AI for Climate Modeling: Present and Future
AI-driven weather forecast models are now more accurate and much faster than the best physics-based models. The open-source Ai2 Climate Emulator (ACE) uses similar technology to accurately emulate both daily weather variability (including extremes) and climate of historical reanalysis or of a reference global atmospheric model. ACE runs 100x faster than a physics-based model of similar grid resolution – 1600 years per day on a 100 km grid with a single GPU. ACE can be forced by specified sea-surface temperatures or coupled to a slab ocean model (SOM). When trained on SOM-coupled reference model simulations spanning multiple climates forced by changed CO2 concentrations, ACE can accurately emulate the reference model climate change response. We have coupled ACE to an AI-based ocean emulator called Samudra; when trained on a pre-industrial control simulation of a reference physically-based GCM simulation, the resulting emulator reproduces its mean climate and ENSO characteristics. Lastly, we discuss key remaining challenges to general-purpose use of ACE for climate modeling applications.
Agenda
3:00pm – 3:45pm: Presentation
3:45pm – 4:00pm: Q&A
4:00pm – 4:30pm: Reception
