发表机构
Johns Hopkins University; Singapore Management University; University of Houston; University of Connecticut（约翰斯·霍普金斯大学; 新加坡管理大学; 休斯顿大学; 康涅狄格大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
BridgeCast提出一种基于条件流匹配的物理-人工智能混合框架，用于校正海洋波浪预报的系统性偏差，通过Transformer架构融合外生变量，在真实数据上优于现有基线。
AI 中文摘要
海洋波浪预报对于海上安全、离岸作业和海岸韧性至关重要，但由于基于物理的模型存在系统性偏差，预报仍然具有挑战性。物理模型虽然被广泛使用，但依赖于近似和参数化，这限制了它们在复杂海洋-大气条件下的准确性。为了增强海洋波浪预报，我们提出了BridgeCast，这是一个物理-人工智能混合框架内的偏差校正方法。BridgeCast是一种基于条件流匹配（CFM）的概率模型，学习将物理模型预报转换为类似再分析的场。它将物理预报视为损坏的观测，并采用连续时间生成过程将其分布桥接到再分析数据的分布。BridgeCast由基于Transformer的架构参数化，该架构支持时空建模、外生大气变量的纳入，以及通过常微分方程和随机微分方程公式进行灵活推理。在真实世界数据集上的大量实验表明，BridgeCast在不同区域和预报提前时间上始终优于最先进的基线方法。
英文摘要
Ocean wave forecasting is essential for maritime safety, offshore operations, and coastal resilience, yet remains challenging due to systematic biases in physics-based models. Physical models, while widely used, rely on approximations and parameterizations that limit their accuracy under complex ocean-atmosphere conditions. To enhance ocean wave forecasting, we propose BridgeCast, within a physics-AI hybrid framework for bias correction. BridgeCast is a probabilistic model based on conditional flow matching (CFM) that learns to transform physical model forecasts into reanalysis-like fields. It treats physical forecasts as corrupted observations and employs a continuous-time generative process to bridge their distribution toward that of reanalysis data. BridgeCast is parameterized by a Transformer-based architecture that enables spatiotemporal modeling, incorporation of exogenous atmospheric variables, and flexible inference via both ordinary and stochastic differential equation formulations. Extensive experiments on real-world datasets demonstrate that BridgeCast consistently outperforms state-of-the-art baselines across regions and forecast lead times.
Comments23 pages, including references and appendix
