Resumen
Across the globe, water quality standards have been implemented to protect receiving waters from stormwater pollution, motivating regulators (and consequently designers) to develop tools to predict the performance of stormwater control measures such as constructed stormwater wetlands (CSWs). The goal of this study was to determine how well the relaxed tanks-in-series (P-k-C*) model described the performance of CSWs in North Carolina. Storm events monitored at 10 CSWs in North Carolina were used for calibrating the model, and statistical evaluations concluded the model could adequately predict the performance for all pollutants except organic nitrogen. Nash?Sutcliff calibration/validation values were determined to be 0.72/0.78, 0.78/0.74, 0.91/0.87, 0.72/0.62, 0.88/0.73, and 0.91/0.63 for total nitrogen, total ammoniacal nitrogen, oxidized nitrogen, organic nitrogen, total phosphorus, and total suspended solids, respectively. Sensitivity analysis revealed only one calibration parameter with strong sensitivity, the Arrhenius coefficient (temperature dependent model coefficient). With this model, CSWs can be optimized to treat watershed-specific influent concentrations to meet effluent targets. In general, the current design technique used in North Carolina and many other locations (a first flush volume detention method) oversizes CSWs for water quality vis-à-vis the method herein, suggesting improved designs for water quality may be possible through scientifically-informed methods.