BEGIN:VCALENDAR
VERSION:2.0
BEGIN:VEVENT
DTSTART:20251223T113000
DTEND:20251223T123000
SUMMARY:Climate modeling in the era of AI
DESCRIPTION:<p>While AI has been disrupting conventional weather</p><p>forecasting, we are only beginning to witness the</p><p>impact of AI on long-term climate simulations. The</p><p>fidelity and reliability of climate models have been</p><p>limited by computing capabilities. These limitations</p><p>lead to inaccurate representations of key processes</p><p>such as convection, cloud, or mixing or restrict the</p><p>ensemble size of climate predictions. Therefore, these</p><p>issues are a significant hurdle in enhancing climate</p><p>simulations and their predictions.</p><p>Here, I will discuss a new generation of climate</p><p>models with AI representations of unresolved ocean</p><p>physics, learned from high-fidelity simulations, and</p><p>their impact on reducing biases in climate</p><p>simulations. The simulations are performed with</p><p>operational ocean model components. I will further</p><p>demonstrate the potential of AI to accelerate climate</p><p>predictions and increase their reliability through the</p><p>generation of fully AI-driven emulators, which can</p><p>reproduce decades of climate model output in seconds</p><p>with high accuracy</p>
LOCATION:
TRANSP:OPAQUE
URL:https://www.weizmann.ac.il/EPS/events
END:VEVENT
END:VCALENDAR