将奇异谱分析(SSA)法应用于新疆维吾尔自治区伊犁河雅马渡站年径流预报,分析了流域多年径流时间序列奇异谱,提取年径流序列的时间主成分量,重构了原序列中的周期震荡成分,获得了径流时间序列规律,并采用时间序列分析中的AR(p)自回归模型预测了流域年径流序列。实例结果表明,SSA法具有较好的实用性,为深入研究流域径流规律提供了依据。 The method of singular spectrum analysis (SSA) was applied to the annual runoff forecast of Yalongdu station in Xinjiang Uygur Autonomous Region. The analysis of the singular spectrum of multi-year runoff time series of watershed, the time principal component of annual runoff sequence was extracted, Periodic turbulence components were obtained, and the runoff time series law was obtained. The AR (p) autoregressive model in time series analysis was used to predict the annual runoff series. The experimental results show that the SSA method has good practicability and provides the basis for further study on the runoff laws in the watershed.