Conclusion

The results partially supported the first hypothesis, that Models C1, C2, and C3 which had the largest water volumes would release the largest quantity of methane, and that Models A1, A2, A3 and D1, D2, D3, with the smallest water volumes, would consequently release the smallest quantity of methane. However due to additional decomposition that took place in models C1-3, the models with the largest water volumes also demonstrated the greatest increase in carbon dioxide over time (see Figures 1 & 3 in "Analysis of Data" section).

In accordance with the second hypothesis, all the carbon dioxide fluxes dropped with the addition of ammonium nitrate to the models (see Figure 2). Although the CO2 fluxes produced by models D1-3 were also reduced, this was likely the result of the loss of plant biomass that occurred between the first set of gas samples and the addition of the NH4NO3. Fewer plants would have released smaller quantities of CO2 gas. Contrary to what was originally predicted, the addition of the NH4NO3 nutrient resulted in an increase in all the methane fluxes (see Figure 4). This increase is a direct result of the anaerobic conditions in which the decomposition of this plant matter took place, which results in the production of methane gas by methanogenic bacteria.

Similar to what was predicted in the third hypothesis, Models C1 and C2 demonstrated the largest nitrous oxide flux, as a result of their high saturation (see Figure 5). As a result of its much larger biomass, Model C3 released only a small quantity of nitrous oxide gas; the majority of the nitrates were taken up by the plants, rather than being denitrified. Models A1-3 produced very unexpected nitrous oxide fluxes. Rather than releasing nitrous oxide, they acted as sinks and took in the gas instead. Possible sources of error include the distribution of the gas within the covered models, or any possible inaccuracies in the analysis of the sample, whether by human or machine error.

The fourth hypothesis was entirely correct. Two hours following the addition of the nutrient, Models C1-3 demonstrated the greatest reduction in nitrates. On average, Models A1-3 showed the lowest decrease in nitrate concentration. The decrease in pollutant was dependent on the water volume of the wetland model, the presence of healthy plants, and any naturally-occurring nitrates. Small quantities of nitrate were also measured in Models D1 and D3; which were probably already present in the soil.

The Global Warming Potentials (GWP) of each of the models were calculated, using their plant uptake, CO2, CH4, and N2O fluxes. Because of the many factors affecting the fluxes, there was a large amount of variability between the GWPs of all the models, even those with identical water volumes. Model C1 demonstrated the highest GWP, due to its high nitrous oxide and methane fluxes. As a result of its low methane, carbon dioxide and nitrous oxide fluxes, Model A2 was in fact taking in more gas than it was emitting; the outcome was a net cooling effect.

Therefore, the saturation in a nitrate-filtering constructed wetland system has a measurable effect on its overall Global Warming Potential. The water volumes of the models in this experiment affected the carbon dioxide flux, the methane flux, the nitrous oxide flux, and the uptake flux of each model. However, other variables, such as the health of the plants and the amount of decomposition that took place in each model, created differences in the GWPs of models with identical water volumes. Thus, it is possible to construct a highly saturated nitrate-filtering model with a maximum biomass consisting of hardy macrophytes that filters a large quantity of nitrate, while maintaining a low Global Warming Potential.