Deep learning weather models for subregional ocean forecasting: A case study on the Canary current upwelling system

Giovanny A. Cuervo-Londoño

doi:10.1016/j.ocemod.2026.102782

Deep learning weather models for subregional ocean forecasting: A case study on the Canary current upwelling system

2026-06-18

0

PRE

AI

G

Giovanny A. Cuervo-Londoño *

J

Javier Sánchez

Á

Ángel Rodríguez-Santana

DOI：10.1016/j.ocemod.2026.102782

Original

Original request for help

Save

Abstract

En 中文

• Adapted GraphCast for subregional SST forecasts in Canary Current Upwelling. • Model performs well year-round, excelling in complex mesoscale coastal areas. • Outperforms GLORYS and ConvLSTM, cutting RMSE by up to 76% in SST forecasts. • Delivers 20-day SST forecasts in 2.3 min, nearly 100 × faster than traditional models.

Deep learning weather models for subregional ocean forecasting: A case study on the Canary current upwelling system

Abstract

Journal

Organization

Cited Papers

Citing Papers