Environmental controls over tropical cyclone formation

热带气旋(TCS，也被称为北大西洋盆地飓风)是每年对经济和社会构成高风险的重大极端天气事件之一。尽管在过去的几十年里，对TC发展的研究取得了很大的进展，但我们目前对TC过程的了解仍然不够。对TC形成缺乏了解的一个直接后果是我们在实践中预测TC形成的能力有限。因此，加强对热带气旋形成过程和后续发展的认识，对于科学进步和社会安全与风险管理都具有重要意义。利用数值模拟和最新的机器学习技术，该项目引入了一种新的TC研究方法，可以帮助揭示不同海洋盆地TC形成的主要机制。与目前热带气旋研究和运行中传统的涡旋跟踪方法不同，本项目的方法为寻找影响热带气旋形成的新的环境因素提供了一种替代途径。通过探索这些机器学习方法在热带气旋研究中的实用性，该项目不仅将引入一种研究热带气旋发生的创新方法，还将有助于为大气研究培训人工智能应用领域的新一代劳动力。这个项目结合了几种新的模拟和机器学习方法来研究两个关键的科学问题：1)控制不同大洋盆地不同TC形成途径的关键环境因素是什么；2)全球热带气旋发生是否具有阶段性发展，如果是，什么环境因素决定了这种全球TC形成的频率。利用模式模拟、物理诊断分析和机器学习算法的组合，该项目将研究热带大气中的主要环境条件及其动力学约束，这些条件在当前的热带气旋成因潜力指数框架中没有得到充分的体现。这些环境约束将有助于确定热带气旋形成的不同有利条件，以及如何利用这些因素来研究过去和未来气候中热带气候学的转变。该项目的成果还将使我们更深入地了解热带大气每年在我们当前的地球大气中产生TC的最大能力背后的物理机制。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Tropical cyclones (TCs, also known as hurricanes in the North Atlantic basin) are among the major extreme weather events that pose a high risk to the economy and society every year. Despite great progress in studying TC development over the last several decades, our current understanding of TC processes is still inadequate. A direct consequence of this lack of understanding of TC formation is our limited capability in predicting TC formation in practice. Thus, improving the understanding of TC formation processes and TC subsequent development is of significance for both the advance of science and social safety and risk management. Using numerical simulations along with recent machine learning techniques, this project introduces a new approach to TC research that can help unfold the dominant mechanisms underlying TC formation in different ocean basins. Unlike the traditional vortex tracking methods in the current TC research and operation, the approach in this project offers an alternative way to search for new environmental factors governing TC formation. By exploring the practicality of these machine learning methods for TC research, this project will not only introduce an innovative way of studying tropical cyclogenesis but also help train a new generation of workforce in the area of artificial intelligence applications for atmospheric research. This project combines several new modelling and machine learning methods to investigate two key scientific questions: 1) what are the key environmental factors that control different TC formation pathways in different ocean basins, and 2) whether global tropical cyclogenesis possesses an episodic development, and if so, what environmental factors determine the frequency of this global TC formation. Using a combination of model simulations, physical diagnostic analyses, and machine learning algorithms, this project will look into major environmental conditions and their dynamical constraints in the tropical atmosphere that are insufficiently represented in the current TC genesis potential index framework. These environmental constraints will help determine different favorable conditions for TC formation and how to use these factors for studying the shift of TC climatology in the past and future climates. The outcomes from this project will also bring more insights into the physical mechanisms behind the maximum capacity of the tropical atmosphere in generating TCs annually in our current Earth’s atmosphere.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.