An improved hydrological forecast method based on chaos and trend term

In this paper, an improved hydrological medium and long-term forecast method based on chaos and trend term is presented. First, the periodic term and trend term of the hydrological time series in the nearest years are analyzed and then using chaotic characteristic analysis for a difference of the residual time series. Finally, the adaptive network-based fuzzy inference system is applied to predict the residual series for obtaining the prediction result. The computation on an actual case shows that the new method is reliable to not only natural flat flow year but also special low flow year.