arXiv：cs.LG· Chenyu Dong, Gianmarco Mengaldo·· 5 小时前AI 评分43
S2S-JEPA：在次季节到季节尺度上预测可预测的部分
S2S-JEPA: Predicting the Predictable at Subseasonal-to-Seasonal Timescales
AI 导读
S2S-JEPA 将 JEPA 范式引入次季节到季节（S2S）预报，通过在潜空间预测缓慢变化分量、舍弃不可预测的细节，在第 5 至 6 周多项指标上超过 ECMWF 物理集合预报，整体技巧与之相当。该工作针对两周至两个月的"可预测性荒漠"窗口，借鉴最先进 AI 天气模型的设计要素。
正文
Abstract:The subseasonal-to-seasonal (S2S) timescale, roughly from two weeks to two months ahead, is a critical forecast window for sectors such as agriculture, energy, and water management. Yet, it is widely known as the `predictability desert'. Recent AI weather models excel up to two weeks ahead but deteriorate beyond, largely because they are trained to predict fine-scale details that are neither predictable nor essential at S2S timescales. We argue that a more physically grounded objective is to forecast only the slowly varying components that remain predictable. Computer vision reached the same conclusion with the Joint-Embedding Predictive Architecture (JEPA), which predicts in latent space, discarding unpredictable details. In this work, we introduce S2S-JEPA, which brings the JEPA paradigm to S2S forecasting. It is tailored to this task through design elements from state-of-the-art AI weather models. S2S-JEPA achieves comparable skill to the gold-standard ECMWF physics-based ensemble and surpasses it on multiple metrics at weeks 5 to 6.
| Subjects: | Atmospheric and Oceanic Physics (physics.ao-ph); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.03106 [physics.ao-ph] |
| (or arXiv:2610.03106v1 [physics.ao-ph] for this version) | |
| https://doi.org/10.48550/arXiv.2610.03106 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Chenyu Dong [view email]
[v1]
Fri, 2 Oct 2026 10:26:05 UTC (2,278 KB)
来源：arXiv：cs.LG · arxiv.org