China has introduced a powerful new artificial intelligence weather model, “Fengyuan,” aimed at improving the prediction of extreme weather events at home and abroad. Developed by the China Meteorological Administration (CMA), this system sits at the center of a broader upgrade to the country’s AI-based forecasting tools, with implications for disaster preparedness, energy security, transportation, and international climate services.
Fengyuan: China’s New AI Engine for Extreme Weather Forecasting
Fengyuan is an AI model designed to enhance the early detection and forecasting of high-impact weather, from severe storms and extreme rainfall to heatwaves and other climate-related hazards. Released in Xiongan New Area, Hebei Province, it marks a strategic step in China’s long-term plan to modernize meteorological services using advanced data science.
An Open-Source, End-to-End Weather Intelligence Platform
The CMA has positioned Fengyuan as an open-source platform, inviting contributions from researchers, institutions, and developers. This approach aims to accelerate innovation in AI-driven meteorology while enabling the broader scientific community to test, validate, and improve the model.
Unlike many traditional forecasting systems, Fengyuan is described as an end-to-end solution, meaning it can ingest raw observational data and directly produce weather forecasts. According to its development team, the long-term vision is an observation-driven, multi-sphere forecasting system that integrates:
This multi-sphere integration is essential for capturing the complex physical mechanisms behind extreme events in a warming climate.
Technological Self-Reliance and Intellectual Property
CMA officials emphasize that Fengyuan is independently developed with full proprietary intellectual property rights. This reflects a broader national push for technological self-reliance in critical scientific infrastructure, reducing dependence on foreign models and data systems.
By controlling the entire stack—from model architecture to training, deployment, and data assimilation—China can tailor Fengyuan to regional characteristics, rapidly update it as new observations arrive, and align it with national priorities in disaster risk reduction and climate resilience.
Sector-Wide Benefits: From Low-Altitude Aviation to Public Health
Extreme weather is no longer a niche concern; it cuts across nearly every sector of the economy. Fengyuan is intended to support a wide range of applications where accurate, timely forecasts are crucial.
Operational Uses Across Key Industries
According to CMA officials, Fengyuan’s improved forecasting capabilities are expected to benefit:
By strengthening early detection of high-impact events, Fengyuan can help authorities shift from reactive emergency response to proactive risk management.
Support for Belt and Road Countries
Beyond China’s borders, the CMA plans to use Fengyuan to provide meteorological services to countries participating in the Belt and Road Initiative. This includes:
Such international applications underscore how AI-based weather systems are becoming tools of both scientific cooperation and global climate resilience.
Upgraded AI Forecasting Suite: Fengqing, Fenglei, and Fengshun
Fengyuan’s launch comes alongside significant upgrades to three existing CMA AI models—Fengqing, Fenglei, and Fengshun—covering different forecasting timescales and applications.
Fengqing: Enhanced Global Medium-Range Forecasting
Fengqing is the CMA’s global medium-range forecasting system, targeting weather predictions over several days to a couple of weeks. It has been upgraded to incorporate 11 additional physical parameters, such as:
By enriching the model with more physical detail, Fengqing is better equipped to resolve complex weather patterns and improve the reliability of medium-range outlooks.
Fenglei: From Radar Echoes to Direct Rainfall Forecasts
Fenglei operates on the nowcasting timescale—minutes to a few hours ahead—where rapid, high-resolution updates are essential. It has advanced from simply predicting radar echoes to performing direct quantitative precipitation forecasting.
This shift means Fenglei can now provide more accurate estimates of how much rain will fall, where, and when, dramatically improving early warnings for:
Fengshun: Better Subseasonal-to-Seasonal Outlooks
The subseasonal-to-seasonal system Fengshun focuses on forecasts from weeks to months ahead. Recent upgrades have:
These improvements strengthen support for:
Toward a Next-Generation, AI-Driven Weather Ecosystem
Collectively, Fengyuan and the upgraded Fengqing, Fenglei, and Fengshun form an integrated AI ecosystem spanning timescales from minutes to months.
This system covers local storms to global circulation patterns.
By combining AI methods with physical mechanisms, the CMA is positioning its forecasting infrastructure for a climate-challenged future.
Expanding international collaboration is also a key part of this effort.
The performance of these evolving systems in real-world extreme events will be closely watched by the global meteorological community.
They serve as scientific testbeds and as models for how AI can be responsibly embedded into operational weather and climate services.
Here is the source article for this story: China boosts AI forecasting for extreme weather events