Development of the CMA-GFS-AERO 4D-Var assimilation system v1.0- Part 2: Evaluation through cycling assimilation experiments

Abstract. In Part 1 of this study (Liu et al., 2025), a strongly coupled aerosol–meteorology four-dimensional variational (4D-Var) data assimilation system, CMA-GFS-AERO 4D-Var, was developed based on the incremental analysis framework of the China Meteorological Administration Global Forecasting System (CMA-GFS), with black carbon (BC) selected as the initial assimilated aerosol species. In this second part, nearly three months of cycling data assimilation experiments from 10 October 2016 to 1 January 2017 were conducted to evaluate the practical performance of the system using BC surface observations from the China Atmosphere Watch Network (CAWNET). The impacts of BC assimilation on BC analyses, forecasts, and meteorological variables were systematically investigated. The results show that assimilating BC surface observations substantially improves the quality of both the BC background and analysis fields, with the analysis exhibiting a high degree of consistency with the observations. BC concentration forecasts are also significantly improved, particularly during severe pollution episodes over eastern China. The forecast benefits are mainly concentrated within the first 2–3 forecast days, with the largest improvements occurring during the initial 24 h forecast period. Moreover, assimilating BC surface observations yields a measurable positive impact on forecasts of near-surface 2 m air temperature (T2m) over heavily polluted regions, primarily through reductions in both warm and cold biases. These meteorological benefits are most pronounced during the early forecast period and gradually diminish with increasing forecast lead time. This study provides a comprehensive quantitative evaluation of the practical performance of CMA-GFS-AERO 4D-Var in real-data cycling assimilation experiments and demonstrates the effectiveness of strongly coupled aerosol–meteorology data assimilation in improving aerosol analyses and forecasts as well as meteorological predictions.