Quantitative synthesis of the denitrifying bioreactor hydraulic retention time-nitrate removal efficiency relationship

Abstract

Denitrifying woodchip bioreactors consistently exhibit a strong positive relationship between hydraulic retention time (HRT) and nitrate (N) removal efficiency (NRE), expressed as a percent. This robust HRT-NRE linear regression provides an empirical tool for simply estimating bioreactor performance, but regression parameters (slope, y-intercept) vary across studies. The objective of this quantitative synthesis was to assess the impact of study conditions (e.g., water temperature, N concentration, woodchip porosity, study scale) on the HRT-NRE linear regression slope and y-intercept. A total of 60 HRT-NRE linear regression slopes and y-intercepts were extracted from 26 woodchip bioreactor studies. The individual regression slopes ranged from 0.13 to 18 %/h (median: 2.9 %/h) and y-intercepts ranged from −19 to +47 % (median: 0.37 %). The HRT-NRE slopes were positively correlated with water temperature (Spearman correlation coefficient ρ: +0.39) and negatively correlated with inflow N concentration (ρ: −0.46). The regression y-intercepts were not correlated with any parameter compiled here; they may be a function of other water or wood chemistry parameters. Generalized HRT-NRE models were developed for the entire database (n = 60) and for a subset of data most representative of subsurface drainage (“Tile Proxy”, n = 24). The simple linear regression linking denitrifying bioreactor HRT and NRE is a straightforward empirical tool that will benefit from further: (1) assessment of varying the HRT using the flow rate versus sampling along the reactor length and (2) increased transparency in methods for estimating woodchip porosity.