Water Resources Management (2005) 19: 37–62 C  Springer 2005 Field Applicability of the SCS-CN-Based Mishra–Singh General Model and its Variants S. K. MISHRA 1 , M. K. JAIN 2 , P. K. BHUNYA 2∗ and V. P. SINGH 3 1 Water Res. Dev. Training Centre, Indian Institute of Technology, Roorkee, India; 2 National Institute of Hydrology, Roorkee-247 667, U.P., India; 3 Department of Civil and Environmental Engineering, Louisiana State University, Baton Rouge, LA 70803-6405, U.S.A. ( ∗ author for correspondence, e-mail: pkb@nih.ernet) (Received: 21 November 2003; in final form: 12 May 2004) Abstract. The general soil conservation service curve number (SCS-CN)-based Mishra and Singh (Mishra and Singh, 1999, J. Hydrologic. Eng. ASCE, 4(3), 257–264) model and its eight variants were investigated for their field applicability using a large set of rainfall-runoff events, derived from a number of U.S. watersheds varying in size from 0.3 to 30351.5 ha, grouped into five classes based on the rainfall magnitude. The analysis based on the goodness of fit criteria of root mean square error (RMSE) and error in computed and observed mean runoff revealed that the performance of the existing version of the SCS-CN method was significantly poorer than that of all the model variants on all the five data sets with rainfall ≤38.1 mm. The existing version showed a consistently improved performance on the data with increasing rainfall amount, but greater than 38.1 mm. The one-parameter modified SCS-CN method (a = 0.5 and λ = a median value) performed significantly better than the existing one on all the data sets, but far better on rainfall data less than 2 inches. Finally, the former with λ = 0 was recommended for routine field applications to any data set. Key words: agriculture research service, ars water database, curve number, initial abstraction coefficient, soil conservation service, SCS-CN method Notations A parameter of the general model AMC antecedent moisture condition B Mockus parameter B b ln(10) C loss coefficient C runoff factor CN curve number D maximum difference of empirical cumulative distribution D α critical D-value for significance level α F cumulative infiltration F max maximum possible infiltration depth when Q corresponds to P e max F ∗ max normalized F max