A multiple-realizations chance-constrained model for optimizing nutrient removal in constructed wetlands

High nutrient loads in surface waters are a cause of hypoxia in coastal ecosystems. Constructed wetlands offer a means of nutrient removal, but their design is complicated by environmental fluctuations. In the present work, a chance-constrained model is developed on the basis of satisfying performance constraints for a subset of the period of hydrologic and meteorological record. The model incorporates a nonsteady state wetland model and is used to determine the most cost-efficient wetland design given the nutrient removal requirements. Three candidate wetlands are considered. For two of them a comparison is made between operating the wetlands individually versus operating them concurrently. Results show that the factors limiting wetland performance are low flow availability and low inflow nutrient concentrations. Further, for the case considered, a two-wetland design is found to be more cost-effective than a single-wetland design. On average, wetlands have huge potential for nutrient removal but have a higher risk of failure than conventional treatment and could, under unfavorable conditions, be more expensive.