摘要
当前设计年径流的计算,一般都采用适线法做参数估计。考虑到适线法是以水文序列独立性为前提,而实际的年径流序列存在一定的相关关系,对此,应用适线法估计到底精度如何?本文采用AR(1)模型生成符合P-Ⅲ分布的非独立样本,以参数、设计值的无偏性和有效性为评价标准,通过大量蒙特卡罗试验,对以上问题进行了研究。结果表明,在总体参数值Cv、Cs较小的地区,自相关系数即使达到0.5,对估计效果的影响也很小,但对于Cv、Cs较大的地区,自相关系数的增大会使估计效果趋于负偏。本文同时也对Cv、Cs、点据比例、序列长以及保证率等的变化对设计值无偏性、有效性存在的综合影响趋势进行了分析,希望对实际生产实践有一定的指导意义。
The curve-fitting method is a common way to estimate parameter in the calculation of design annual runoff values base on hydrology series in theory,but the actual annual runoff series often have a certain extent auto-correction. This paper researched whether the auto-correction affects the precision of estimation result or no and how great the influence is by using Monte-Carlo method. This paper produced some non-independent samples which have P-Ⅲ distribution by using AR（1） model, and analyzed the estimation result from the curve-fitting method with the criteria of the un-hiasness and efficiency of parameter and design value. The result shows that the effects of correction coefficient to estimation result is very small in the area where the parameter Cv and Cs of the population are less, even the correction coefficient reach 0.5, but the estimation result will become less than the real one obviously when the parameter Cv and Cs of the population are both big. Meanwhile, this paper has also discussed the influence of other factors to the design value of non-independent annual runoff.
出处
《水文》
CSCD
北大核心
2010年第6期10-13,共4页
Journal of China Hydrology
关键词
适线法
设计年径流
AR(1)模型
自相关系数
无偏性
有效性
P-Ⅲ分布
curve-fitting method
design annual runoff
AR （1） model
correction coefficient
un-hiasness
efficiency
P-Ⅲ distribution
