Abstract
Two subsurface flow constructed wetlands were tested for pollutant removal performance in conjunction with an on-line stormwater detention pond, in Kingston Township, Ontario. The 4.9 m2 wetland cells were filled with 9 mm limestone gravel, and planted with cattail, common reed and spike rush. Changes in nutrient (total organic carbon, PO43- and NH4+), suspended solids and metal (Cu, Pb, Zn) concentrations were used to assess performance. Contaminant removal occurred through a combination of physical, chemical and biological means. As with any biological system, variation in performance of stormwater wetlands can be expected to occur as a result of fluctuations in contaminant loading, contact time and ambient environmental conditions.
Storm pond effluent was delivered in continuous flow through the wetlands (during baseflow and event conditions), with a detention time of 1 to 3 days. The wetlands were able to maintain removal rates of up to 39% for orthophosphate even during the more severe conditions of fall dieback. Average removal of suspended solids (46%) and dissolved metals (Cu 50%) remained similar throughout all tests. Organic carbon was reduced by less than 10% during these tests.
Low nutrient levels in the pond effluent were supplemented by spiking with sources of carbon, nitrogen and phosphorus during pulsed loading conditions. Daily sampling produced a time series, which illustrated the rates of decline in concentration of nutrients. First order kinetic assimilation rates ranged from 1.7 d-1 for NH4002B to 0.12 d-1 for organic carbon, which were noticeably lower when compared with municipal and industrial wastewater treatment rates.
Three methods of sizing stormwater wetlands (impervious surface area, volumetric load and kinetic reaction rates) were compared using the same design storm and data from this study. From this comparison it was seen that the kinetic sizing approach proved to be the most versatile, and allowed for adaptation to northern climatic conditions and anticipated nutrient loading.