Design and application of the intelligent nautical buoy sea surface temperature prediction system based on ConvGRU network

• Ying Ye,
• Feng Pan,

Vol. 19, No. 9, September 30, 2025
10.3837/tiis.2025.09.007, Download Paper (Free):

• ConvGRU network
• nautical buoy structure design
• Stability analysis
• sea trial application
• sea surface temperature monitoring and prediction

Abstract

In marine science, sea surface temperature (SST) constitutes a critical data for marine environmental monitoring. Prior studies predominantly focused on discrete spatial points when addressing SST prediction tasks, resulting in suboptimal predictive stability due to neglected spatial correlations inherent in two-dimensional marine environments. To address this limitation, this study proposes a deep learning model based on convolutionally gated loop cells is proposed and applied to an innovatively designed buoy observation system for SST prediction of 2D smart ocean. The framework directly processes spatiotemporal SST data, with both its input and output being spatio-temporal in nature. Through embedded spatial attention mechanisms, the model dynamically enhances temperature pattern recognition while preserving spatial coherence across multiple spatial scales. Experimental results demonstrate that the proposed model outperforms three other state-of-the-art deep learning models in terms of prediction performance. The proposed model was implemented through an innovative 3-meter miniaturized nautical buoy system, enabling SST observation and forecasting. This advancement significantly enhances the precision of SST predictions while preserving essential spatial dependencies, providing a robust solution for real-time marine environmental monitoring systems.

Statistics

Cite this article