Exploring the relationship between groundwater geochemical factors and denitrification potentials on a dairy farm in southeast Ireland

Abstract

Nitrate (NO 3 −) loss from agriculture to shallow groundwater and transferral to sensitive aquatic ecosystems is of global concern. Denitrifying bioreactor technology, where a solid carbon (C) reactive media intercepts contaminated groundwater, has been successfully used to convert NO 3 − to di-nitrogen (N 2) gas. One of the challenges of groundwater remediation research is how to track denitrification potential spatially and temporally within reactive media and subsoil. First, using δ 15N/δ 18O isotopes, eight wells were divided into indicative transformational processes of ‘nitrification’ or ‘denitrification’ wells. Then, using N 2/argon (Ar) ratios these wells were divided into ‘low denitrification potential’ or high denitrification potential’ categories. Secondly, using falling head tests, the saturated hydraulic conductivity ( K sat) in each well was estimated, creating two groups of ‘slow’ (0.06 m day −1) and ‘fast’ (0.13 m day −1) wells, respectively. Thirdly, two ‘low denitrification potential’ wells (one fast and one slow) with high NO 3 − concentration were amended with woodchip to enhance denitrification. Water samples were retrieved from all wells using a low flow syringe to avoid de-gassing and analysed for N 2/Ar ratio using membrane inlet mass spectrometry. Results showed that there was good agreement between isotope and chemical (N 2/Ar ratio and dissolved organic C (DOC)) and physio-chemical (dissolved oxygen, temperature, conductivity and pH) parameters. To explain the spatial and temporal distribution of NO 3 − and other parameters on site, the development of predictive models using the available datasets for this field site was examined for NO 3 −, Cl −, N 2/Ar and DOC. Initial statistical analysis was directed towards the testing of the effect of woodchip amendment. The analysis was formulated as a repeated measures analysis of the factorial structure for treatment and time. Nitrate concentrations were related to K sat and water level ( p < 0.0001 and p = 0.02, respectively), but did not respond to woodchip addition ( p = 0.09). This non-destructive technique allows elucidation of denitrification potential over time and could be used in denitrifying bioreactor technology to assess denitrification hotspots in reactive media, while developing a NO 3 − spatial and temporal predictive model for bioreactor site specific conditions.