Modeling daily soil water dynamics during vertical drainage using the incoming flow concept

Crop management models require simulation of daily soil water dynamics. The objective of this study was to develop a model to simulate the daily soil water dynamics during vertical drainage with reasonable accuracy using the incoming flow concept. The execution of this model, which has been developed based on the conservation of mass law, consists of two steps. First, calculating the potential daily change of soil water content (Δθp) for each soil layer in the profile assuming each one receives no water from the above layer. Then, calculating the actual daily change of soil water (Δθa) for each soil layer in the profile by adjusting Δθp using the incoming water flow, which can be defined as the amount of drainage water that reaches a layer in a soil profile from the above layer. The model was compared with the Suleiman and Ritchie [Suleiman, A.A., Ritchie, J.T., 2004. Modifications to the DSSAT vertical drainage model for more accurate soil water dynamics estimation. Soil Sci. 169 (11), 745–757] vertical drainage model (SRVDM) and HYDRUS-1D for diverse soils and was tested using drainage experimental data of a Eutric Regosol in Bekkevoort, Belgium and a sandy soil in Georgia, U.S. The difference in Δθp between the new model and HYDRUS-1D for diverse soils ranged from − 0.01 to 0.016 m3 m− 3 for the first day and from − 0.005 to − 0.025 m3 m− 3 for the second day while the difference in Δθp between the SRVDM and HYDRUS-1D for these soils ranged from 0.014 to 0.062 m3 m− 3 for the first day and from − 0.01 to 0.026 m3 m− 3 for the second day. The relative maximum absolute errors in Δθa between the new model and HYDRUS-1D was 10% while the relative maximum absolute errors in Δθa between the SRVDM and HYDRUS-1D was 112%. In the experiments, the root mean square difference of the soil water content for the new model was lower than that for the SRVDM at the different soil depths. These results indicated that the new model outperformed the SRVDM in simulating Δθp and Δθa for diverse soil. It can be concluded that the new model was robust and reasonably accurate for diverse soils at different soil depths. The implementation of such model will improve the accuracy and applicability of regional soil water dynamics simulation and will reduce considerably the computational time and the required inputs.