As extreme weather becomes a global challenge, China has  shared its early warning practice under the framework of MAZU, a joint action for the United Nations' Early Warnings for All, launched by the China Meteorological Administration (CMA).

Along the coastal areas of China, Shanghai's smart meteorology, Zhejiang's coordinated “mountain-to-sea” response, and Fujian's progressive meteorological service mechanism have all painted unique strokes on the canvas of disaster prevention and mitigation. This initiative not only safeguards people's livelihoods but also radiates the strength of Chinese wisdom to the world.

Fujian: from Mazu to progressive meteorological service mechanism

In Meizhou Island, Fujian, the Mazu (Lin Mo) culture has endured for thousands of years, which is an Intangible Cultural Heritage by United Nations Educational，Scientific and Cultural Organization (UNESCO).

In July 2025, as Typhoon "Danas" approached, Fujian activated its "1262" progressive meteorological service mechanism: designating key prevention areas 12 hours in advance, pre-positioning rescue forces 6 hours ahead, and pinpointing high-risk townships 2 hours prior to potential impact.

"Early meteorological warning serves as the 'vanguard' for urban drainage and flood prevention, as well as the 'forerunner' for water system regulation. It drives the transformation of flood control and water management from passive defense to proactive preparedness, and from blanket protection across the entire area to precise, targeted response," said XUE Yubin, Deputy Director of the Fuzhou Urban Water System Joint Operation and Scheduling Center.

And at the abalone farming base in Pinghai Town, Putian, meteorological warnings guide seedling nurseries in temperature control and water regulation, mitigating risks of abrupt water quality changes caused by heavy rainfall.

"The value of early meteorological warnings for the aquaculture industry can be summarized as precise forecasting, proactive prevention and control, and disaster reduction with enhanced efficiency," said YANG Jianbin, Agronomist from the Xiuyu District Aquatic Technology Extension Station in Putian.

The data speaks volumes about the effectiveness. In the last five years, Fujian's meteorological departments have issued more than 1,300 warnings concerning ocean waves, storm surges, and other hazards to 237 marine fishery entities across 30 cities and counties, while delivering over 27 million safety alerts to 33,000 fishing vessel groups.

From weather stations to forecasts for the "blue tears" bioluminescent phenomenon, Fujian has integrated meteorological services into the entire chain of marine fisheries, port shipping, tourism, etc.

Zhejiang: harnessing technology to form more effective“mountain-to-sea” response

In July this year, after intensifying offshore, Typhoon Co-way lingered, subjecting Yuyao in Ningbo to severe precipitation tests. Within just a few hours, over a thousand residents in high-risk areas of key villages were successfully evacuated, safe and sound.

"Each mountain torrent ditch has its own 'dedicated file'," said JIN Xufeng, a forecaster from Yuyao Meteorological Department, pointing at the platform. "This platform visually displays the terrain, real-time monitoring data, and water level forecasts for each ditch, triggering conspicuous alarms once thresholds are reached."

Through on-site terrain surveys and analysis of the correlation between rainstorms and mountain torrents, Yuyao has set disaster-causing thresholds for each ditch. Combined with three-dimensional (3D) monitoring from X-band radar and millimeter-wave cloud radar, it achieves "visible risks and early warnings."

Port management in Zhejiang is another case in point. Ningbo-Zhoushan Port, the world's largest, used to halt operations for 800 to 1,000 hours annually due to sea fog and gale.

To address this situation, Ningbo Meteorological Department has independently developed a "Programmatic Early Warning and Pre-control System" by integrating real-time data from multiple sectors, including meteorology, maritime affairs, ports, and hydrology. This system not only enables real-time access to multi-departmental data on a single platform but also delivers minute-scale real-time forecasts and early warning information via mobile devices. Furthermore, it features dedicated modules for different disaster weather conditions, addressing traditional weaknesses in port disaster defense.

Nowadays, the System achieves an 86% accuracy rate in predicting gale and nearly 70% for sea fog forecasts. In the first half of 2025, the System provided an additional 200 hours of berthing windows for the port, generating direct economic benefits of 160 million yuan.

Besides, the municipal meteorological department has significantly enhanced its forecasting and early warning capabilities for sea fog alerts, upgrading from "single-point monitoring" to "sector-wide scanning" and achieving regional-specific warnings.

Shanghai: A "two-way sprint" between health and meteorological technology

Currently, the Fengxian District Meteorological Office in Shanghai, in collaboration with the district's Center for Disease Control and Prevention (Health Supervision Institute) and community health service center, is leveraging the "Health + Meteorology" project to transform "early meteorological warnings" into warm health advisories, shifting the focus from treatment to prevention.

