NVIDIA AI Podcast NVIDIA’s Mike Pritchard Shares How Applying AI to Climate Simulation is Helping Forecast the Future - Ep. 266
45 snips
Jul 23, 2025 Mike Pritchard, Director of Climate Simulation Research at NVIDIA, shares insights on how AI is transforming climate science. He discusses developing high-resolution simulations that improve predictions of extreme weather and long-term climate patterns. AI enables better data analysis, revealing hidden patterns in historical climate records. The conversation emphasizes the need for advanced computational resources and how innovative modeling techniques can inform policy decisions, helping communities prepare for climate challenges.
AI Snips
Chapters
Transcript
Episode notes
Climate Modeling Challenges
- Climate prediction models focus on averaging weather statistics over time, unlike daily weather forecasts.\n- Simulating future climate requires enormous computational power due to complexity and the multiple scenarios needed.
AI Revolutionizes Climate Simulation
- AI helps accelerate climate simulators by approximating complex physics like clouds through learned neural networks.\n- This approach now enables full AI atmospheric simulations with interactive, high-resolution, and counterfactual weather forecasts.
Understanding Climate Model Resolution
- Resolution in climate models refers to the spatial detail of the simulation mesh, often at county or state-level scales.\n- Important atmospheric phenomena, like clouds and storms, are usually unresolved explicitly due to computational limits.