Integration of farm-scale and catchment-scale models to quantify the effect of mitigation practices on nitrogen load in the Tararua catchment : a thesis submitted in partial fulfilment of the requirements for the degree of Master of Environmental Management, Massey University, New Zealand
Leached nitrogen from pastoral lands has been identified as a key contaminant contributing to detrimental water quality in New Zealand agricultural landscapes. However, the effectiveness of various management and mitigation measures at reducing the overall catchment nitrogen load in rivers is poorly understood, especially at catchment scale. A robust evaluation of the impact of mitigation measures on water quality is constrained by the limitations of current modelling techniques, particularly their capability to model spatial and temporal variability in nitrogen transport, loads and the potential attenuation of nitrate in flow pathways between the root zone on farms and the receiving waters. This thesis aimed to develop a methodology to integrate a farm-scale nutrient budgeting model, Overseer, with a catchment-scale hydrology model, eWater SOURCE (SOURCE), to determine the effectiveness of farm- and catchment-scale mitigation practices at reducing the dissolved inorganic nitrogen (DIN) load in rivers. A SOURCE model for the Tararua Catchment was set up and calibrated. This model defined and mapped 3,996 functional units (i.e., hydrologic response units) to model the spatial variability of relevant catchment characteristics including climate, land use, soils, and underlying geology. This informed the parameterisation of model inputs related to: rainfall runoff, water flow pathways, and nutrient generation and attenuation. The Overseer estimates of root zone nitrogen (predominantly nitrate) losses for different combinations of land use, soil, and climate (rainfall) were integrated using a look-up table approach with SOURCE to predict river DIN loads across different sub-catchments in the Tararua Catchment. A simple mixing model, which assumed that the average annual nitrate losses from the farm root zone (modelled by Overseer) were mixed into interflow and percolation to groundwater (modelled by SOURCE) from the soil profile, was used to calculate the DIN concentrations in slow flow (groundwater) and quick flow (surface runoff and interflow) components from different functional units. A comparison of the modelled annual average root zone nitrate losses with the measured average annual river DIN loads suggested that between 10 and 90 % of nitrate losses are attenuated, likely through subsurface denitrification, across different sub-catchments. The integrated SOURCE model was calibrated and validated by comparing the modelled and measured daily river flows and monthly DIN loads at six sites in the study catchment. The model was able to predict river average monthly DIN loads across different sub-catchments more accurately when spatially variable nitrogen attenuation factors, based on soil and geological characteristics, were applied to both slow flow and quick flow pathways. The modelling efficiency (NSE) ranged from 0.6 to 0.8 while percent bias (PBIAS) ranged from -2 to 15, indicating an acceptable calibration of the model. The model was then used to investigate the potential effects of various farm- and catchment-scale scenarios to reduce river DIN loads in the study catchment. The modelling results suggest that the catchment-scale scenarios of targeted drainage (quick flow) management, matching intensive land use (dairy farming) to potentially high nitrate attenuation capacity land, and 6 ha of wetlands resulted in a reduction of 11%, 13% and 8%, respectively, in the average annual river DIN load in the study catchment. In contrast, modelling a reduction of 10 to 30% reduction in the average annual root zone nitrate losses from both sheep/beef and dairy farming areas resulted in a 6 to 19% reduction in the overall average annual river DIN load. Interestingly, modelling a reduction of 30% reduction in the average annual root zone nitrate losses only from sheep/beef and dairy farming areas over low to medium nitrogen attenuation capacity lands also resulted into similar level of 15% reduction in the overall average annual river DIN load. This highlights that it is crucial to reduce root zone nitrate losses from headwater catchments and those with highly permeable soils and geology, high rainfall and low subsurface nitrate attenuation capacity. The findings of this thesis clearly suggest that catchment-scale mitigation practices can reduce the river DIN load on a catchment-scale, without significantly impacting farm production, and should be targeted to specific areas. The innovative methodology to integrate farm-scale (such as Overseer) and catchment-scale models (such as SOURCE) described here can be further developed to help identify targeted and effective water quality management measures.