EVALUATING POTABILITY OF GROUNDWATER USING AI-DRIVEN APPROACHES

Abstract

Integrating artificial intelligence (AI) techniques into groundwater quality assessments is a pivotal step toward achieving water resource sustainability and ensuring its safety for human consumption. This is achieved by leveraging the advanced analytical capabilities of these technologies. Parameters such as pH, total dissolved solids (TDS), total alkalinity (T.A), chloride, sulfate (SO4), phosphorus oxide (PO4), Temperature (T) , electrical conductivity(EC), turbidity (Tur), bicarbonate (HCO3), total hardness (T.H), calcium hardness (Ca.H), magnesium hardness (Mg.H), sodium (Na), potassium (K) , and dissolved oxygen were analyzed to assess groundwater quality in Duhok City, Iraq. Water Quality Index (WQI) and Adaptive Neuro-Fuzzy Inference System for Drinking (ANFIS-D) were employed in classifying water quality as excellent, good, poor, very poor, and unsuitable. It is beneficial to infer water quality for the individuals and decision-makers in the region. The WQI and ANFIS-D of the research area are between (37-50) (33-49), respectively. The overall WQI of the research area finds that the groundwater is safe and potable. The statistical metrics such as root mean square error (RMSE), mean bias error (MBE), and correlation coefficient (R) are used to check the validity of the ANFIS-D model. Studies show that the R value, MBE, and RMSE of the ANFIS-D model are (0.851, 2.9 and 3.81) respectively. Based on the results of these Statistical parameters, the ANFIS-D estimation model can predict the groundwater quality index of Duhok City with reasonable accuracy, which is useful and valuable for estimating the groundwater quality index.