A comparison of curve number based methods for runoff estimation from small watersheds in a semi-arid region of India

The performance of five curve number based methods, generally employed for estimation of runoff, viz. National Engineering Handbook (NEH-4), storm event (SE), rank order (RO), lognormal frequency (LF) and s-probability (SP), was evaluated using runoff data recorded from three small watersheds in a semi-arid region of India. The most effective and reliable curve number method for runoff estimation was selected based upon several tests of goodness of fit including coefficient of determination (R-2), index of agreement (D), root mean square error (RMSE) and relative bias (RB). The LF method was adjudged as the most promising curve number method for estimating runoff from small watersheds in the semi-arid regions of India. The runoff predicted by the LF method was in close agreement with the observed runoff for all four tests of goodness of fit. Though the performance of the SE, RO and NEH-4 methods was almost comparable to each other, the SP method registered higher deviation from the observed runoff values. From the analysis, it is concluded that the LF method can be successfully employed for estimation of curve number based runoff from small watersheds in the semi-arid regions of India.