Analysis on complexity of monthly runoff series based on sample entropy in the Dongjiang river

An index of nonlinear dynamics based on sample entropy, the paper has analyzed spatial distribution and dynamic change of runoff series complexity with the long-term monthly runoff series of Longchuang, Heyuan and Boluo stations in the Dongjiang river. The results showed that, complexity of the monthly runoff series has an obvious spatial difference, and shows an increasing trend from upstream to downstream in the Dongjiang river. As the whole, runoff series complexity has a great dynamic change and periodic fluctuation, which reflects the structure change of hydrological dynamics system caused by environmental change, espe- cially influence of human activities. In addition, there are negative relationship between average of runoff sliding window and the corresponding sample entropy, and their peak-to-valley value shows the well corresponding relationships, which has an important significance for research of runoff series evolution and forecast.