China has recently unveiled a new plan for meteorological disaster prevention covering the 2026-2030 period, calling for stronger capabilities in monitoring and forecasting extreme weather, as well as greater use of intelligent forecasting and digital technologies.

As extreme weather becomes more frequent and intense, artificial intelligence (AI) is already helping turn those goals into practical improvements in forecasting and early warning.

The advances are supported by China's expanding meteorological observation network. The country has built what it describes as the world's largest integrated space-air-ground meteorological observation network, combining Fengyun meteorological satellites, new-generation weather radars and automatic weather stations.

The network supplies AI models with massive volumes of real-time data, enabling faster and more accurate weather forecasts. It supports forecasts ranging from seasonal outlooks to minute-level nowcasting.

AI is already transforming severe weather forecasting in east China's Jiangsu Province. An AI nowcasting model developed by Jiangsu meteorological authorities and research institutions provides forecasts up to 12 hours ahead. During a severe convective weather event in September 2025, it detected signals indicating the development of organized severe convective weather about six hours in advance. The event later produced an EF0 tornado and force-12 thunderstorm winds.

Provincial meteorological authorities say the average lead time for severe convective weather warnings has increased to 53 minutes, while forecast accuracy has improved by 22% over conventional methods.

Another AI-powered warning system further streamlines the warning process by automatically identifying severe weather, assessing risks, generating warning reports and distributing alerts. Tasks that once took around 10 minutes can now be completed in about one minute. During nine major severe convective weather events in Jiangsu in 2025, the system successfully identified storm initiation and organization trends in five cases, providing an average lead time of 5.4 hours.

AI-powered weather services are also being applied in transportation and energy, from detecting highway fog to helping wind farms optimize power generation and reduce operational risks.

China is also extending these technologies through international cooperation. An upgraded version of MAZU, China's early warning solution supporting the United Nations' Early Warnings for All initiative, was released in April integrating AI across the platform.

It offers tailored services covering weather monitoring, early warning and impact-based forecasting. The solution has already been deployed in several countries and provides cloud-based services to users in more than 40 countries.

As extreme weather poses growing challenges, faster and more accurate forecasts are becoming an increasingly important part of disaster preparedness.