摘要
本文在合理的假定条件下,建立了非简单样本的概率权重矩估计公式。大量的统计试验成果表明,本文建议的非简单样本概率权重矩估计公式是合理可行的。概率权重矩法估计的分布参数和设计值,其有效性和无偏性均优于传统矩法和适线法;当 C_s≤2.5时,与极大似然法相当。近六十个站的实际洪水资料的分析计算也表明,建议的公式可以获得满意的成果。
The formulas for computing probability weighted moments(PWM)in terms of non-simple sample(sample with historical flood information)are developed under a reasonable assumption.These formulas make PWM method capable of being applied to estimating parameters of hydrologic frequency distribution when extraodinary values exist in a sample,so that the domain of the opera- tion of PWM method is largely extended.A large number of statistical experi- ment results have shown that these formulas suggested in this paper are feasible and reasonable. The efficiency and unbiasedness of the parameters and quantites of Pearson Type Three(P-Ⅲ)distribution estimated by PWM method are superior to those estimated by the curve fitting method and are almost the same as those estimated by the maximum likelihood method when skewness coefficint of P-Ⅲ distribution is less than 2.5,and the unbiasedness is better but effi- ciency is similar to those estimated by the traditional moment method. flood data observed and investigated at about sixty gauging stations in China have been used to analyse the applicability of PWM method.The re- sults show that the proposed formulas are most satisfactory.
关键词
洪水
概率权重矩
估计公式
probability weighted moments
pearson type three distribution
non-simple sample
statistical experiments
unbiasedness and efficiency
