Simulating Woodchip Bioreactor Performance Using the Mobile-Immobile Flow Model.

Denitrifying bioreactors have proven to be effective in removing nitrate from tile drain water before it discharges to surface waters.  While we have a general understanding of how bioreactors work, we do not have a quantitative understanding of the hydraulic and denitrification processes acting within them, nor accurate models for simulating their performance.  We hypothesized that solute transport through woodchip bioreactors would be best described by a mobile-immobile (MIM) transport model where the water-filled void is effectively divided into a mobile domain between the woodchips where water is free to flow and solute movement is by convection and dispersion; and an immobile domain of water within the woodchips that is stagnant and solute movement is by diffusion alone.  We calibrated the MIM model for a woodchip bioreactor using the results of a Br breakthrough experiment conducted in 2013 assuming that Br is a conservative non-adsorbing tracer.  These model parameters were then used to describe two years of nitrate transport and denitrification within the bioreactor supplied by tile drainage.  The only model parameters fitted to the nitrate data was either the 0-order or 1^st- order denitrification rate, and its temperature dependence (Q₁₀).  The model fit the nitrate data very well, with 1^st-order kinetics fitting 2013 data slightly better and 0-order kinetics fitting 2014 data slightly better.  The fitted 0-order reaction rate was 8.7 and 8.2 mg N L^-1 d^-1 in 2013 and 2014, respectively.  The corresponding 1^st-order reaction rates were 0.99 and 1.02 d^-1.  The temperature dependence of denitrification 0- and 1^st-order kinetics in 2013 corresponded to Q₁₀ = 1.2. While in 2013, Q₁₀ = 1.7 for 0-order kinetics and Q₁₀=8 for 1^st-order kinetics.  There was no consistent trend in the fitted parameters from 2013 to 2014 indicating no measureable change in the bioreactor’s ability to remove nitrate.