Artificial Neural Networks: Intelligent Approach to Simulate Groundwater Level Pattern

To depict hydrogeological variables and understand the physical processes taking place in a complex hydrogeological system, artificial neural networks (ANN) are widely used as a good alternative approach to tedious numerical models. This study devises the dynamic fluctuation of the piezometric level in Nebhana aquifers using ANN. A correlation analysis was first carried out. It revealed that piezometric levels were influenced with monthly rainfall, evapotranspiration, and initial water table level. These informative variables were used as inputs to train the ANN demonstrating that they were convenient. In fact, the maximal error reached was about 19%. It was observed only one time in Ouled Slimen piezometer. To test the generalization capacity of the developed ANN models, monthly piezometric levels were forecasted in the medium term: September 2016-September 2018. The obtained results were satisfactory for all piezometers.