LANS Seminar

Seminar Title: Downscaling with Uncertainty using Data-Driven Surrogate Models

Speaker: Philip Dinenis, Postdoctoral Researcher, Mathematics and Computer Science, Argonne National Laboratory

Date/Time: Thursday, January 30, 2025/ 2:30 PM – 3:30 PM (In-Person)
Location: See Meeting URL on the cels-seminars website which will require an Argonne login.

Description: Super-resolution of climate models (also known as downscaling) is key in prediction and planning for critical infrastructure. The relationship between fine-scale and coarse-scale information is complex and often not well captured with traditional statistical or machine-learning techniques, especially in the context of extreme events. Incorporating knowledge about the physics of the system can increase accuracy but at a computational cost. The advent of data-driven surrogates, however, has shown great promise in accelerating prediction. We utilize one such model with a data assimilation framework to carry out super-resolution with uncertainty quantification on the fine-scale prediction.

Bio: Philip Dinenis is a postdoctoral researcher in Mathematics and Computer Science at Argonne National Laboratory. He is interested in large-scale statistics and optimization problems in atmospheric sciences. He completed his PhD in Applied Mathematics at Columbia University in 2023.

Please note that the meeting URL for this event can be seen on the cels-seminars website which requires an Argonne login.