自然资源学报 ›› 2015, Vol. 30 ›› Issue (5): 824-835.doi: 10.11849/zrzyxb.2015.05.010

• 资源评价 • 上一篇    下一篇

1951—2010年珠江流域洪水极值序列平稳性特征研究

顾西辉1,2, 张强1,2, 王宗志2   

  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. 南京水利科学研究院, 南京210029
  • 收稿日期:2014-04-08 修回日期:2014-10-17 出版日期:2015-05-20 发布日期:2015-05-20
  • 通讯作者: 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 E-mail:zhangq68@mail.sysu.edu.cn
  • 作者简介:顾西辉(1990-),男,河南信阳人,博士研究生,主要从事区域水循环与水资源演变研究。Email:guxihui421@163.com
  • 基金资助:

    国家杰出青年科学基金项目(51425903);香港特别行政区研究资助局(CUHK441313);中央高校基本科研业务费专项资金。

Evaluation on Stationarity Assumption of Annual Maximum Peak Flows during 1951-2010 in the Pearl River Basin

GU Xi-hui1,2, ZHANG Qiang1,2, WANG Zong-zhi2   

  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. Nanjing Hydraulic Research Institute, Nanjing 210029, China
  • Received:2014-04-08 Revised:2014-10-17 Online:2015-05-20 Published:2015-05-20

摘要:

水文序列的平稳性假设是传统水文统计学方法在水文序列分析研究中的基本假设,随着气候变化与人类活动对地表水文过程的影响,这种假设往往存在问题,使水文分析得出误导性结论。以珠江流域28 个测站1951—2010 年年最大洪峰流量序列为例,用Pettitt 方法结合Loess 参考函数检验序列中均值和方差变异,用Mann-Kendall(MK)和Spearman 法检测时间趋势性,用广义可加模型(GAMLSS)和长期持续效应等具体分析序列的平稳性。研究结果表明:①均值/方差变异主要集中在西江和东江流域,变异时间分别集中在1990 年左右和1968—1987年间;②变异点的存在与否对序列趋势检验结果有重要影响,在考虑变异点前提下,珠江流域年最大洪峰流量序列基本无显著趋势性;③在GAMLSS模型中,对于不具有和具有突变点序列,Gamma分布均为选择次数最多的最优极值分布,不具有突变点序列分布参数θ1θ2非平稳模型与平稳模型差距较小,具有突变点序列则相反;④统计上检测出具有突变点或者显著时间趋势性的测站,同样检测出高Hurst 系数,反之亦然。Hurst 系数估计因样本容量较小具有较大不确定性。东江流域受流域内水利工程的剧烈影响,尽管检测出高Hurst 系数,但仍认定为非平稳序列;西江干流主要受支流汇流和气候变化影响,高Hurst 系数表明其水文过程可能是长期稳定过程中局部波动的结果。

关键词: Pettitt分析, 长期持续效应, 平稳性, GAMLSS模型, 珠江流域

Abstract:

The stationarity is the basic assumption in the traditional statistical analysis of hydrological series. Taking the cases of the annual maximum peak flows at 28 stations across the Pearl River basin during 1951-2010, this study attempts to evaluate the abrupt changes of the mean and the variance of peak floods with Pettitt method. Besides, two nonparametric (Mann-Kendall and Spearman) tests are used to detect the temporal trends. Generalized additive models for location, scale and shape (GAMLSS) and long-term persistence are used to test the stationarity assumption. The results suggest that: 1) The mean/variance variation are detected mainly in the basins of the West River and the East River where the change points mainly happened in 1990 and during 1968-1987, respectively. 2) The existence of change points greatly affected the trend test results. The annual peak flood flow series are free of significant trends if change points are taken into account. 3) For both the annual peak flow series with and without abrupt changes, the results based on GAMLSS model indicate that gamma distribution are the best extreme value distribution, however the differences between the parameters of non-stationary and stationary models in terms of θ1 or θ2 is small for series without change point, while it is on the contrary for the series with change points. 4) Higher Hurst coefficient is detected at the stations where the peak flood flow series show significant abrupt changes or significant trends, and vice versa. It should be noted here that the Hurst coefficient is subject to larger uncertainty due to limited sample size in this case. It can be confirmed that the peak flood flow series of the East River Basin is still non-stationary even though the Hurst coefficient is high; and the high Hurst coefficients of the peak flood flows within the West River Basin could be attributed to the short-term variability in the backdrop of long-term stationary processes. The results of this study are crucial for the risk assessment of flood events and the design practice of hydraulic engineering facilities.

Key words: long-term persistence, Pettitt test, stationarity assumption, GAMLSS model, the Pearl River Bsin

中图分类号: 

  • TV122