arXiv：cs.AI· Mutian Xi, Xinyu Su, Xiaohui Zhong, Hao Li·· 6 小时前
WeatherBridge：面向预训练天气预报模型的预报感知多源输入适配
WeatherBridge: Forecast-Aware Multi-Source Input Adaptation for Pretrained Weather Forecasting Models
AI 导读
WeatherBridge 是一种源条件适配器，可将 GFS、HRES、CMA 等异构业务 NWP 初始场映射为 ERA5 参考状态，且不更新预训练预报模型。它用 Swin-UNet 预测结构化残差修正并在多个空间尺度注入源嵌入，训练目标结合纬度加权 ERA5 重建与 One-step Rollout Alignment（ORA），通过冻结的 FuXi 回传一步预报误差。
正文
Abstract:Data-driven weather forecasting models are commonly trained on ERA5, yet operational deployment must use timely initial states from physics-based numerical weather prediction (NWP) systems whose distributions differ substantially from ERA5. Direct initialization with these states can degrade autoregressive forecasts, motivating input adaptation that keeps the pretrained forecasting model unchanged. Here, we introduce WeatherBridge, a unified, source-conditioned adapter that maps heterogeneous operational NWP initial states to ERA5 reference states without updating the pretrained forecasting model. WeatherBridge predicts the structured residual corrections with a Swin-UNet and injects source embeddings across multiple spatial scales, allowing the model to capture shared atmospheric structure and source-specific discrepancies. Its training objective combines latitude-weighted ERA5 reconstruction with One-step Rollout Alignment (ORA), which backpropagates one-step forecast error through a frozen FuXi model to optimize the adapted states for both initial state fidelity and downstream forecast compatibility. We evaluate WeatherBridge using daily GFS, HRES, and CMA initializations from 1 January to 22 November 2025 across six variables and lead times up to 240 hours. Relative to raw source initialization of FuXi, WeatherBridge reduces the mean RMSE of FuXi over the 6- 240-hour rollout for 17 of 18 source-variable combinations, with a median reduction of 10.2% and a maximum reduction of 46.3%. The adapter trained with frozen FuXi also improves FengWu forecasts on GFS inputs without retraining either the adapter or FengWu. These results demonstrate that forecast-aware input adaptation can improve forecasts from heterogeneous operational NWP inputs and transfer across pretrained ERA5-based forecasting models.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.14863 [cs.AI] |
| (or arXiv:2610.14863v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.14863 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Xinyu Su [view email]
[v1]
Sat, 10 Oct 2026 12:32:56 UTC (11,997 KB)
来源：arXiv：cs.AI · arxiv.org