地理学报 ›› 2011, Vol. 66 ›› Issue (1): 3-12.doi: 10.11821/xb201101001

• 气候变化 •    下一篇

基于Copula 函数的新疆极端降水概率时空变化特征

张强1,2, 李剑锋1,2, 陈晓宏1,2, 白云岗3   

  1. 1. 中山大学水资源与环境系, 广州510275;
    2. 中山大学华南地区水循环与水安全广东省普通高校重点实验室, 广州510275;
    3. 新疆水利水电科学研究院, 乌鲁木齐830049
  • 收稿日期:2010-09-12 修回日期:2010-10-22 出版日期:2011-01-25 发布日期:2011-03-11
  • 作者简介:张强(1974-), 男, 山东沂水人, 博士, 教授, 博士生导师, 主要从事流域气象水文学研究、旱涝灾害机理、流域地表水文过程及其对气候变化的响应机制与机理以及流域生态需水等领域的研究工作。E-mail: zhangq68@mail.sysu.edu.cn
  • 基金资助:

    新疆自治区科技攻关项目(200931105); 国家自然科学基金面上项目(41071020); 中山大学理工科青年教师重点培育计划项目(2009-37000-1132381)

Spatial Variability of Probability Distribution of Extreme Precipitation in Xinjiang

ZHANG Qiang1,2, LI Jianfeng1,2, CHEN Xiaohong1,2, BAI Yungang3   

  1. 1. Department of Water Resources and Environment, Sun Yat-sen University, Guangzhou 510275, China;
    2. Guangdong University Key Laboratory of Water Cycle and Security in South China, Sun Yat-sen University, Guangzhou 510275, China;
    3. Xinjiang Research Institute of Water Resources and Hydropower, Urumqi 830049, China
  • Received:2010-09-12 Revised:2010-10-22 Online:2011-01-25 Published:2011-03-11
  • Supported by:

    Xinjiang Technology Innovative Program, No.200931105; National Natural Science Foundation of China, No.41071020; Program for Outstanding Young Teachers of the Sun Yat-sen University, No.2009-37000-1132381

摘要: 依据新疆地区53 个雨量站1957-2009 年日降水资料,根据研究需要,定义了8 个极端降水指标。运用K-S 法确定降水指标最适概率分布函数,确定十年一遇极端降水量值;在此基础上,采用Copula 非参数估计方法,通过Akaike Information Criterion (AIC) 法确定两降水指标联合分布函数,系统分析极端降水单变量极值及降水极值二维联合概率分布特征,研究新疆地区降水极值概率变化的空间演变特征。研究结果表明:(1) 北疆比南疆湿润,北疆发生极端强降水的概率大,而南疆发生极端弱降水的概率较大,另外,相比较而言,山区要比平原降水多;(2) 极端强、弱降水同年发生的概率分布特征复杂,从降水天数来看,一年内同时发生长历时强降水与弱降水事件的概率山区较平原大;从极端降水总量来看,同时发生强降水与弱降水事件的概率在平原区较山区为大;从极端降水强度来看,同时发生强度较大的强降水与弱降水事件的概率在天山南坡较其他地区为大;(3) 洪旱发生概率与地形有关,天山是洪旱发生的分界线,山区发生洪旱灾害的概率比平原小。

关键词: 极端降水, 联合分布, Copula, 干旱半干旱地区, 新疆

Abstract: Daily precipitation of 53 rain stations in Xinjiang during 1957-2009 is analyzed and 8 extreme precipitation indices are defined in this study. We use Kolmogorov-Smimov method to confirm the most fitted probability distributions and evaluate the ten-year return periods values. Based on that, a nonparametric estimation procedure for Copula and Akaike Information Criterion (AIC) method are used to calculate joint distribution of 2 precipitation indices. Then we comprehensively analyze the spatial variability of probability distribution of one precipitation index and joint distribution of two indices after 1980. The results show that: (1) North Xinjiang is wetter than South Xinjiang. The probability of extreme heavy precipitation is great in North Xinjiang, while that of extreme slight precipitation is great in South Xinjiang. In addition, the precipitation in the Tianshan Mountains is more than that in plain areas. (2) The spatial distribution of the probability of the event that extreme heavy precipitation and extreme slight precipitation occur in the same year is very complex. In terms of the days of precipitation, the probability in Tianshan Mountains is greater than that in plain areas. In terms of the total extreme precipitation, it is greater in plain areas. In terms of the precipitation intensity, it is greater on the southern slope of Tianshan Mountains. (3) There are relations between probability distribution of drought-flood and terrain: Tianshan Mountains is the dividing line of the occurrence of drought-flood, and the plain areas are prone to drought-flood disasters than the mountain regions. This study is of significance to get a better understanding of the droughtflood and scientific water resources management in arid and semi arid areas.

Key words: extreme precipitation, joint distribution, Copula, arid and semi arid areas, Xinjiang