arXiv：cs.LG· Noelia Otero, Atahan \"Ozer, Miguel-\'Angel Fern\'andez-Torres, Jackie Ma·· 4 小时前AI 评分33
基于 Vision Transformer 的欧洲次季节土壤湿度预测：闪旱预测的前景与局限
Skillful Data-Driven Subseasonal Soil Moisture Forecasting: Prospects and Limits for Flash Drought Prediction
AI 导读
研究用双路径时空注意力的 Vision Transformer 架构做欧洲次季节土壤湿度预测，2021-2022 年对比深度学习与 ECMWF S2S 基线，在所有预见期取得最高的确定性及概率性技巧，并能可靠识别根系层异常干燥状态（低于第 20 百分位）。
正文
Abstract:Despite substantial progress in short-to-medium-range weather forecasting, predicting high-impact events such as flash droughts remains a key challenge for both early warning operations and physically-based subseasonal-to-seasonal (S2S) prediction systems. Here we demonstrate that, for S2S soil-moisture forecasting over Europe, forecast skill depends as much on how the prediction problem is formulated as on the forecasting model itself. Using a Vision Transformer-based architecture with dual-pathway temporal and spatial attention, we show that residual learning is essential to outperform persistence. This advantage is realized only when forecasting root-zone soil moisture in physical units rather than standardized anomalies, revealing that the target representation itself constrains predictability. A probabilistic extension via quantile-head fine-tuning further provides well-calibrated predictive distributions. Benchmarked against deep-learning and operational ECMWF S2S baselines over 2021-2022, our model achieves the highest deterministic and probabilistic skill at all lead times and reliably detects anomalously dry root-zone states (below the 20th percentile). Yet flash drought onset, defined by multi-pentad intensification criteria, remains a fundamental challenge shared across all current S2S systems. These findings advance data-driven S2S soil-moisture forecasting while highlighting the remaining challenge of predicting rapid drought development.
| Comments: | 27 pages, 8 figures, 5 tables. Accepted for publication in npj Hydrosphere. Supplementary information available with the published version |
| Subjects: | Machine Learning (cs.LG); Atmospheric and Oceanic Physics (physics.ao-ph) |
| Cite as: | arXiv:2610.07060 [cs.LG] |
| (or arXiv:2610.07060v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.07060 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Noelia Otero [view email]
[v1]
Mon, 5 Oct 2026 06:34:10 UTC (10,460 KB)
来源：arXiv：cs.LG · arxiv.org