基于GEV分布模型参数与历时关系的暴雨强度公式推求

国内现常用的暴雨强度公式一般基于对数Pearson-Ⅲ概率分布模型和Gumbel概率分布模型,采用经验Horner公式推求。该方法具有参数推求复杂,无确定解释解,并且存在非常大的不确定性。将GEV概率分布模型引入到东莞市城区年最大降雨强度频率分析中,并与对数Pearson-Ⅲ概率分布模型和Gumbel概率分布模型进行了比较,通过拟合优度检验,确定GEV概率分布模型优于上述2种分布模型。然后,利用GEV分布模型参数与历时关系推导出一种新型的暴雨强度公式,并与经验Horner公式法进行了比较。结果显示,本文方法对观测数据的拟合效果更好,并且参数易推求,推导的暴雨强度公式具有更高的确定性和鲁棒性。

Deriving Storm Intensity Formula Based on the Relationship between the GEV Parameters and Durations

Horner’s equation with Log-Pearson-Ⅲ or Gumbel model was commonly used for deriving storm intensity formula in China.However,the method was uncertainty and complicated to estimate parameters.There are no explanation solutions for the parameters.In order to solve this problem,the Generalized Extreme Value(GEV) distribution model was fitted to the annual maximum rainfall intensity data for Dongguan city,and was compared with Log-Pearson-Ⅲ and Gumbel model.The goodness-of-fit test was used to decide which model fitted best.According to the result of test,the GEV distribution model was better than the other two models.Then,a new method to derive storm intensity formula by the relationship of the GEV parameters with durations was proposed and compared with the Horner’s equation.The result showed that the method is better fit to the observed data.In addition,It is robust and convenient to estimate parameters.