Enhancing Accuracy in Climate Forecasting
In a significant breakthrough for climate research, a team of scientists from the University of Chicago and ETH Zurich have created a comprehensive dataset of extreme weather events, known as ClimateNetLarge. This dataset, which is the largest expert-labeled resource of its kind, contains over 49,000 examples of intense rain bands, powerful cyclones, and persistent atmospheric blocks.
The dataset, designed to support deep learning segmentation and tracking of these events, includes extreme weather events from 1980 to 2022. It was drawn from climate data collected by the European Centre for Medium-Range Weather Forecasting. Each event in the dataset was annotated by scientists, offering a consistent, high-quality reference for tracking how major weather patterns form and move.
Currently, direct public access to ClimateNetLarge is not readily available. However, research articles describing the dataset can be found through academic databases or directly from the researchers involved. Keep an eye on platforms like arXiv, ResearchGate, or Academia.edu, where researchers often share their work and datasets.
If the dataset is not publicly available through a specific repository, consider reaching out to the authors or institutions involved. They might provide access for research purposes or direct you to any relevant repositories. Look for updates or announcements from the University of Chicago's Urban Theory Lab and ETH Zurich's Institute of Geodesy and Photogrammetry, as they are the primary institutions behind the dataset.
With access to this dataset, researchers can build more accurate tools to predict floods, heatwaves, and other climate-related risks. The image associated with this article is credited to Dominik Schroder, likely depicting a weather-related event, given the context of the article.
Scientists at the University of Chicago and ETH Zurich have conducted research in environmental science and climate-change, developing ClimateNetLarge, a dataset of extreme weather events from 1980 to 2022. This resource, which contains over 49,000 examples of intense rain bands, powerful cyclones, and persistent atmospheric blocks, is designed to support AI-based prediction and tracking of such events. While public access to ClimateNetLarge isn't immediate, it can be discovered through academic databases or by contacting the authors or institutions involved.
