Distinguishing Initial-Condition and Model Error Using Machine-Learning Weather Models
P. Trent Vonich, Gregory J. Hakim
Abstract
Gated and ungated optimizations evaluated in GraphCast and Pangu-Weather. The first panel (upper-left) shows the mean over the 36 dates; remaining panels show first-of-month 12 UTC optimization descents from January-November 2017. The y-axis metric is the summed day 1-10 Z500 RMSE verified against ERA5, expressed as percent change from each model's date-matched control. The gated arm optimizes the initial conditions jointly with a LoRA weight correction; its GraphCast curve uses the initial conditions alone (LoRA removed), the same state Pangu receives. Stars and diamonds mark the lowest Pangu errors in the 5- and 10-day windows, respectively. Percentages above each panel show the best Pangu improvement relative to control for gated (green) and ungated (red) optimization. Ungated outperforms gated on half of dates when stopped early. Gating does not guarantee a better Pangu performance, but it does produce more comparable GraphCast and Pangu-Weather improvements on all 36 cases.
Create a lesson
Related papers
Evaluating the prediction of subgrid eddy kinetic energy from a machine learning model
Thomas Wilder, Ranjini Swaminathan, David Storkey et al.
Shifts in Rossby wave regimes via the change of group velocity with zonal jets
Zhixiang Li, Jianhua Lu, Yimin Liu et al.
A Multidimensional Evaluation Framework for Artificial Intelligence-Based Ocean State Forecasts
Ziqing Zu, Ye Yuan, Tiejun Ling et al.
RadarFill: Observation-Constrained Gap Filling of Systematic Blind Zones in 3-D Weather Radar Reflectivity Mosaics
Xin Huang, Haofei Sun, Wei Han et al.
Machine learning emulators of dynamic global vegetation models for fast gridded carbon budget predictions
Piyu Ke, Xiaofan Gui, Stephen Sitch et al.
From Sea Surface Height to Velocity - A Machine Learning Approach
Yaswanth Reddy Yaradoddi, Jinbo Wang, Elyes Serghine