Predicting Concentrations of Selected Ions and Total Hardness in Groundwater Using Artificial Neural Networks and Multiple Linear Regression Models

Abstract

Assessing the quality of groundwater in a given aquifer can be an expensive and time-consuming process. An effort is made in this thesis to predict several water quality parameters, namely Fe, Cl, SO4, and total hardness (as CaCO3), from the easily measured properties total dissolved solids (TDS) and electrical conductivity (EC). This is achieved by establishing functional relationships between the four quality parameters, TDS, and EC using multiple linear regression (MLR) and artificial neural network (ANN) models. Data for these variables were gathered from five unrelated groundwater quality studies. Results indicate that the ANN models produced more accurate functions than MLR, showcasing the strength of ANN’s in predictive applications. Analysis of the relative importance of each parameter illustrates that total hardness (CaCO3) is most influential in determining TDS, while sulphate is most influential on EC. These results could have a significant impact on the future of groundwater quality assessments.