Breakthrough in Weather Forecasting with FuXi-Subseasonal Model

Scientists from the Shanghai Academy of Artificial Intelligence for Science, Fudan University, and China’s National Climate Center have made a groundbreaking advancement in weather forecasting with the development of their new model, “FuXi-Subseasonal.” This innovative model utilizes artificial intelligence (AI) technology to extend the forecasting range to 42 days, representing a significant leap forward in AI climate modeling.

The unveiling of the FuXi-Subseasonal model took place at the China Pavilion during the recent Conference of the Parties to the United Nations Framework Convention on Climate Change (COP28) in Dubai, United Arab Emirates. This new model addresses the pressing need for improved sub-seasonal forecasts, as sub-seasonal anomalies greatly contribute to high-impact climate events.

The FuXi-Subseasonal model integrates the transformer architecture with guided random samples in a latent space, effectively accounting for underlying uncertainty in sub-seasonal forecasts. As a result, this model generates forecasts with superior accuracy compared to those of the European Centre for Medium-Range Weather Forecasts (ECMWF). Qi Yuan, director of the Shanghai Academy of Artificial Intelligence for Science and a professor of Fudan University, states that the new model has overcome challenges related to inaccurate initial conditions and inadequate external forcing signals at sub-seasonal timescales, addressing notable gaps in forecasts.

“This achievement addresses a long-standing technical challenge in climate change research. It offers the potential for more timely and accurate assessment of climate-related risks,” says Qi.

Additionally, the FuXi-Subseasonal model has significantly advanced its ability to predict the atmospheric phenomenon of Madden-Julian Oscillation (MJO), extending the prediction skill of MJO from 30 to 36 days. Accurate MJO prediction is crucial for agricultural planning, disaster preparedness, risk mitigation, and long-term climate research.

With its sub-seasonal forecasts of major weather processes such as intense heat, severe cooling, and heavy precipitation, the FuXi-Subseasonal model achieves an unprecedented level of precision that traditional methods cannot match, according to Qi.

As this model continues to evolve with its cutting-edge AI capabilities, it is set to make significant progress in addressing a broader spectrum of climate-related challenges. This progress will contribute to the empowerment of renewable energy development, the construction of new types of electric power systems, ensuring agricultural food security, and achieving sustainable socio-economic transformation.

Wu Libo, a professor at Fudan University, emphasizes the vital role that AI plays in climate change risk management. “With such high-end technology, we can better address the risks of climate change,” he affirms.

The development of the FuXi-Subseasonal model marks a significant milestone in weather forecasting and climate science. By leveraging the power of AI, scientists are paving the way for more accurate and timely assessment of climate-related risks, ultimately enhancing our ability to prepare for and mitigate the impact of climate events.