Simulating Woodchip Bioreactor Performance Using a Dual-Porosity Model

There is a general understanding in the scientific community as to how denitrifying bioreactors operate, but we lack a quantitative understanding of the details of the denitrification process acting within them and comprehensive models for simulating their performance. We hypothesized that nitrate transport through woodchip bioreactors would be best described by a dualporosity transport model where the bioreactor water is divided into a mobile domain (i.e., the water between the woodchips where it is free to flow and solute movement is by advection and dispersion) and an immobile domain of water (i.e., the water mostly within the woodchips that is stagnant and where solute movement is by diffusion alone). We calibrated the dual-porosity model contained in the HYDRUS model for a woodchip bioreactor using the results of a Br breakthrough experiment where we treated Br as a conservative nonadsorbing tracer. We then used the resulting model parameters to describe 2 yr of NO3 transport and denitrification within a bioreactor supplied by tile drainage. The only model parameters fitted to the NO3 data were either the zero-or first-order denitrification rate and its temperature dependence. The bioreactor denitrified 2.23 kg N (38%) of the NO3 entering it in 2013 and 3.73 kg N (49%) of the NO3 that entered it in 2014. The dual-porosity model fit the NO3 data very well, with fitted zero-order reaction rates of 8.7 and 6.8 mg N L-1 d(-1) in 2013 and 2014, respectively, and corresponding first-order reaction rates of 0.99 and 1.02 d(-1). For the 2-yr data set, both reaction rate models fit the data equally well. Consistent model parameters fitted for the 2 yr indicated that the model used was robust and a promising approach for modeling fate and transport of NO3 in woodchip bioreactors.