AI-Driven Second Revolution in Weather Forecasting: Monsoons, Gray Swans, and Butterflies

About

As part of our Schiller Catalyst Seminar Series, Pedram Hassanzadeh will give the presentation: "AI-driven Second Revolution in Weather Forecasting: Monsoons, Gray Swans, and Butterflies."

Artificial intelligence (AI) is driving a second revolution in weather forecasting. AI weather models now match or beat the best physics-based models for short- and medium-range forecasts, at a tiny fraction of the computing cost. There is growing interest in operationalizing these models, including expanding access to actionable, high-quality forecasts in regions that have long been underserved. To increase trust in these systems, there is also interest in understanding why they work so well and when and how they fail. In this talk, Dr. Pedram Hassanzadeh will present his team's work on forecasting the onset of the Indian monsoon, a complex and consequential weather event for millions of farmers. He will describe the development of a decision-oriented benchmarking and real-time probabilistic forecasting system. In 2025, as part of a Government of India project, these AI-based forecasts of the monsoon onset were sent to 38 million farmers, providing a timely warning of an upcoming dry spell. In 2026, this work was extended to forecasting the rainy season of Ethiopia, bringing new challenges and opportunities. Finally, Dr. Hassanzadeh will discuss two failure modes of AI weather models: their inability to forecast unprecedented weather extremes (the so-called "gray swans") and the missing "butterfly effect," a fundamental property of the chaotic atmosphere. He will present results on how to mitigate these shortcomings and close with what these lessons mean for creating AI forecasts that are both accurate and trustworthy.

Pedram Hassanzadeh is an Associate Professor in the University of Chicago’s Department of Geophysical Sciences, Committee on Computational and Applied Math, and Data Science Institute. His research sits at the intersection of climate dynamics, scientific machine learning, and computational and applied math. He is the faculty director of the AI for Climate Initiative (AICE) and co-director of the Human-centered Weather Forecasting Initiative. Dr. Hassanzadeh received his MA and PhD from UC Berkeley and was a Ziff Environmental Fellow at Harvard University before joining Rice University in 2016 and the University of Chicago in 2024.

The lecture is open to the public and will take place from 12:00 PM – 1:00 PM on Monday, November 9, in the Schiller Institute Convening Space (Room 501 in Hayden Hall). Lunch will be provided.