Author(s): A. Massoudieh; M. Kayhanian
Linked Author(s):
Keywords: Mechanistic model; Hybrid genetic algorithm; First flush; Pollutograph; Totalsuspended solids
Abstract: Urban and non-urban stormwater are a main source of pollution in various water bodies, including rivers, streams, estuaries, lakes and coastal areas, as well as groundwater. In this research a stepwise approach is used, along with a hybrid genetic algorithm optimization technique, to identify the optimal, physically based model for prediction of highway storm-water pollutographs. Several mechanistic models are applied to the hydrograph and pollutograph data for total suspended solids (TSS) collected from an urban highway in California during a threeyear period. Model complication is added using a stepwise approach. The optimization algorithm is used to estimate the values of controlling parameters in the model, while the model selection at each step is conducted based on the goodness of fit for both a dataset used for calibration and a dataset not used for calibration. The effects of incorporating advective-dispersive transport, kinetic attachment and detachment of contaminants to the highway surface during the event, flow shear stress and raindrop impact, and non-linear build-up of contaminants during the dry period are investigated using this approach. The effect of considering that attached contaminants to occur in one phase versus two phases, in terms of the detachment rate, is also studied.
Year: 2009