淮河上游紫罗山流域枯水频率及退水规律分析

Analysis of dry season runoff frequency and regression law in Ziluoshan catchment,Huaihe River

应用枯季径流退水规律预报枯季径流总量和径流过程是枯季径流预报最常用、有效的方法。但目前,尚未有一种通用的分布可以很好地模拟所有枯水径流。以淮河上游紫罗山流域为研究对象,以1952~2010年各年最小7 d平均流量作为枯季径流统计特征序列,采用P-Ⅲ,LP-Ⅲ,GLO,GPA,LN3和GEV 6种分布,利用适线法、矩法等进行参数估计,推求经验频率对应的理论线型计算值。利用概率点据相关系数、四阶线性矩系数、确定性系数,均方根误差和平均绝对误差5个特征统计量,比较各分布线型拟合情况,确定该流域枯水频率最优拟合线型为P-Ⅲ分布。此外,利用退水曲线法推求各年份退水系数与多年综合退水系数。结果表明:流域年退水系数在0.007~0.200范围内波动,其变化与降水无关,随着最小7 d平均流量的变小而增大,流域多年综合退水系数为0.08。研究成果对合理开发利用水资源、保护河流生态系统平衡具有现实意义。

The runoff regression law is the commonly used to forecast runoff and process in dry season, but at present, there is not yet a general distribution that can well simulate all dry runoff. This paper took Ziluoshan catchment in the upper reaches of Huaihe River Basin as the research area. The annual minimum 7-day average discharges from 1952 to 2010 were used as the statistical characteristic sequence of runoff in dry season. The curve-fitting method, the moments method and six three-parameter distributions including Pearson-III distribution(P-III), Logarithmic Pearson-III distribution(LP-III), Generalized Logistic distribution(GLO), Generalized Pareto distribution(GPA), Logarithmic Normal distribution(LN3), and Generalized Extreme Value distribution(GEV) were used to estimate the parameters and calculate the theoretical linear value of the empirical frequency. Five characteristic statistics including probability plot correlation coefficient, fourth-order linear moment coefficient, deterministic coefficient, root mean square error and average absolute error were used to compare the line fitting for the six distributions. Then P-III was regarded as the optimal distribution for dry season runoff frequency in this catchment. In addition, the annual and multi-year-comprehensive coefficients of flow recession were calculated by using regression hydrograph method. The results showed that the annual flow recession coefficients fluctuated from 0.007 to 0.2, showing no correlation with annual precipitation. It increased with the decrease of minimum 7-day average flow. The multi-year-comprehensive coefficient of flow recession in this catchment was 0.08. This study has practical significance for developing and utilizing water resources rationally and protecting river ecosystem balance.