CSCS Joins EU WeatherGenerator Project to Advance AI-Driven Climate Prediction
Feb. 13, 2025 -- On February 5, 2025, the European Union launched the WeatherGenerator project, a new initiative aimed at transforming weather and climate forecasting using artificial intelligence (AI). Funded by the Horizon Europe program, the project brings together 16 leading organizations to develop a cutting-edge AI model for Earth system science.
Led by the European Centre for Medium-Range Weather Forecasts (ECMWF), the initiative seeks to improve the accuracy of climate predictions, optimize renewable energy planning, and enhance disaster response to extreme weather events like floods and heatwaves.
Switzerland is back in the game and is contributing to the WeatherGenerator project, with ETH Zurich and MeteoSwiss among the participating organizations. ETH Zurich is represented by the Swiss National Supercomputing Centre (CSCS), the Center for Climate Systems Modeling (C2SM) and the Scalable Parallel Computing Laboratory (SPCL).
CSCS will develop a data transfer service to facilitate the efficient movement of large datasets between supercomputing centers for training and deployment on high-performance computing (HPC) resources. CSCS will also contribute to identifying and cataloging relevant training datasets. C2SM will play a key role in developing the machine learning architecture, aligning it with the advancements of SwissAI, and conducting verification and fine-tuning over the Alpine region in close collaboration with MeteoSwiss. Lastly, SPCL will focus on data compression, exploring techniques to compress multidimensional weather and climate data while maintaining high compression ratios and preserving data integrity.
Key objectives include advancing AI-driven climate research, improving disaster forecasting, and providing open-source tools for global collaboration. By merging expertise in AI, meteorology, and energy, WeatherGenerator aims to establish Europe as a leader in climate technology. Stay tuned for further updates on this pioneering initiative.
More information about the WeatherGenerator project here.
Source: Maria Grazia Giuffreda, CSCS
