102. Improved Global Weather Prediction with GFDL’s FV3 Dynamical Core
Authors: Shannon Rees (Engility Corporation)
Abstract: Global weather models are extremely computationally intensive, especially when run at high resolutions. With increasing HPC resources and the desire for better forecasts, the resolutions these models are demanded to run at is always increasing. The current U.S. National Weather Service operational Global Spectral Model (GSM) is a hydrostatic model, and it has reached its end-of-life. The field of global weather modeling has reached a point where resolutions are so high that non-hydrostatic, or cloud-resolving, models are required. The Next Generation Global Prediction System (NGGPS) is being built to replace the GSM with a model that will improve forecasts and extend the predictability range out to 30 days, while keeping pace with changes in HPC resources over the next two decades. The Geophysical Fluid Dynamics Laboratory’s Finite Volume dynamical core was chosen, through a two-year program of rigorous testing, to be part of this NGGPS.
Two-page extended abstract: pdf