Performance, denitrification activity, and microbial community dynamics of denitrifying biofilters under fluctuating water level

Agricultural runoff through tile drainage systems is a significant source of nitrogen to coastal water bodies, causing water quality degradation. Denitrifying biofilters have been identified as a technology that reduces the export of nitrate from tile-drained fields. A better understanding of the microbial communities that allow the biofilter to function may provide insight to improve performance. Two studies were performed to achieve this aim. Laboratory-scale biofilters were operated for two years with variations in water level. Nitrate removal performance was measured. Denitrifying enzyme assays (DEAs) were performed regularly to measure the denitrification potential in different parts of the two reactors. A comparison of the microbial communities of agriculture, natural wetland, restored wetland, and biofilter habitats was also carried out to determine the relative effects of niche and legacy on the structure of the microbial communities. For both studies, total bacterial communities were analyzed using automated ribosomal intergenic spacer analysis (ARISA), and denitrifying bacterial communities were analyzed using terminal restriction fragment length polymorphism (tRFLP) of nosZ, the gene encoding the catalytic subunit of nitrous oxide reduction. In the laboratory biofilter study, the performance was sensitive to changing water level, with 31% nitrate removed at high water level and 59% at low water level. The denitrification potential was the same under almost all conditions, ranging from 0.000015-0.004 mg N2O-N/hour/dry g woodchip; the microbial communities were still able to carry out denitrification, unaffected by the water level, unless dried out for an extended period of time. The denitrifying bacterial communities were not significantly different from each other, regardless of water level. This indicates resistance of these bacteria, meaning that the bacterial communities did not change in response to disturbance. The total bacterial communities became more distinct between the two reactors once the regular disturbance period began, with a stronger effect on the most severely disturbed port. The communities were not distinguishable based on high or low water level, though, in the same place during the regular disturbance period, indicating that disturbance communities were created, rather than high or low water level communities. In the habitat comparison, the biofilter microbial communities were distinct from those of the other habitats. This was true for both the total and denitrifying bacterial communities. This suggests a stronger influence of environmental parameters on the microbial community structure than the legacy effects of the agricultural bacteria that were initially present and continued to enter with the influent water. These results show the biofilters to be stable systems with resistant and functionally redundant bacterial communities. Water level and HRT should be considered in the design of the biofilters, as they influenced the nitrate removal. Overall, the biofilters show promise as a method to reduce the amount of nitrogen pollution from agricultural fields to surface water