The Study Of Removal Efficiencies And Mechanism For Nitrogen And Phosphorous In Constructed Wetlands

In this paper, we firstly study the removal efficiencies of the unplanted and uninterrupted integrated vertical constructed wetland for stormwater runoff. It shows that the wetlands had a steady removal efficiency for COD and TP, and the removal efficiencies were 38% and 42%, respectively. However, the wetlands had a weak removal efficiency for TN, the main reasons were these DO was quite high in wetland so that restrained the process of denitrification and nitrate represented a low percentage of TN so that did not play a lead role in removing nitrogen. And with the experiment, the substrate gradually reached to saturation and temperature gradually declined, the removal efficiencies for COD and TP also declined.With the increasing number of constructed wetlands being built, the modelling of wetland function and performance is valuable. This work examines the efficacy of applying a numeric model(SubWet 2.0) originally designed for horizontal subsurface flow wetlands to model wastewater treatment within vertical subsurface flow constructed wetlands. The treatment efficiencies of two constructed wetlands with substantially different influent characteristics, one in Canada and one in China, were modelled with SubWet 2.0 and simulated values were then compared to observed values to determine how closely SubWet 2.0 reflects the actual observed performance of these wetlands. The model was calibrated to each wetland with observed data that had been collected prior to the simulations. The correlation coefficient(R) and Nash-Sutcliff coefficient of efficiency(NSE) were used to evaluate the modelling performance for 5-day biochemical oxygen demand(BOD5), ammonium nitrogen, nitrate nitrogen and total phosphorous(TP). The results showed that the modelling performance for TP and BOD5 was better for these parameters than that observed for ammonium nitrogen and nitrate nitrogen for either of the two wetlands. For TP and BOD5, the correlation coefficient R achieved a value of 0.79 for the wetland receiving stormwater and exceeded this value for the Canadian wetland receiving domestic wastewaters. For nitrate nitrogen, the wetland treating domestic waste showed a correlation coefficient R as high as 0.97, while the wetland treating stormwater runoff had a correlation coefficient R of 0.48. For ammonium nitrogen, both wetlands showed low correlation coefficients with values of 0.70 and 0.60 for domestic wastewater and for stormwater runoff, respectively. This study demonstrated that SubWet 2.0 is suitable for the modelling of constructed wetlands. The two case studies, with substantial differences in the characteristcs of the influents, demonstrated that Subwet 2.0 is a versatile and robust tool for modelling of constructed wetlands.