Coupled processes in seasonally frozen soils : Merging experiments and simulations

Sammanfattning: Soil freezing/thawing is of importance in the transport of water, heat and solute, with coupled effects. Due to complexity in soil freezing/thawing, uncertainty could be influential in both experimentation and simulation work in frozen soils. Solute and water in frozen soil could reduce the freezing point, resulting in uncertainty in simulation water, heat and solute processes as well as in estimation of frozen soil evaporation. High salinity and groundwater level could result in high soil evaporation during wintertime. Seasonal courses in energy and water balance on surface have shown to be influential to soil water and heat dynamics, as well as in salt accumulation during wintertime. Water and solute accumulated during freezing period resulted in high evaporation during thawing period and enhanced surface salinization. Diurnal changes in surface energy partitioning resulted in significant cycle of freezing/thawing as well as in evaporation/condensation in surface layer, which could in turn affect atmosphere. Uncertainties in experiments and simulations were detectable in investigation of seasonally frozen soils with limited methods and simplified representations of reality in two agricultural fields in northern China. Soil water and solute contents have shown to be more uncertain than soil temperatures in both measurements and simulations. The combination of experiments with process-based model (CoupModel) has proven to be useful in understanding freezing/thawing processes and in identification of uncertainty, when Monte-Carlo based methods were used for evaluation of simulations. Correlations between parameters and model performance indices needed to be taken into account carefully in calibration of the process-based model. Parameters related to soil hydraulic processes and surface energy processes were more sensitive when using different datasets for calibration. In using multiple model performance indicators for multi-objective evaluation, the trade-offs between them have shown to be a source of uncertainty in calibration. More proper representations of the reality in model (e.g., soil hydraulic and thermal properties) and more detailed measurements (e.g., soil liquid water content and solute concentration) as input would be efficient in reducing uncertainty. Relationships between groundwater, soil and climate change would be of high interest for better understanding of cold regions water and energy balance.