Author(s): Ping Su; Cuiping Kuang; Jie Gu; Jianwen Qi; Binyu Wang; Congrui Qian
Linked Author(s): Cuiping Kuang
Keywords: Datong station; Wavelet transform; Runoff series; Periodic characteristics
Abstract: Runoff series of the Yangtze River presents an intricate variation tendency under the reinforced influence of human activities. The Morlet Wavelet Transform method has been applied to analyze the annual runoff data from 1950 to 2011 at Datong Station in the Yangtze River, it can clearly reveal the multi-time scales structure, break point, change and distribution of periodic variation in the different time scales of the runoff series. The main conclusions are that: 1) Repeated periodic oscillations accompanied a extremely large fluctuation are presented in the runoff series with an obvious wet and dry year change, the major periods of the time series are about 3, 8, 16 and 23 years respectively. Among them, the presented maximum periodic oscillation is the 23 years scale. The 16 and 8 years scales also present intense periodic variation. 2) The variant process of the runoff series show that there are several periods emerging in the different time scales. In the entire study period, there are roughly 2, 4 and 8 periods corresponding to the 23, 16 and 8 year time scales respectively. The fluctuations of 16 and 23 year scales are quite smooth and the 8 year scale’s fluctuation has weakened visibly since the 1990s. It indicates that the 16 and 23 year scales play the dominant roles in the recent years. 3) In the 23 years time scale, the wet periods are 1950-1958, 1969-1980 and 1992 -2003, the dry periods are 1959-1968, 1981-1991 and 2004-2011. Accordingly, the break points are 1959, 1969, 1981, 1992 and 2004 respectively. In the 16 years time scale, 1957, 1965, 1972, 1980, 1987, 1995 and 2004 are the turning points of wet and dry years. 4) It can be predicted from the view of long time scales that the low annual runoff will successively occur in the recent future, and this study also indicates that the Morlet Wavelet Transform method has a significant capability on the analysis of long-term time-series runoff and the potential forecasting in the future.
Year: 2013