This work proposes to adjust mathematical models to predict the parameters of van Genuchten equation (thetaR, thetaS, alpha and n), besides a model to predict the soil moisture corresponding to the field capacity. The models express the equation parameters as dependent variables and sand, clay, organic matter, bulk density and depth as independent variables. All independent variables can be easily and routinely determined in laboratory. The samples were collected at 12 points in and three layers (0-30; 30-60; 60-90 cm) of a Cambisol, using a Uhland sampler. The soil moisture at matric potentials of 2, 6 and 10 kPa were determined using a porous plate apparatus and those at matric potentials of 33, 100, 500 and 1500 kPa were determined using a pressure plate apparatus. The models were adjusted with the SAS for Windows program, ProcReg routine and Backward procedure to select the variables significantly different from zero at the probability level adopted. The coefficient of determination and prediction errors for the parameters predicted and soil moisture generated by van Genuchten equation adjusted with parameters predicted by models, were analyzed to evaluate the models. The coefficient of determination for models was high and the errors were small justifying the application of models.
irrigation; soil moisture; water retention curve
