An empirical method that estimates direct runoff depth from a storm's rainfall using one number, the curve number, set by soil and land cover.
The SCS Curve Number method estimates how much of a storm's rainfall becomes direct runoff. It was developed by the US Soil Conservation Service, now the Natural Resources Conservation Service (NRCS), and is set out in its National Engineering Handbook. It is one of the most widely used runoff methods for small catchments, urban drainage design and models such as SWAT and HEC-HMS. The method uses a single parameter, the curve number (CN), which runs from about 30 for very permeable, well vegetated ground to 100 for water surfaces and fully sealed areas. CN is read from tables using the hydrologic soil group (A, B, C or D, from high to low infiltration), the land use and cover, and how wet the soil is before the storm (the antecedent moisture condition). CN sets the potential maximum retention S. Part of the rain, the initial abstraction Ia, is lost to interception, infiltration and surface storage before any runoff starts, and is usually taken as 0.2S. For example, with CN = 80 and a storm of 100 mm: S = 25,400 / 80 - 254 = 63.5 mm, Ia = 0.2 × 63.5 = 12.7 mm, and Q = (100 - 12.7)² / (100 - 12.7 + 63.5) = 7,621.3 / 150.8 = 50.5 mm. About half the rain runs off. The method is simple, needs little data and is easy to map from soil and land cover layers. However, it gives the runoff depth for a storm, not its timing, and ignores rainfall intensity and duration. It was built from data on small US agricultural catchments, so results elsewhere depend on how well the tables fit local soils. The 0.2 ratio for Ia has been questioned, and several studies suggest a smaller value such as 0.05 fits observed data better. Where flow records exist, CN should be calibrated rather than taken straight from tables.