Geographic Dependency of the Curve Number Method’s Initial Abstraction Ratio

Abstract

The community of hydrologic engineers that uses the curve number method of the Natural Resources Conservation Service (NRCS, formerly the Soil Conservation Service, SCS) has had a spirited debate over the last two decades on the proper value of the initial abstraction ratio, λ, the ratio of the initial abstraction to watershed storage. Historically, this value has been set at 0.2, but many in the community advocate that a value of 0.05 is more appropriate. This work builds on the recent argument to use hydrologic conservatism to inform the proper selection of λ for hydrologic design. Analyses in the State of Maryland and in Clark County (Las Vegas Valley), Nevada, are used to illustrate high resolution variation in curve number and rainfall magnitude to determine zones where each λ value is more conservative. These high-resolution analyses are used to develop simplifying assumptions so that similar large-scale analyses of the conterminous United States (CONUS) can be performed and presented. The CONUS results show a clear shift in the more conservative λ value from 0.05 to 0.2 as rainfall depth increases. The humid eastern United States favors λ=0.2 as being more conservative, especially for larger rainfall events. The arid west and southwest favor λ=0.05 as more conservative, even for longer return periods. The high-resolution results highlight how the more conservative λ value shifts from 0.05 to 0.2 as curve number increases. This work invokes hydrologic conservatism as the rationale for choosing the initial abstraction ratio in the NRCS curve number method. Here, conservatism is defined simply by the larger estimated runoff depth resulting from a design rainfall event. Results are mapped as zones showing the more conservative initial abstraction ratio depending on rainfall frequency and geographic location. Two high-resolution studies covering Maryland and Clark County, Nevada, were carried out to illustrate the dependencies on spatially varying land use and hydrologic soil type in addition to rainfall information. A coarser resolution study presenting results at the continental scale is also presented. Results show that smaller-magnitude rainfall associated with more frequent rainfall events or an arid climate favor the selection of the smaller initial abstraction ratio as more conservative. For larger rainfall events, the larger initial abstraction ratio tends to produce more conservative runoff estimates. Practitioners using the curve number method for design can use these results to set expectations for which initial abstraction ratio will be more conservative in their unique design application.