the haze event but fails to capture the intensely oscillation of instantaneous PM2.5 concentration. However, the ensemble forecasting shows a considerable advantage to some extent. Compared with CTRL run, the root mean square error (RMSE) of PM2.5 concentration decreases by 4.33%, 6.91% and 8.44% in INDE, BDDE and INBDDE, respectively. It’s found that the RMSE decreases of wind direction (-5.19%, -8.89% and -9.61%) seem to be the dominant reason for the improvement of PM2.5 prediction skill in three ensemble experiments. Generally, the ensemble forecasting scheme is an effective method to improve the prediction skills of wind direction and PM2.5 concentration by using WRF-Chem model in this case.