CMA-MESO/CUACE CW化学天气耦合模式是自主研发的大气化学耦合模式， 目前CMA-MESO/CUACE CW 3DVar同化系统实现了地面气溶胶观测可吸入颗粒物PM_2.5和PM₁₀的同化， 为增强耦合同化系统非常规观测的同化能力， 文中在CMA-MESO大气化学天气耦合三维变分同化框架基础上， 利用查表法获得气溶胶消光系数， 然后建立气溶胶光学厚度（Aerosol Optical Depth， AOD）和气溶胶组分之间关系的观测算子、 切线性观测算子和伴随观测算子， 实现AOD观测资料的同化应用。针对2016年12月18 -20日华北、 黄淮地区一次污染天气过程进行同化预报试验， 试验结果表明同化葵花-8卫星（Himawari-8）气溶胶光学厚度观测后， PM_2.5分析的重污染区范围有所扩大， 山西东南部分析与实况分布更为接近， 但是山东大部地区PM_2.5分析偏强， 与观测相比PM_2.5质量浓度存在高估。同时同化Himawari-8 AOD观测和地面气溶胶站点观测的PM_2.5分析最优， 分析与观测距平相关系数最高， 平均偏差、 均方根误差及标准差最小。重污染区的PM_2.5预报检验结果表明， 同化Himawari-8 AOD观测对大于350 μg·m^-3量级PM_2.5预报正贡献可以持续到48 h， 但整体来说， 同时同化Himawari-8 AOD观测和地面气溶胶站点观测对各个量级的PM_2.5质量浓度预报质量最优。

The CMA-MESO/CUACE CW is an atmospheric chemistry coupled model independently developed by China.At present， the CMA-MESO/CUACE CW 3DVar system realized the assimilation of ground-based aerosol observations of PM_2.5 and PM₁₀.In order to enhance the assimilation capability of the coupled assimilation system for non-conventional observations， based on the CMA-MESO three-dimensional variational assimilation framework of atmospheric chemical weather coupled system， the aerosol extinction coefficient is obtained by using the look-up table method.And then the observation operator， the tangent linear operator and the adjoint operator are established using the relationship between aerosol optical depth （AOD） and aerosol components.The assimilation tests were carried out for a haze process in North China on December 18 -20， 2016.The results show that after assimilating the AOD observations from the Himawari-8 satellite， the heavily polluted area of the PM_2.5 is enlarged， and the analysis in southeastern Shanxi is closer to the actual situation， but the analysis of PM_2.5 in the majority of the Shandong region is overestimated.The simultaneous assimilation of Himawari-8 AOD observations and ground-based aerosol concentration observations test is the best for the analysis of PM_2.5， it has the highest correlation coefficients （ACCs） and the smallest mean bias， root-mean-square error and standard deviation.The PM_2.5 forecast results in the heavily polluted area showed that the positive contribution of the assimilated Himawari-8 AOD observations to the PM_2.5 forecasts for the >350 μg·m^-3 magnitude could be sustained up to 48 h.However， the simultaneous assimilation of the Himawari-8 AOD observations and the ground-based aerosol site observations had the best test scores of PM_2.5 forecasts for each magnitude.