DP-BICNN: A Bidirectional Information Compensation Neural Network Coupled With Data-Driven and Physical Information for Sea Surface Temperature Prediction

Sea surface temperature (SST) changes in the Mediterranean Sea have profound impacts on both the Mediterranean regions and remote areas. Previous studies show that the Mediterranean SST has significant decadal variability that is comparable with the Atlantic multidecadal variability (AMV). However, few studies have discussed the characteristics and sources of the decadal predictability of Mediterranean SST based on observations. Here for the first time we use observational datasets to reveal that the decadal predictability of Mediterranean SST is contributed by both external forcings and internal variability for both annual and seasonal means, except that the decadal predictability of the winter mean SST in the eastern Mediterranean is mostly contributed by only internal variability. Besides, the persistence of the Mediterranean SST is quite significant even in contrast with that in the subpolar North Atlantic, which is widely regarded to have the most predictable surface temperature on the decadal time scale. After the impacts of external forcings are removed, the average prediction time of internally generated Mediterranean SST variations is more than 10 years and closely associated with the multidecadal variability of the Mediterranean SST that is closely related to the accumulated North Atlantic Oscillation forcing. Significance Statement Decadal prediction of sea surface temperature (SST) in the Mediterranean Sea is very important due to its profound climate impacts. Understanding the spatial/seasonal characteristics and sources of the decadal predictability of the Mediterranean SST are crucial for skillful decadal prediction of the Mediterranean SST and related variations. This study for the first time shows that the decadal predictability of the Mediterranean SST is quite significant even in comparison with the AMV. The decadal predictability of the Mediterranean SST is both region and season dependent. The internally generated decadal predictability of the Mediterranean SST is most probably related to the local ocean dynamics.

The prediction of sea surface temperature has always been the focus of ocean meteorological research. However, existing sea surface temperature prediction methods have limited ability in spatial information extraction of sea surface temperature and also perform poorly in long-term prediction. To address these problems, we propose a multi-point long-term prediction method for sea surface temperature, which is named graph multi-head self-attention neural network (GMSAN). Firstly, GMSAN constructs an adaptive graph neural network, which can adaptively extract the spatial information between points and realize multi-point prediction of sea surface temperature. GMSAN then uses the multi-head self-attention mechanism to extract global dependence point by point from the historical series of sea surface temperature to improve its long-term forecasting ability. We use 35 years of sea surface temperature data from Bohai Sea for the experiment and predict the sea surface temperature in three time windows of 120 days, 180 days and 240 days. The experimental results show that our model can meet the demand of multi-point long-term prediction of sea surface temperature. In addition, compared with the best-performing sea surface temperature prediction model, GMSAN improves the average prediction accuracy by 4.4% at 28 points and 3.3% at 112 points.

Short-term prediction of sea surface temperature (SST) is commonly achieved through numerical models. Numerical approaches are more suitable for use over a large spatial domain than in a specific site because of the difficulties involved in resolving various physical sub-processes at local levels. Therefore, for a given location, a data-driven approach such as neural networks may provide a better alternative. The application of neural networks, however, needs a large experimentation in their architecture, training methods, and formation of appropriate input–output pairs. A network trained in this manner can provide more attractive results if the advances in network architecture are additionally considered. With this in mind, we propose the use of wavelet neural networks (WNNs) for prediction of daily SST values. The prediction of daily SST values was carried out using WNN over 5 days into the future at six different locations in the Indian Ocean. First, the accuracy of site-specific SST values predicted by a numerical model, ROMS, was assessed against the in situ records. The result pointed out the necessity for alternative approaches. First, traditional networks were tried and after noticing their poor performance, WNN was used. This approach produced attractive forecasts when judged through various error statistics. When all locations were viewed together, the mean absolute error was within 0.18 to 0.32 °C for a 5-day-ahead forecast. The WNN approach was thus found to add value to the numerical method of SST prediction when location-specific information is desired.

