Hydraulic characteristics of a subsurface flow constructed wetland for winery effluent treatment

Similar to batch reactors, wastewater treatment by constructed wetlands depends on the detention time within the system. Unlike reactors, however, internal detention times within constructed wetlands are largely unknown because they are a function of wastewater flowpaths and extent of wastewater interaction with constructed wetland porous media and vegetation. Internal hydraulic characteristics of two pilot-scale constructed wetland tanks designed to reclaim winery wastewater are evaluated to link treatment performance to hydraulic properties and compare these properties between new and old tanks. Each tank contained pea gravel to a 0.95-m depth. The old tank had established bulrush and cattail vegetation and had operated as a winery wastewater treatment system for more than 3 years, while the new duplicate tank had no vegetation. Two impulse-type, bromide-tracer studies were conducted, and the results were analyzed, assuming that traditional one-dimensional flow breakthrough-curve models could be applied to the three-dimensional flow system. Because of variable recovery of the tracer at different locations along the tanks, the method of moments was used to analyze bromide residence-time distributions (RTDs) to determine detention times, water velocities, and dispersion coefficients. While tank outflow tracer RTDs could not be described using either plug-flow reactor or simple continuously mixed-system models, 20 to 35 tanks-in-series models adequately replicated the measured outflow RTDs. No preferential flows were observed in the constructed wetland tanks because outflow RTDs were nearly symmetric and surprisingly similar. Overall, outflow RTD-based dispersion coefficients (D) were small and only slightly different between the old (D = 1.5 mm(2)/s) and new (D = 1.2 mm(2)/s) tanks. However, the mean detention times in the new and old tanks were approximately 15% greater and 30% less, respectively, than those estimated using the average gravel porosity and assuming plug flow.