MODIFIED SCS CURVE NUMBER METHOD FOR PREDICTING SUBSURFACE DRAINAGE FLOW

Abstract

Agricultural fields in Central Illinois are predominantly drained by subsurface drainage systems, which often are in depressional areas. The prediction of outflow from drainage systems is a challenge. The SCS curve number (SCS–CN) meth-od is a popular method for evaluating direct runoff from rainfall. The concept behind this method can be applied to subsurface drainage flow. The SCS–CN method was modified through theoretical analogy and term redevelopment to estimate subsurface drainage flow from rainfall. The analogical theory is that when accumulated subsurface drainage flow is plotted versus accu-mulated infiltration, subsurface drainage flow starts after some infiltration has accumulated and the relationship becomes asymptotic to a line of 45
slope, just as the generalized SCS rainfall–runoff relationship. Procedures are introduced for modification of the SCS–CN method and determination of curve numbers for subsurface drainage flow. In the process of defining curve numbers for drainage flow, it was found that the curve number varied not only with season but also with rainfall amount. In addition, the curve number varied with previous rainfall condition as with the traditional SCS–CN method. The best previous rainfall adjustment is based on the 10–day previous rainfall for subsurface flow prediction. The curve number is sensitive to the fraction of initial abstraction (k). However, this does not mean that predicted flows resulting from different k values are significantly different, since the equation for flow prediction changes when k is changed; in fact, the predicted flow is not sensitive to the fraction of initial abstraction to potential maximum retention of the watershed. The modified SCS–CN method was applied to estimate subsurface drainage flow for five drainage monitoring stations in the Little Vermilion River (LVR) watershed in East–Central Illinois. Predicted subsurface flows using the modified SCS–CN method were compared with observed subsurface flows. Statistical tests showed that the predicted subsurface flows using the modified SCS–CN method were not significantly different from the observed subsurface flows. Validation performed on two of the sites using modified curve number relationships from the other sites showed that the predicted subsurface drainage flows were not significantly different from the observed subsurface drainage flows.