ExtremeWeather

ExtremeWeather数据集是由蒙特利尔大学MILA实验室创建的大型气候数据集，旨在通过半监督学习方法检测和理解极端天气事件。该数据集包含从1979年至2005年的27年气候模拟输出，总计78,840张16通道的768x1152图像，每张图像代表全球大气状态的快照。数据集的创建过程涉及使用CAM5模拟模型，该模型以25公里空间分辨率模拟全球气候。ExtremeWeather数据集的应用领域包括风险管理、政府政策决策支持以及气候系统的基本理解，旨在解决极端天气事件的检测、定位和理解问题。

The ExtremeWeather dataset is a large-scale climate dataset developed by MILA Lab at the University of Montreal, designed to detect and understand extreme weather events using semi-supervised learning methods. It includes 78,840 16-channel 768×1152 images from 27 years of climate simulation outputs covering the period from 1979 to 2005, with each image representing a snapshot of the global atmospheric state. The development of this dataset employed the CAM5 simulation model, which simulates global climate at a spatial resolution of 25 kilometers. The application scenarios of the ExtremeWeather dataset span risk management, government policy decision support, and fundamental understanding of the climate system, targeting the resolution of issues related to the detection, localization, and understanding of extreme weather events.