Radford, J. T., I. Ebert-Uphoff, J. Q. Stewart, K. D. Musgrave, R. DeMaria, N. Tourville, and K. Hilburn, 2024: Accelerating community-wide evaluation of AI models for global weather prediction by facilitating access to model output. Bull. Amer. Meteor. Soc., https://doi.org/10.1175/BAMS-D-24-0057.1, in press.

Summary: This article documents a 3-year reforecast archive of the most prominent AI for global weather prediction (AIWP) models (https://noaa-oar-mlwp-data.s3.amazonaws.com/index.html) as well as a visualization webpage to view real-time output of these models (https://aiweather.cira.colostate.edu/). AIWP models have shown tremendous promise, with performance competitive to traditional numerical weather prediction (NWP) in a fraction of the time. However, more research needs to be done on AIWP model performance for extremes, specific meteorological phenomenon, and specific regions. This reforecast archive was developed to help facilitate that additional research without users needing access to powerful GPU resources.

(Jacob Radford; jacob.radford@noaa.gov)