Prediction of Soil Water Content Using Multiple Regression Model with Time Series Data

A multiple regression model was used to predict the averaged soil water content in root zone instead of solving a Richards' equation numerically, which of applicability was examined. (1) The soil water content of the depth of 5cm (model I) or (2) the effective rainfall and evapotranspiration (model II) were used as the explanatory variables of the multiple regression model. As a result, it became clear that both models could predict the averaged soil water content in the root zone. In addition, it was necessary to convert the original data to its difference and to apply them to the multiple regression model because it was judged that the decrease of the autocorrelation coefficient for the time series data of soil water contents was slow, and therefore, the explanatory variables were correlated each other.