Average total nitrogen was similar across all hydrologic zones and decreased with depth. The C:N ratio ranged from 8.9 to 11.7 and was relatively consistent with depth in all hydrologic zones . Average KCl-extractable NO3 and NH4 were highest in the flow path zone and generally decreased with depth.There was no significant difference in DNP for any wetland zone between ambient conditions and glucose C-source amendment. However, adding glucose and nitrate significantly increased DNP in all three hydrologic zones . The largest increase was seen in the flow path zone soils at 10 cm . In the upper 10 cm, it is also notable that the maximum measured DNP under non N-limiting conditions was much higher in the flowpath zone. DNP was relatively low and similar for the 50- and 100-cm depths for all hydrologic zones and amendments. Amending the 50- and 100-cm depth soils with glucose or nitrate produced no significant response in DNP. When glucose and nitrate were added there was a slight increase in DNP in a few instances .The first statistical model tested was the most complex and hypothesized that log10 DNP was the result of wetland zone, sample depth, amendment, and soil organic carbon content. The two-way interactions of depth and hydrologic zone, depth and amendment,blueberry pot and hydrologic zone and amendment were also included in this preliminary model. Soil organic carbon content had no significant effect on DNP. There were also no significant interaction effects between depth and hydrologic zone, or zone and amendment.
Depth had a highly significant effect on DNP . DNP was an order of magnitude higher at 10 cm than at the 50- or 100- cm depths for all treatment/wetland zone combinations . There was no significant difference in DNP between the 50- or 100-cm depths. Because of the huge disparity in DNP among depths, the 10-cm depth was separated from the 50- and 100-cm depths for further analysis. At the 10-cm depth, both wetland zone and amendment had a significant effect on DNP, but there was no significant interaction between amendment and wetland zone.With all amendments, DNP showed the following pattern among hydrologic zones: flow path > fingers > uplands . This same pattern was observed with the nitrate removal rates that were calculated from piezometer/pore water depth profiles. The calculated removal rates were similar in magnitude to DNP removal rates.The wetland was highly effective at removing nitrate with an estimated 75% total NO3-N removal efficiency for surface and subsurface flow paths combined . The wetland received 5127 kg of NO3-N from input water originating from agricultural return flows and exported 714 kg of NO3-N in output water during the 6-month irrigation season . Approximately 4122 kg NO3-N infiltrated into the wetland soil as seepage, and of this amount, 547 kg NO3-N was lost as seepage below 100 cm. Thus, 3866 kg of the in flowing NO3-N load was either immobilized biologically via plant and microbial uptake or lost from the system via biotic and abiotic transformations . Patterns of N loading from inflows were similar among zones; increasing in the middle of the season. Outflow N loads were consistently low throughout the study period .
NO3 loads lost via deep seepage were low during the beginning of the season and remained low in flow path and upland zones. In the finger zone, however, a dramatic increase in NO3 seepage loads occurred from late June through September . Seasonal retention efficiencies for NO3-N loads in seepage water were 95, 81, and 70% for the flow path, finger, and upland zones, respectively . A moderate decrease in surface water NO3-N concentration between inflow and outflow locations indicates some NO3-N removal via surface processes , however, high measured rates for DNP in surface soil and significantly lower pore water NO3- N concentrations at 50- and 100-cm depths indicate that subsurface denitrification was a dominant nitrogen removal mechanism . Notably, the NO3-N removal rate estimated via non-nitrate limited DNP values considering all wetland zones was similar to that calculated from the mass balance. Considering all hydrologic zones, NO3-N removal estimated from DNP was 5085 kg NO3-N. This estimate was slightly higher than the estimate of NO3-N removed via the mass balance .Despite the large amount of water lost as vertical seepage , overall NO3-N removal washigh in this restored wetland and comparable to that of other regions with temperate climates. Other studies of wetlands receiving agricultural runoff report NO3-N removal efficiencies ranging from0toashigh as99%.Comparisons of wet land-Ntreatment capability, however, is challenging in agricultural settings, because climate, flow characteristics , N species and N load vary across a wide range of temporal and spatial scales. Wetland characteristics also vary widely . The fact that there was no significant difference between NO3 concentrations at the 50- and 100-cm depths for a given wetland zone suggests that nearly all NO3 removal in this system occurred at depths above 50 cm.
