Machine learning approach to evaluate the groundwater quality and human health risk for sustainable drinking and irrigation purposes in South India

Abstract

The main purpose of this study was to evaluate the suitability of groundwater for sustainable drinking and irrigation purposes using various indices, such as the nitrate pollution index, agriculture suitability index (ASI), non-carcinogenic human risk assessment (NCHRA), and radial basic function (RBF) model. The novelty of the present study is to develop the ASI model and integrate with RBF model to identify the highly dominating parameter in chemical equilibrium of groundwater. Results showed that >85% of sample locations were suitable for drinking purposes, and the nitrate concentration in groundwater had a negative impact on the overall quality of water. Approximately 12 and 19 sample locations were contaminated owing to the high nitrate concentrations in the study region. The NCHRA study identified that approximately 8.5%, 27.28%, 29.54%, 40.40%, and 28.20% of area was excessively affected during the winter compared to summer season for people 6 to 12 y, 13 to 19 y, 20 to 29 y, 30 to 65 y, and >65 y of age. The RBF model shows that the R2 values for each season were 0.84 and 0.85 during summer and winter, respectively. The north-east and central parts of the study region were found to be more contaminated. The present study identified that, pathway of nitrate contaminant from the agriculture field towards to the sample locations. Overall, parent rock weathering, carbonate ion dissolution, and infiltration of rainwater and leachate from municipal waste dumping yards were the dominant factors influencing the chemical composition of groundwater. The present study achieved the vibrant knowledge about source of contamination, health effect on human body and impact on agriculture uses to develop the cleaner water supply system. The study results will be helpful in enhancing the sustainable action plan for water management in the study area.