Evaluation of Mechanistic Models for Nitrate Removal in Woodchip Bioreactors

Abstract

Woodchip bioreactors (WBRs) are increasingly being applied to remove nitrate from runoff. In this study, replicate columns with aged woodchips were subjected to a range of measured flow rates and influent nitrate concentrations with an artificial stormwater matrix. Dissolved oxygen (DO), nitrate, and dissolved organic carbon (DOC) were measured along the length of the columns. A multispecies reactive transport model with Michaelis-Menten kinetics was developed to explain the concentration profiles of DO, nitrate, and DOC. Four additional models were developed based on simplifying assumptions, and all five models were tested for their ability to predict nitrate concentrations in the experimental columns. Global sensitivity analysis and constrained optimization determined the set of parameters that minimized the root-mean-squared error (RMSE) between the model and the experimental data. A k-fold validation test revealed no statistical difference in RMSE for predicting nitrate concentrations between a zero-order model and the other multispecies reactive transport models tested. Additionally, the multispecies reactive transport models demonstrated no significant differences in predicting DO and DOC concentrations. These results suggest that denitrification in an aged woodchip bioreactor at constant temperature can effectively be modeled using zero-order kinetics when nitrate concentrations are >2 mg-N L-1. A multispecies model may be used if predicting DOC or DO concentrations is desired.