Toward usable predictive climate information at decadal timescales

Sea surface temperature (SST) prediction has received increasing attention in recent years due to its paramount importance in the various fields of oceanography. Existing studies have shown that neural networks are particularly effective in making accurate SST predictions by efficiently capturing spatiotemporal dependencies in SST data. Among various models, the ConvLSTM framework is notably prominent. This model skillfully combines convolutional neural networks (CNNs) with recurrent neural networks (RNNs), enabling it to simultaneously capture spatiotemporal dependencies within a single computational framework. To overcome the limitation that CNNs primarily capture local spatial information, in this paper we propose a novel model named DatLSTM that integrates a deformable attention transformer (DAT) module into the ConvLSTM framework, thereby enhancing its ability to process more complex spatial relationships effectively. Specifically, the DAT module adaptively focuses on salient features in space, while ConvLSTM further captures the temporal dependencies of spatial correlations in the SST data. In this way, DatLSTM can adaptively capture complex spatiotemporal dependencies between the preceding and current states within ConvLSTM. To evaluate the performance of the DatLSTM model, we conducted short-term SST forecasts in the Bohai Sea region with forecast lead times ranging from 1 to 10 days and compared its efficacy against several benchmark models, including ConvLSTM, PredRNN, TCTN, and SwinLSTM. Our experimental results show that the proposed model outperforms all of these models in terms of multiple evaluation metrics short-term SST prediction. The proposed model offers a new predictive learning method for improving the accuracy of spatiotemporal predictions in various domains, including meteorology, oceanography, and climate science.

Sea Surface Temperature (SST) prediction has attracted increasing attention due to its critical role in climate change. Traditional SST prediction methods can be mainly divided into two types, the physics-based numerical methods and the data-driven methods. However, the above methods have certain limitations, the former type can not perform well when the physical prior information is incomplete, while latter type can not perform well when the training data is insufficient. This paper uses a deep neural network to extract some valuable information from the data, and then introduces the space-time partial differential equation (PDE) to model the prior physical information referring to SST. By incorporating them together, a new Space-Time PDE-guided Neural Network (STPDE-NET), which can better deal with the prior physical information incompleteness and data insufficiency problems mentioned above is proposed. In the experiments, we compare our STPDE-NET with several famous or state-of-the-art SST prediction methods. The experimental results show that STPDE-NET outperforms the compared methods in most SST prediction circumstances, especially when the training data is insufficient.

Sea Surface Temperature (SST) prediction is a hot topic that has received tremendous popularity in recent years. Existing methods for SST prediction usually select one sea area of interest and conduct SST prediction by learning the spatial and temporal dependencies and patterns in historical SST data. However, global SST is a unified system of high regionality, and the SST in different sea areas shows different changing patterns due to the influence of various factors, e.g., geographic location, ocean currents and sea depth. Without a good understanding of such regionality of SST, we cannot quantitatively integrate the regionality information of SST into SST prediction models to make them adaptive to different SST patterns around the world and improve the prediction accuracy. To address this issue, we proposed the Multi-Stage Spatio–Temporal Clustering (MuSTC) method to quantitatively identify sea areas with similar SST patterns. First, MuSTC sequentially learns the representation of long-term SST with a deep temporal encoder and calculates the spatial correlation scores between grid ocean regions with self-attention. Then, MuSTC clusters grid ocean regions based on the original SST data, encoded long-term SST representation and spatial correlation scores, respectively, to obtain the sea areas with similar SST patterns from different perspectives. According to the experiments in three ocean areas, i.e., the North Pacific Ocean (NPO), the South Atlantic Ocean (SAO) and the North Atlantic Ocean (NAO), the clustering results generally match the distribution of ocean currents, which demonstrates the effectiveness of our MuSTC method. In addition, we integrate the clustering results into two representative spatio–temporal prediction models, i.e., Spatio–Temporal Graph Convolutional Networks (STGCN) and Adaptive Graph Convolutional Recurrent Network (AGCRN), to conduct SST prediction. According to the results of experiments, the integration of regionality information leads to the reduction of Root Mean Square Error (RMSE) by 1.95%, 1.39% and 1.28% in NPO, SAO and NAO, respectively, using the STGCN model, and the reduction of RMSE by 4.94%, 0.74% and 1.43% by using the AGCRN model. Such results indicate that the integration of regionality information could notably improve the prediction accuracy of SST.

