A new dynamical forecasting model of Western Pacific subtropical high ridge line index based on the improved self-memorization principle and forecast experiments

Aiming at tackling the problem of inaccurate long-term Western Pacific Subtropical High (WPSH) forecasts, a new dynamical forecasting model of WPSH ridge line index (RI) is developed based on the concept of dynamical model reconstruction and improved self-memorization principle. To overcome the problem of single initial prediction value, the largest Lyapunov exponent is introduced to improve the traditional self-memorization function, making it more appropriate to describe the chaotic systems, such as WPSH; and the equation reconstructed by actual data is used as its dynamical core, getting rid of the problem existing in traditional equation that dynamical core set is relatively simple. The developed dynamical forecasting model of RI index is used to predict WPSH strength in the long term. 17 experiments of the WPSH abnormal years and normal years are performed, and forecast results within 25 months are found to be good, with a correlation coefficient of about 0.80 and root mean square error under 8%, showing that the improved model has yielded satisfactory long-term forecasting results. Additional experiments for predicting the area index (AI) and the west ridge point index (WI) are also performed to demonstrate that our method is effective for complete prediction of WPSH, especially the aberrance of the subtropical high can be drawn and forecast. The mechanism for the occurrence and development of WPSH is complex, and this paper offers a new thought for WPSH forecast research.