UniOcean: A Unified Framework for Predicting Multiple Ocean Factors of Varying Temporal Scales

Accurate prediction of ocean factors (e.g., temperature and salinity) is crucial for plenty of applications, including weather forecasting, storm tracking, and ecosystem protection. Meanwhile, it is well-known that the ocean is a unified system and various ocean factors usually influence each other. For example, the changes in temperature would affect the distribution of salinity in ocean. However, the existing studies for ocean factor prediction mainly focus on designing individual models for predicting specific factors and ignore the correlations between different factors, thus having potentials to be further improved. Therefore, we propose a unified framework UniOcean to predict multiple ocean factors simultaneously, and capture the correlations between them to improve the prediction accuracy. First, considering that ocean factors are usually collected with different temporal scales, we develop the fine-grained multiscale data fusion module to integrate multiple ocean factors with different temporal scales, and effectively learn their hierarchical patterns at different levels. Then, since the correlations between ocean factors may vary across different time periods, the multifactor correlation learning module is constructed to adaptively learn the dynamic correlations between different factors. Finally, we utilize the factor-specific towers to predict multiple ocean factors simultaneously. Experimental results on five real-world remote sensing datasets demonstrate that UniOcean significantly improves the prediction accuracy by 11%–53% in terms of MSD for different ocean factors.

I. Introduction

Ocean covers more than two-thirds of the Earth and plays a crucial role in regulating the global climate. Various ocean factors, e.g., temperature, salinity, and ocean heat content, have been studied for decades to identify extreme climate events (e.g., El Niño phenomena and typhoons) and protect the ocean ecosystem [1]. With the development of sensing technology, more and more remote sensors have been deployed to monitor the changes of these ocean factors [2]. Various ocean datasets are thus collected and published by different organizations [e.g., National Aeronautics and Space Administration (NASA), European Centre for Medium-Range Weather Forecasts (ECMWF), and the Chinese Academy of Sciences (CAS)], which lays the solid foundation for analyzing the changing patterns and regularity of the ocean. For example, since 1979, NASA has launched a series of satellites [3], [4] to continually monitor the sea surface temperature (SST) and sea surface salinity (SSS) globally. Over the past decades, there have been numerous studies utilizing the collected data for predicting various ocean factors, which hold great significance for various applications, e.g., weather forecasting [5], fishing detection [6], and storm tracking [7].