论文部分内容阅读
通过构建多变量联合分布进行干旱分析,可揭示干旱的演变规律。根据新疆乌鲁木齐和石河子气象站的长系列月降水资料,提取干旱历时、干旱烈度和烈度峰值3个干旱特征变量,基于4种对称的Archimedean Copulas函数分别构建二维、三维干旱变量的联合分布;基于5种非对称的Archimedean Copulas函数构建三维干旱变量的联合分布,以进一步推求各自的重现期。经拟合优度检验,Frank Copula函数对干旱历时和干旱烈度、干旱历时和烈度峰值的二维联合分布的拟合度最好;Clayton Copula函数对于干旱烈度和烈度峰值的二维联合分布以及干旱历时、干旱烈度和烈度峰值的三维联合分布拟合效果最佳。单变量的重现期介于二维、三维变量联合重现期与同现重现期之间。表明Copulas函数能描述多维干旱特征变量的联合分布。
By constructing a multivariate joint distribution for drought analysis, the evolution of drought can be revealed. According to the long series of monthly precipitation data of Urumqi and Shihezi stations in Xinjiang, three drought characteristic variables such as drought duration, drought intensity and peak intensity were extracted. Based on the four symmetric Archimedean Copulas functions, the joint distribution of 2D and 3D drought variables was constructed respectively. Five Asymmetric Archimedean Copulas Functions Construct Joint Distribution of Three-dimensional Drought Variables to Further Estimate Their Recurrences. According to the goodness-of-fit test, Frank Copula function is the best fit for the two-dimensional joint distribution of drought duration and drought intensity, drought duration and intensity peak; the two-dimensional joint distribution of Clayton Copula function to drought intensity and intensity peak and drought The three-dimensional joint distribution of duration, drought intensity and intensity peak fitted the best. Recurrence of univariate between the two-dimensional and three-dimensional variables between the joint recurrence and co-occurrence period. It shows that Copulas function can describe the joint distribution of multidimensional drought characteristic variables.