Simulation of Nitrate Distribution Pattern in Drip Irrigation Systems using artificial neural network

Surplus application of nitrate can lead to contamination of groundwater resources. Accurate knowledge of nitrate distribution in the root zone is essential for the design and management of drip irrigation systems. In this study, artificial neural network (ANN) model is used based on strong pattern recognition technique that made reasonable relations between input and output parameters. In order to simulate the nitrate distribution pattern, the Experiments were set on three different soil textures for surface and subsurface drip irrigation systems. The drippers were installed at 3 different soil depths (i.e. 15, 30 and 45 cm). The emitter outflows were considered as 2.4, 4 and 6 lit/hr. The fertigation treatments include treatments with nitrate concentrations of 125, 250 and 375 mg/liter. The effective variables include the initial nitrate of soil, nitrate concentrations in fertigation, initial soil moisture content, radial distance of points, applied water volume, saturated hydraulic conductivity, emitter discharge, emitter installation depth, soil bulk density, and the proportions of sand, silt and clay of soil. The mentioned variables were utilized for estimation of nitrate distribution pattern of surface and subsurface drip irrigation systems using ANN model. The comparisons results of simulated and measured values indicated that ANN models were capable methods for prediction of nitrate distribution. The values of correlation coefficient (R) were ranged 0.9-0.94 and 0.8-0.96 for surface and subsurface drip irrigation systems, respectively.