A Customized Multi‐Scale Deep Learning Framework for Storm Nowcasting
Problem
Deep learning models perform poorly when nowcasting high-impact storms due to complex multi-scale spatiotemporal variations.
Why it matters
Storm nowcasting is described as critical and urgently needed.
What the AI did
Proposes a customised multi-scale deep learning framework with a flexible attention module and customised loss function to ensure multi-scale spatiotemporal consistency.
Results
All models applying the framework improved skill scores by 8.5%–42.6% for 1-hr nowcasting on the SEVIR dataset.
Impact
Improved 1-hour nowcasting skill for intense storms on a benchmark dataset.
Record
| Venue | Geophysical Research Letters |
|---|---|
| Published | 2023 |
| Work type | article |
| Authors | Shangshang Yang, Huiling Yuan |
| Citations | 21 (OpenAlex, at import time) |
| References | 58 |
| Model or method | Customised multi-scale (CM) deep learning framework |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1029/2023gl103979 |
| Open access | gold |
| Keywords | deep learning, storm nowcasting, multi-scale framework, SEVIR, attention module |
Importance rating
significant — thousands to millions affected
Abstract states storm nowcasting is critical and urgently needed; high-impact storm forecasting bears on public safety, with 8.5%-42.6% skill improvement on SEVIR benchmark.
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
|