Prediction of sea surface temperature (SST) is desired for several applications ranging from climate studies to maintenance of coastal eco-system. Such prediction with the help of artificial, or simply, neural network has by now fairly stabilized. However corresponding studies are mostly applicable only to a specified single location. In this study we have expanded them to cover an entire sea basin. The basin under consideration is Bay of Bengal (BoB) located on the east side of the Indian peninsula. We have predicted SST at the daily time scale using time series approach in which we feed a selected length of past daily SST observations to the neural network and derive the predicted value of SST at multiple lead times (days) as output. The gridded NOAA v2 high resolution dataset derived from satellites was used for this purpose. At every grid in the BoB feed forward back propagation type of neural network was developed. The networks were trained using 70% of data and tested with the help of remaining 30%. The performance in testing of such large spatial-scale networks was judged on the basis of the error statistics of correlation coefficient, ‘r’, and root mean square error, RMSE. The prediction skill of ANN models were found to be very good at shorter lead times (1-3 days) and reasonably good at higher lead times (4-7 days). Apart from that, these ANN models were also evaluated for their performance during extreme weather events which are peculiar to BoB region and found to be capturing such events in advance with sufficient time. Overall therefore it is claimed that the basin-scale neural networks developed in this study can not only carry out multiple time step predictions of daily SST at individual grid points simultaneously but can also predict basin scale weather phenomena in advance.

We develop a memory graph convolutional network (MGCN) framework for sea surface temperature (SST) prediction. The MGCN consists of two memory layers: one graph layer and one output layer. The memory layer captures SST temporal changes via temporal convolution units and gate linear units. The graph layer encodes SST spatial changes in terms of characteristics derived from graph Laplacian. The output layer encapsulates information from the previous layers and produces SST prediction results. The MGCN characterizes both the temporal and spatial changes, rendering a comprehensive SST prediction strategy. We use daily mean SST data for two areas near the Bohai Sea and the East China Sea for experimental evaluations and validate that the MGCN performs better than other traditional machine learning methods for nearshore SST prediction. In addition, we test the MGCN on weekly and monthly mean SST datasets and validate that the MGCN is robust and suitable for SST prediction.

Sea surface temperature (SST) prediction plays an important role in scientific research, environmental protection, and other marine-related fields. However, most of the current prediction methods are not effective enough to utilize the spatial correlation of SSTs, which limits the improvement of SST prediction accuracy. Therefore, this paper first explores spatial correlation mining methods, including regular boundary division, convolutional sliding translation, and clustering neural networks. Then, spatial correlation mining through a graph convolutional neural network (GCN) is proposed, which solves the problem of the dependency on regular Euclidian space and the lack of spatial correlation around the boundary of groups for the above three methods. Based on that, this paper combines the spatial advantages of the GCN and the temporal advantages of the long short-term memory network (LSTM) and proposes a spatiotemporal fusion model (GCN-LSTM) for SST prediction. The proposed model can capture SST features in both the spatial and temporal dimensions more effectively and complete the SST prediction by spatiotemporal fusion. The experiments prove that the proposed model greatly improves the prediction accuracy and is an effective model for SST prediction.

The accurate prediction of sea surface temperature (SST) is the basis for our understanding of local and global climate characteristics. At present, the existing sea temperature prediction methods fail to take full advantage of the potential spatial dependence between variables. Among them, graph neural networks (GNNs) modeled on the relationships between variables can better deal with space–time dependency issues. However, most of the current graph neural networks are applied to data that already have a good graph structure, while in SST data, the dependency relationship between spatial points needs to be excavated rather than existing as prior knowledge. In order to predict SST more accurately and break through the bottleneck of existing SST prediction methods, we urgently need to develop an adaptive SST prediction method that is independent of predefined graph structures and can take full advantage of the real temporal and spatial correlations hidden indata sets. Therefore, this paper presents a graph neural network model designed specifically for space–time sequence prediction that can automatically learn the relationships between variables and model them. The model automatically extracts the dependencies between sea temperature multi-variates by embedding the nodes of the adaptive graph learning module, so that the fine-grained spatial correlations hidden in the sequence data can be accurately captured. Figure learning modules, graph convolution modules, and time convolution modules are integrated into a unified end-to-end framework for learning. Experiments were carried out on the Bohai Sea surface temperature data set and the South China Sea surface temperature data set, and the results show that the model presented in this paper is significantly better than other sea temperature model predictions in two remote-sensing sea temperature data sets and the surface temperature of the South China Sea is easier to predict than the surface temperature of the Bohai Sea.

