February 20
|
3:30 pm
–
4:30 pm
Speaker – Greg Hakim, Professor, University of Washington. This semester’s special Braham Seminar.
Seminar Title – Using Machine-Learning Weather Models to Study Predictability and Extreme Events
Abstract – Recently developed machine-learning (ML) weather models have been widely recognized for revolutionizing weather prediction, producing forecasts more skillful than traditional models at a fraction of the computational cost. Here I will argue that the next phase of the revolution involves the adjoints of these models, applied to a wide range of problems, including novel exploration of dynamical process in weather and climate variability. Adjoints, which derive from gradient operations on a model, are useful for measuring the sensitivity of model outputs to inputs and parameters. The ubiquitous availability of adjoints for ML models makes these tools easily accessible and available for a wide range of applications. Specific examples I will discuss include shadowing trajectories for testing the limit of predictability and exploring gray swan extreme events.
About the Speaker – Greg Hakim is well known as a dynamic and synoptic meteorologist, and is a co-author on the most recent edition of the classic textbook “An Introduction to Dynamic Meteorology” with Jim Holton. His most recent contributions have focused on novel uses of AI/ML for weather and climate prediction, so we anticipate a presentation of very broad appeal and applicability.