Prediction of Sea Surface Temperature in the China Seas Based on Long Short-Term Memory Neural Networks

Abstract

Sea surface temperature (SST) in the China Seas has shown an enhanced response in the accelerated global warming period and the hiatus period, causing local climate changes and affecting the health of coastal marine ecological systems. Therefore, SST distribution prediction in this area, especially seasonal and yearly predictions, could provide information to help understand and assess the future consequences of SST changes. The past few years have witnessed the applications and achievements of neural network technology in SST prediction. Due to the diversity of SST features in the China Seas, long-term and high-spatial-resolution prediction remains a crucial challenge. In this study, we adopted long short-term memory (LSTM)-based deep neural networks for 12-month lead time SST prediction from 2015 to 2018 at a 0.05° spatial resolution. Considering the sub-regional differences in the SST features of the study area, we applied self-organizing feature maps (SOM) to classify the SST data first, and then used the classification results as additional inputs for model training and validation. We selected nine models differing in structure and initial parameters for ensemble to overcome the high variance in the output. The statistics of four years’ SST difference between the predicted SST and Operational SST and Ice Analysis（OSTIA）data shows the average root mean square error (RMSE) is 0.5 °C for a one-month lead time and is 0.66 °C for a 12-month lead time. The southeast of the study area shows the highest predictable accuracy, with an RMSE less than 0.4 °C for a 12-month prediction lead time. The results indicate that our model is feasible and provides accurate long-term and high-spatial-resolution SST prediction. The experiments prove that introducing appropriate class labels as auxiliary information can improve the prediction accuracy, and integrating models with different structures and parameters can increase the stability of the prediction results.

High Precision Sea Surface Temperature Prediction of Long Period and Large Area in the Indian Ocean Based on the Temporal Convolutional Network and Internet of Things

“…Generally speaking, the error increased with the increase of model prediction time. Lei G. [9] found that the accuracy of prediction one month in advance was the highest, with an error of 0.5 °C; When predicting 2 months in advance, RMSE increased to 0.59 °C; When predicting 12 months in advance, RMSE reached a peak of about 0.66 °C. So, it is obvious that with the increase of lead time, RMSE increases slowly.…”

“…Xiao et al (2019) [8] established an LSTM model in the East China Sea using 36 years of spaceborne sea surface temperature data; the model is accurate for the daily prediction of the short-term and medium-term sea surface temperature field. L. Guan et al (2020) [9] divided the entire China Sea and its adjacent area into 130 small regions using the self-organizing map algorithm, constructed an LSTM model for each region to predict its SST, and found that the root-mean-square error (RMSE) of the forecasts at 1 month in advance was 0.5 • C. In summary, previous studies have mostly used data sets to make short-term predictions of regional SST, in which the selected feature is rather simple. The prediction of sea temperature over large areas of water and a long period has been rarely investigated.…”

“…The prediction of sea temperature over large areas of water and a long period has been rarely investigated. Moreover, as the forecast period is extended, the accuracy of the existing methods decreases [9]. Compared with the above, this paper used a TCN model to accurately predict long-term SSTs of a large area in the Indian Ocean.…”

Marine Data Prediction: An Evaluation of Machine Learning, Deep Learning, and Statistical Predictive Models

“…In [ 37 ], the authors proposed building a predictive model for predicting the SST of the entire China Sea. They utilized collected data over 12 months.…”

Sustainable Marine Ecosystems: Deep Learning for Water Quality Assessment and Forecasting

“…The proposed scheme is able to outperform other conventional benchmarks with a reduced input sequence length. The authors in [99] adopt an LSTM for predicting sea surface temperature over the China Seas for 12-month lead time. Considering the sub-regional feature differences within the study area, they use self-organizing feature maps to classify the data first, and then use the classification results as additional inputs for the DL network.…”