Carbon dosing of denitrifying bioreactors to remove nitrate from agricultural drainage

Abstract

Several studies have demonstrated that denitrifying woodchips bioreactors effectively remove nitrate from a broad spectrum of wastewater. Even though bioreactors have a comparatively small spatial footprint and are straightforward to construct, the low long-term nitrate removal efficiency they provide compared to other mitigation options, such as constructed wetlands, may limit their utilization by catchment managers and landowners. This thesis investigated the use of external carbon dosing in bioreactors to improve nitrate removal and quantified some potential adverse effects of carbon dosing on bioreactors and downstream aquatic environments. Through observations ranging from field to mesocosm experiments, this thesis sought to (1) determine whether and to what extent external carbon dosing could improve nitrate removal in bioreactors and (2) assess the potential adverse effects of carbon dosing, such as reduced hydraulic performance and pollution swapping. The first component of the study evaluated the carbon dosing of a pilot-scale (58 m³) bioreactor treating drainage waters from a 0.65 ha paddock on a dairy farm. The effects of carbon dosing on the bioreactor's nitrate removal rates were assessed following a constant dosage of 8% (v/v) methanol solution at 10 mL min⁻¹ to the bioreactor's inlet distributor. Sulfate reduction and losses of added methanol from the bioreactor’s outlet were quantified as the potential adverse effects of external carbon dosing. Methanol dosing increased seasonal nitrate removal rates to 8 g N m⁻³ day⁻¹ in 2020 and 5 g N m⁻³ day⁻¹ in 2021 (with a halved dosing rate), compared to 1 g N m⁻³ day⁻¹ in the undosed bioreactor in 2018. Even under nitrate-limiting conditions, the added methanol was effectively removed in the bioreactor, with a mean removal rate of 106 g CH₃OH-C m⁻³ day⁻¹ . Methanol concentrations decreased by order of magnitude along the bioreactor (lengthwise) and never exceeded 50 mg CH₃OH-C L⁻¹ at the outflow. The added carbon caused sulfate reduction with a mean removal rate of 8.5 g SO4²⁻-S m⁻³ day⁻¹ in 2020 (higher dosing rate) and 0.5 g SO4²⁻-S m⁻³ day⁻¹ in 2021 (halved dosing rate). Carbon dosage at a continuous rate was shown to be a viable method for increasing nitrate removal while minimizing capital and operational costs. The simplicity of this approach is a significant advantage, and it could be used by a variety of industries, including farmers. While field observations showed that carbon dosage could improve nitrate removal, the very variable flows and nitrate concentrations made determining nitrate, sulfate, and methanol removal rates difficult. Therefore, the second component of the thesis was to assess nitrate, sulfate and added carbon removal rates more accurately using mesocosms. An empirical model was also developed to calculate the amount of methanol required to remove specific loads of nitrate. Mesocosm bioreactors were designed to share some characteristics with the field bioreactor (for example, added carbon component and dosage ratio) but diverged in that they were hydrologically regulated to eliminate the impact of transient operating circumstances on bioreactor performance. Carbon dosing increased nitrate removal rates from 7 g N m⁻³ day⁻¹ in the control treatment bioreactors (without methanol addition) to 27 g N m⁻³ day⁻¹ in the methanol-dosed treatment bioreactors. Added methanol was removed in either the absence or presence of nitrate, with mean methanol removal rates of 23 g CH₃OH-C m⁻³ day⁻¹ in nitrate prevailing conditions and 17 g CH₃OH-C m⁻³ day⁻¹ in nitrate limiting conditions. The empirical model suggested that a methanol to nitrate ratio of 0.7 resulted in complete nitrate removal. In the third component of this thesis research, observations from both mesocosm and full-scale bioreactors were used to assess the hydraulic characteristics of bioreactors under various carbon dosage regimes. Field measurements revealed a decrease in hydraulic conductivity from 4601 m day⁻¹ in 2018 (season without carbon dosing) to 1600 m day⁻¹ in 2021 (second year of dosing). Based on tracer tests of the mesocosm bioreactors, I investigated the effects of added carbon on the internal hydraulic performance of bioreactors. The results of the tracer tests revealed that carbon dosing had no significant effect on the hydraulic parameters of the mesocosm bioreactors (p-value > 0.05). The findings of this thesis have implications for improving the performance of existing bioreactors or for developing new, externally dosed bioreactors with smaller spatial footprints. The constant dosage strategy, demonstrated in the full-scale bioreactor, was effective in increasing nitrate removal rates while also being safe regarding added carbon losses to the receiving environments. The ease of implementation of the constant dosage strategy could increase take-up from landowners and managers. The thesis also provides an understanding of whether and to what extent carbon dosing affects the hydrology of bioreactors, which can be used to develop new bioreactor designs with less backflow and bypass.