考虑历史洪水资料的EMA参数估计方法

EMA Parameters Estimation Method with Consideration of Historical Flood Information

摘要: 为了更有效地应用历史资料信息，提高具有历史洪水资料序列参数的估计精度，研究expected moments algorithm (EMA)法进行具有历史洪水资料的参数估计方法。本文在介绍EMA原理的基础上，以实测洪峰序列为例，选用P-III型分布，进行年最大洪峰序列拟合分布的参数估计，并根据估计参数推求洪峰流量频率曲线图。与传统参数估计方法比较，结果表明：大洪水段，EMA最接近经验点据，而在一般洪水序列段中，三种方法拟合结果相差不大，大部分与实测值接近。因此，在考虑历史洪水的情形下，EMA能有效地进行P-III型分布参数估计，且较传统方法精度高。

Abstract: In order to more efficiently use historical information and improve the precision of distribution parameter estimates of historical flood information, this paper summarized the method of using expected moments algorithm (EMA) to calculate flood quantile estimates when historical flood information is available. Based on EMA principle, a annual peak discharge data with P-III type distribution was employed to illustrate to parameters estimation of the distribution, the annual peak discharge frequency curve with the estimated parameters was given. Comparing with traditional methods, the results show that all the three method’s results are similar and close to empirical probabilities for lower floods segment, but in the large floods segment, EMA is the closest to empirical probabilities. It also indicates that EMA has considerable merit in the P-III type distribution of flood frequency parameter estimation when historical flood information is available and has higher precision.