ABSTRACTIncreasing interest in solute transport phenomena in agricultural systems on a sub‐annual basis necessitates a better understanding of seasonal changes in natural systems and how these changes can be incorporated into modelling. A better understanding of the seasonal timing of nutrient loading in tile drained agricultural systems in particular is essential for efforts trying to replicate or predict the occurrence of harmful algal blooms. Literature exists showing there are seasonal dynamics (freeze–thaw, plant‐root processes, land management practices, etc.) that may cause changes in the hydraulic properties of the soil zone including hydraulic conductivity and porosity. To test whether these changes are important in an agricultural system, a MODFLOW‐6 model using the unsaturated zone flow package was constructed. The simulation was comprised of separate, seasonal models to be run sequentially with each year being broken into a winter and summer seasons. As part of this architecture, model parameters representing soil hydraulic properties were allowed to vary by season. The model was calibrated against soil moisture observations at multiple depths using a genetic algorithm machine learning technique. The parameters of the sub‐models were compared for the winter and summer seasons. Brook‐Corey epsilon, saturated vertical conductivity, saturated volumetric water content and residual volumetric water content were found to be consistently different between the modelled summer and winter periods. A more traditional model which did not allow hydraulic properties to vary seasonally was also run and compared to the seasonal architecture and the seasonal architecture was found to improve simulation results. The hydrologic dynamics of the unsaturated zone—particularly in tile drained agricultural systems—control the residence time for water and solutes, which is critical for in‐field chemical processes such as denitrification. This work has important implications for seasonal transport phenomena in agricultural systems and improving the simulation and prediction of harmful algal blooms.
Read full abstract