Agriculture is one of the main sources of phosphorous and nitrogen (P and N) contributing to cultural eutrophication of freshwater lakes and estuaries. In cold regions, the effects of agricultural management practices used to mitigate the runoff loss of these nutrients remain uncertain. In particular, the use of forage crops and minimum tillage, have not reduced some forms of P and N in runoff to streams, in part, as a result of freeze-thaw induced losses of mobile P and N from forages and crop residues. The purpose of this research is to improve the current understanding of the controls on P and N loss from Canadian Prairie fields to ultimately aid in the development and evaluation of beneficial agricultural management practices that perform predictably in cold regions. This study aims to provide new insights into the effects of cold regions hydrological processes on runoff quality through the development and application of a novel inductive - deductive modelling approach. Runoff flowpaths resulting from the three infiltration regimes identified for frozen soils (Granger et al., 1984) are hypothesized to impact the chemistry of field scale meltwater runoff by varying meltwater interaction with agricultural soils and vegetation. Hydrochemistry data from six intensively monitored minimum tillage and forage cropped fields in South Tobacco Creek, Manitoba were used to develop a nutrient model to integrate with a physics-based hydrological modelling platform that can represent the frozen soil infiltration regimes, in addition to other important cold region hydrological processes. The inductive development of a nutrient model, integrated with a deductive physics-based hydrological platform, enabled the modelling of meltwater flowpaths and freeze-thaw induced losses from vegetation. Further testing of the developed model and field experimentation are required to test the hypothesis that runoff generated over a basal ice layer eliminates the transfer of soil nutrients to runoff. Comparison of predicted and observed field scale runoff concentrations and masses suggest that this method of inductive-deductive model development has potential to predict the performance of agricultural management practices in cold regions.