Predicting the interannual variability of the subtropical high over the western Pacific Ocean based on the improved information diffusion model

Abstract

Because of its special location and structure, the Western Pacific Subtropical High (WPSH) influences greatly the climate and weather in East Asia, especially the summer precipitation. To clarify how the interannual variability (IAV) of the WPSH is related to anomalies in the tropical sea surface temperature (SST) and atmospheric circulation, time series of the intensity index of the WPSH are subjected to wavelet analysis, showing IAV in the index. Characteristic indexes are defined for three key sea areas and the equatorial-latitude westerly region. After a continuous wavelet transform, the oscillation period of them is similar to that of the WPSH. The cross-wavelet transform of the four regional and two WPSH indexes is used to obtain the corresponding time-delay correlation. Regarding the potential correlation, WPSH weakening leads to strengthening of the westerly wind and then affects the rise of SST in the eastern equatorial Pacific Ocean. At the same time, warm water moves eastward. This gradually increases the SST in the equatorial central Pacific and warm pool area and then strengthens the WPSH under the action of the Hadley circulation. From the above analysis, a model for predicting the IAV of the WPSH intensity index is established based on the information diffusion model improved by a genetic algorithm. An experiment is conducted to predict the IAV of the WPSH intensity index, and the results show that the prediction model is accurate in predicting the IAV trend, with good prediction for 84 months. The mean absolute percentage error is 14.44% and the correlation coefficient is 0.8507. Also, the normal and abnormal years of the WPSH are used as different starting points for different prediction experiments. However, the different starting points have little influence on the predictions, showing the stability of the model. Studying the IAV of the WPSH provides a strong theoretical and scientific basis for predicting its abnormal interannual behavior and offers the prospect of socially important disaster prevention and mitigation.