Abstract

Storm water is considered as a significant source of contaminants to receiving rivers and the constructed wetland has been used to treat storm water before the discharge. In this study, a hybrid fuzzy-stochastic modeling approach is developed to examine the wetland treatment efficiency, to analyze the environmental impact associated with the wetland effluents into the receiving water, and to quantify system uncertainties. The proposed approach first incorporates a water quality model to simulate storm water flow going through the wetland and the fate and transport of nutrients in the wetland. A Monte Carlo modeling method is next developed to extend the water quality model, providing a stochastic simulation of the concentration distribution of nutrients in the wetland effluents. It is intended for the analysis of probabilistic environmental risks associated with wetland effluents on the receiving waters. The fuzzy membership functions are further used to quantify the variability or suitability of regional surface water guidelines, which is incorporated into the Monte Carlo modeling framework to identify the integrated risks from the discharge on the river.

The developed modeling approach has been applied to the Kennedale wetland, a storm water treatment system, in the city of Edmonton, Canada. Before the environmental risk assessment, the HEC-RAS (Hydrologic Engineering Centers River Analysis System) model and the QUAL2K (River and Stream Water Quality) model are applied to simulate the flow and nutrients removal efficiency in the wetland. According to the simulation results from the HEC-RAS model and the QUAL2K model, the removal efficiencies of TN (Total Nitrogen) by the wetland are 25.64% and 13.59%, respectively. The removal efficiencies of TP (Total Phosphorus) are 50% and 50.91%, respectively. The differences between the HEC-RAS simulation results and on-site field data are 0.05% for TN and 6.1% for TP. The differences between the QUAL2K simulation results and on-site field data are 13.99% for TN and 4.35% for TP based on this study. The water quality simulation results from the two models are both acceptable compared to the monitoring data. It is seen that the HEC-RAS model has better performance on modeling this field case, and is integrated with the environmental risk assessment process. Consequently, the results of the integrated risk assessment referring to different guidelines in the North America show that the concentrations of TN at the wetland discharge port have a high possibility of violating the TN guidelines in both Alberta, Canada and the US EPA (Environmental Protection Agency). Similarly, the concentrations of TP at the wetland discharge port have a high possibility to violate the Canadian and US TP guideline during this study period. Therefore, the nutrients in storm water discharges from the Kennedale wetland may have a great risk to adversely affect the receiving river (North Saskatchewan River) at the time of this study. The analysis results of nutrient guidelines have supported the management of decision making process, and the study results indicate that the developed hybrid fuzzy-stochastic modeling approach is a useful tool for the practical managing of wetland systems and the impact of the wetland discharges on the receiving waters.