Mathematical modelling technique to mitigate soil moisture measurement inaccuracies under the conditions of drip irrigation

The paper studies the application of mathematical modelling of moisture transport as a tool for assessing the accuracy of measurements in the process of drip irrigation management in production conditions. We propose a novel modelling framework aimed at solving the problem of automated detection and correction of measurement inaccuracies, which can be implemented in decision support systems in irrigation. At first, the problem of predicting the dynamics of soil moisture content is studied on the base of the two-dimensional Richards’ equation stated in terms of water head. The values of model’s parameters are identified using the particle swarm optimization algorithm. The ability to assess the reliability and accuracy of measurements was tested by studying deviations of the simulated values from Watermark sensor readings along with the variations within the growing season and between several seasons of identified values of model’s parameters that can serve as indirect indicators of changes in soil structure. The results of computational experiments prove the stability of the considered technique to the amount of initial data and to the level of potential errors, and the ability to identify and correct errors of soil moisture measurements in the conditions of drip irrigation.