WeatherBrains
WeatherBrains
Big Brains Media LLC
WeatherBrains 1022: Tighter Than A Tick's Butt
1 hour 38 minutes Posted Aug 19, 2025 at 3:58 pm.
Tonight's Guest WeatherBrain is a professor of mechanical engineering and applied mathematics at the University of Pennsylvania. He previously served as a principal scientist at Microsoft Research, where he led the development of Aurora. This was the groundbreaking AI foundation model for earth system forecasting. Dr. Paris Perdikaris, welcome to WeatherBrains! Our email officer Jen is continuing to handle the incoming messages from our listeners. Reach us here: email@weatherbrains.com. What is Aurora and the motivation for the program?
Practical experience and forecasting with tropical forecasting with Aurora
Success stories with Aurora
Convection-allowing models (CAMS)
Convective feedback problems with models
Emergence of the Google DeepMind ensemble and other AI models
Aurora's hardware architecture and its importance
Continuing need of physics-based models in the age of AI
Ethical considerations and biases in model training data
Role of US agencies and potential loss of funding
Communicating AI forecasts to the public and the ensuing ethical issues
Broad risks of using AI
Aurora business model
The Astronomy Outlook with Tony Rice
This Week in Tornado History With Jen
E-Mail Segment
0:00
1:38:40
Download MP3
Show notes
Tonight's Guest WeatherBrain is a professor of mechanical engineering and applied mathematics at the University of Pennsylvania. He previously served as a principal scientist at Microsoft Research, where he led the development of Aurora. This was the groundbreaking AI foundation model for earth system forecasting. Dr. Paris Perdikaris, welcome to WeatherBrains! Our email officer Jen is continuing to handle the incoming messages from our listeners. Reach us here: email@weatherbrains.com. What is Aurora and the motivation for the program? (08:15) Practical experience and forecasting with tropical forecasting with Aurora (12:30) Success stories with Aurora (14:00) Convection-allowing models (CAMS) (17:15) Convective feedback problems with models (18:45) Emergence of the Google DeepMind ensemble and other AI models (20:00) Aurora's hardware architecture and its importance (23:30) Continuing need of physics-based models in the age of AI (28:00) Ethical considerations and biases in model training data (33:30) Role of US agencies and potential loss of funding (37:55) Communicating AI forecasts to the public and the ensuing ethical issues (42:30) Broad risks of using AI (44:30) Aurora business model (58:45) The Astronomy Outlook with Tony Rice (01:05:25) This Week in Tornado History With Jen (01:07:25) E-Mail Segment (01:09:30) and more! Web Sites from Episode 1022: Penn's Predictive Intelligence Lab Alabama Weather Network on Facebook Picks of the Week: James Aydelott - Dr. Cameron Nixon on YouTube: Cell Mergers and Nudgers Jen Narramore - "Significant Tornadoes 1680-1991" by Thomas Grazulis Rick Smith - USA Jobs Troy Kimmel - Foghorn Kim Klockow-McClain - Out John Gordon - Live Recon in the Atlantic Basin Bill Murray - Foghorn James Spann - Weather Lab: Cyclones The WeatherBrains crew includes your host, James Spann, plus other notable geeks like Troy Kimmel, Bill Murray, Rick Smith, James Aydelott, Jen Narramore, John Gordon, and Dr. Kim Klockow-McClain. They bring together a wealth of weather knowledge and experience for another fascinating podcast about weather.