Using of pH as a tool to predict salinity of groundwater for irrigation purpose using artificial neural network

Monitoring of groundwater quality is one of the important tools to provide adequate information about water management. In the present study, artificial neural network (ANN) with a feed-forward back-propagation was designed to predict groundwater salinity, expressed by total dissolved solids (TDS), using pH as an input parameter. Groundwater samples were collected from a 36 m depth well located in the experimental farm of the City of Scientific Researches and Technological Applications (SRTA City), New Borg El-Arab City, Alexandria, Egypt. The network structure was 1–5–3–1 and used the default Levenberg–Marquardt algorithm for training. It was observed that, the best validation performance, based on the mean square error, was 14819 at epoch 0, and no major problems or over-fitting occurred with the training step. The simulated output tracked the measured data with a correlation coefficient (R-value) of 0.64, 0.67 and 0.90 for training, validation and test, respectively. In this case, the network response was acceptable, and simulation could be used for entering new inputs.