Topic: Emulating Physical Parameterizations using Artificial Intelligence (105th AMS Annual Meeting)

Tuesday, 14 January 2025: 4:30 PM-6:00 PM

245 (New Orleans Ernest N. Morial Convention Center)

Host: 24th Conference on Artificial Intelligence for Environmental Science

Submitters:

Kyle A. Hilburn; Christina E. Kumler, CIRA, Boulder, CO and Tiago Quintino, ECMWF, Forecast Department, Reading

Cochairs:

Waylon G. Collins, National Weather Service, Corpus Christi, TX and Diana R. Gergel

Dynamical cores of numerical weather prediction (NWP) models are extremely powerful tools. Rather than completely replacing NWP models with artificial intelligence, another approach is to replace just the physical parameterizations with artificial intelligence. This can achieve dramatic speed-up of model integration, without sacrificing interpretability. This session solicits research topics related to emulating physical parameterizations using artificial intelligence.

Papers:

4:30 PM

8B.1

Robust Machine Learning-Based Albedo Parameterization: Generalization Across Land Cover Schemes

Akarsh Ralhan, University of Maryland, College Park, MD; and X. Z. Liang

4:45 PM

8B.2

Shedding new light on radiative transfer: Deep learning for multispectral emulation of the RRTM

Ryan A. Lagerquist, CIRA and NOAA/ESRL/GSL, Boulder, CO; and D. D. Turner, J. Q. Stewart, and I. Ebert-Uphoff

5:00 PM

8B.3

An Updated Treatment of Aerosol Optical Properties in E3SM using a Neural Network Mie Scattering Emulator