Accurate sea surface temperature (SST) prediction is vital for disaster prevention, ocean circulation, and climate change. Traditional SST prediction methods, predominantly reliant on time-intensive numerical models, face challenges in terms of speed and efficiency. In this study, we developed a novel deep learning approach using a 3D U-Net structure with multi-source data to forecast SST in the South China Sea (SCS). SST, sea surface height anomaly (SSHA), and sea surface wind (SSW) were used as input variables. Compared with the convolutional long short-term memory (ConvLSTM) model, the 3D U-Net model achieved more accurate predictions at all lead times (from 1 to 30 days) and performed better in different seasons. Spatially, the 3D U-Net model’s SST predictions exhibited low errors (RMSE < 0.5 °C) and high correlation (R > 0.9) across most of the SCS. The spatially averaged time series of SST, both predicted by the 3D U-Net and observed in 2021, showed remarkable consistency. A noteworthy application of the 3D U-Net model in this research was the successful detection of marine heat wave (MHW) events in the SCS in 2021. The model accurately captured the occurrence frequency, total duration, average duration, and average cumulative intensity of MHW events, aligning closely with the observed data. Sensitive experiments showed that SSHA and SSW have significant impacts on the prediction of the 3D U-Net model, which can improve the accuracy and play different roles in different forecast periods. The combination of the 3D U-Net model with multi-source sea surface variables, not only rapidly predicted SST in the SCS but also presented a novel method for forecasting MHW events, highlighting its significant potential and advantages.

Sea surface temperature (SST) prediction has attracted increasing attention, due to its crucial role in understanding the Earth’s climate and ocean system. Existing SST prediction methods are typically based on either physics-based numerical methods or data-driven methods. Physics-based numerical methods rely on marine physics equations and have stable and explicable outputs, while data-driven methods are flexible in adapting to data and are capable of detecting unexpected patterns. We believe that these two types of method are complementary to each other, and their combination can potentially achieve better performances. In this paper, a space-time partial differential equation (PDE) is employed to form a novel physics-based deep learning framework, named the space-time PDE-guided neural network (STPDE-Net), to predict daily SST. Comprehensive experiments for SST prediction were conducted, and the results proved that our method could outperform the traditional finite-difference forecast method and several state-of-the-art deep learning and physics-guided deep learning methods.

Sea surface temperature (SST) prediction is not only theoretically important but also has a number of practical applications across a variety of ocean-related fields. Although a large amount of SST data obtained via remote sensor are available, previous work rarely attempted to predict future SST values from history data in spatiotemporal perspective. This letter regards SST prediction as a sequence prediction problem and builds an end-to-end trainable long short term memory (LSTM) neural network model. LSTM naturally has the ability to learn the temporal relationship of time series data. Besides temporal information, spatial information is also included in our LSTM model. The local correlation and global coherence of each pixel can be expressed and retained by patches with fixed dimensions. The proposed model essentially combines the temporal and spatial information to predict future SST values. Its structure includes one fully connected LSTM layer and one convolution layer. Experimental results on two data sets, i.e., one Advanced Very High Resolution Radiometer SST data set covering China Coastal waters and one National Oceanic and Atmospheric Administration High-Resolution SST data set covering the Bohai Sea, confirmed the effectiveness of the proposed model.

The effective prediction of sea surface temperature has become a hot topic of scientific research. The hidden characteristics and time characteristics of sea surface temperature are studied, and a new sea surface temperature prediction model SLA based on decomposition is proposed in this paper. Firstly, the trend term and periodic seasonal term of sea surface temperature were decomposed by STL decomposition algorithm, which improved the problem of single characteristics of sea surface. Secondly, on the basis of the calculation method of the improved LSTM model and the attention mechanism, the problem that the prediction of sea surface temperature cannot be remembered in the long term and the influence weight of historical information is ignored is solved. Finally, the sea surface temperature of the South China Sea (112.125E,10.125N) was taken as the research object, and the experimental verification was carried out in three data sets. The experimental results show that the SLA model proposed in this paper has good prediction accuracy.