Prediction of Future Groundwater Elevation by Using Artificial Neural Network South of Riyadh, KSA

Saudi Arabia is considered an arid country with very limited water supplies.  About 80% of water supplies come from nonrenewable water resources, groundwater. Natural replenishment of the groundwater is far less than the annual extraction of groundwater. Consequently, groundwater elevations have sharply decreased through the last two decades. Neural Network has proved to be an effective tool for future prediction for a set of data. For a set of data of groundwater elevations for a reasonable period of time, Artificial Neural Network (ANN) can effectively predict future groundwater levels. The main objective of the current study is to predict the elevations of groundwater in four wells south of Riyadh for 20 years in future. In order to achieve that, nearly daily data of water level for a period of more than 30 years have been used for the prediction. A complete analysis of the 30 years data has been conducted. The results gave a warning to the governments to begin a general strategy in the area to reduce using excessive water for irrigation. The ANN was used to develop a prediction model to predict the water elevation in the well. About 80% of the data are used for training the network while the rest of the data are used for validating and testing the model. A sigmoid function is used at the hidden layer which consisted of 1-6-1. Results from both measurements and ANN were compared which gives a very good matching. The future prediction offered by the ANN says that the future reduction of the well water level in the coming 20 years will not exceed 30% of the reduction done during the last 30 years.