气溶胶粒子吸水性会很大程度上改变其光学特性。然而，不同的气溶胶粒子由于其化学和物理性质不同，它们的吸湿性也不相同。气溶胶粒子吸湿特性的差异影响气溶胶液态水含量(Aerosol Liquid Water Content，ALWC)。迄今为止，ALWC的计算方法主要基于地面站点观测数据，但是ALWC的垂直分布对于分析、研究垂直方向上气溶胶与大气气象要素之间的相互作用至关重要。为此，北京师范大学气溶胶研究团队多年来在全国不同地区开展了以气溶胶为核心的大型科学实验。本研究利用在三个超大城市观测实验获取的激光雷达观测数据并结合气溶胶光学及物化特性观测和气象资料，较深入系统地研究了气溶胶的吸湿特性，发展了ALWC估算算法，反演了ALWC垂直廓线。利用ALWC廓线产品，分析了超大城市不同污染背景下大气能见度与ALWC之间的关系。三个超大城市包括北京(2018.4-2019.9)，广州(2019.11-2020.4)，南京(2020.9-2021.6)。

首先基于在北京南郊站点对两个不同类型气溶胶吸湿性增长事件的深入研究，发现不同类型气溶胶的吸湿特性差异显著。它们分别是以沙尘为主的粗模态气溶胶(Case I)和以非沙尘为主的细模态气溶胶(Case II)。计算得到在相同湿度范围内，气溶胶吸湿增强因子f_β(RH，λ_532)分别约为1.4和3.1，细模态为主的气溶胶的吸湿性比沙尘型气溶胶高得多。为了研究沙尘型气溶胶对总体大气气溶胶吸湿性的影响贡献，在Case I中采用偏振激光雷达光度计联网方法(Polarization Lidar Photometer Networking，POLIPHON)识别并分离了沙尘型气溶胶。在Case I中，分离后得到的无沙尘粒子的细模态气溶胶的吸湿增强因子有明显的加强，与Case II中的大气总体气溶胶的吸湿增强因子接近。为了验证将沙尘型气溶胶分离之后结果的可靠性，利用同期2.5μm切割头的气溶胶质谱仪(Aerosol Chemical Speciation Monitor，ACSM)观测的细模态气溶胶化学化学组分，计算得到的吸湿性参数κ (Case I非沙尘气溶胶：0.357±0.024；Case II细模态气溶胶：0.344±0.026) 进行验证，结果表明二者均具有很强的吸湿性，与吸湿增强因子表现一致。这部分研究表明，北京地区细模态气溶胶在吸湿性增长过程中的贡献最大，而非吸湿性矿物粉尘气溶胶可能会降低单位体积的大气总体的吸湿特性。

再利用微脉冲激光雷达(Micropulse Lidar)获取的气溶胶后向散射系数等光学参数廓线、微波辐射计获取的大气温湿廓线，并结合气溶胶吸湿增强因子等参数，提出了计算气溶胶液态水含量(ALWC)垂直廓线反演的新方法。该方法基于梯度提升决策树模型(Gradient-Boosted Decision Tree，GBDT)的机器学习模型。反演模型的准确性和稳定性的评估通过与地面站点和532 m的广州塔站点的ALWC数据进行比较来进行的。验证结果一致性较好：全部数据的验证结果为R² = 0.904，均方根误差为RMSE = 3.06μg /m³；与塔上数据的验证结果为R² = 0.669，RMSE = 2.48μg /m³；与地面数据的验证结果为R² = 0.917，RMSE = 3.32μg /m³。利用ALWC廓线反演模型获取了广州地区ALWC的垂直分布，发现下边界层的气溶胶液体含水量较高。在大气环境中气溶胶颗粒中液态水含量与相对湿度的关系最为密切，这也会导致气溶胶粒子的形态及规则程度(水分占比每增加10％，体积退偏比降低0.02)。该模型可以用于研究大气污染和二次气溶胶的产生。

为了继续探究中国地区ALWC、气溶胶颗粒物等因素对于大气能见度的影响。在上述研究结果的基础上，对北京、广州、南京三个超大城市分别进行了分析。针对污染事件频发的秋冬季节，利用GAM(Generalized Additive Model)模型对各个影响大气能见度的因素进行分析，计算得到了各自对大气能见度的贡献。在秋冬季节，南京地区细模态颗粒物对于大气能见度的贡献更大；广州地区ALWC对于大气能见度的贡献更大；在北京地区两者均有贡献且差异较小。因此针对北京重污染事件进行了个例分析。综合分析得到北京的重污染事件中ALWC对于大气能见度的影响更大，当ALWC大于60μg /m³时，大气能见度处于较低的水平；当ALWC小于60μg /m³时，大气能见度会快速增加。

