Alibaba DAMO Academy, the research arm of Alibaba Group, has introduced “Baguan,” an AI-driven weather forecasting model designed to improve weather predictions across a wide range of applications, from renewable energy management to natural disaster planning.
“Baguan,” derived from a term meaning “observing from different perspectives,” offers weather forecasts with a high spatial resolution, pinpointing predictions to a 1 x 1 kilometer area and updating them hourly. The model’s forecasting range spans from one hour to 10 days in advance, making it adaptable to both immediate and short-term forecasting needs.
“Baguan represents a significant advancement in our dedication to harnessing technology for the greater good,” said Wotao Yin, director of Decision Intelligence Lab at Alibaba DAMO Academy. “Its sophisticated technology not only helps elevate climate science but also benefits sustainable practices across diverse sectors such as renewable energy and agriculture.”
Baguan uses the Siamese Masked Autoencoders (SiamMAE) structure to analyze vast data sets from the European Centre for Medium-Range Weather Forecasts (ECMWF) along with local indicators like temperature and wind speed. This blend of global and local data allows Baguan to fine-tune its forecasts to regional conditions, improving precision.
In China’s Shandong province, Baguan’s accuracy was demonstrated during an unexpected temperature drop in August, predicting a 20% reduction in electricity demand one day ahead with 98.1% accuracy. This helped grid operators adjust power dispatch, increasing efficiency and lowering operational costs.
“We will continue to enhance performance for key weather indicators,” Yin added, citing plans to expand Baguan’s capabilities for climate scenario analysis and other sectors, including civil aviation and agriculture.
Thumbnail Image by wal_172619 from Pixabay