Development of Improved Stormwater Quality Models

It is not possible to derive general pollutant washoff functions, for either storm event total loadings or instantaneous fluxes, that contain a given set of independent variables and give reasonable results for most areas, even with calibration. Therefore, a methodology is outlined that facilitates the development of models that can be used with increased confidence to predict urban stormwater pollutant washoff. The methodology consists of testing several linear and linear transform multiple regression models for estimating pollutant washoff loadings as a function of storm and basin characteristics, using measured data. The best model can be chosen for each pollutant using the coefficient of determination (R²) and other criteria. This methodology eliminates the need to estimate both pollutant buildup and washoff.