-
- [ONLINE] GC42A-02: AI-based climate model evaluation through the lens of pseudo-global warming storylines
-
OnlineOnline
Author(s):Generic 'disconnected' Message
Shiheng Duan, Lawrence Livermore National Laboratory (First Author, Presenting Author)
Jishi Zhang, Lawrence Livermore National Laboratory
Celine Bonfils, Lawrence Livermore National Laboratory
Giuliana Pallotta, Lawrence Livermore National Laboratory
Paul Ullrich, Lawrence Livermore National Laboratory
Artificial intelligence is transforming weather and climate modeling by providing fast and accurate predictions at much lower cost than traditional models. These advances open the door for new ways to study how extreme events, like heatwaves and powerful storms, might change in a warming world. In this work, we test AI-based models using a method called “pseudo-global warming,” which applies future climate conditions to historical extreme events. We focus on two cases: a major heatwave in the Pacific Northwest and a strong atmospheric river in California. Our results show that AI models can realistically reproduce these events and provide future projections in a fraction of the time required by traditional models. However, some differences remain in the details of future changes, highlighting the need for further improvements. This research demonstrates the potential of AI to support climate science and help understand how extreme events may evolve under climate change.
Scientific DisciplineSuggested ItinerariesNeighborhoodTypeWhere to Watch
Enter Note
Go to previous page in this tab
Session
