China unveils AI models to assist Fengyun meteorological satellites to detect extreme weather more effectively

China unveiled a new generation of AI models for its Fengyun meteorological satellite system on Thursday, aimed at improving warnings for typhoons and extreme weather, in a push to move satellite use from observation toward short-term forecasting.

According to industry experts, the models are expected to significantly enhance the country's ability to monitor and provide early warnings for extreme weather such as typhoons and severe convective storms.

The first batch of Fengyun extreme-weather AI models was launched at the 2026 FENGYUN Satellite User Conference (FYSUC) held in Changsha, Central China's Hunan Province on Thursday.

The models mark a new phase in which the satellites are used not only to monitor weather systems, but also to support trend forecasting and disaster prevention, according to the organizers.

The new tools include a typhoon cloud-image interpretation model and a severe-convection nowcasting and warning model. Both combine Fengyun satellite data with AI algorithms to boost monitoring and early-warning capabilities for extreme weather.

Frequent extreme weather events have increased demand for faster and more accurate alerts. Fengyun satellites, a core part of China’s meteorological observation network, provide continuous tracking of weather systems as they develop.

According to the China Meteorological Administration (CMA), the typhoon model integrates multi-channel Fengyun imagery and related products to automatically analyze a storm’s location, structure, intensity and development trend, and generate interpretive text. It supports seven types of typhoon features, uses more than 20 real-time satellite products, runs every 15 minutes and can generate an analysis in 27 seconds.

The severe-convection model uses the previous two hours of continuous observations from Fengyun-4 satellites, combined with AI, to forecast future changes. It separates cloud evolution into the movement of large systems and the birth and decay of smaller ones, extending warning lead time to four hours, the CMA said.

At the event, the National Satellite Meteorological Center released data products from the Fengyun-3H and Fengyun-4C satellites. Officials said the products will expand Fengyun observation data and improve capabilities in greenhouse gas detection, numerical weather prediction, ecological monitoring and space weather forecasting.

The data are also expected to support disaster prevention, ecological efforts, civil aviation and aerospace safety, and global meteorological governance, while further expanding the MAZU-FengYun Satellite AI Box.

According to Cao Xiaozhong, deputy director of the CMA, the Fengyun meteorological satellites are a major space infrastructure asset and have reached internationally advanced levels after more than 50 years of development. He said they play a key role in extreme-weather response, as well as in food security, national defense, energy security and low-altitude economy applications.

Cao added that the Fengyun system is helping China expand its "MAZU" intelligent meteorological early warning system overseas, with services now reaching 141 countries and regions, chinanews.com.cn reported on Thursday.

The new release comes as meteorological authorities worldwide are increasingly turning to AI to improve the speed and precision of forecasts, especially for fast-changing hazardous weather events where minutes can make a difference.

Analysts said the extension of the Fengyun satellites' application capabilities toward forecasting future weather trends has offered a brand-new technological tool for disaster prevention and mitigation work.

China is one of the few countries that operate both geostationary and polar-orbiting satellites. It is providing satellite data and products to around 2,700 users from over 100 countries or regions, according to data released on FYSUC’s website.