The Gaussian copula model for the joint deficit index for droughts.
The Gaussian copula model for the joint deficit index for droughts.
复制标题
DOI：
10.1016/j.jhydrol.2018.03.064
复制
发表时间：
2018-06
影响因子：
6.4
通讯作者：
H. Vyver;J. V. D. Bergh
中科院分区：
文献类型：
--
作者：
H. Vyver;J. V. D. Bergh
干旱及其影响的特征在很大程度上取决于所涉及的时间尺度。为了获得全面的干旱评估，需要一起审查不同时期缺水的累积影响。例如，最近开发的联合赤字指数(JDI)是基于从1个月到12个月的各种时间尺度上的多变量降水概率，并且是根据经验的Copula构建的。在本文中，我们研究了JDI的高斯Copula模型。我们用两参数函数来模拟时间尺度上的协方差，该函数通常用于空间统计或地质统计学的特定背景下。用长期降水量序列验证了协方差模型的有效性。Bootstrap实验表明，在干旱严重程度评估中，高斯Copula模型比经验Copula方法具有优势：(1)它能够量化经验Copula范围之外的干旱；(2)提供了充分的干旱量化；(3)更好地理解了估计中的不确定性。
The characterization of droughts and their impacts is very dependent on the time scale that is involved. In order to obtain an overall drought assessment, the cumulative effects of water deficits over different times need to be examined together. For example, the recently developed joint deficit index (JDI) is based on multivariate probabilities of precipitation over various time scales from 1- to 12-months, and was constructed from empirical copulas. In this paper, we examine the Gaussian copula model for the JDI. We model the covariance across the temporal scales with a two-parameter function that is commonly used in the specific context of spatial statistics or geostatistics. The validity of the covariance models is demonstrated with long-term precipitation series. Bootstrap experiments indicate that the Gaussian copula model has advantages over the empirical copula method in the context of drought severity assessment: (i) it is able to quantify droughts outside the range of the empirical copula, (ii) provides adequate drought quantification, and (iii) provides a better understanding of the uncertainty in the estimation.