The interface between rivers and groundwater is a key driver for the turnover of reactive nitrogen compounds, that cause eutrophication of rivers and endanger drinking-water production from groundwater. Molecular-biological data and omics tools have been used to characterize microorganisms responsible for the turnover of nitrogen compounds. While transcripts of functional genes and enzymes are used as measures of microbial activity it is not yet clear how they quantitatively relate to actual turnover rates under variable environmental conditions. We developed a reactive-transport model for denitrification that simultaneously predicts the distributions of functional-gene transcripts, enzymes and reaction rates. Applying the model, we evaluate the response of transcripts and enzymes at the river–groundwater interface to stable and dynamic hydrogeochemical regimes. While functional-gene transcripts respond to short-term (diurnal) fluctuations of substrate availability and oxygen concentrations, enzyme concentrations are stable over such time scales. The presence of functional-gene transcripts and enzymes globally coincides with the zones of active denitrification. However, transcript and enzyme concentrations do not directly translate into denitrification rates in a quantitative way because of non-linear effects and hysteresis caused by variable substrate availability and oxygen inhibition. Based on our simulations, we suggest that molecular-biological data should be combined with aqueous chemical data, which can typically be obtained at higher spatial and temporal resolution, to parameterize and calibrate reactive-transport models.

Bioretention cells are a Low Impact Development (LID) technology that is being promoted as a green solution to attenuate urban stormwater nutrient loadings. Despite extensive implementation of bioretention cells in Canada, the mechanistic understanding of phosphorus (P) cycling in bioretention cells is still limited. We conducted detailed analyses of (geo)chemical and hydrological data coupled to numerical reactive transport modeling to simulate the fate and transport of P in a bioretention cell located in Mississauga (Ontario, Canada) within the Credit River watershed. Our objective is to utilize the model to predictively understand the accumulation and speciation of P in the bioretention cell under long-term field operation. Unlike existing bioretention models, our model incorporates a detailed representation of the biogeochemical processes that control P cycling in the bioretention cell. We further compare the model predictions with data from sequential chemical extractions of P from soil samples taken from the bioretention cell. The model correctly estimates the cumulative TP (total P) and SRP (soluble reactive P) outflow loadings from the bioretention cell, as well as the TP accumulation rate and observed partitioning of P over the different pools in the bioretention cell. The relative importance of various processes controlling P retention are assessed using mass balance calculations and sensitivity analyses of the model. The results show that filtration of fine P-containing particles and slow sorption are the main processes retaining P in the bioretention cell.