Modelling nitrate & bacterial contamination in the lower Waitaki alluvium aquifer, North Otago, New Zealand


Heller, Thomas B. (Thomas Brendan)


Project type:




Nitrate and bacterial contamination in the Waitaki Alluvium poses a threat to groundwater bore users, connected surface water systems and to further development of the unconfined aquifer. The existing land use is predominantly dairying, and a large volume of dairy shed effluent along with on-farm nitrogen from urea patches and animal dung is irrigated with the existing Waitaki River based irrigation scheme, providing a surface based recharge source to the aquifer. Potable (household) and dairy shed bore-water supplies are very susceptible to degradation from excessive concentrations of nitrate and bacterial counts. Moreover, the ·overlying soils are very porous and depth to the water table is relatively small. Discharges to land may inevitably threaten groundwater and be difficult to remedy under natural conditions. This research involved a physical and chemical modelling approach to determine nitrate (NNN) and faecal coliform (FC) contamination of the aquifer. The modelling process established the geo-hydrological parameters of the groundwater system, a detailed water balance, and geo-chemical analysis of nitrate and bacterial dispersion and decay within the aquifer. The variability in source concentrations of NNN and FC were calibrated on existing land use inventory database information. Initially steady state, then transient model simulations using finite difference numerical modelling calibrated to field evidence, were developed to provide sound resource management criteria for the sustainable management of dairy effluent discharges and ultimately land use nitrate-loading to the Lower Waitaki Alluvium. Modelling consisted of Visual MODFLOW simulations, utilising approximately 60,000, 100 m by 100 m cells containing lumped estimates for hydraulic conductivity, thickness, porosity, dispersion and decay values. Initially the physical MODFLOW model was required to examine stresses and boundaries of the Lower Waitaki aquifer system. The mass-transport model MT3D was then used to simulate nitrate and bacterial flux through the aquifer for steady state, then transient simulations for the period of 1998-2002 (5 years) with monthly time steps/stress periods. The results for calibration and sensitivity analysis of the models were mixed, with good calibration to field data obtained for the physical flow model of root mean squared (RMS) error in heads of <1 m or 2.5%, and target outflows to springs < 5% error. The sensitivity of target heads to percentage variations in recharge and hydraulic conductivity in the models was m similar. The errors associated with NNN and FC concentrations reflected the variability in the data, and the limitations of the modelling process to replicate actual conditions. RMS errors in concentrations for the models were in the order of 25% reflecting 1.5 mg/L NNN and 0.32 cfu/100 mL FC respectively. For the transient water quality model, the overall RMS error for NNN was compared to Kendall trend tests on the data. A slope of +0.232 mg/L/year and a p-value of >0.05 was obtained for the data. The test however was significant at the 10% level (p=0.1). FC data were highly variable and trends were very weak. For the Kendall trend tests, a slope of 0 and a pvalue> 0.2 were obtained for the data. Hence, the model also gave a poor fit to most of the field data. The median values however, showed a very good relationship with the annual flux (seasonal trend) of irrigation water throughout the aquifer system. The modelling results were therefore not statistically significant. However, the modelling process detected a weak, increasing trend for NNN levels, and forward prediction modelling suggested that for a 25% increase in source concentrations, peak aquifer concentrations in some areas would only rise to 6 mg/L by the year 2012. FC concentrations marginally increased in the forward prediction model and were generally widespread in the aquifer at low levels. However, there was significant decay of FC concentrations through the soil profile and FC also decayed readily in groundwater, consistent with published decay rates. Of note in this model was the large buffering capacity of surface waters associated with theW aitaki River, which essentially diluted the nitrate flux from the aquifer to wetlands and springs. The dilution of the irrigation scheme alone is 20: 1 in the current nitrate recharge source.

Thesis description:

x, 104, [16] leaves : ill., maps ; 30 cm.


OU geology Identifier:


Author last name:

OURArchive handle:

OURArchive access level:

Location (WKT, WGS84):

POLYGON ((171.200904436660295 -44.810253486013281,171.193155032467018 -45.057837043510993,170.779292057027817 -45.049941077960078,170.79030280156104 -44.807212399869989,171.200904436660295 -44.810253486013281))




Heller, Thomas B. (Thomas Brendan), “Modelling nitrate & bacterial contamination in the lower Waitaki alluvium aquifer, North Otago, New Zealand ,” Otago Geology Theses, accessed May 22, 2024,

Output Formats