The purpose of this study was to develop an optimization model for the optimal placement of structural best management practices (BMPs) at the watershed scale for a reservoir. In this research, the authors considered three types of structural BMPs (i.e., a wet detention pond, a grass swale, and an infiltration trench). The complete model consists of three interacting submodels: a watershed water quality simulation model [the Hydrologic Simulation Program-FORTRAN (HSPF) model], a reservoir water quality model (the CE-QUAL-W2 model); and an optimization model that uses a well-developed genetic algorithm (GA) software. This research used the Feitsui Reservoir and its watershed in northern Taiwan as the area of application. For the optimization model, the objective function was to minimize the total annual cost of BMPs, within the constraints of achieving minimum water quality standards for the concentrations of phosphorus, ammonium, nitrate-nitrite and for the total suspended solids in the reservoir. The study first simulated the discharge and pollutant loads for the watershed in 2002–2003 by using HSPF. The reservoir model CE-QUAL-W2 was then used to simulate variations in pollutant concentrations in the reservoir on the basis of the input discharge and pollutant loads from HSPF. The optimization model was finally used to search for the near-optimal selection and placement of BMPs in the watershed. The results obtained from the completed model effectively demonstrate its viability in improving water quality in the reservoir by adopting optimal BMPs placement strategies.
Journal of Irrigation and Drainage Engineering 138(8), pp.736–747