Artificial Neural Network (ANN) Model for Ground Water Quality Simulation (Case study: Kashan aquifer)

Kashan aquifer is adjacent to Salt Lake. Because of this adjacency، the saline water of the lake has moved to the aquifer. In this study groundwater quality of the aquifer was simulated using Artificial Neural Network (ANN) model. For this purpose، the dominant ion of water was first determined by Piper diagram. Results showed that the sodium chloride is the dominant ion of water and so it was selected as the target variable to be simulated So the output variable of the ANN model was the concentration of sodium chloride in current year while the input variables were the water table of groundwater، yearly rainfall and the concentration of sodium chloride in previous year. Result showed that Multilayer Perceptron ANN model has better result in predict of chlorine concentration compared to Radial Basis ANN model. The sensitivity analysis showed that concentration of chloride in previous year and water table of groundwater are the most important variables in the ANN model respectively.