Simulations of water, heat, and solute transport in partially frozen soils

Abstract. Experiments for soil freezing/thawing were conducted in two seasonally frozen agricultural fields in northern China during 2011/2012 and 2012/2013 wintertime, respectively. Mass balance was checked based on measured data at various depths. Simulation work was conducted by combining CoupModel with Monte-Carlo sampling method to achieve parameter sets with equally good performance. Uncertainties existed in both measurements and model due to complexity in freezing/thawing processes as well as in surface energy partitioning. Parameters related to surface radiation and soil frost were strongly constrained with datasets available in two sites combining multi-criterion on outputs. Simulated soil heat processes were better described than soil water processes given the data obtained for calibration. Model performance was improved with consideration of solute effects on freezing point depression. More detailed solute transport processes in CoupModel needed to be improved by taking more processes such as diffusion and expulsion into consideration based on more precise experimental results, to reduce uncertainty in model. Generally, combination of measurement with process-based model and Monte-Carlo sampling method provided an approach for understanding of solute transport as well as its influences on soil freezing/thawing in cold arid agricultural regions. Incorporating more detailed descriptions of processes for frozen soil in the model can be justified if uncertainties in measurements can be reduced by introducing of high-precision novel technologies.