BridgeCast: Bridging Ocean Wave Forecasts to Reanalysis via Flow Matching with Exogenous Variables
Abstract
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.
Related
- Self-supervised temporal reconstruction of reanalysis fields at subhourly scales
- Satellites and Reanalysis Reveal the Dominance of Dynamics over Precipitation in Marine Cold-Air Outbreak Cloud Break-Up
- All-weather lake-ice phenology from Sentinel-1 and reanalysis thermodynamics: a validated, transferable method applied to 131 Adirondack lakes
- Rigorous uncertainty quantification of probabilistic AI weather forecasts with conformal prediction
- Divergent responses of streamflow reanalysis errors to precipitation reanalysis errors modulated by catchment heterogeneity
BibTeX
@misc{gan2026bridgecast,
title = {{BridgeCast: Bridging Ocean Wave Forecasts to Reanalysis via Flow Matching with Exogenous Variables}},
author = {Siyu Gan and Dongsheng Luo and Kunxiaojia Yuan and Dongjin Song and Jingchao Ni},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.03759},
eprint = {2610.03759},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.03759},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - BridgeCast: Bridging Ocean Wave Forecasts to Reanalysis via Flow Matching with Exogenous Variables
AU - Siyu Gan
AU - Dongsheng Luo
AU - Kunxiaojia Yuan
AU - Dongjin Song
AU - Jingchao Ni
PY - 2026
DA - 2026/10/20
PB - arXiv
DO - 10.48550/arxiv.2610.03759
UR - https://arxiv.org/abs/2610.03759
AB - 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.
ER -