Determining the initial abstraction ratio of the upper Sawaga watershed, Bukidnon

Abstract

The Curve Number (CN) method has been widely used for estimating runoff from rainfall. However, some uncertainties in the method have been recognized by various researchers all over the world. One of which is the NRCS-assumed initial abstraction ratio (Ia/S) of 0.20. In this study, the Ia/S for the upper catchment area of Sawaga River in Bukidnon was determined using event analysis of different rainfall events with a total precipitation depth ranging from 21.50 to 57.90 mm. The resulting representative values of Ia/S and CN are 0.03 and 62.3, respectively. Evaluating the runoff prediction performance of both the original NRCS (Qesti0.20) and locally derived parameter values (Qesti0.03), Qesti0.03 had a lower error rate with a Standard Error of 2.2639 compared to Qesti0.20’s 2.6760. For accuracy in predicting runoff, Qesti0.03 also showed better performance with a Nash-Sutcliffe Efficiency of 0.7854 versus 0.6942 for Qesti0.20, and its Percent Bias was lower at 2.6410 compared to 36.4201 for Qesti0.20. Both methods had very good coefficients of determination, with Qesti0.20 at 0.8459 and Qesti0.03 at 0.8343. Overall, Qesti0.03 proved to be more reliable and effective for accurate runoff prediction in this watershed. These findings suggest that the CN method, with certain adjustments on Ia/S and CN, is appropriate for the direct runoff estimation in the upper Sawaga watershed, Bukidnon.