On May 6, 2025, researchers from the Chinese Academy of Sciences and Tsinghua University introduced FengShun-CSM, an artificial intelligence-driven climate system model capable of providing 60-day global daily forecasts across 29 critical environmental variables. This development marks a significant advancement in climate prediction, offering enhanced accuracy over traditional models.

FengShun-CSM distinguishes itself by outperforming the European Centre for Medium-Range Weather Forecasts (ECMWF) subseasonal-to-seasonal (S2S) model, particularly in predicting precipitation, land surface, and oceanic components. The model's improved representation of intra-seasonal variability modes, notably the Madden-Julian Oscillation (MJO), contributes to its superior forecasting capabilities.

The collaborative effort between the Chinese Academy of Sciences and Tsinghua University underscores China's commitment to advancing climate science through interdisciplinary research. By integrating machine learning algorithms, FengShun-CSM processes vast datasets efficiently, enabling more accurate and timely predictions compared to traditional models that rely on complex physical equations and substantial computational resources.

The model's enhanced performance is largely attributed to its ability to accurately represent the MJO, a significant driver of tropical weather patterns. This capability allows FengShun-CSM to provide more precise forecasts of extreme weather events, which is crucial for various applications.

Potential applications of FengShun-CSM include:

• Meteorological Disaster Mitigation: Improved forecasting can enhance early warning systems for extreme weather events, facilitating better preparedness and response strategies.

• Marine Ecosystem Conservation: Accurate oceanic predictions can aid in monitoring and protecting marine biodiversity and managing fisheries more effectively.

• Agricultural Productivity Enhancement: Reliable climate forecasts can inform planting schedules, irrigation planning, and pest control measures, leading to increased agricultural yields and food security.

FengShun-CSM is part of a broader trend of integrating artificial intelligence into climate modeling. Other notable AI-based models include FuXi-Subseasonal, developed by scientists from the Shanghai Academy of Artificial Intelligence for Science, Fudan University, and China's National Climate Center, which extends forecasting capabilities to 42 days and has demonstrated superior accuracy compared to traditional models. Additionally, Fengwu, a machine learning model developed by the Shanghai Artificial Intelligence Laboratory, has advanced the effective forecast time of global mid-term weather to 11.25 days, surpassing some international equivalents.

The integration of AI into climate action has also prompted discussions on ethical principles and governance. In April 2022, the Institute for AI International Governance at Tsinghua University released the "Principles on AI for Climate Action," emphasizing values such as energy conservation, privacy protection, fairness, and promoting education. These principles aim to guide the responsible development and deployment of AI technologies in addressing climate change.

The introduction of FengShun-CSM marks a significant milestone in climate modeling, demonstrating the transformative potential of AI in enhancing forecasting capabilities. By outperforming traditional models and offering practical applications in disaster mitigation, ecosystem conservation, and agriculture, FengShun-CSM exemplifies the benefits of interdisciplinary collaboration and technological innovation in addressing global climate challenges.