Urbanization has a substantial effect on urban meteorology. It can alter the atmospheric diffusion capability in urban areas and therefore affect pollutant concentrations. To study the effects of Hangzhou’s urban development in most recent decade on its urban meteorological characteristics and pollutant diffusion, 90 weather cases were simulated, covering 9 weather types, with the Nanjing University City Air Quality Prediction System and high-resolution surface-type data and urban construction data for 2000 and 2010. The results show that the most recent decade of urban development in Hangzhou substantially affected its urban meteorology. Specifically, the average urban wind speed decreased by 1.1 m s ⁻¹; the average intensity of the heat island increased by 0.5°C; and the average urban relative humidity decreased by 9.7%. Based on one case for each of the nine weather types, the impact of urbanization on air pollution diffusion was investigated, revealing that the changes in the meteorological environment decreased the urban atmosphere’s diffusion capability, and therefore increased urban pollutant concentrations. For instance, the urban nitrogen oxides concentration increased by 2.1 μg m ⁻³ on average; the fine particulate matter (diameter of 2.5 μm or less; PM_2.5) pollution concentration increased by 2.3 μg m ⁻³ on average; in highly urbanized areas, the PM_2.5 concentration increased by 30 μg m ⁻³ and average visibility decreased by 0.2 km, with a maximum decrease of 1 km; the average number of daily hours of haze increased by 0.46 h; and the haze height lifted by 100–300 m. The “self-cleaning time” of pollutants increased by an average of 1.5 h.

In numerical model simulations, data assimilation (DA) on the initial conditions and bias correction (BC) of model outputs have been proven to be promising approaches to improving PM_2.5 (particulate matter with an aerodynamic equivalent diameter of ≤ 2.5 μm) predictions. This study compared the optimization effects of these two methods and developed a new scheme that combines DA and BC simultaneously. Four parallel experiments were conducted during winter 2019: a control experiment directly forecasted by WRF-Chem (experiment name: WRF-Chem); an experiment that assimilated in situ observations based on the GSI (Gridpoint Statistical Interpolation) system (WRF-Chem_DA); an experiment with deep-learning-based BC (WRF-Chem_BC); and an experiment considering the combination of DA on the initial conditions and BC (WRF-Chem_DA_BC). Statistically, the accuracy of PM_2.5 predictions could be optimized by both DA and BC for the first 24-h period, and WRF-Chem_BC performed better than WRF-Chem_DA in the initial field, especially in the period of 10-24 h, while the best performance was achieved by combining BC and DA. Throughout the initial 24-h period, compared with the control experiment, the results of WRF-Chem_DA_BC (WRF-Chem_DA, WRF-Chem_BC) showed an improvement in terms of root-mean-square error, with reduction proportions varying from 38.90 % to 48.86 % (18.88 % to 32.44 %, 30.10 % to 46.08 %). Besides having the best optimization effect over the whole domain, the combined method also performed well in different regions: during the forecasting period of 0-24 h, the RMSEs decreased from 32 % to 62 %, 39 % to 57 %, 28 % to 40 %, and 30 % to 49 % in the Beijing-Tianjin-Hebei, Yangtze River Delta, Central China, and Sichuan Basin urban agglomerations, respectively.