OceanCastNet: A Deep Learning Ocean Wave Model with Energy Conservation

2024-06-06

arXiv.org

AI Summary

This study addresses the long-term energy instability and physical uninterpretability prevalent in deep learning–based wave forecasting. To this end, we propose OceanCastNet—the first energy-conserving deep learning model for ocean wave prediction. Methodologically, we (i) explicitly embed an energy balance equation into the network architecture; (ii) design an adaptive Fourier neural operator to capture spatiotemporal dynamics of multi-source wind and wave fields; and (iii) introduce a land–sea boundary mask loss to enhance numerical robustness. Experiments on ERA5 data demonstrate that OceanCastNet achieves short-term forecast accuracy comparable to state-of-the-art physics-based models, outperforms WaveWatch III under both typical and extreme sea states, and maintains strict energy conservation even in 100-hour forecasts—significantly improving long-term reliability. Our core contribution is establishing a new end-to-end wave forecasting paradigm that is simultaneously physically interpretable, energetically stable, and highly accurate.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Computer Vision: Low Level & Physics-based Vision
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Traditional wave forecasting models, although based on energy conservation equations, are computationally expensive. On the other hand, existing deep learning geophysical fluid models, while computationally efficient, often suffer from issues such as energy dissipation in long-term forecasts. This paper proposes a novel energy-balanced deep learning wave forecasting model called OceanCastNet (OCN). By incorporating wind fields at the current, previous, and future time steps, as well as wave fields at the current and previous time steps as input variables, OCN maintains energy balance within the model. Furthermore, the model employs adaptive Fourier operators as its core components and designs a masked loss function to better handle the impact of land-sea boundaries. A series of experiments on the ERA5 dataset demonstrate that OCN can achieve short-term forecast accuracy comparable to traditional models while exhibiting an understanding of the wave generation process. In comparative experiments under both normal and extreme conditions, OCN consistently outperforms the widely used WaveWatch III model in the industry. Even after long-term forecasting, OCN maintains a stable and energy-rich state. By further constructing a simple meteorological model, OCN-wind, which considers energy balance, this paper confirms the importance of energy constraints for improving the long-term forecast performance of deep learning meteorological models. This finding provides new ideas for future research on deep learning geophysical fluid models.

Problem

• Developing a deep learning model for ocean wave forecasting
• Predicting wave height, period and direction using wind data
• Comparing machine learning performance against conventional forecasting systems

Innovation

• Deep learning model incorporates wind and wave fields
• Machine learning approach outperforms conventional forecasting systems
• Computationally efficient method captures extreme weather patterns