IPCC第六次评估报告指出，气溶胶-云相互作用（ACI）是当前气候评估中最主要的不确定性来源，这很大程度上与当前全球气候模式对云滴数浓度和云滴谱谱型的简化计算和经验性参数化方法有关。本研究以云物理学经典理论为参考，运用新的观测数据和分析方法，对两种参数化方案进行了改进和评估，分析并量化了气溶胶影响这两个参数的不同效应对ACI的影响。首先，本研究基于气溶胶活化的经典云物理学理论方程，采用数值求解方法开发了一个考虑垂直次网格变化、计算更为准确、且具有可调参数的新云滴数浓度方案（QDGE方案），并以精确求解的数值模型和飞机/卫星观测数据为参考，评估了该方案在区域和全球尺度的模拟效果。结果显示，相比于以往的参数化方法，新方案虽然计算量略大，但却显著提高了对云滴数浓度的计算准确性，整体表现更佳。其次，本研究基于物理方程和理论推导，从云滴谱谱型对云微物理量和辐射通量的影响出发，提出了一种定量评估离散效应（谱型参数对气溶胶或云滴数浓度的敏感度）及其对ACI影响的新方法，并基于此评估了五种在气候模式中广泛使用的云滴谱谱型参数化方案。研究结果发现，基于理论推导的LiuLi15方案在计算谱型参数、离散效应及其对ACI影响等方面的综合表现最优，而基于区域观测数据经验性拟合的几种参数化方案则存在较大偏差。随后，我们将QDGE和LiuLi15方案加入到清华气候模式CIESM中，并对气候模式中的气溶胶和云微物理模块进行了多处改进，然后利用多源观测数据评估了CIESM在模拟气溶胶、云、降水和辐射上的表现。结果显示，改进后的CIESM对云滴数浓度的模拟更为准确，与观测数据一致性更高。最后，本研究运行上述改进的CIESM，利用偏辐射扰动法分离了气溶胶影响云滴数浓度和云滴谱谱型的两种效应，并采用多源观测数据进行约束评估。结果表明，两种效应的辐射强迫分别占ACI总辐射强迫的66%和10%。这表明气溶胶-云滴数浓度效应对气候模式中的ACI估算具有主要影响，而气溶胶-云滴谱谱型效应的贡献相对次要但也不可忽略。因此，在未来的ACI研究中，有必要谨慎选择两种参数化方案，以减少由此而引入的不确定性。
The Sixth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) emphasizes Aerosol-Cloud Interactions (ACI) as the primary source of uncertainty in current climate assessments. This is largely attributed to the simplified calculations and empirical parameterizations for the cloud droplet concentration and cloud droplet spectral parameter in current global climate models. This study improves and evaluates the parameterizations of cloud droplet number and spectral parameter, analyzing and quantifying the aerosol effects on the two cloud parameters and their impacts on ACI, based on classical cloud physics theory, new observational data, and novel analysis methods.Firstly, this study developed a new scheme (QDGE scheme) to calculate the cloud droplet number using a numerical method according to the classical cloud physics theory equations. This scheme considers vertical sub-grid variations of cloud supersaturation, improves accuracy in the calculation of the aerosol activation, and includes tuning factors. The QDGE scheme is evaluated on regional and global scales using a numerically accurate parcel model and aircraft/satellite observations. Results indicate that the QDGE scheme significantly enhances the accuracy of the calculated cloud droplet number compared to previous empirical parameterizations in climate models, despite a slightly higher computational cost.Secondly, we developed a new method to quantitatively assess the dispersion effect (sensitivity of droplet spectral parameter to aerosol or cloud droplet number, hereafter DE) and its impact on ACI. This method is applied to quantify and evaluate five cloud droplet spectral parameterizations widely used in current climate models. Our results indicate that the semi-theoretical LiuLi15 scheme shows the best overall performance in calculating spectral parameters, DE, and the impact of DE on ACI, while other empirical parameterizations have noticeable discrepancies. We then implemented both QDGE and LiuLi15 schemes into the Tsinghua climate model CIESM and made several modifications to aerosol and cloud microphysical modules of CIESM. We evaluated the performance of CIESM in terms of aerosol, cloud, precipitation, and radiation using multi-source observational datasets. Results show that the improved CIESM performs better in simulating the cloud droplet number concentration, showing higher consistency with observations.Finally, we investigate the aerosol effects on cloud droplet number and spectral parameter and their impacts on ACI. By applying the Partial Radiative Perturbation method to divide the two aerosol effects on ACI, we reveal that the radiative forcing of aerosol effects on the droplet number concentration and cloud droplet spectral parameter contribute to 66% and 10% of the total ACI, respectively. The results highlight the predominant role of the aerosol-droplet number effect on ACI, and the secondary but non-negligible significance of the aerosol-spectral parameter effect on ACI. Therefore, it is necessary to carefully choose the two parameterization schemes in future model studies to reduce the uncertainties in ACI estimation.
