Using the data of a large number of observed rainfall (P)/runoff (Q) events for an agricultural experimental field located in Roorkee, Haridwar, Uttarakhand, India, the effects of land features such as slope and antecedent moisture were evaluated on the parameter curve number (CN) of the SCS-CN methodology used for runoff estimation. The plot of 5% slope yielded the largest runoff and, in turn, CN compared to the plots of 3% and 1% grades for the same soil type and land use. The derived CN values from observed P-Q data were considerably different from conventional NEH-4 table values. The optimized initial abstraction ratio (lambda) values showed that the original assumption of the lambda of 0.20 is unusually high. The median and mean lambda values were respectively 0 and 0.034 for natural P-Q data and 0.033 and 0.108 for ordered P-Qdata. The employment of a modified equation of sharpley and Williams (1990) for deriving the slope-corrected CN-values improved runoff estimation significantly. As expected, CN (or, potential maximum retention, S) values showed a higher degree of dependence on the physically observed 1-day antecedent soil moisture (theta(o1)) than other duration antecedent soil moisture values. (C) 2015 Elsevier B.V. All rights reserved.