Predicting the Long-Term Performance of a Structural Best Management Practice with the BMP ToolBox Model

Abstract

It is costly to constantly sample to monitor the performance of a structural best management practice (BMP). Alternatively, occasional sampling might not be adequate. The BMP ToolBox model, developed by Tetra Tech and Prince George's County in Maryland, USA, assesses the performance of a structural BMP treatment site. The study applied the BMP ToolBox model to a BMP site in Taiwan to test its validity. The case study site was designed to remove pollution from nonpoint sources (tea gardens) in order to maintain water quality in the Feitsui Reservoir. The BMP ToolBox model was calibrated and verified using two years of sample data. Results were satisfactory with the coefficient of determination (R2) for calibration and verification being 0.87 and 0.8, respectively. Furthermore, the one-factor-at-a-time method (OFAT) was applied in a sensitivity analysis to identify sensitive model parameters. Several hydrographs were created to predict BMP performance. The positive relationship between the total phosphorus (TP) removal rate and the recurrence interval was observed: rainfall with longer durations showed increased removal rates compared to shorter periods of rainfall. The BMP Toolbox model was successfully applied, and a process for evaluating the long-term operation of structural BMP sites was established.