arXiv：cs.LG· Chaeyeon Yi, Yun Am Seo·· 4 小时前AI 评分38
CLARA：仅凭站点观测学习联合概率天气预报
Learning joint probabilistic weather forecasts from station observations alone
AI 导读
CLARA（Calibrated Advection-Routing Attention）仅用站点观测即可学习五种地面变量的联合高斯预测分布，无需数值天气预报或再分析数据，模型约 28,000 参数，支持 CPU 训练与预测。
正文
Abstract:Assessing compound weather risks requires forecasts representing dependence between variables. CLARA (Calibrated Advection-Routing Attention) learns joint Gaussian predictive distributions of five surface variables from station observations alone, without numerical weather prediction or reanalysis; the approximately 28,000-parameter model supports CPU training and prediction. Across six multi-year folds on 96 stations, its lead-mean energy score is 4.9% lower than that of a learned comparator with matched temporal inputs (4.7% with a similar parameter count) and 11-65% lower than those of statistical baselines. Holding marginal variances fixed, removing learned correlations worsens joint negative log-likelihood by 1.0-2.8 nats per station. A covariance-scale estimator, proved consistent under stated assumptions, improves short-lead calibration but over-corrects at long leads. Synthetic interventions show an attention-bias coefficient alone does not measure forecast influence. Retrained in ten regions on six continents, CLARA outperforms persistence in all 60 multi-year region-lead comparisons and a similarly sized learned model in 57 of 60.
| Comments: | 62 pages, 6 figures, 3 Extended Data figures, 7 Extended Data tables; includes Supplementary Information |
| Subjects: | Applications (stat.AP); Machine Learning (cs.LG); Atmospheric and Oceanic Physics (physics.ao-ph) |
| Cite as: | arXiv:2610.09898 [stat.AP] |
| (or arXiv:2610.09898v1 [stat.AP] for this version) | |
| https://doi.org/10.48550/arXiv.2610.09898 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yun Am Seo [view email]
[v1]
Wed, 7 Oct 2026 11:57:10 UTC (2,602 KB)
来源：arXiv：cs.LG · arxiv.org