Improvement and evaluation of SCS model based on a modified antecedent runoff condition

Abstract. The accurate prediction of rainfall runoff is one of the most important basis of water resource management and soil environmental quality assessment. However, the insufficient description of antecedent runoff condition (ARC) effect on runoff amount in Soil Conservation Service (SCS) model limits the accuracy of runoff prediction. On the basis of the recurrence relation of the daily effective influence rainfall formed by the potential initial abstraction and the effective rainfall influence coefficient, ARC effect on runoff is attributed to the difference between potential initial abstraction and initial abstraction, and then forms the improved SCS model. Application results of the runoff prediction in different land uses and regional scales show that the improved SCS model can accurately forecast runoff changes. Compared with the original SCS model, the improved SCS model increased the NSE and R2 of the runoff prediction under three land uses by more than 20%, and the NSE and R2 under three regional scales are up to 8.3%−23.3%. The effective rainfall influence coefficient K is the highly sensitive parameter that improves the runoff prediction accuracy of the improved SCS model.