| 引用本文: | 廖显薇,高 峰,魏 婷,等.基于Copula函数的松花江流域水文干旱频率分析[J].灌溉排水学报,2021,(3):134-141. |
| LIAO Xianwei,GAO Feng,WEI Ting,et al.基于Copula函数的松花江流域水文干旱频率分析[J].灌溉排水学报,2021,(3):134-141. |
|
| 摘要: |
| 【目的】基于Copula函数对松花江流域进行水文干旱频率分析。【方法】利用可变阈值法识别日尺度水文干旱事件的历时和烈度，并进行融合处理；选择6种常用的分布函数拟合干旱特征变量，以Kolmogorov-Smirnov 检验法优选单变量边缘分布；从3种Copula函数中优选函数模型，建立了干旱历时和烈度二维联合分布，并计算联合分布概率和重现期，再利用蒙特卡洛方法，讨论基于Copula干旱研究的不确定性。【结果】融合标准取pc=0.1和tc=5，流域干旱历时均值为81~105 d；对数正态分布是大赉站、扶余站和哈尔滨站干旱历时的最优边缘分布，而各个站点干旱烈度最优分布各异；Frank Copula为流域干旱二维联合分布的最优模型，流域大多数干旱的重现期不超过20 a；在同现重现期为20 a的条件下，大赉、扶余、哈尔滨和佳木斯4个水文站点的最大可能设计值表现出较大的不确定性。【结论】Copula函数能够较好地拟合松花江流域水文干旱特征变量间的联合分布，但需考虑其不确定性影响。 |
| 关键词: 水文干旱；Copula函数；干旱事件融合；松花江流域 |
| DOI：10.13522/j.cnki.ggps.2020330 |
| 分类号: |
| 基金项目: |
|
| Using Copula Method to Analyze Drought Frequency in Songhua River Basin |
|
LIAO Xianwei, GAO Feng, WEI Ting, SONG Xiaoyan, SONG Songbai
|
|
1. College of Water Resources and Architectural Engineering, Northwest A&F University, Yangling 712100, China 2. Inner Mongolia Water Conservancy Research Institute, Hohhot 010051, China
|
| Abstract: |
| 【Background】Drought could significantly impact hydrological processes and water resources, and understanding frequency and uncertainty of drought occurrence is hence imperative to water resources planning. The copula method has emerged as an improved multivariate analysis over the univariate analysis for quantitative analysis of drought.【Objective】This paper is to present the results of the frequency and severity of droughts in Songhua river basin calculated using the copula multivariate method.【Method】The analysis was based on archived data, and duration and severity of the pooled daily drought events were identified using the variable threshold level method. The drought index variables were fitted by six common distribution functions respectively, and they were then evaluated using the Kolmogorov Smirnov test method. Optimal model selected from three Copula functions was used to establish a two-dimensional joint distribution for the drought index variables, from which we calculated the joint distribution probability as well as the return period. The uncertainty of the drought was analyzed using the Monte Carlo method.【Result】With the pooling criteria set to be pc=0.1 and tc=5, the average drought duration was 81~105 days in the basin. The lognormal distribution can adequately describe the marginal distribution of the drought duration in Dalai, Fuyu and Harbin stations in the basin, despite the differences in optimal distribution of the drought severity between them. The Frank copula was the best model for the two-dimensional joint distribution of the drought, and the return periods of the droughts were less than 20 years. When the return period was 20 years, the maximum design values for Dalai, Fuyu, Harbin and Jiamusi hydrological stations were most uncertain.【Conclusion】The copula multivariate model can adequately describe the joint distribution of the drought index variables in Songhua river basin, and its application should consider drought uncertainty. |
| Key words: hydrological drought; Copula function; pooling of drought events; Songhua river basin |
