28 February 2023 in IEEE Geoscience and Remote Sensing Letters
Accurately predicting sea surface temperature (SST) plays an important role in the study of marine ecosystems and global climate. The SST prediction problem is usually formulated as a time-series regression problem; i.e., the future SST is predicted based on the historical SST. However, the existing methods are typically devoted to modeling the highly nonlinear temporal correlations in SST data. They often ignore the dynamic spatial correlations that exist. This can limit the performance of these models, making accurately predicting SST challenging. To address this challenge, by combining graph neural networks (GNNs) that have a clear advantage in modeling spatial correlations, we propose a global spatiotemporal graph attention network (GSTGAT). Specifically, we capture the global dynamic spatial correlations of nodes through a global graph attention network (GGAT) module that fuses the static adjacency matrix learned adaptively by the graph learning (GL) module with the dynamic attention coefficients. A gated temporal convolutional network (GTCN) module is used to capture the nonlinear temporal correlations. Then, the above modules are integrated into a unified neural network to predict SST. We conduct experiments on multiple time-scale datasets in the Bohai Sea and the South China Sea. The experimental results show that GSTGAT achieves the best performance and consistently outperforms other methods for different prediction horizons in the two sea areas.
20 May 2025 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
There is a growing interest in using spatio-temporal graph networks to predict sea surface temperature (SST). Nevertheless, numerous existing methodologies struggle to accurately capture both local and global spatial dependencies and are red frequently vulnerable to data noise. To tackle these challenges, we introduce a novel framework known as Spatio-Temporal Graph Contrastive Learning Network (STGLN), which enhances the precision of sea surface temperature forecasts. Notably, STGLN employs both feature-level and topology-level data augmentation techniques to alleviate the influence of data noise. Subsequently, the refined SST feature matrix and adjacency matrix are analyzed through a dual-branch network with shared parameters, specifically designed to effectively capture both local and global spatial dependencies within the SST time series data. Furthermore, we propose an auxiliary task for spatio-temporal contrastive learning, which aims to mitigate the sparsity of supervised signals and to efficiently extract essential spatio-temporal information. Comprehensive experiments conducted on real-world SST datasets demonstrate that STGLN significantly surpasses other advanced methodologies.