5 Conclusions
In this study, an equation of the CN value combining with CN2 value and three introduced factors of slope, soil moisture and storm duration was developed to improve the SCS-CN method. Six models including the original SCS-CN method, Huang et al.(2006), Huang et al.(2007) and the proposed method with different optimized parameters (Methods 1-3) were tested the reliability with data of three experimental plots in the XDG watersheds on the Loess Plateau. Subsequently, using the parameters derived from the initial three experimental plots, the proposed method was used to predict runoff from the remaining three experimental plots. large NSE values and lowRMSE values for the calibration, validation, and application of the proposed method indicated that it could accurately predict runoff for six plots of two watersheds and had greater reliability than the original SCS-CN, Huang et al. (2006) and Huang et al. (2007) methods. Moreover, Method 2 with initial abstraction ratio λ =0.001 instead of the standard value of 0.2 and slope parameters obtained from Huang et al.(2006) with data from slopes ranging from 14 % to 140 % appears to be the most promising for runoff prediction in the study area of the Loess Plateau. Furthermore, a sensitivity analysis of the parameters in the proposed method indicated that the parameters of soil moisture (b1 and b2 ) and storm duration equations (c ) are more sensitive than those parameters of slope equation (a1 anda2 ) andλ .