Assessing and modeling the spatial variability of soil water redistribution and wheat yield along a sloping landscape

Assessing and modeling the spatial variability of soil water redistribution and crop yield along a sloping landscape is a prerequisite for a better understanding of site-specific-management. The objectives of this research were to (1) evaluate and improve, when appropriate, the vertical soil water dynamics in the water balance portion of the CERES model, (2) develop a simple functional model to simulate lateral downslope soil water flow, and (3) combine remote sensing and crop modeling to predict the spatial variability of wheat yield grown on a sloping landscape. The daily change of soil water content (SWC) in a recent version of the crop model CERES is estimated from the difference between the initial and residual SWC of the water balance components multiplied by a transfer coefficient representing the fraction of the remaining soil water that can be removed in the processes of soil evaporation, vertical drainage and root water uptake. The transfer coefficients for the three processes is assumed to be fixed for all soils. The residual SWC values depend on the input soil properties for the SWC at air dry, drained upper limit, and lower limit of plant availability, respectively. Testing the dependancy of the drainage and evaporation transfer coefficients on soil characteristics was done by monitoring the SWC. The drainage and evaporation transfer coefficients were found to be soil specific and highly correlated with the drained upper limit SWC. Refinements in the drainage and second stage evaporation models improved their accuracy for different soils.; A functional model was developed to simulate downslope lateral soil water flow based on Darcy's Law and equations for estimating unsaturated and saturated hydraulic conductivity. The model requires inputs of the drained upper limit and saturated SWC, water table level, slope, and the amount of incoming flow. Soil water profiles and water table levels were monitored at 15 locations along a sloping landscape to test the model. The model performed reasonably well in estimating lateral soil water drainage. A wheat crop grown on the 6 ha field where lateral flow was studied provided an opportunity to assess spatial variation in yield as influenced by soil spatial variability of differences in water supply related to position in the landscape. Remote sensing from an aircraft helped to quantify spatial variability in leaf area index (LAI) in the field. When this spatial variation in LAI at anthesis was input in the CERES-Wheat model as an alternative to predicting LAI uniformly for the whole field, the modeled spatial variation in yield agreed quite well with the variation in yields monitored for the entire field. The experiments done for this research demonstrated the need to take both vertical and horizontal water flow into account for sloping land in humid regions in order to adequately describe causes of spatial variability in crop yields.