【摘要】 为了合理评估区域干旱灾害风险,以岷江流域14个气象站1961-2012年逐月降水数据为研究对象,以阈值为变量设定3种游程截取水平识别干旱,然后利用经拟合优选的Copula函数构建干旱历时与干旱烈度的联合分布,并进行概率分析和重现期计算。结果表明:Gumbel Copula函数对整个流域的拟合效果最优;利用Gumbel Copula函数分析二维条件概率P(S≤60mm|D≥3月),3种截取水平的发生概率从流域东南到西北递增;联合重现期结果显示流域东北部和南部一带的旱情比中西部严重,而同现重现期结果则显示流域西部的旱情比东部严重;联合重现期比同现重现期的空间分布更稳定,任意一种截取水平下的联合重现期都比较接近。研究结果可为游程理论截取水平设定、Copula函数优选以及多变量干旱分析提供依据。更多还原

【Abstract】 In order to assess the risk of regional drought disaster reasonably,we took the monthly precipitation data of 14 meteorological stations in Minjiang River Basin from 1961 to 2012 as the research object,and,with threshold as a variable,set three cutting levels according to the theory of runs to identify drought events.Then we used the best fit Copula function to construct the joint distribution of drought duration and severity and analyzed the probabilistic characteristics and return period.The results showed that Gumbel Copula function demonstrated the best fitness for Minjiang River Basin.After the analysis of the spatial distribution of conditional probability P(S≤60mm|D≥3 months),the results indicated that the probability of occurrence on the three cutting levels increased from the southeast to the northwest of the basin.The results of joint return period demonstrated that the drought in the northeastern and southern parts of the basin was more serious than that in the central and western regions,while the results of co-occurrence return period showed that the drought in the west of the basin was more serious than that in the east.The spatial distribution characteristics of joint return period were more stable than those of co-occurrence return period.The joint return periods of the whole basin on each cutting level were relatively close.This study could provide a basis for setting the cutting level of the theory of runs,choosing optimal Copula function,and conducting multivariate drought analysis.更多还原