The CMA Chemistry–Weather Coupled Model and Its Operational Forecast of Fog–Haze and Sand/Dust Storms

The China Meteorological Administration (CMA) Chemistry–Weather Coupled Model (CMA-CW), combined with a two-option chemical data assimilation (DA) module and an anthropogenic emission processing module, has been developed for operational prediction of fog–haze and sand and dust (SD) storms (SDSs). The chemistry–weather (CW) two-way feedback is realized through aerosol–cloud–radiation interactions. The fog–haze and SD predictions are evaluated and compared with those by CMA Unified Atmospheric Chemistry Environment model (CUACE) and Asian Dust Aerosol Model 3 (ADAM3) models based on the WMO platform. Analyses show that the 24–72-h threat scores (TSs) of PM_2.5 and visibility from CMA-CW are obviously higher than those from CMA-CUACE, especially for the high PM_2.5 and low visibility levels. Spatiotemporal variations of PM_2.5 and visibility in some subregions, e.g., Northeast China (NEC), Beijing–Tianjin–Hebei (Jing–Jin–Ji or JJJ), Sichuan basin (SCB), Yangtze River Delta (YRD), and Pearl River Delta (PRD), demonstrate that CMA-CW successfully reproduces most regional fog–haze events and basically captures the peaks, troughs, and variation timings of PM_2.5 and visibility. The prediction accuracy for fog–haze processes is about 82%. Ten out of the total 11 SDSs predicted by CMA-CW are rated as “qualified”, including 5 cases graded as “excellent” or “good”. This proves the model’s capability in predicting the life cycle of SDS events, including dust emission, transport, movement, duration, and affected area. Model comparison indicates that CMA-CW and ADAM3 generally reproduce the coverage, arrival time, and intensity of SDSs in most downwind areas, demonstrating better performance than CMA-CUACE. Superior performance in SD forecasts by CMA-CW is demonstrated for the 96–120-h prediction, with the 24–120-h prediction showing higher prediction consistency. Data assimilation (DA) can improve 24-h SD forecasts by 10%–13%. The impact of the CW coupling on model meteorology prediction is examined and it is found that CMA-CW produces generally improved meteorological predictions from surface to 500 hPa in varying degrees than the uncoupled model.