China has launched an upgraded version of its MAZU early warning platform, a cloud-based meteorological system designed to help countries better anticipate extreme weather, and delivered customized versions to Jordan and Sri Lanka to strengthen disaster preparedness.

The launch took place on Wednesday during a parallel forum of the Third High-Level Conference of the Forum on Global Action for Shared Development, where officials and meteorological experts discussed ways to boost global cooperation on climate resilience and early warning systems.

Developed by the China Meteorological Administration, MAZU is part of China's effort to support the Early Warnings for All initiative promoted by the United Nations.

"The MAZU solution is a global public good created through long-term scientific accumulation and artificial intelligence empowerment," said Zhang Xingying, director of the Department of International Cooperation at CMA.

"Meteorological disaster prevention and climate response are key foundations for sustainable development," he said, adding that China hopes to make early warning systems more accessible to developing countries, especially in the Global South.

Dai Kan, deputy director of the National Meteorological Center, said the upgraded system now integrates multiple AI forecasting models, including the AI-based seasonal prediction system Fengshun and the medium-range forecasting system Fengqing, improving accuracy in predicting extreme rainfall and other severe weather events.

New overseas cloud nodes have boosted solution performance, increasing loading speeds sixfold and reducing data response time sevenfold, thus improving access stability for users in regions such as Africa, Dai said.

The system has also expanded from traditional weather forecasts to impact-based forecasting, incorporating hydrometeorological products such as global rainfall predictions and flood risk warnings for major river basins. New sectorspecific forecasts now support decision-making in agriculture, transportation and public health.

Andrea Celeste Saulo, secretary-general of the World Meteorological Organization, said initiatives like MAZU help provide early warning systems as a global public good because they are based on sharing knowledge, technology and innovation.

"China has the openness to recognize that the system cannot be replicated exactly as it works here, but should be adapted to different environments," she said.

During the forum, China delivered customized MAZU systems to meteorological authorities of Sri Lanka and Jordan.

Athula Kumara Karunanayake, director of Sri Lanka's Department of Meteorology, said the system will help the island nation strengthen disaster preparedness as it faces growing climate risks.

As a tropical island in the Indian Ocean, Sri Lanka frequently experiences extreme weather events including floods and droughts.

The MAZU system can provide forecasts ranging from one-day warnings for disaster response to monthly and seasonal predictions, which are valuable for agriculture, fisheries, aviation and water management, Karunanayake said.

By linking meteorological forecasts with disaster management agencies and other departments, the system could help deliver early warnings more effectively and reduce economic losses and casualties, he added.

Officials from Jordan said the customized solution will support the country's efforts to cope with climate-related risks such as sandstorms, heat waves and drought.

Raed Rafid, director of the Jordan Meteorological Department, said the MAZU system will help forecasters issue warnings to the public by integrating multiple models, satellite data and AI tools.

"The system has many models, including AI and satellite information, which help forecasters make better decisions," Rafid said.

Cooperation between Jordan and China has included training programs and technical workshops for forecasters and engineers. He added that the customized version could open the door for further collaboration.

Since its initial release in 2025, the MAZU solution has been deployed in seven countries and supports cloud-based use in more than 40 nations, providing meteorological data products to 153 countries and regions, according to CMA.

International training programs linked to the system have also benefited participants from 89 countries, it added.

zhaoyimeng@chinadaily.com.cn