CGD/CISL Seminar - Using High-Resolution Earth System Simulations to Improve Predictions and Projections of Extreme Weather Events
1:00 – 2:00 pm MDT
Abstract
Extreme weather events are among the most consequential manifestations of Earth system variability and long-term change, yet they remain one of the greatest challenges for accurate model prediction. Many of the physical processes that govern tropical cyclones, organized convection, and extreme precipitation occur at mesoscale resolutions, beyond the reach of most current-generation Earth system models. In this seminar, I will present recent advances from the MESACLIP (Understanding the Role of MESoscale Atmosphere-Ocean Interactions in Seasonal-to-Decadal CLImate Prediction) project, which aims to improve our understanding and prediction of extreme weather across timescales using high-resolution Earth system simulations. The presentation begins with an overview of the MESACLIP project and its high-resolution simulation archive. I will then demonstrate how increased model resolution substantially improves the simulation of extreme precipitation through a more realistic representation of mesoscale dynamics and convective organization, leading to more credible projections of future precipitation extremes. The seminar further examines the response of tropical cyclone (TC) hazards to long-term environmental changes, showing that although TC frequency is projected to decline under warming condition, compound wind–rainfall hazards can become substantially more severe, particularly over inland regions, owing to enhanced rainfall intensity and changes in storm characteristics. These results emphasize the importance of evaluating climate risks using physically based hazard metrics rather than storm frequency alone. Finally, I will highlight the ongoing efforts to leverage the MESACLIP simulation archive to develop AI-based Earth system emulators, including TC-CNN and CREDIT-MESACLIP. By combining high-resolution physics-based numerical simulations with emerging AI techniques, these efforts aim to accelerate Earth system prediction while preserving the physical realism provided by process-based models. Together, these results illustrate how high-resolution Earth system simulations can improve both our scientific understanding of extreme weather and the application of AI in Earth system science.
The visiting faculty is sponsored by the EdEC Faculty Fellowship Program.
Here is the public livestream link. Staff members can look for a Google Calendar invitation for the talk or check the "Staff-only" calendar.
Please reach out to Sam Scalice (sscalice@ucar.edu) with any questions you may have.
Name
Dan Fu
Dr. Dan Fu is an Assistant Professor in the Department of Atmospheric Sciences at Texas A&M University. His research focuses on high-resolution Earth system modeling, extreme weather, seasonal-to-multidecadal prediction, and artificial intelligence (AI) for Earth system science. His work integrates physics-based numerical modeling with AI to improve the understanding, prediction, and projection of a wide range of extreme weather phenomena. Dr. Fu is a co- Principal Investigator of the MESACLIP project, a collaborative effort between Texas A&M University and
NSF NCAR supported by the National Science Foundation’s Climate and Large-Scale Dynamics (CLD) Program. The project investigates the role of mesoscale atmosphere–ocean interactions in seasonal-to-decadal climate prediction using high-resolution Earth system simulations. His current research also explores AI-based Earth system emulators and machine-learning approaches for accelerating Earth system prediction.