Modeling the washoff of pollutants from urban catchments is a difficult task. Although the exponential washoff model has been in use for decades, there is a general lack of data on model parameters and much less is known about how the model parameters are influenced by storm and climate characteristics. This study was done to understand how model parameters vary between different catchments having vastly different rainfall characteristics. Such a study will elucidate model response to rainfall, which in turn provides better knowledge on model function. High quality rainfall, runoff, and water quality data for a total of 117 storm events were sourced from the literature, which included 48 storm events from temperate (Bellevue, Washington, US; Dublin, Ireland) and 69 storm events from tropical (Skudai, Malaysia; Singapore; Pampano Beach, Florida, US) catchments, respectively. All the data were sourced from catchments with residential land use. This dataset was supplemented by additional measurements done in two residential catchments in Geelong (temperate) where a total of 16 storm events were monitored. Monte Carlo analysis was used to obtain the best-fit values of the washoff model parameters consisting the washoff coefficient, C₃, washoff exponent, C_4, and the initial mass on surface, B_in. It was found that the spread of these parameter values was greater for the dissolved pollutants ortho-phosphate (OP) and ammonium- nitrogen (NH₄ - N) than for total suspended solids (TSS). There are strong suggestions that B_in represents the mass available on the surface at the start of a washoff event, as evidenced by observations that B_in is significantly higher for the tropical dataset, a reflection of the greater storm energy associated with tropical rainfall. The small variation in C₄ across all (temperate and tropical) sites is related to its dependence on local flow conditions, linked with the friction velocity and critical shear stress conditions. The parameters C₃ and B_in on the other hand, are correlated with rainfall, with C₃ having a larger uncertainty. Of all the rainfall characteristics investigated, the rainfall depth for the current event, d was found to be the single parameter that correlated well with B_in and C₃ for TSS. The strength of this dependence diminished for pollutants in mixed and dissolved forms. In temperate regions, B_in correlates strongly with both the average and maximum intensities whereas in tropical catchments, maximum intensity and depth were found to be significant. This correlation for the temperate catchments can be understood by the dependence of rainfall energy on intensity. Rainfall energy ceases to increase once intensity exceeds 3 in/hr (76.2 mm/hr), which is uncommon in temperate catchments, but is common in the tropics.

模拟城市集水区污染物的冲刷是一项艰巨的任务。尽管指数冲刷模型已经使用了几十年，但普遍缺乏模型参数的数据，并且对模型参数如何受风暴和气候特征的影响知之甚少。进行这项研究是为了了解模型参数如何在具有截然不同降雨特征的不同集水区之间变化。这样的研究将阐明模型对降雨的响应，从而提供更好的模型功能知识。总共 117 场风暴事件的高质量降雨、径流和水质数据来自文献，其中包括来自温带（美国华盛顿贝尔维尤；爱尔兰都柏林）的 48 场风暴事件和来自热带（士姑来，马来西亚；新加坡；佛罗里达州潘帕诺海滩，US) 流域。所有数据均来自有住宅用地的集水区。该数据集得到了在吉朗（温带）的两个住宅集水区进行的额外测量的补充，共监测了 16 次风暴事件。蒙特卡罗分析用于获得由冲刷系数组成的冲刷模型参数的最佳拟合值，Ç ₃，冲刷指数，Ç _4，和上表面的初始质量，乙_在。发现溶解污染物正磷酸盐 (OP) 和铵态氮 (NH ₄ - N)的这些参数值的分布大于总悬浮固体 (TSS)。强烈建议B _in代表冲刷事件开始时地表可用的质量，正如观测所证明的那样，热带数据集的B _in明显更高，这反映了与热带降雨相关的更大风暴能量。C ₄的小变化所有（温带和热带）地点都与其对当地流动条件的依赖有关，与摩擦速度和临界剪切应力条件有关。另一方面，参数C ₃和B _in与降雨量相关，其中C ₃具有更大的不确定性。在调查的所有降雨特征中，当前事件的降雨深度d是与TSS 的B _in和C ₃相关性良好的单个参数。对于混合和溶解形式的污染物，这种依赖性的强度减弱。在温带地区，乙_在与平均和最大强度密切相关，而在热带集水区，最大强度和深度被发现是显着的。温带集水区的这种相关性可以通过降雨能量对强度的依赖性来理解。一旦强度超过 3 英寸/小时（76.2 毫米/小时），降雨能量就会停止增加，这在温带集水区并不常见，但在热带地区很常见。