新粒子生成事件对全球颗粒物数浓度的贡献几乎超过一次排放，全球平均占比达到了60%，这不仅为大气环境提供了高数浓度的气溶胶粒子、也为云滴形成提供了凝结核（CCN），而且使得 PM_2.5 在短时间内增长，严重加剧了我国频繁发生的二次颗粒物污染甚至导致灰霾天气。本研究基于本研究组在我国大气污染严重的京津冀、珠三角及长三角三大城市群的核心城市（北京、广州和上海）开展的综合大气、气溶胶增强观测实验（实验包括不同季节和新冠疫情），并集成地面站点的各种常规大气和气溶胶理化特性综合观测，获得了大量大气新粒子生成事件与同期丰富的大气边界层结构信息，通过颗粒数浓度谱分布的观测计算几何平均粒径和凝结汇，并结合美国和芬兰类似气溶胶观测资料以及大量的辅助观测，分析了新粒子的生成和发展属性，识别了不同季节内各要素对新粒子生成的贡献和影响，揭示并解释了大气湍流对新粒子生成和增长的相对贡献。

通过参加三个超大城市观测实验，掌握了气溶胶物理、化学、吸湿等特性和湍流参数的观测、质量控制、资料处理和分析等方法，比较了各种方法的差别。结合以上数据的质控和联合分析，研究新粒子生成的主要因素如何通过改变成核和其后的增长过程来决定新粒子生成的增长速率和持续时间，揭示了近地面湍流发展强度与地面新粒子生成发生之间潜在的联系。当可凝结的蒸汽受到湍流的支配时，产生更多的局部过饱和度利于形成团簇，更多的团簇继续自发增长，为新粒子生成的发生和持续发展提供了丰富的来源；同时增强的湍流还稀释了原有颗粒的浓度，导致凝结汇减小有利于新生颗粒的继续增长，这也延长了新粒子生成事件的持续时间。

本研究进一步利用分子动力学模型将新粒子生成的成核阶段进行微观层面的模拟，结合宏观的外场观测发现的实例，可以比较完整地揭示气溶胶成核和增长过程受到湍流发展的影响。发现并提出了一个可能的物理机制：近地面湍流发展可能为新粒子生成创造一个有利的环境，导致新粒子生成更频繁发生。这一物理机制是基于分析北京观测资料提出的，并使用美国南大平原站点（SGP）和芬兰北极森林（HYY）的长期观测数据进行验证，在全球不同排放背景下发现该机制也同样适用。另外利用广州塔和地面同步观测比对新粒子生成和受边界层演变的影响，发现新粒子生成在垂直空间内和其他气象要素之间的相关关系，进一步深入讨论在亚热带气候背景下的新粒子生成形成过程和分析结果。最后，结合机器学习和随机森林算法，尝试采用相关变量建模的方法解析直接影响新粒子生成的关键因子和影响程度，为预报新粒子生成进行探索研究。

The contribution of new particle formation (NPF) to global particle number concentration exceeds that of primary emissions, accounting for 60% at the global average. It does not only lead to high proportion of aerosol particles influencing atmospheric environment especially PM_2.5 outbreak, but also the cloud condensation nuclei (CCN) affect the formation of cloud droplets. Note that some newly formed particles can grow very fast to become secondary particle pollutants and cause haze. This study focuses on sever atmospheric pollution hotspots in our country, the Beijing-Tianjin-Hebei region, the Pearl river delta and the Yangtze river delta region's three megacities of Beijing, Guangzhou and Shanghai as research background. intensity experiment was carried out during different seasons and lockdown period, integrate various observation of aerosol physical and chemical properties, gained new particle formation characteristics and boundary layer structure simultaneously. Calculate the PNSD to get GMD and the condensation sink combined with auxiliary variable attributes at U.S. and Finland similar observation to identify different factor's impact on the NPF by their contribution, comprehensive parameters are used to identify the NPF seasonal variation, and the source that influences each factor explained the correlation between each factor and the relative contribution to NPF.

Based on these measurements, gained the aerosol physics, chemistry, hygroscopic characteristics, and turbulence development observation, quality control, data processing, and analysis methods are sorted out, and the differences of the relevant process methods are compared.  Determine the formation and duration of NPF by changing the nucleation and subsequent growth process, our findings reveal the ubiquitous relationship between the development tendency of surface-layer turbulence and the NPF growth rate and duration. When the condensable vapors are dominated by turbulence flow, more local supersaturation is generated which conducive to the formation of clusters, more clusters continue to grow spontaneously, providing a rich source to trigger the occurrence and maintain the growth of NPF. At the same time, the enhanced turbulence also dilutes the pre-existing particle concentration, leading to the reduction of condensation sink, which is favor to the continued growth of new particles, these prolong the duration of new particle formation events.

By using the molecular dynamics model to mimic the nucleation stage of new particles at the microscopic level, combining with the macroscopic field observation, revealed the process of aerosol nucleation and growth process impacted by turbulence. Our findings proposed a possible physical mechanism: the development of surface-layer turbulence may create a favorable environment for the formation of new particles, leading to more frequent NPFs. This physical mechanism was proposed based on the analysis of the Beijing observations and was verified using long-term observations from the SGP station in the United States and the arctic forest in Finland (HYY), this mechanism is also found to be applicable in different global emission situation. Besides, the characteristics of NPF in Guangzhou city where a field experiment was conducted on the ground and at a tower of 500m, the synchronous comparison is made between the tower and the ground, relationships were analyzed between NPF and other vertical meteorological elements, analysis results of NPF in the subtropical climate background are further discussed. Finally, approached by machine learning and random forest algorithm, we attempt to find the key influential factors affecting the occurrence of NPF for the sake of predicting the new particle formation events.