GIS-based slope-adjusted curve number methods for runoff estimation

Accurate estimation of surface runoff and determination of susceptible lands to runoff generation in ungauged watersheds were the problems for hydrologic engineering which could be predicted through a simple model such as Soil Conservation Service Curve Number (SCS-CN). Due to the slope effects on this method, slope adjustment for curve number was developed to improve its precision. So, the main objectives of this study were to apply GIS-based slope SCS-CN approaches for surface runoff estimation and compare the accuracy of three slope-adjusted models including: (a) model with three empirical parameters, (b) model with two parameters slope function, and (c) model with one parameter in the region located in the central part of Iran. For this purpose, soil texture, hydrologic soil group, land use, slope, and daily rainfall volume maps were utilized. In order to provide the curve number map of the study area, land use and hydrologic soil group layers built in Arc-GIS were intersected and the curve number was determined. Then, three slope adjustment equations were used to modify curve numbers of AMC-II by employing slope map. Finally, recorded runoff data of the hydrometric station were applied to assess the performance of the models through four statistical indicators of the root mean square error (RMSE), the Nash–Sutcliffe efficiency (E), the coefficient of determination (R2)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${(R}^{2})$$\end{document}, and percent bias (PB). Land use map analysis showed that rangeland was the dominant land use, whereas the soil texture map specified the greatest and smallest area belonging to loam and sandy loam textures, respectively. Although the runoff results showed the overestimation of large rainfall values and underestimation for rainfall with less than 40 mm volume in both models, the values of E (0.78), RMSE (2), PB (16), and R2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${R}^{2}$$\end{document} (0.88) revealed that eq. (a) with three empirical parameters was the most accurate equation. The maximum percent of runoff generated by rainfall for eqs. (a), (b), and (c) were 68.43, 67.28, and 51.57% which showed that bareland located in south part with the slope of more than 5% was susceptible to runoff generation and should be paid attention to watershed management.