论文部分内容阅读
在石羊河上游山区采集流域树木年轮样芯,建立高分辨率宽度年表。分析年表、降水和径流量间的相关关系,在它们间良好相关性基础上,分别利用多元逐步回归和提取主分量再回归方法重建金塔河和西营河历史时期上年9月至当年8月年径流量序列,以逐一剔除和交叉检验验证回归方程的稳定性和可靠性。1880年以来,石羊河年径流量共经历了4个枯水阶段和3个丰水阶段,而枯水年与丰水年出现的概率大致相同,且远低于平水年出现的概率。用多窗谱分析径流量序列周期变化,发现在不同置信水平上,年径流量存在2.58、2.76、3.28、3.53、4.14和22.2 a的周期。
In the upper reaches of the Shiyang River mountain watershed sampling tree-like cores, the establishment of high-resolution width chronology. Analyze the correlation between chronology, precipitation and runoff. On the basis of good correlation between them, use multi-step stepwise regression and principal component retrieval respectively to rebuild the Jinta River and Xiyinghe Historical Period From September to the current year August Annual runoff series, with one by one elimination and cross-validation of the regression equation to verify the stability and reliability. Since 1880, the annual runoff of Shiyang River has experienced 4 dry periods and 3 wet periods, while the probabilities of dry years and wet years are about the same, much lower than that of normal years. Using multi-window analysis to analyze the periodic variation of runoff series, we find that the annual runoff has periods of 2.58, 2.76, 3.28, 3.53, 4.14 and 22.2 a at different confidence levels.