| Contents Notes |
Treatment efficiency in a constructed wetland is related in part to the amount of time that a wastewater remains in the system. Current design methods idealize the system as a plug flow reactor and use a "residence time" based solely on the volume of the cell and the flow rate. Under this assumption, every element of wastewater entering the wetland cell experiences the same residence time. It is understood that this idealization ignores the existence of longitudinal dispersion, short circuiting and stagnant regions within the wetland cell. The result of these phenomena is a distribution of residence times. In other words, portions of the effluent exit the cell earlier than predicted, resulting in undertreatment, and portions exit late, resulting in excess treatment. The average concentration of treated wastewater at the outlet is a function of this distribution and the reaction kinetics associated with the waste. The overall effect of a distribution of residence times is reflected in a reduction of treatment efficiency at the outlet. Hydraulic regimes of constructed wetland systems were investigated at a pilot project site providing tertiary treatment of a pulp mill wastewater. Two vegetation types, bulrush and cattail, were investigated and compared to nonvegetated and rock-filter cells with identical configurations. Tracer studies used a fluorescent dye and were performed over the course of a year. Dye was input as a pulse at the inlet end of the cell and sampled over time at the outlet end to obtain concentration breakthrough curves. From these curves, time to peak, actual mean detention times, degree of dispersion, and extent of dead space were calculated, as well as predicted treatment efficiency. Results indicated varying degrees of dispersion, short circuiting, and dead space in the individual cells. Analysis of the residence time distributions provided estimates of the "active" volume of the treatment cell and the degree of short circuiting in the system. Effective volume of the planted cells ranged from 15 to 25% of full volume. Early arrivals of the peaks of the distributions, indicative of short circuiting, ranged from 30% to 80% of the theoretical detention times. A first order treatment model and a kinetic coefficient, k, were assumed, and corresponding treatment efficiencies were compared to the theoretical treatment of an ideal plug flow reactor. Reduced treatment efficiencies for the planted systems ranged from 2 to 20 %, by this estimation. Many references attempt to analyze wastewater treatment systems by refering to two models: dispersed plug flow and an approximation of tank-in-series. These models were investigated as potential descriptions of the hydraulic regime present in constructed wetlands. Residence time distributions of the constructed wetlands in this study indicated flow was not exclusively dispersed plug flow. This simplified model does not account for the exchange of material with "dead" space in the wetland cell. The data suggest a combination model of dispersed plug flow with a transient storage zone component may be more appropriate. |