Depth profiles suggest that nitrate removal is uniformly low at depth across all wetland zones . Trends in DNP for N-unlimited conditions were consistent with the nitrate losses observed in piezometer water samples.Denitrification potentials in this wetland were highly variable depending on amendment, depth and wetland environment. DNP measured in this study, ranged from non-detectable to over 15,000 mg NO3-N m 2 d 1 , which spans the range of DNP rates reported by several studies . Average DNP in the main flow path zone was higher than rates reported from wetlands receiving agricultural runoff in other regions, however, DNP in fingers and uplands was similar to that in other studies . Wetland soil properties that influence spatial patterns in denitrifying bacterial communities are pH, redox potential, temperature, soil texture, labile organic carbon, and nitrogen . With the exception of KCl-extractable N these properties were similar throughout the wetland in the upper 10 cm, so it is hard to assess the apparent differences in denitrifier activity based on these soil properties alone . Organic carbon, KCl-extractable N, and pore water nitrate were substantially lower at the 50- and 100-cm depths, so it is possible that denitrifier activity was limited at the lower depths by lack of substrate. The observed lower denitrification potentials at depth are consistent with other studies of constructed wetlands . Some studies in constructed wetlands have found DNP to be spatially uniform . In contrast, we found large differences in space,nursery pots with DNP being higher in the main flow path. Differences in DNP between hydrologic zones at the 10-cm depth may be explained by spatial variability in organic carbon content, differences in redox potential, sedimentation and organic matter quality. Highly variable inflow water fluxes resulted in fluctuating water depths across the wetland, with brief dry-down periods in the finger zones and long dry periods in the upland zones . Higher redox potentials in the upland and finger zones may have contributed to spatial differences in DNP . Many studies report that DNP is more strongly correlated with available carbon rather than total organic carbon . Only organic carbon was measured for this study, so it is possible that there may be substantial differences in carbon availability between the environments that may affect DNP. Also, DOC, which may serve as an important energy source for denitrifiers, was relatively constant across hydrologic zones and soil depths. Sediment deposition in the flow path zone was substantially higher than in the fingers or upland zones, which offers a possible explanation for the disparities in DNP despite similar soil conditions . Areas of active sediment deposition may receive organic matter of different quality compared to that of the native soil from which the wetland was constructed . It is also possible that the sediment, which originated from surrounding farmland, is a seed source of denitrifying bacteria. In fact, studies have shown that frequently tilled agricultural soils in the region have more facultative anaerobes and higher denitrification rates compared to untilled soils .
Since this wetland has only received tail waters for two seasons, it is plausible that we are witnessing the initial stages of recruitment of microbial populations and the associated evolution of wetland bio-geochemical processes. Thus, in older wetlands DNP may be expected to be more uniform. Other studies have shown that spatial variation in denitrification corresponds to patterns in nitrate concentration, increasing in areas of high N loading . Hernandez and Mitsch found higher denitrification potentials in constructed wetland soils where emergent macrophytes were present, when compared to unvegetated constructed wetland sediments. Since vegetation was sparse in both the finger and the flowpath zones, it is unlikely that the relative amount of vegetation had much effect on the observed denitrification potentials. A disproportionately high amount of the nitrogen was removed in the flow path zone compared to the fingers and uplands . This trend was a result of higher nitrate loading rates and significantly higher DNP rates in the flow path compared to other hydrologic zones. The higher mean N-amended DNP rates in the flow path suggest a larger denitrifier microbial population in this zone. The finger zone, although accounting for 40% of seepage, is responsible for the majority of NO3-N lost via deep seepage. This was the result of significantly lower DNP rates relative to the flow path zone and significantly higher pore water NO3-N concentrations at the 100-cm depth . As with any biological process, temperature strongly regulates denitrification rate . Lab incubations were performed at the mean field temperature , which was similar to the mean temperature of the flow path and 0.5 C higher than that of the fingers . Warm daytime temperatures are likely to substantially increase denitrification rate over diurnal timescales.Other NO3-N removal pathways may play an important role in this wetland. NH4 accumulation in pore water, and elevated KCl-extractable NH4 concentrations in the soil at 10 cm suggests that sulfur or ferrous iron-driven nitrate reduction may play a role in nitrogen cycling in this system . Redox potential frequently reached the sulfate reduction level suggesting the presence of free sulfide. Anecdotal evidence such as H2S smell in groundwater samples, as well as visual identification of iron monosulfides in the sediment verifies the presence of sulfide in the system . At high concentrations, free sulfide is known to inhibit the final two reduction steps in the denitrification sequence, which may drive the reduction to ammonium rather than N2O and N2 . Sorption of ammonium from seepage water to cation exchange sites in the soil may also account for accumulation of ammonium in the upper 10 cm of sediment . Equilibrium with the sediment bound ammonium would result in elevated ammonium concentrations in the associated pore water. Despite the predominately unvegetated main flow path, plant uptake may play a substantial role in nitrogen cycling in this wetland. There may be diffusion of NO3-N from surface water into the upland areas via the shallow water table in the upland zone located approximately at the same elevation as the wetland water surface. The dense vegetation in the upland areas may assimilate a significant amount of N thereby increasing N removal rates.This study demonstrated that soils of recently restored wetlands have the capacity to remove large nitrate loads from vertically percolating water with low risk to groundwater in California’s Central Valley. Bio-geochemical processes in this wetland facilitated significant removal of nitrate inputs from agricultural tail waters. The active flow path of the wetland had the highest DNP at the 10-cm depth under all N amended conditions, and also experienced the greatest sediment deposition rates, nitrogen load and seepage volume . While the flow path had a significantly higher DNP relative to the other zones; the finger environments had a significantly higher DNP relative to the upland environments. These significant differences in DNP between zones may have resulted in the substantial differences in NO3 removal efficiencies, with 95, 81 and 70% reduction in NO3 seepage load in the flow path, finger and upland zones, respectively . Nevertheless, high NO3 removal efficiency in the flow path resulted in a high overall net decrease of NO3 load from seepage water for the entire wetland.