中文摘要：

【目的】初损率是SCS-CN方法进行流域地表径流预报的基础输入参数之一,影响着径流模拟精度。论文通过研究确定黄土丘陵沟壑区典型小流域初损率取值,为SCS-CN方法在该地区的适用性评价提供参考。【方法】SCS-CN方法中,初损率定义为初损量（Ia）与流域最大蓄水能力（S）的比值,通常取标准值0.2。然而,已有研究表明不同区域初损率取值存在差异。论文选取黄土丘陵沟壑区桥子西沟小流域1987—2006年14场典型降雨事件,采用反算法（BC）和事件分析法（EA）确定流域初损率。【结果】结果表明,两种方法计算初损率均小于标准值0.2,事件分析法计算值（0.17）与反算法计算值（0.1）相比略偏大。初损率取0.1、0.17和0.2分别计算研究降雨事件径流深,采用皮尔逊相关系数（r）、模型效率系数（E）、相对误差（RE）、绝对误差（AE）和图形拟合5种评价标准分析模拟结果,确定研究流域初损率为0.1。论文采用两种算法计算桥子西沟流域初损率,结果与标准值显著不同,初步推断是研究流域与SCS-CN方法最初研发区域地质地貌及气候条件存在较大差异所致。【结论】不同区域应用SCS-CN方法时,需利用当地观测数据率定研究流域初损率,才能进行流域地表水文模拟。

英文摘要：

【Objective】 The initial abstraction ratio,as one of the basic inputs of SCS-CN method,plays an important role in watershed rainfall-runoff simulation.In this study,attempt was made to determine the initial abstraction ratio of a typical small watershed in the Loess hilly-gully region,in order to supply some references for regional applicability assessment of SCS-CN method.【Method】In SCS-CN method,the proportion of initial abstraction（Ia） to the maximum retention（S） is defined as initial abstraction ratio,which was believed to have a standard value of 0.2.However,many studies indicate that variations of the ratio exist in different regions.In this study,14 typical rainstorms in 1987-2006 in Qiaozi-West watershed were selected to calculate the initial abstraction ratio using Back Calculation（BC） and Event Analysis（EA） methods.【Result】Results showed that most of the values calculated by the two methods were lower than the standard value of 0.2.Moreover,the value（0.17） determined by EA method was a little higher than that（0.1） by BC method.By assigning 0.1,0.17 and 0.2 to initial abstraction ratio,the runoff depth of each study event was calculated,respectively.Through five evaluation criteria of Pearson correlation coefficient（r）,model efficiency coefficient（E）,relative error（RE）,absolute error（AE） and curve fitting,0.1 was found to be the appropriate value for the watershed initial abstraction ratio.The results indicated that the derived initial abstraction ratio by the two methods differed markedly from the standard value.This may be due to the discrepancies in geologic,geomorphologic and climatic circumstances between the loess hilly-gully region and the originally intended areas of SCS-CN method.【Conclusion】 It is essential to calibrate the initial abstraction ratio using locally observed data prior to SCS-CN method application.