Sub‐Seasonal Forecasting With a Large Ensemble of Deep‐Learning Weather Prediction Models
Problem
Producing sub-seasonal (six-week) global weather forecasts with large ensembles is computationally prohibitive with traditional numerical models.
Why it matters
Longer-range forecasts improve preparedness for droughts, heatwaves, and storm planning.
What the AI did
A Deep Learning Weather Prediction (DLWP) model using CNNs on a cubed sphere grid recursively predicts six atmospheric variables, with ensemble spread from 32 randomly-trained model variants.
Results
320-member six-week forecasts at 1.4° resolution in 3 minutes on a single GPU; reasonable 4.5-day deterministic forecast of Hurricane Irma; spontaneously generates tropical cyclones in one-year simulation.
Impact
Demonstrated computationally cheap sub-seasonal forecasting but does not forecast precipitation and abstract is truncated before final skill metric.
Record
| Venue | Journal of Advances in Modeling Earth Systems |
|---|---|
| Published | 2021 |
| Work type | article |
| Authors | Jonathan A Weyn, Dale Richard Durran, Rich Caruana, Nathaniel Cresswell‐Clay |
| Citations | 190 (OpenAlex, at import time) |
| References | 31 |
| Model or method | DLWP (convolutional neural networks on cubed sphere grid) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1029/2021ms002502 |
| Open access | gold |
| Keywords | deep learning, sub-seasonal forecasting, ensemble prediction, CNN, tropical cyclones |
Importance rating
significant — thousands to millions affected
Sub-seasonal forecasting is a recognized global capability gap; abstract notes current systems have low skill at S2S timescales and offers 320-member six-week forecasts in minutes, a genuinely new cheap capability.
More in Climate & Environment
| # | Paper | Year | Citations | Importance |
|---|---|---|---|---|
| 1 | Learning skillful medium-range global weather forecasting | 2023 | 1,416 |
8
|
| 2 | Accurate medium-range global weather forecasting with 3D neural networks | 2023 | 1,631 |
8
|
| 3 | FuXi: a cascade machine learning forecasting system for 15-day global weather forecast | 2023 | 355 |
8
|
| 4 | Skilful nowcasting of extreme precipitation with NowcastNet | 2023 | 451 |
8
|
| 5 | Forecasting Renewable Energy Generation with Machine Learning and Deep Learning: Current Advances and Future Prospects | 2023 | 378 |
8
|