Coupled the DSSAT and Hydrus-1D for Soil Water Dynamics Simulation in the Soil-Plant-Atmosphere System.

The accurate estimation of the soil water balance is important for determining availability of water resources and for the optimal management of limited water resources. Evapotranspiration and leaching are the main water loss processes in the soil-plant system that determines soil moisture status. Accuracy in soil moisture change and evapotranspiration simulations affects crop yield simulations as well. The DSSAT is a suite of field‐scale, process‐based crop models that simulate the phenological development of crops and detailed yield components from emergence till maturity on the basis of crop genetic properties and environmental conditions. The hydrologic models can estimate the soil water profile with high accuracy. HYDRUS-1D is a hydrological model that simulates water flow in one-dimensional variably-saturated soils using numerical solutions of Richards’ equation. Both the DSSAT and HYDRUS-1D have been widely used and extensively tested. The tipping bucket water balance approach is currently used in DSSAT for hydrologic cycle processes and water redistribution. The objectives of our study were: (1) to couple a hydrological model (HYDRUS-1D) and use it in a crop models (the DSSAT) for soil water dynamics simulation; (2) to evaluate coupled models using field experimental datasets distributed with DSSAT for different environments and (3) to compare the HYDRUS-1D water simulation results to that of the tipping bucket approach using the same measured datasets. Modularity of these packages made it easy to couple them. The coupling was done with code wrapping. Time steps and soil profile grids were synchronized. The pairing provided to the DSSAT interface an ability to use both approaches. Evaluation of these approaches was based on their ability how accurately they estimated soil moisture and evapotranspiration. The effects of soil moisture simulations on other simulated hydrological variables and crop yield components were also assessed. The results that demonstrate where two approaches were most different in their simulations are given in the presentation.