Nature Communications: Tsinghua University's Li Yong team uses AI to decode global climate coupling, extending ENSO prediction lead time to 19 months
Researchers from Tsinghua University, led by Professor Li Yong, have developed a unified AI model called UniCM to understand the coupled dynamics of global climate modes. This model simultaneously learns how physical fields like sea surface temperature influence climate modes such as ENSO, IOD, and TNA, and how these modes, in turn, affect future climate patterns. UniCM demonstrates superior predictive capabilities over baseline methods, notably extending the effective prediction lead time for ENSO to approximately 19 months and improving the forecasting of non-ENSO modes. AI
IMPACT Enhances early warning systems for extreme weather events by improving the prediction of coupled climate dynamics across multiple ocean basins.