GraphCast: Learning skillful medium-range global weather forecasting
About
Global medium-range weather forecasting is critical to decision-making across many social and economic domains. Traditional numerical weather prediction uses increased compute resources to improve forecast accuracy, but cannot directly use historical weather data to improve the underlying model. We introduce a machine learning-based method called "GraphCast", which can be trained directly from reanalysis data. It predicts hundreds of weather variables, over 10 days at 0.25 degree resolution globally, in under one minute. We show that GraphCast significantly outperforms the most accurate operational deterministic systems on 90% of 1380 verification targets, and its forecasts support better severe event prediction, including tropical cyclones, atmospheric rivers, and extreme temperatures. GraphCast is a key advance in accurate and efficient weather forecasting, and helps realize the promise of machine learning for modeling complex dynamical systems.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Global Weather Forecasting | ERA5 lite | RMSE0.72 | 108 | |
| Precipitation forecasting | CONUS Dec 2022 Winter Storm Elliott | Correlation Coefficient (r)0.89 | 12 | |
| Precipitation forecasting | CONUS Jun 2022 Summer Convective | Correlation (r)0.55 | 12 | |
| Precipitation forecasting | CONUS Mar 2023 Spring Transition | Pearson Correlation Coefficient (r)0.83 | 12 | |
| 10-m u-wind prediction | CONUS Summer Jun 2022 | Correlation (r)0.8118 | 9 | |
| 10-m u-wind prediction | CONUS Winter Dec 2022 | Correlation Coefficient (r)0.8462 | 9 | |
| 10-m u-wind prediction | CONUS Spring Mar 2023 | Correlation Coefficient (r)0.8441 | 9 | |
| 10-m v-wind prediction | CONUS Summer Jun 2022 | Correlation (r)0.8182 | 9 | |
| 10-m v-wind prediction | CONUS Mar 2023 (Spring) | Correlation (r)0.8253 | 9 | |
| 2-m temperature prediction | CONUS Dec 2022 Winter | Correlation Coefficient (r)0.9919 | 9 |