长江和黄河流域降水丰枯遭遇及不确定性分析

Analysis of precipitation wetness-dryness encountering and its uncertainty in the Yangtze River and Yellow River basins

  • 摘要: 长江和黄河流域降水遭遇概率分析对中国南水北调工程水安全保障具有重要的意义。重构长江和黄河流域上、中、下游不同区域500 a降水资料,采用Copula函数理论及多维方差分析方法,分析长江、黄河流域不同区域降水丰枯遭遇概率,并解析丰枯遭遇结果的不确定性及其主要来源。结果表明:长江、黄河流域间不同区域降水枯-枯遭遇概率均较小,均未超过20%,流域内部上、中、下游降水枯-枯的概率均未超过10%;丰-丰、枯-枯、丰-枯和枯-丰4种组合的遭遇概率计算结果不确定性较大,较上游和中游而言,下游的不确定性总体较小;边缘分布、Copula函数、数据样本长度间的相互作用是长江流域上、中、下游降水遭遇概率结果不确定性的主要来源,黄河流域则主要受不同数据样本长度的影响。

     

    Abstract: The probability analysis of precipitation encountering in the Yangtze River and Yellow River basins is of great significance for ensuring water security in China's South-to-North Water Diversion Project. Based on the reconstruction of precipitation data in different regions of the upper, middle, and lower reaches of the Yangtze River and Yellow River basin over the past 500 years, this paper uses the Copula function theory and the Multidimensional Variance Analysis Method to analyze the probability of precipitation encountering in different regions of the Yangtze River and Yellow River basins, with the uncertainty and main sources of the results illustrated. The results show that the probability of dryness-dryness encountering in different regions of the Yangtze River and Yellow River basins is relatively small, not exceeding 20%, and that the probability of dryness-dryness encountering in the upper, middle, and lower reaches of the basin does not exceed 10%. There is significant uncertainty in the calculation results of the encounter probability for the four combinations of wetness-wetness, dryness -dryness, wetness-dryness and dryness-wetness; compared with that in the upstream and midstream, the uncertainty in the downstream is generally less. The interaction between marginal distribution, the Copula function, and data sample length is the main source of uncertainty in the probability of precipitation encountering in the upper, middle, and lower reaches of the Yangtze River basin, with the Yellow River basin mainly affected by different sample sizes of data.

     

/

返回文章
返回