Subsurface-Flow Constructed Wetlands and Sand Filters Performance Evaluation Using Multivariate Statistical Analysis

A constructed wetlands treatment system consisting of subsurface flow constructed wetland (CW) cells and sand filters was studied for its effectiveness as a small community wastewater treatment alternative. Monitoring data collected from wetland cells and the sand filter cells, which began from June 1996 to the end of 2000, were evaluated using multivariate statistical analysis for the purpose of differentiating the performance between CW and CW followed by sand filters. Sand filtration of subsurface flow constructed wetland (CW) effluent was found to be a promising way to polish the treated wastewater. Wastewater passing through wetlands followed by sand filters resulted in an average ammonia removal efficiency of 67.1% while the CW alone resulted in an average ammonia removal of only 27.0%. Not only did sand filters improve ammonia removal in constructed wetlands, but also enhanced the removal efficiencies of chemical oxygen demand (COD), total Kjeldahl nitrogen (TKN) and total phosphorus (TP). Discriminant analysis was performed to analyze overall performance of the CW and the CW with the sand filters. It showed that the general performance of the CW and the CW with the sand filters was significantly different at a confidence level of 0.05. The results showed that sand filters can significantly upgrade the CW effluent to meet more stringent water quality standards. Sand filters can increase the dissolved oxygen (DO) concentration in the CW effluent, especially during the initial period of the wetland system, and improve the nitrification mechanism in the CW. However, it was found that the difference of the DO level between the sand filter effluent and the CW effluent began to decrease along with the operation of the wetlands. Ultimately, the DO of sand filter effluent was lower than that of CW effluent after about 3 years of wetland operation, and thus, nitrification efficiency of the CW effluent was affected.