ASSESSMENT OF FUNCTIONS AND PARAMETER-ESTIMATION METHODS IN ROOT WATER-UPTAKE SIMULATION

Limited information is available on the influence of estimates of parameters of the root water uptake (RWU) function on RWU simulation in macroscopic RWU modeling. This study was conducted to assess the influence of (i) parameter estimation methods, linear vs. nonlinear, and (ii) the type of RWU functions, continuous vs. discontinuous, on the estimates of RWU parameters and RWU simulation. The model calibration and validation were carried out using the data for soybean [Glycine mar (L.) Merr.] grown in large soil columns. The estimates of the RWU parameters varied with the estimation method used or the type of function used. The simulations in soil water content (theta(v)) and RWU distribution across space and time varied with the estimates of the parameters used in the flow model. The simulations in theta(v) produced with the estimates from the nonlinear estimation method agreed more closely with the corresponding measured values than did those produced with the estimates from the linear method. A similar trend was observed for the discontinuous vs. continuous functions. The differences in residual sum of squares between the linear vs. nonlinear estimation methods, and that between the continuous vs. discontinuous functions, were significant. The accuracy of the simulations in theta(v) and RWU increased significantly using the nonlinear parameter estimation method and a discontinuous RWU function.