EFFECT OF WATERSHED CHARACTERISTICS ON RUNOFF CURVE NUMBER

Abstract

Estimation of runoff and sediment yield is essential for solution of a number of problems such as the design of reservoirs, dams, planning of soil conservation structures and river morphology studies. A number of models have been developed to compute runoff and sediment yield generated from rainfall event. The Soil Conservation Service Curve Number (SCS-CN) method is one of the most popular and widely used event-based methods for runoff estimation and recently it has been also coupled with the popular soil erosion models such as Universal Soil Loss Equation (USLE) and Revised Universal Soil Loss Equation (RUSLE) for sediment yield estimation. This study explores the effects of the watershed characteristics such as soil type, land use, antecedent moisture condition, and watershed slope on the Curve Number (CN) and, in turn, on watershed runoff and sediment yield using the rainfall-runoff and sediment yield data of an experimental watershed located at Toda Kalyanpur, Roorkee, Haridwar, India (lat. 29º 50´ 6´´ N and long. 77º 50´ 17´´ E). The experimental field is sub-divided into 9 equal agricultural plots having the dimension of 12 x 3 m each. These plots were further divided into 3 groups having different slopes of 8%, 12% and 16%. 3 plots of a slope were planted with two different crops, i.e., Maize, Finger millet and one plot left as fallow land to study their comparative impacts on hydrological responses in terms of runoff and sediment yield. In this study, a total of nineteen storm events were carefully observed for runoff and sediment yield. The sediment samples collected in the filed were analyzed in WRDM laboratory for suspended sediment concentration using oven drying method. Experiments were also conducted for estimation of initial soil moisture before storm, soil texture analysis and infiltration capacity tests using double ring infiltrometer. The hydrological soil group was however the same for all plots, as established from infiltration tests. The results show that as slope is increased from 8% to 16%, the runoff is also found to increase and hence CN. While keeping the rest of watershed characteristics the same, the Maize crop showed the highest runoff and CN as compared to the other crops. The SCS-CN parameter S showed an inverse relation with the physically measured antecedent moisture content. The trend line of the plotting of rainfall against the runoff shows that runoff increases with the increase in rainfall for the given plot. Sediment yield in the representative sample increases with increase in runoff and relatively high value of R2 is obtained with these relations. The sediment rating curves iii were also plotted between suspended sediment concentration and runoff volume with coefficient of determination greater than 0.5 for all crops and land slopes. The runoff was estimated using the existing SCS-CN method and for estimation of the sediment yield, the model developed by (Mishra et al., 2006) was coupled with the MUSLE model. The model performance was evaluated using Nash-Sutcliffe efficiency (NSE %), the root mean square error (RMSE) and percent bias (PBIAS). The overall values of NSE, RMSE and PBIAS were found to be 90.57%, 60.55%; 0.37, 0.41; -17.34, -14.47, respectively for runoff and sediment yield prediction. Based on the criteria of (Moriasi et al., 2007), the model was found to perform very good in runoff prediction and good in sediment yield prediction. Due to non-linearity between runoff and suspended sediment yield, the CNs estimated for runoff cannot be directly used for sediment yield estimation. An effort was also made in this study to develop a relationship between CNs estimated by using observed sediment yield and CN estimated using observed runoff for all nine plots. The developed relationships can be directly used for estimation of CNs for application in sediment yield estimation by using CNs obtained from observed rainfall-runoff data.