Scientists from the Chinese Academy of Meteorological Sciences (CAMS) have developed the world’s first Artificial Intelligence-driven Global Aerosol-Meteorology Forecasting System (AI-GAMFS), a major technological breakthrough that could transform global air-quality prediction and climate monitoring. The research was published on 5 March in the journal Nature.
The system, developed under the China Meteorological Administration (CMA) in collaboration with international research institutions, combines advanced artificial intelligence with decades of atmospheric data to deliver rapid, high-precision global forecasts of aerosol conditions.
Researchers say the innovation could significantly improve early warning systems for dust storms, air pollution events and climate-related environmental risks worldwide.
AI Forecasts Global Aerosols in Under One Minute
According to GUI Ke, associate researcher at CAMS and a core member of the project team, development of the system began in May 2024.
The model was trained using 42 years of global aerosol reanalysis data, covering 120,000 time instances, enabling the system to generate five-day global forecasts in less than one minute.
AI-GAMFS produces forecasts with:
-
Three-hour temporal resolution
-
50-kilometre spatial resolution
-
Eight rolling forecasts per day
This represents a major leap forward in forecasting speed compared with traditional atmospheric modelling systems that often require hours of supercomputer processing.
Monitoring Five Major Aerosol Components
The forecasting system analyzes 54 atmospheric variables, providing detailed information about aerosol distribution and meteorological interactions.
Key aerosol components monitored include:
-
Dust
-
Sulfate
-
Black carbon
-
Organic carbon
-
Sea salt
Outputs include aerosol optical properties, surface concentrations and related meteorological factors, offering a comprehensive picture of atmospheric pollution and its environmental impacts.
Already Deployed Across China
According to CHE Huizheng, another leading scientist on the project, the system has already moved from research to real-world application.
AI-GAMFS is currently operational at the National Meteorological Centre (NMC) of CMA and meteorological departments in more than ten provinces, including Shaanxi and Ningxia.
The platform has been used to support:
-
Sand and dust storm forecasting analysis
-
Environmental meteorology consultations
-
Regional air-quality monitoring
Some regional meteorological agencies, including those in Gansu Province, have further enhanced the system by applying downscaling optimization, improving local forecast resolution to five kilometres.
Global Forecast Services Now Available
The system is also being deployed internationally through the World Meteorological Centre Beijing (WMC-BJ) via the International Meteorological Early Warning Operational Support Platform.
Through this platform, AI-GAMFS is providing global aerosol forecasting services, helping meteorological agencies monitor atmospheric pollution and environmental hazards worldwide.
Open-Source Access for Developing Countries
In line with international open-science standards, the forecasting system has been fully released under open-source requirements.
This means it can be deployed on a single server, offering a low-cost, high-precision forecasting solution for developing countries that may lack advanced meteorological infrastructure.
Experts say this could significantly improve air-quality monitoring and early warning systems in regions vulnerable to dust storms, pollution and climate-related environmental changes.
Future Applications in Climate and Environmental Protection
Looking ahead, the research team plans to continue upgrading the system and developing regional AI-driven environmental meteorology models.
Future applications are expected to include:
Researchers say the system demonstrates how artificial intelligence can revolutionize atmospheric science, providing faster and more accurate tools to address global environmental challenges.