The Green Roof at the OnCenter in Syracuse, New York: Measurements and Modeling with Curve Number and Rational Methods

Green roofs are a modern stormwater management tool with an increasing number of applications across the globe. Implementing these technologies requires conformance with regulatory requirements, for which two methods dominate: the curve number method and the rational method. Universally accepted model inputs (curve number CN and runoff coefficient CV, respectively) do not exist for green roofs, and calculated values vary greatly based on roof composition and location. In this paper, CN and CV are calibrated using 6.3 years of rainfall-runoff data (313 events>2 mm depth) measured on a large extensive green roof in Syracuse, NY. Using calculations recommended by the National Resources Conservation Services (NRCS) to determine curve number, the median event CN and a least-squares estimate both result in a CN of 96. Removing small events (rain depth<12.5 mm) does not change the CN integer value. CN also varies by season: 98 in winter (identical to the value for an impervious surface), 94 in summer, and 96 in spring and fall. Use of the step function method gives CN=94 for the 313 events. Based on the rational method, event CV ranges from 0 to 0.99 with a mean of 0.31 and a median of 0.22; when small events are removed (P<10 mm), the values of CV have a mean of 0.46 and a median of 0.46. Overall, the values skew toward the higher side of what is reported in the literature. This probably reflects the impact of a shallow depth of growth medium, 76.2 mm, and the presence of drain conduits which short-circuit water flows to the drains. The results of this study suggest that the curve number method and rational method are too simple to reflect the complex hydrologic processes taking place on a green roof. However, when a large set of rainfall and runoff data is available, the data can be used to estimate a curve number and runoff coefficient that can then be used to roughly estimate total runoff expected over a long period of time.