The Curve Number hydrology community has debated, without resolution, the proper value for the initial abstraction ratio (IAR) for 10-20 years. This presentation resolves this debate by turning to engineering conservatism as the controlling criterion. We make use of how calibration to observed data for different IARs results in a threshold precipitation value. Below this threshold, one IAR produces larger runoff estimates. Above this threshold the other IAR produces larger estimates. With design precipitation and curve number as the inputs, we present a diagram that shows how threshold precipitation varies, and indicates where each IAR will produce the more conservative estimate. Hence, there is not a single best IAR. Instead, the selection of IAR is driven by watershed curve number and the precipitation frequency where the watershed is located.
