MUST: A Multi-source Spatio-Temporal data fusion Model for short-term sea surface temperature prediction. (1st September 2022)

• Record Type: Journal Article
• Authors: Hou, Siyun; Li, Wengen; Liu, Tianying; Zhou, Shuigeng; Guan, Jihong; Qin, Rufu; Wang, Zhenfeng
• Is Part Of: Ocean engineering. Volume 259(2022)
• Publication Date: 2022-09-01
• Subjects: Sea surface temperature (SST) -- SST prediction -- Multi-source data fusion -- Deep learning -- Spatio-temporal features; Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2022.111932
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: Sea surface temperature (SST) plays an important role in various oceanic applications, including climate prediction, ocean environment monitoring and marine disaster warning. Although many approaches have been developed for predicting SST, most of them conduct the prediction only based on historical SST data. However, SST is essentially affected by many external factors, e.g., the short-wave radiation from the sun and the long-wave radiation from the atmosphere and ground, which are ignored by existing approaches. In this work, we proposed a Multi-source Spatio-Temporal data fusion model (MUST) to fuse multi-source data, including SST data and external factors, to improve the accuracy of short-term SST prediction. Concretely, MUST first introduces Bicubic Convolutional Interpolation (BCI) to address the issue of inconsistent spatial resolutions of multi-source data, then employs the Spatio-Temporal Dilated ConvLSTM (ST-DC) to learn the spatio-temporal features of SST and external factors, and finally fuses the learned features from multi-source data to predict SST with a Cross Data Fusion (CDF) component. As validated on two real datasets, MUST achieves much better performance than existing SST prediction approaches. Highlights: Multiple sources of data improve the prediction of short-term SST. A MUST model was proposed to predict the SST for future seven days. A case study was conducted in the regions of Nino3.4 and East China Sea. The proposed MUST model outperforms CNN, LeNet and ConvLSTM.