Correcting Systematic and State‐Dependent Errors in the NOAA FV3‐GFS Using Neural Networks
Problem
Imperfect weather models contain systematic, state-dependent errors that degrade forecasts and are only partially corrected by data assimilation.
Why it matters
State-dependent model errors directly degrade weather forecast quality.
What the AI did
Neural networks were trained on large sets of data assimilation analysis increments to predict corrections to systematic error in the FV3-GFS model.
Results
Simple column NN significantly improves model error correction vs linear baseline; convolutional NN adds no further skill; no numeric metrics given.
Impact
Peer-reviewed evidence that NNs can extract model-error corrections from analysis increments.
Record
| Venue | Journal of Advances in Modeling Earth Systems |
|---|---|
| Published | 2022 |
| Work type | article |
| Authors | Tse‐Chun Chen, Stephen Gregory Penny, Jeffrey S. Whitaker, Sergey Frolov, Robert L. Pincus, Stefan N. Tulich |
| Citations | 34 (OpenAlex, at import time) |
| References | 19 |
| Model or method | Neural networks (simple column NN; convolutional NN) |
| Organization | NOAA (FV3-GFS model owner) |
| Primary source | https://doi.org/10.1029/2022ms003309 |
| Open access | gold |
| Keywords | neural networks, model error correction, FV3-GFS, data assimilation, weather forecasting |
Importance rating
real but narrower advance
Correcting state-dependent model errors improves weather forecast quality, which broadly matters, but the abstract quantifies only forecast skill gains, not the scale of impact.
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
|