Recently, the world's first Artificial intelligence-driven Global Aerosol-Meteorology Forecasting System (AI-GAMFS), developed by a team led by CHE Huizheng, researcher from the Chinese Academy of Meteorological Sciences (CAMS), has been available on the International Meteorological Early Warning Operational Support Platform (hereinafter referred to as “the Early Warning Platform”) at the World Meteorological Centre Beijing (WMC-BJ).

As human activities and the impacts of climate change intensify, issues such as dust storms and wildfires—which contribute to atmospheric aerosol pollution, have become increasingly severe, heightening the urgent need for precise forecasts. The research team overcame key technical challenges in aerosol-meteorological spatiotemporal coupling intelligent forecasting, developed an AI large-scale model encompassing 54 forecast variables based on a 42-year global advanced aerosol reanalysis dataset, enabling more precise characterization of the complex interactions between aerosols and meteorological phenomena. Previously, AI-GAMFS has been implemented for operational work of National Meteorological Centre (NMC) of China Meteorological Administration (CMA) and meteorological services in 12 provinces (autonomous regions) across China, including Xinjiang, Ningxia, Inner Mongolia, Gansu, and Shaanxi.

The main advantage of AI-GAMFS lies in its exceptionally high computational efficiency, enabling global-scale forecasts of optical properties, ground concentrations, and related meteorological elements for key aerosol components such as dust, sulfates, black carbon, organic carbon, and sea salt over a 5-day period with 3-hour intervals, at a spatial resolution of 50 kilometers, all within just 36 seconds per run. After integrating with the Early Warning Platform, it is expected to significantly enhance the operational capabilities and timeliness of global aerosol pollution forecasting.

Editor: JIANG Zhiqing