SCS Lag Time of Concentration

The SCS lag method answers the same question as Kirpich and answers it differently, because it lets land cover into the calculation. The curve number appears as a retardance term: $S = 1000/CN - 10$ in inches, and the lag is $L^{0.8}(S+1)^{0.7}/(1900 Y^{0.5})$ with flow length in feet and watershed slope in percent. Time of concentration is the lag divided by 0.6. Seven thousand feet of flow path at CN 75 on a three percent slope gives a lag of 1.01 hours and a time of concentration of 1.68 hours, which is 101 minutes.

Compare that against Kirpich on the same geometry and the difference is instructive. Kirpich sees only length and slope; the lag method sees a curve number, and a change from CN 90 to CN 60 stretches the answer by more than a factor of two on identical terrain. That is the honest part of the model. Water moves slowly across a permeable, rough, vegetated surface and quickly across a compacted one, and no length-and-slope formula can know the difference.

The limits are specific and often ignored. The lag equation was developed on agricultural watersheds up to about 8 square kilometres with curve numbers between 50 and 95, and NRCS explicitly warns that it under-predicts for urban catchments unless you apply the urban adjustment factors in TR-55 for channel improvement and imperviousness. It also uses the average watershed slope rather than the slope along the flow path, which are not the same number and are frequently confused. When this method and Kirpich disagree by a factor of two on a real site, that disagreement is information: it usually means the land cover is doing something the geometry alone cannot express, and the sensible move is to compute the travel time segment by segment with the TR-55 velocity method instead of trusting either lump formula.