ORIGINAL RESEARCH article
Front. Mar. Sci.
Sec. Physical Oceanography
This article is part of the Research TopicExtreme Seas: Next-Generation Forecasting of Significant Wave HeightView all articles
Ocean wave conditions forecasting using convolutional neural networks in the Yantai Fishing Zone, China
Provisionally accepted
Liangming Zhou1Qingjie Li2Xin Hong2Chawei Hou2
Feifei Jiang3*Shuang Wu2Jie Yan2Jianting Cheng2Mengke Wang3Xuelian Mao2- 1Yazhou Bay Innovation Institute of Hainan Tropical Ocean University, Sanya, China
- 2Yantai Marine Center of the Ministry of Natural Resources (Yantai Marine Forecasting Station of the Ministry of Natural Resources), Yantai, China
- 3College of Engineering, Ocean University of China, Qingdao, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Ocean wave conditions forecasting is crucial for reducing wave-related disasters and enhancing disaster prevention and mitigation capabilities in China's coastal regions. This study employs the physics-based nearshore wave model SWAN as the data source to develop a convolutional neural network (CNN)-based intelligent forecasting model for significant wave height (Hs) and mean wave period (Tm), which is applied to the Yantai Fishing Zone, China. First, wave simulations were completed using the SWAN numerical model. Then, a deep learning model for Hs and Tm was established using CNN, analyzing the impact of different input step lengths on forecasting performance. The optimal performance was achieved with an input step length of 3 hours of historical wave data, yielding correlation coefficients (CC) of 0.9997 for Hs and 0.9969 for Tm, with mean absolute errors (MAE) of 0.0075 meters (m) and 0.0562 seconds (s), and root mean square errors (RMSE) of 0.0149 m and 0.2014 s, respectively. Finally, using an input step length of 3, forecasts were made for wave height and period at lead times of 3h, 6h, 9h, and 12h. As the forecast lead time increased, model accuracy decreased, but it still effectively captured the main trends in Hs and Tm. These errors remain within acceptable ranges, demonstrating the model's good applicability.
Keywords: Ocean wave forecasting, Convolutional Neural Network, Input step lengths, Short-term wave forecasting, Yantai Fishing Zone
Received: 07 Nov 2025; Accepted: 28 Nov 2025.
Copyright: © 2025 Zhou, Li, Hong, Hou, Jiang, Wu, Yan, Cheng, Wang and Mao. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Feifei Jiang
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.