Water uptake by aerosol particles alters aerosol optical properties significantly. However, the hygroscopicities of different aerosol particles are not the same due to their different chemical and physical properties. The difference in the hygroscopicities of aerosol particles also affects the aerosol liquid water content (ALWC). Using ample observation data obtained from the field experiments in three megacityies, the hygroscopic properties of aerosols were analyzed, and a lidar-based machine-learning model was developed to retrieve the profile of ALWC. On the basis of obtaining the ALWC profile, the relationship between atmospheric visibility and ALWC under different pollution backgrounds in megacities is investigated. Data used in the study are aerosol optical, physical and chemical characteristics and comprehensive atmospheric variables measured during field experiments in three megacities in Beijing (2018.4-2019.9), Guangzhou (2019.11-2020.4), and Nanjing (2020.9 -2021.6).

Differences in the hygroscopicity of different types of aerosols are first investigated using measurements made at the Beijing site for two hygroscopic growth events of two different types of aerosols . They are coarse modal aerosols dominated by dust (Case I) and fine modal aerosols dominated by non- dust (Case II). Aerosol hygroscopicity behaves rather distinctly for mineral dust coarse-mode aerosol (Case I) and non-dust fine-mode aerosol (Case II) in terms of the hygroscopic enhancement factor,fβRH,λ532, calculated for the same humidity range. The hygroscopicity for non-dust aerosol was much higher than that for dust conditions with thefβRH,λ532 being around 1.4 and 3.1, respectively, at the same relative humidity of 86% for the two cases. To study the effect of dust particles on the hygroscopicity of the overall atmospheric aerosol, the two types of aerosols were identified and separated by applying the polarization lidar photometer networking method (POLIPHON) in Case I. The hygroscopic enhancement factor of separated non-dust fine-mode particles in Case I is significantly strengthened, approaching to that of the total aerosol in Case II. These results were verified by the hygroscopicity parameter, κ (Case I for non-dust particles: 0.357±0.024; Case II total: 0.344±0.026), based on the chemical species measured by an aerosol chemical speciation instrument (ACMS), both of which showed strong hygroscopicity. It was found that non-dust fine-mode aerosol contributes most during hygroscopic growth and that non-hygroscopic mineral dust aerosol lowers the total hygroscopicity in Beijing.

Next, using the profile of optical parameters such as aerosol backscattering coefficient (β) obtained by Micropulse Lidar, the atmospheric temperature and humidity profile obtained by microwave radiometer, and combined with parameters such as aerosol moisture absorption enhancement factor, a new method of aerosol liquid water concent (ALWC) profile is proposed. This method is based on the machine learning model of the Gradient-Boosted Decision Tree (GBDT) model. The assessment of the accuracy and stability of the inversion model was carried out by comparison with ALWC data from the ground site and the 532 m Canton Tower site. The verification results are in good agreement: the verification results of all data are R2 = 0.904, and the root mean square error is RMSE = 3.06; the verification results with the tower data are R2 = 0.669, RMSE = 2.48; the verification results with the ground data are R2 = 0.917, RMSE = 3.32. From the vertical distribution of the retrieved ALWC in Guangzhou, it is found that the aerosol liquid water content is higher in the lower boundary layer, especially when air pollution is severe. In the atmospheric environment, the content of liquid water in aerosol particles is most closely related to relative humidity, which also leads to the shape and regularity of aerosol particles (every ALWC 10% increase in the proportion of water, the volume depolarization ratio decreases by 0.02). The model may be of general usage for studying air pollution and secondary aerosol generation.

In order to continue to explore the influence of ALWC, aerosol particles and other factors on atmospheric visibility in China. Based on previous research, three mega-cities of Beijing, Guangzhou and Nanjing in China were analyzed respectively. For the autumn and winter seasons with frequent pollution incidents, the GAM model is used to analyze the degree of influence of various factors on atmospheric visibility, and the respective contributions to the atmospheric visibility are calculated. In autumn and winter, fine-mode particulate matter in Nanjing area contributed more to atmospheric visibility; ALWC in Guangzhou area contributed more to atmospheric visibility; in Beijing area, both contributed and the difference was small. Therefore, a case analysis was carried out for the heavy pollution incident in Beijing. Comprehensive analysis shows that ALWC has a greater impact on atmospheric visibility in heavy pollution events in Beijing. When ALWC is greater than 60μg /m³, atmospheric visibility is at a low level; when ALWC is less than 60μg /m³, atmospheric visibility will increase rapidly.