China unveils first Fengyun satellite AI models for extreme weather

1

30th October 2026 – (Changsha) China has launched its first batch of Fengyun meteorological satellite artificial intelligence models, presented at a users’ conference in Changsha on Thursday, 29th October, to bolster detection and warning capabilities for extreme weather. The China Meteorological Administration said the suite includes an intelligent typhoon cloud imagery interpretation model and a severe convection nowcasting warning model, designed to enhance surveillance of typhoons and violent storm systems and to speed decision‑making for forecasters.

Nowcasting, used for very short‑range forecasting, lies at the core of the upgrade. The intelligent typhoon model can automatically analyse position, structure, intensity and evolution, while simultaneously producing standardised interpretation text and visual outputs to support monitoring and forecast operations. It currently covers seven categories of typhoon features, performs an interpretation every 15 minutes and generates the accompanying text within 27 seconds, aiming to improve situational awareness during rapidly changing conditions.

The severe convection model draws on continuous observations from the Fengyun‑4 satellite over the preceding two hours to project the next phase of cloud development and identify areas at risk of intense convection. According to the administration, this extends warning lead time to four hours, giving emergency managers and the public more time to prepare for hazardous weather.

Fengyun satellites form a key pillar of China’s integrated meteorological observing network. With artificial intelligence increasingly vital for extracting signal from vast satellite datasets and for predicting short‑term weather evolution, officials said the new models mark a step forward in operational readiness. China currently has 10 Fengyun meteorological satellites in orbit, and data access is available to 141 countries and regions, with further applications of AI expected as capabilities mature.