OceanLight: Efficient Global Ocean Forecasting via Geometry-Adaptive Unstructured Mesh Representation
2026-08-17 • Machine Learning
Machine LearningArtificial Intelligence
AI summaryⓘ
The authors developed OceanLight, a new method to predict ocean conditions more efficiently and accurately. Unlike older methods that use fixed grids and waste resources on land areas, OceanLight uses flexible meshes that better match the ocean's varying details. It uses graph neural networks to improve forecast accuracy and capture important ocean features like swirling currents called eddies. Their approach uses less computer memory and power while providing more reliable results compared to existing AI and traditional models.
ocean forecastingstructured gridunstructured meshgraph neural networkkinetic energy spectrumgeostrophic balancemesoscale eddiescomputational efficiencynumerical ocean models
Authors
Wei Wu, Xiang Wang, Hongze Leng, Qingye Min, Junxing Zhu, Junqiang Song
Abstract
Reliable global ocean forecasting is critical for climate monitoring, marine navigation, and extreme event early warning. Physics-based ocean forecasting models impose prohibitive computational costs, while existing deep learning approaches predominantly rely on structured-grid architectures, incurring unnecessary computation on masked land cells and enforcing uniform resolution across dynamically heterogeneous ocean regions regardless of local flow complexity. Here we present OceanLight, an efficient global ocean forecasting framework innovatively combining geometry-adaptive unstructured mesh tokenization with a graph neural network (GNN) backbone. OceanLight achieves pointwise forecast accuracy and kinetic energy spectral fidelity exceeding both operational numerical analyses and state-of-the-art AI-based models, while surpassing all AI-based ocean models in geostrophic balance consistency. Furthermore, OceanLight demonstrates reliable mesoscale eddy representation, capturing coherent ocean structures beyond pointwise statistical optimization. These capabilities are delivered with a 62% reduction in GPU memory consumption and 70\% reduction in FLOPs relative to structured-grid baselines. Our unstructured mesh representation establishes a generalizable paradigm for scalable data-driven oceanography.