以渭河流域西安站77年的月降水资料为例,采用7种单参数族Copulas函数,建立干旱特征变量之间的联合分布。经拟合优度评价,Plackett Copula函数对干旱历时和干旱烈度、干旱历时和干旱烈度峰值之间拟合程度最好;而Clayton Copula函数对于干旱烈度和烈度峰值拟合程度最好。分别应用这2种Copulas函数对西安站进行干旱特征变量之间的二维条件概率分布及组合重现期计算。结果表明,Copulas函数能较好的描述干旱特征变量间的联合分布;并且具有灵活性和应用范围广等特点,在多变量干旱特征分析中具有广泛的应用前景。
This paper taking 77 years of monthly precipitation data of Xi'an gauge Station which locates in Wei River Basin as an example,using seven kinds of one-parameter family Copula function to construct the joint distribution of drought variables.The goodness-of-fit test shows that Plackett Copula can best describe the relationship between drought duration and drought severity as well as the relationship between drought severity and drought severity peak;Clayton Copula can best describe the relationship between drought severity and drought severity peak.Using Plackett Copula and Clayton Copula respectively,these two dimensional conditional probability distribution and joint return period between drought variables are calculated.The results show that,Copula function can better describe the joint distribution of drought variables;and has the flexibility and wide range of application,which has a broad application prospects in the multivariate analysis of drought characterization.