StretchCast: Global-Regional AI Weather Forecasting on Stretched Cubed-Sphere Mesh

Jin Feng · 2026

Read StretchCast: Global-Regional AI Weather Forecasting on Stretched Cubed-Sphere Mesh by Jin Feng (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-27288.txt.

Description

Feng introduces StretchCast, a global-regional AI weather forecasting framework on a stretched cubed-sphere mesh. Uniform global models struggle to refine one region. The stretched mesh keeps a closed globe while packing cells over a target region. A one-step base model and a multistep rollout model test the design. Training used ERA5 data with 69 variables from 1998 to 2022. The larger rollout model stayed competitive on regional anomaly correlation. Typhoon and spectral checks showed realistic, smooth structure across mesh faces.

Catalog record

CREATORJin Feng
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE79022
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.27288

Browse the full library · About the Tower of Babel