Regionalization of Default Parameters for Urban Stormwater Quality Models

Abstract

AbstractAccurate parameterization of water quality routines within stormwater models can improve simulation, prediction of system behavior, and lead to more informed decision making. This work recognizes that factors controlling runoff concentrations and treatment in stormwater best management practices (BMPs) vary with region, assumes region is a proxy for these factors, and seeks to improve tool parameterization by (1) identifying a regionalization scheme that best explains variability in national datasets of runoff water quality and BMP performance, (2) generating region‐specific model parameter values, and (3) demonstrating how model output varies when using national vs. regional parameters. Of the four regionalization schemes tested, the National Climatic Data Center’s Regions best explained variance in pollutant runoff concentrations from the National Stormwater Quality Database, accounting for more data variability (1.1%) than the watershed’s land use (0.4%). For BMP performance extracted from the International Stormwater BMP Database, the United States (U.S.) Environmental Protection Agency’s Rain Zones explained the most variance (1.1%), which is one‐fifth of the variance explained by BMP type (5.0%). These results were used to generate regional parameter lookup tables for stormwater quality modeling. Test cases from the 100 most populous cites in the U.S. showed national parameters predicted BMP effluent concentrations that were 15% lower than the regional parameters for five pollutants.