摘要
枯水是包含枯水历时、缺水量和最小枯水流量等多个特征变量的水文事件,现行方法常常选取其中某一个特征值假定其服从某一指定分布来进行单变量频率分析,使得枯水事件的真实特性难以得到全面反映。采用多变量核密度估计方法对枯水径流进行两变量的频率分析,从而避开了水文变量分布线型和相依型式的选择。以长江宜昌站126 a径流系列为例,研究长江上游枯水频率的特征。结果表明,近10 a长江上游枯水发生的频次和严重性都高于多年平均水平,但还未达到历史最枯年组的水平。综合考虑生态和航运的要求,确定河道在枯水期允许的最小流量,分析不同枯水历时条件下出现低于河道允许最小流量的概率。
Low flow is a hydrologic event that contains several characteristics, and its severity can be represented by low flow duration, deficit and the minimum flow during the dry period. Conventional approach to analyze low flow frequency is the univariate method by assuming an alternative probability distribution for one selected variable. Bivariate kernel estimation is proposed to analyze the low flow frequency, avoiding the prior assumptions of probability distribution and the form of dependence. The 126 - year daily flow data series at the Yichang station in the Yangtze River is selected as case study. Results suggest that both the low flow frequency and severity for the last ten years exceed the average annual level, but not yet reach the most severity level of the observed continues dry years. The permitted minimum flow during the dry period is obtained for the consideration of both ecology and shipping. Occurrence probabilities of runoff lower than permitted minimum now under different low flow durations are analyzed.
出处
《人民珠江》
2017年第4期15-19,共5页
Pearl River
基金
国家重点研发计划项目(2016YFC0402708)
中央高校基本科研业务费资助(HUST:2015QN216)
关键词
两变量
核估计
枯水
频率分析
长江上游
bivariate
kernel estimation
low flow
fi＇equency analysis
upper Yangtze River basin
