China shares multi-hazard early warning system globally

The China Meteorological Administration (CMA) launched an AI-powered integrated meteorological system to provide early warnings for all, or "MAZU", at the opening ceremony of the 2025 World Artificial Intelligence Conference in Shanghai on Saturday. The aim of the system is to address global climate challenges and share China's expertise and technological achievements with the world, especially developing countries.

In the presence of Celeste Saulo, secretary-general of the World Meteorological Organization, head of the China Meteorological Administration Chen Zhenlin donated "MAZU-Urban", a multi-hazard early warning intelligent system for urban applications, to representatives from Djibouti and Mongolia during the ceremony. This will enable the system, which integrates advanced algorithms and multi-source data to enhance early warning practices and disaster mitigation efforts globally, to be used internationally for the first time.

"Ensuring universal access to meteorological early warning systems is not only a shared vision of the global community, but also an important mission of China's meteorological departments," said Chen.

MAZU's mission includes providing early warning technical support, enhancing capacity building, strengthening risk identification and assessment systems, and developing cooperation mechanisms and models, the CMA said.

Named after the ancient Chinese goddess of the sea, MAZU embodies a spirit of protection and preparedness, according to the CMA, with the acronym standing for multi-hazard, alert, zero-gap and universal.

"MAZU-Urban" is the first globally shared product developed and promoted by the Shanghai Meteorological Service in collaboration with other institutions, including the National Meteorological Center and the Shanghai Academy of AI for Science.

Core technologies of the intelligent system include flexible multi-hazard monitoring tools and forecasting analytical applications in monitoring and early warning. The smart system can also generate disaster bulletins during the warning release phase automatically, and use AI-empowered large language models to generate role-based, disaster-specific defense guidelines and emergency plans, while supporting Q&A with users to enhance response efficiency.

The system also integrates a three-tiered structure, catering to meteorological and emergency management departments, industry-specific users and the general public. It offers real-time disaster monitoring, personalized risk assessments and localized emergency response guidance.

The intelligent system has been used on a trial basis in 35 countries and regions across Asia, Africa and Oceania since January, receiving widespread acclaim, the CMA said.

In recent years, the CMA has jointly developed cloud-based early warning systems with the meteorological departments of Pakistan, Ethiopia and the Solomon Islands, among others.

"The Ethiopian Meteorological Institute and the CMA have carried out fruitful cooperation," said Fetene Teshome, director of the Ethiopian Meteorological Institute.

"Through the joint development of early warning systems, we have enhanced the capabilities in disaster prevention and mitigation, which served economic and social development."

Through international training courses, scholarship programs and visiting scholar programs, the CMA has also collaborated with its counterparts in other countries to facilitate cross-border experience sharing and technological innovation, and support developing countries in cultivating local professionals. The CMA has also been sharing China's practices in disaster risk survey and assessment, and supporting other countries in establishing a scientific basis for risk decision-making.

"I hope we can continue to deepen cooperation in supporting such initiatives to accelerate global actions for early warnings for all," said David Hiba, director-general of the Meteorological Service of the Solomon Islands.