Title: Prediction of sea surface height based on recurrent neural network

Authors: Ying Li; Yong Li; Danya Xu; Jingyu Jiang

Addresses: College of Computer Science and Technology, China University of Petroleum, Qingdao, China ' College of Oceanography and Space Informatics Science and Technology, China University of Petroleum, Qingdao, China ' Southern Marine Science and Engineering Guangdong Laboratory, Zhuhai, China ' College of Computer Science and Technology, China University of Petroleum, Qingdao, China

Abstract: The prediction of sea surface height (SSH) is not only theoretically important, but also has many practical applications in various ocean-related fields. Due to the temporal nature of SSH variation, recurrent neural network (RNN) models in deep learning have the ability to capture long-range dependent information in temporal data. Therefore, this paper will attempt to build a SSH prediction model using RNN and compare it with the traditional BP neural network prediction results. The CORA reanalysis of SSH data in the South China Sea is selected for the experiment. The one-day average accuracy of the RNN prediction reached 90.22%.

Keywords: prediction of sea surface height; recurrent neural network; RNN; BP neural network; deep learning.

DOI: 10.1504/IJAIS.2021.117878

International Journal of Adaptive and Innovative Systems, 2021 Vol.3 No.1, pp.43 - 57

Received: 12 Oct 2020
Accepted: 11 Nov 2020
Published online: 04 Oct 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article