Nitrogen Attenuation along Delivery Pathways in Agricultural Catchments

Abstract

Anthropogenic application of inorganic and organic nitrogen (N) to agricultural landscapes has pervasive consequences, including human health implications, eutrophication, loss of habitat biodiversity and greenhouse gas emissions. In light of the negative impact of surplus N losses, intensification of agriculture must be compatible with environmental sustainability. In order to fulfil European Union Water Framework Directive requirements, a fundamental understanding of the processes governing the mobilisation, transport and transformation of reactive N is required. At the catchment scale, agronomic, meteorological, hydrogeological and physicochemical parameters control the spatiotemporal occurrence of N. In this thesis, quantitative field study, chemical analyses and numerical modelling offers a solution to quantify the form and fate of N under a range of climatic, geological and agronomic settings. The project was undertaken from 2012 to 2016 (following on earlier monitoring from 2010) in two ca. 10km2 Irish catchments, characterised by well drained soils. The slate catchment is dominated by arable land overlying Ordovician slate bedrock. The sandstone catchment is characterised by grassland overlying Devonian sandstone bedrock. Both catchments contain two instrumented hillslopes, each of which has a stream channel at its base. Multi-level monitoring wells were targeted to intercept both shallow and deeper groundwater pathways. Average annual N loading (2010-2016) to the sandstone hillslopes (430 kgN/ha/yr.) greatly exceeded the slate hillslopes (145 kgN/ha/yr.). FracLEACH describes the quantity of reactive N leached from soil to groundwater as a proportion of the total N applied. In the sandstone hillslopes, a mean FracLEACH of 18% was calculated. In the slate hillslopes, the mean FracLEACH of 43% highlighted a propensity for greater N leaching losses in arable catchments. Average annual stream NO3- (2010-2016) in the slate catchment (7.0 mgN/L) closely reflected shallow groundwater NO3- (7.2 mgN/L). In the sandstone catchment, average annual stream NO3- concentrations (4.8 mgN/L), approximately 30% lower than the streams of the slate catchment. Groundwater denitrification has the capacity to mitigate stream NO3- enrichment by returning N to the atmospheric pool. The reaction is sequential, with several intermediary products. The differentiation between which reaction product dominates is of environmental concern: dinitrogen (N2) gas is environmentally benign whereas nitrous oxide (N2O) is a potent greenhouse gas. Denitrification was assessed by measuring the concentrations of N species, aquifer hydrogeochemistry, aquifer hydraulic properties and stable isotope signatures from 2013 to 2015. Physical factors (agronomy, water table elevation and permeability) determined the hydrogeochemical signature of the aquifers, which acted as the dominant control on denitrification rates. High permeability, aerobic conditions and a lack of bacterial energy sources in the slate catchment resulted in low denitrification rates (0?32%), high NO3- and comparatively low N2O emission factors (EF5g1). In the sandstone catchment, denitrification rates ranged from 4 to 94% and were correlated to aquifer permeability, water table elevation, dissolved oxygen concentrations and the presence of solid phase bacterial energy sources. Denitrification of NO3- to N2 occurred in anaerobic conditions, while at intermediate dissolved oxygen, N2O was the dominant reaction product. EF5g1 (mean: 0.0018) in the denitrifying sandstone catchment was 32% less than the IPCC default. Denitrification observations across catchments were supported by stable isotope signatures. The effects of agronomic, meteorological and hydrogeological drivers on the spatiotemporal occurrence of groundwater NO3- were measured from 2010-2016, using a multiple linear regression model. Responses in groundwater NO3- concentrations over the main recharge period of each year were averaged at each piezometer and regressed on driver variables that were integrated over a 6-month period preceding the groundwater NO3- response. Soil moisture deficit (SMD) was the dominant meteorological driver influencing groundwater NO3- concentrations. Increases in groundwater NO3- were linked to mineralisation during high SMD periods, followed by flushing of NO3- during high rainfall. The effects of significant hydrogeological drivers on groundwater NO3- were linked to aquifer vulnerability and denitrification capacity. A significant increase in average groundwater and stream NO3- concentrations was measured in the slate catchment over the seven-year period (p < 0.05). Groundwater travel times (tS) and the contribution of groundwater pathways to the stream were characterised using MODFLOW. Numerical 2D ?slice? models for groundwater flow were constructed for one hillslope per catchment. At the sandstone hillslope, tS to the stream ranged from less than one day to 5.1 years over. At the slate hillslope tS ranged from less than one day to 1.3 years. At both the sandstone and slate hillslopes, the minimum tS represented the fastest possible pathways where the source was adjacent to the stream receptor, and groundwater flow was through shallow, high velocity pathways. The relative contribution of shallow pathways to the stream increased during high recharge at both hillslopes. The design and implementation of successful N mitigation strategies requires a process-based understanding of the production, consumption and transport of N, across a range of agronomic and hydrogeological settings. It is hoped that the research presented herein has contributed to that understanding.