arXiv cs.AI· Minjong Cheon·· 12 小时前AI 评分52
Rescene：用冻结天气模型实现稳定的百年气候模拟
Rescene: stable century-long climate emulation from a frozen weather model
AI 导读
Rescene 是一个仅 0.4M 参数的包装器，把冻结的 1.5°、6 小时、69 变量天气模型改造成随机气候模拟器。100 年 8 成员运行未检测到漂移（每 100 年 -0.004 ± 0.014 K），2022 测试年 16 成员集合的 spread/skill 在 Z500、T2m、MSLP 上分别为 0.89、0.97、0.93。
正文 · 原文
Abstract:Machine-learning weather models now match or exceed operational numerical prediction in the medium range, but when they are run freely for years they drift, blow up or lose their variability, and dedicated climate emulators are expensive to train. Here we present Rescene, a 0.4 M-parameter wrapper that turns a frozen 1.5$^\circ$, 6-hourly, 69-variable weather model into a stochastic climate emulator. A deterministic slow clock anchors the frozen model to a calendar climatology, and a band-limited AR(1) perturbation, whose spectral envelope is fixed to the observed anomaly spectrum, restores the weather that anchoring removes. A 100-year, 8-member Rescene run shows no detectable drift ($-0.004 \pm 0.014$ K per 100 years), and on the 2022 test year the 16-member ensemble is calibrated (spread/skill 0.89, 0.97 and 0.93 for Z500, T2m and MSLP) with an extratropical synoptic amplitude of 97% of ERA5 on the tuning year. The free run reproduces the geography of day-to-day variability (pattern correlation 0.83 for Z500) and 71% of the observed blocking frequency, where the deterministic wrapper alone has none. Interventions on the frozen operator show a scale-local, approximately linear response in which Z500 perturbations grow upscale 6--14 times more than downscale per 6 h. All constants are fitted on a training year against a rule declared beforehand, and we show that common evaluation shortcuts would have overstated variability and skill. Rescene makes an existing weather model usable for century-long ensembles at negligible training cost.
| Comments: | 24 pages, 5 figures, 1 table; supplementary information with 12 figures and 3 tables. A shorter version was accepted at the NeurIPS 2026 Workshop on Tackling Climate Change with Machine Learning |
| Subjects: | Atmospheric and Oceanic Physics (physics.ao-ph); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2608.09971 [physics.ao-ph] |
| (or arXiv:2608.09971v2 [physics.ao-ph] for this version) | |
| https://doi.org/10.48550/arXiv.2608.09971 arXiv-issued DOI via DataCite |
Submission history
From: Minjong Cheon [view email]
[v1]
Thu, 30 Jul 2026 09:42:39 UTC (4,478 KB)
[v2]
Wed, 14 Oct 2026 23:27:12 UTC (6,092 KB)
来源：arXiv cs.AI · arxiv.org