基于1991 年- 2008 年香溪河流域降雨量和径流量资料, 将最大熵2Copula(M EC) 方法应用于流域降雨- 径流 相关性分析中。结果表明: MEC 方法生成的边缘分布能够有效拟合降雨量和径流量的观测值, 且在边缘分布确定 的过程中无需对线型进行假设, 以均值、标准差、偏度和峰度作为约束也能够很好的体现变量的统计特征; Gumbel2 Ho ug aard Copula 函数对于降雨量和径流量联合分布的拟合效果较好; 香溪河流域降雨量和径流量存在上尾相关 性, 这对防洪排涝有着重要的作用, 同时也为流域工程规划设计工作提供了参考依据。

Based on the rainfall and runo ff in Xiang xi River Wat er shed, this paper applied a maximum ent ro py2co pula ( MEC) metho d to study the cor relation between the rainfall and r unoff o f wat er shed. Results show that: the generated marg inal distr i2 butio ns o f the rainfall and runo ff can effectiv ely fit the obser ved v alues w ithout any hy po thesis, and t he statistical character istics of r andom v ariables can be r eflected w ell as the constraints o f mean, standar d dev iatio n, skew ness and kur tosis; Among all the Co pula functio ns used, Gumbel ho ug aard copula can be used to generate the jo int distributio n of the rainfall and r unoff as the best effect o f fitting; the upper tail dependence ex ists betw een the rainfall and runo ff in Xiang xi River Water shed, w hich not only can play a gr eat ro le in f loo d and drainag e contr ol in Xiang x i River Wat ershed, but also pr ov ide refer ence basis fo r water shed project planning and desig n w o rk.

福建省教育厅A 类项目( JA15371) ; 厦门理工学院科学技术研究项目( YKJ3018R, XYK201419)