Researchers copy NASA climate data to Swiss supercomputer amid US funding cuts (www.rfi.fr)

Swiss researchers have successfully transferred around 100 petabytes of NASA's climate and environmental data to the Swiss National Supercomputing Centre (CSCS) as a safeguard against potential US funding cuts in climate science. This data, comprising roughly six billion files, is essential for training AI models to enhance weather forecasting and climate prediction. As ETH professor Reto Knutti pointed out, this data is critical for understanding Earth systems like greenhouse gases and precipitation, and the initiative aims to integrate the vast datasets with the Alps supercomputer’s computing power to leverage advanced AI techniques. This move underscores a significant shift in the AI/ML landscape, especially as data-driven forecasting models—known as foundation models—begin to outperform traditional mathematical simulations. These models can generate forecasts much faster, with the capability to run global weather predictions in minutes rather than hours. The ETH’s initiative reflects ongoing concerns about the sustainability of funding for climate research, while simultaneously ensuring that vital data remains accessible for scientific advancements. By harnessing AI in data analysis, researchers hope to make more accurate predictions about climate change impacts, thereby enhancing our preparedness for natural disasters.