Hybrid improved empirical mode decomposition and BP neural network model for the prediction of sea surface temperature

Abstract

Abstract. Sea surface temperature (SST) is the major factor that affects the ocean–atmosphere interaction, and in turn the accurate prediction of SST is the key to ocean dynamic prediction. In this paper, an SST-predicting method based on empirical mode decomposition (EMD) algorithms and back-propagation neural network (BPNN) is proposed. Two different EMD algorithms have been applied extensively for analyzing time-series SST data and some nonlinear stochastic signals. The ensemble empirical mode decomposition (EEMD) algorithm and complementary ensemble empirical mode decomposition (CEEMD) algorithm are two improved algorithms of EMD, which can effectively handle the mode-mixing problem and decompose the original data into more stationary signals with different frequencies. Each intrinsic mode function (IMF) has been taken as input data to the back-propagation neural network model. The final predicted SST data are obtained by aggregating the predicted data of individual series of IMFs (IMFi). A case study of the monthly mean SST anomaly (SSTA) in the northeastern region of the North Pacific shows that the proposed hybrid CEEMD-BPNN model is much more accurate than the hybrid EEMD-BPNN model, and the prediction accuracy based on a BP neural network is improved by the CEEMD method. Statistical analysis of the case study demonstrates that applying the proposed hybrid CEEMD-BPNN model is effective for the SST prediction. Highlights include the following: Highlights. An SST-predicting method based on the hybrid EMD algorithms and BP neural network method is proposed in this paper. SST prediction results based on the hybrid EEMD-BPNN and CEEMD-BPNN models are compared and discussed. A case study of SST in the North Pacific shows that the proposed hybrid CEEMD-BPNN model can effectively predict the time-series SST.