Rising nitrate in groundwater is a global environmental and health concern. This is a growing issue for the Canterbury plains in New Zealand, as increasing groundwater nitrate levels have been linked to intensification of land use. It is important to capture information about the trends in groundwater from when they were pristine and unaffected in order to isolate the cause of the increasing trends. It is difficult to establish trends in nitrate concentrations without historic data, however other parameters (such as iron and manganese) that have a more extensive sampling history may be used with statistical approaches to fill these gaps. Here, a model was created to predict nitrate prior to it being included in water quality testing, when iron or manganese was sampled. This study uses a parsimonious modelling approach which includes the use of multiple linear regression (MLR), random forest (RF) and boosted regression tree (BRT) models. A number of features (year, well depth, coordinates, dissolved manganese, dissolved iron, land use, soil type, season and rainfall) are used in the model as they were expected to influence trends in nitrates. But the only significant factors were well depth, year, land use, and latitude (following the direction of groundwater flow). The model's predictive accuracies were measured through use of an R² value. The R² values are highest for the BRT meaning that it is the best model for predicting nitrate, followed by RF models which also have relatively high R² scores compared the MLR model. While this study agrees with correlations between nitrate, iron, and manganese observed in previous research, using these relationships for predictive purposes proved ineffective and thus this study was unable to fill the historic data gaps in nitrate trends. A limitation of this study is that the groundwater is mainly oxic in the study area and further work is needed in mixed or reduced groundwater environments. Oxic environments have high oxygen content, while reduced environments refers to environments where oxygen is absent or very low.