基于功率谱周期和去趋势波动分析的河川径流特性分析

周期性和长程相关性是河川径流序列演变和波动的重要特性.论文以泾河流域杨家坪、雨落坪和张家山三个水文站1956—2001 年天然径流序列为研究对象,采用功率谱周期和去趋势波动分析方法对天然径流序列进行分析,结果表明:① 张家山、杨家坪和雨落坪站的天然径流序列均为“白噪声”过程,且具有0.05 置信水平下显著的2.2 a 周期,径流量与降水量有较好的相关性;② 三个水文站的天然径流具有相同的标度不变区间(τ介于6~30 个月),标度指数α分别为0.549 3(张家山)、0.555 2(雨落坪)、0.554 9(杨家坪),均大于0.5,表明三个水文站的天然径流具有长程相关性(持久性),可能影响着径流2.2 a显著周期的形成.

Runoff Characteristics Analysis Based on Power Spectrum Method and Detrended Fluctuation Analysis

Periodicity and long-range correlation are two important characteristics of the evolution and fluctuation of runoff series. This study took the natural runoff sequences over the period of 1956-2001 at three hydrologic stations (Yangjiaping, Yuluoping and Zhangjiashan) in the Jinghe Watershed as research objects. The power spectrum method and the detrended fluctuation analysis were employed to analyze the natural runoff sequences at the three stations. Results demonstrate that all of the natural runoff sequences from the three stations are“white noise”processes. Under the 0.05 confidence level, there exists a prominent period of 2.2 years for the natural sequences from the three stations. There is a good correlation between the runoff series and the precipitation. The natural runoff sequences at the three stations have the same scale invariant interval (the variable τ is between 6 and 30 months). The related scaling exponents α are greater than 0.5 (0.5493 for Zhangjiashan Station, 0.5552 for Yuluoping Station, and 0.5549 for Yangjiaping Station), which shows that there is long-range correlation or persistence of the natural runoff at the three stations. The long-range correlation may associate with the formation of 2.2-year prominent period of natural runoff sequences.