The district meteorological office prepares health forecast information, which is then reviewed by the disease control department and pushed to the mobile phones of community family doctors. Family doctors harness the "Fengxian Intelligent Follow-up Platform" to directly reach out to contracted residents. This relay process of "meteorology-disease control-community" has been integrated into Fengxian's "unified management network" super brain, delivering tailored "meteorological health reminders" to key populations 1-2 days in advance.

In October this year, the Fengxian District Meteorological Bureau, Fengxian District Center for Disease Control and Prevention (Health Supervision Institute), and Shanghai Key Laboratory of Meteorology and Health jointly launched the "Health + Meteorology" pilot project. The initial focus is on over 2,000 cardiovascular and cerebrovascular disease patients who have signed up with family doctors at two community health service centers in Nanqiao and Tairi. By precisely delivering health protection information regarding adverse weather conditions, the project aims to guide and assist chronic disease patients in self-protection, thereby reducing the risk of illness and medical expenses.

"When the minimum temperature drops by more than 6℃ within 48 hours during autumn and winter, we will promptly activate the 'health warning' mechanism," introduced YIN Liyang, Chief of the Disaster Prevention and Mitigation Division at the Fengxian District Meteorological Office.

In the future, relying on the "Health + Meteorology" project, the Office will collaborate with the health commission to explore expanding the scope of health meteorology services based on early warnings, injecting strong momentum into the construction of a "Healthy Shanghai."

Chinese Solutions with a Global Perspective

In 2025, the average lead time for 13 local severe convective weather forecasts reached 4 hours and 20 minutes, while Typhoon "Ragasa" over the Northwest Pacific was successfully predicted 3-4 days in advance to make landfall in Yangjiang, Guangdong. Behind the dual improvements in severe convective weather prediction accuracy and typhoon forecast lead time lies robust support from the Shanghai Meteorological Service's deep integration of artificial intelligence (AI) technologies with meteorological forecasting.

"Severe convective weather develops rapidly, posing key challenges for forecasting. Our team incorporated AI to develop the 'Yushi' and 'Fuyao' AI-based forecasting models for severe convective weather," explained CAO Yuan, Head of the AI Weather Forecasting Team at the Shanghai AI Meteorological Application Innovation Centre. Taking “Fuyao” as an example, it utilizes the AI-driven Nowcastformer architecture, integrating multi-source data from weather radars, meteorological satellites, numerical models, and wind lidars to build an auto-regressive forecasting system. This enables hourly street-level precipitation and wind forecasts, increasing update frequency from hourly to every 10 minutes and refining warning accuracy to the sub-district level.

During the opening ceremony of the World Artificial Intelligence Conference (WAIC) on July 26,CMA launched MAZU, a Joint Action for Early Warnings for All, to build an Early Warning Service Network and to share practical experience and technological achievements with global partners.

"MAZU-Urban incorporates Shanghai's independently developed AI-based meteorological model, and has developed customized regional forecasting and early warning generation-and-distribution capabilities based on a global forecasting framework, forming a full-chain early warning system that covers 'risk knowledge - monitoring and warning - warning dissemination - emergency response,'" introduced ZHAO Yang, Director of the Shanghai Meteorological Information and Technical Support Center and the Urban Multi-Hazard Early Warning Centre.

In addition, to tackle the complexity of global application scenarios, MAZU-Urban innovatively adopts a "three-in-one" architecture integrating all-in-one machines, tablets, and smartphones, serving specialized departments, key industries, and the general public. It leverages AI models for intelligent Q&A and delivers location-based disaster alerts and evacuation guidance.

"Drawing on the concept of a toolkit, MAZU-Urban develops tailored scenario-specific 'tools' based on the unique conditions and practical needs of different countries, gradually cultivating intelligent early warning capabilities with 'country-specific and city-specific strategies,'" explained LIU Haobo, a senior engineer at the Shanghai Meteorological Information and Technical Support Centre & Urban Multi-Hazard Early Warning Centre.

For instance, to meet Mongolia's specific needs, the R&D team of MAZU-Urban developed role-based and disaster-specific workflow prompts for disaster prevention and mitigation guidelines, guiding large language models in deep reasoning to dynamically generate localized defense strategies.

Since January in 2025, MAZU-Urban has been in trial operation in 35 countries and territories across Asia, Africa, and Oceania, receiving widespread acclaim. And on the occasion of 2025 SDG Cities Global Conference, the Shanghai Meteorological Service unveiled the co-construction action plan of MAZU-Urban.

Editor: LIU Shuqiao