Modeling water infiltration rate under conventional tillage systems on a clay soil using artificial neural networks

This study presents the application of artificial neural networks for modeling the parameters of Lewis-Kostiakov infiltration under conventional tillage systems on a clay soil. The conventional tillage systems were moldboard, chisel and rotary plows. Water infiltration rate was defined experimentally by double ring infiltrometer. Artificial neural network estimation indicated strong correlations (R² = 0.999) between the parameters of Lewis-Kostiakov infiltration (I=ktⁿ) and affected variables (soil total porosity, soil moisture content, working index and aspect ratio). The simulated data from the developed artificial neural network formulated the parameters of Lewis-Kostiakov infiltration (k and n) as a function of tillage implement weight and width, speed and depth of plowing, tractor nominal power, soil total porosity and soil moisture content with R2 around 0.60. The developed model can help managers of irrigation systems to modify field practices during growing season to conserve irrigation water. The working index has more contribution on constant (k). Meanwhile, soil total porosity has more contribution on constant (n). Using the developed model, infiltration rate could be optimized during seedbed preparation process.