Towards an Improved Mechanistic Understanding of Dangerous Heat Extremes Affecting US Cities in the Historical Records and Future Climate Projections

极端高温越来越频繁地发生，并对自然环境、社区和公众健康产生负面影响。该项目研究这些极端高温，特别是那些导致美国城市地区死亡率上升的极端高温，目的是提高我们对这些极端高温事件及其未来几十年预期变化的理解。最初，该项目评估了不同的指标，以确定最能代表热风险的气候变量。值得注意的是，它超越了温度，包括湿度、太阳辐射和热辐射以及地表风速等因素。然后，它将历史上主要的极端高温与天气模式和长期气候驱动因素联系起来。分析数据的结果将与使用先进统计技术的全球气候模式的输出进行比较。最后，该项目使用高分辨率气候预测来了解致命的高温事件在未来可能发生的变化。这项工作将扩展社区对极端高温和死亡率的理解，但这项研究的重要性超出了科学理解。易受气候变化影响的社区，特别是那些适应能力有限的社区，面临极端高温的巨大风险。尽管认识到这些极端高温的社会影响，但我们在季节到十年尺度上预测它们的能力是不够的。该项目通过使用数据驱动的方法将致命的极端高温与大规模天气模式联系起来，解决了这一问题。这项研究的结果将传达给公共卫生和城市规划合作伙伴，以帮助他们应对这些威胁。通过这些分析产生的物理见解可以帮助解释基于气候模型的预测，预测这些危险的极端高温的未来发生和严重程度。此外，该项目还为研究生在统计建模、数据可视化、气候建模和气候传播方面提供了宝贵的培训。因此，它不仅通过其成果，而且通过培养下一代气候科学家，为更广泛的科学界和社会做出贡献。该项目由地球科学理事会共同资助，旨在支持人工智能/机器学习在地球科学领域的发展。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Heat extremes are happening more often and negatively affect natural environments, communities, and public health. This project studies these heat extremes, particularly those causing increased mortality in urban areas of the United States, with the goal of enhancing our understanding of these severe heat events and their anticipated changes in the coming decades. Initially, the project evaluates different metrics to identify climate variables that most robustly represent heat risks. Notably, it goes beyond temperature, including factors like humidity, solar and thermal radiation, and wind speed at the surface. It then connects major historical heat extremes to patterns in weather and long-term climate drivers. The results from analyzed data will be compared to output from global climate models using advanced statistical techniques. Lastly, the project uses high-resolution climate predictions to understand how lethal heat events may change in the future.This work will extend the community’s understanding of extreme heat and mortality, but the importance of this research extends beyond scientific understanding. Communities vulnerable to climate change, especially those with limited means to adapt, face substantial risks from heat extremes. Despite recognizing the social implications of these heat extremes, our ability to predict them on a seasonal to decadal scale is insufficient. This project addresses this issue by using a data-driven approach to associate lethal heat extremes with large-scale weather patterns. The findings from this research will be communicated to public health and urban planning partners to aid in their response to these threats. The physical insights generated through these analyses can help guide interpretation of climate model-based projection of future occurrence and severity of these dangerous heat extremes. Additionally, this project offers valuable training for graduate students in statistical modeling, data visualization, climate modeling, and climate communication. Thus, it contributes to the broader scientific community and society, not only through its results but also by training the next generation of climate scientists.This project is co-funded by the Directorate for Geosciences to support AI/ML advancement in the geosciences.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.