Agricultural tile drainage is one of the major causes of increasing nitrate (NO3-) concentrations in surface water bodies thanks to the usage of nitrogen fertilizers and manure. Denitrifying bioreactors are constructed at the edge of agricultural lands in order to remove NO3- from drainage water through labile carbon substrates intended to promote denitrification.
This field-scale study is the examination and modeling of NO3- removal performance of a woodchip bioreactor installed in Corvallis, OR. During flow periods, water samples were collected on a weekly basis for lab analysis of nitrate, nitrite and ammonium. NO3- concentrations measured in the influent varied in the range from 2 to 13 mg/L (on average 8 mg/L) while effluent concentrations averaged 6 mg/L (1 - 12 mg/L). Results showed that mean volumetric NO3- removal rate achieved by the bioreactor throughout the study period was 21 g N/ m3 /d and the average percent NO3- reduction was 26% that fell within the range of reported values. These findings further indicated that
woodchip bioreactors operating under cold-weather environmental settings are effective means to removal of NO3- loads from agricultural landscapes.
The model that integrates simulated drainage discharge into temperature-dependent denitrification rates was validated using observed effluent nitrate with a Nash-Sutcliffe efficiency coefficient (NSE) value of 0.506, indicating the efficacy of the model in line with model evaluation criteria. Based on the univariate analysis conducted, the most sensitive model parameters are influent nitrate, hydraulic retention time and water temperature, respectively.
The MIN3P code was also utilized to predict NO3- concentrations along the length of the bioreactor for a simulation interval of a day and nitrate levels in effluent water through the monitoring period. The simulated nitrate concentration profile suggests that nitrate removal occurs primarily within the first few meters of the denitrification bed. The assessment of MIN3P model resulted in a NSE value of 0.043 and a coefficient of determination (R2) of 0.416.