A new application of K‑means clustering analysis in hazard assessment of nitrate pollution in detrital aquifers

Abstract

Groundwater protection against anthropogenic pollution is an unavoidable task to ensure the sustainability of this important water resource. Evaluating pollution hazards through mapping is essential for aquifer protection. This research aims to delineate the hazard map of an anthropogenic contaminant (nitrate) in a detrital aquifer using K-means cluster analysis to eliminate some of the subjectivity of traditional methods have. Nitrate demonstrated to be a suitable indicator of anthropogenic pollution from nitrogen fertilisers used in crops, with a background threshold estimated at 14.9 mg/L using a probabilistic approach. Three levels of hazard were obtained. The highest hazard occurs to the south of study area, coinciding with the area with the highest recharge from irrigation returns. In these zones is not recommended groundwater for human consumption because the health risk. This demonstrates the strong influence of recharge on aquifer pollution risk. The results indicated that the K-means cluster map obtained a significant statistical correlation of 53% with nitrate concentrations in groundwater obtained from 23 sampling wells, which reliably validates the effectiveness of using this technique. This research shows the advantage of applying a novel data mining method to groundwater hazard assessment, providing a valuable tool for the adequate management and planning of groundwater resources.