Abstract Flood discharge and its frequency is an important input to many infrastructure development, management and decision making models. However, for a number of potential sites, the data used for statistical analysis of flood discharge and its frequency, are either insufficient or unavailable. In such a situation, empirical formulae are the only alternative method to provide an estimate of design flood. Dicken’s formula is one of such empirical formulae used by practicing field engineers for flood estimation. In spite of its simplicity and applicability, Dicken’s formula is not capable of yielding flood volume at different desired frequency. In the present study an attempt has been made to modify the Dicken’s coefficient so as to accommodate the frequency component in flood estimation. For this purpose, observed flood discharge data sixteen gauging points of Ganga basin have been used to develop regional flood frequency curve, using probability weighted method, Index flood method, General extreme value and Wakeby distribution. Modified Dicken’s coefficient CT shows a sharp gradient for sub-catchments draining up to 5000 Km2. Further, CT value, ranges from 0.834 to 6.924. The CT –values are lower when calculated by Index Flood method while these are higher by Wakeby method. In all of these methods, CT –values do not approach the values of 11 or 14 as suggested for northern Indian catchments. An analysis of flood volume corresponding to CT –values of 11 or 14, reveals that these values yield high flood volume that correspond to return period of more than 1000 years and 3000 years respectively. These values are certainly very high for any rational design. Best fit distribution is decided, based on annual maximum discharge peak (ADF), Efficiency (EFF) and Standard Error (SE) parameters. Prediction accuracy of flood frequency analysis using revised Dicken’s formula is found to be above 85% for the best-fit distribution. General Extreme Value is found to be the best fit distribution for the Upper Ganga basin.

中文翻译：

---

上恒河盆地洪水频率估计的修正狄肯方法

摘要 洪水及其频率是许多基础设施开发、管理和决策模型的重要输入。然而，对于一些潜在的站点，用于洪水流量及其频率统计分析的数据要么不充分，要么不可用。在这种情况下，经验公式是提供设计洪水估计的唯一替代方法。狄更斯公式是现场工程师用于洪水估计的经验公式之一。尽管它的简单性和适用性，狄更的公式不能在不同的期望频率下产生洪水量。在本研究中，已尝试修改狄肯系数以适应洪水估计中的频率分量。以此目的，利用恒河流域16个测点的实测洪水数据，采用概率加权法、指数洪水法、一般极值法和Wakeby分布，绘制区域洪水频率曲线。修正狄肯系数 CT 显示排水量高达 5000 Km2 的子集水区的陡峭梯度。此外，CT 值的范围从 0.834 到 6.924。用 Index Flood 方法计算 CT 值较低，而用 Wakeby 方法计算时 CT 值较高。在所有这些方法中，CT 值都不接近印度北部集水区建议的 11 或 14 值。对对应于 CT 值 11 或 14 的洪水量的分析表明，这些值产生了分别对应于 1000 年和 3000 年以上重现期的高洪水量。对于任何合理的设计来说，这些值肯定是非常高的。根据年度最大流量峰值 (ADF)、效率 (EFF) 和标准误差 (SE) 参数确定最佳拟合分布。发现使用修正的狄肯公式进行洪水频率分析的预测准确度在最佳拟合分布的 85% 以上。发现一般极值是上恒河盆地的最佳拟合分布。