Evaluation and classification of water quality of glacier‐fed channels using supervised learning and water quality index

In this study, a series of laboratory tests were performed on 25 water samples obtained from the high‐altitude glacier‐fed water resources of Gilgit‐Baltistan Pakistan. The variations in the measured physico‐chemical parameters were found as follows: conductivity 62–989 (µS/cm), turbidity 0.94–10.86 (NTU), pH 6–8.4, TDS 40–830 (mg/L), hardness 30–305 (mg/L), Ca 2+ 35–135 (mg/L), K + 2–12.2 (mg/L), Mg 2+ 0–17.9 (mg/L), Na + 2–27.6 (mg/L), Zn 2+ 0–3.5 (mg/L), Cu 2+ 0.27–3.51 (mg/L), Cl − 25–162 (mg/L), SO 4 2− 0–83 (mg/L), NO 3 − 0–0.94 (mg/L), and HCO 3 − 15–327 (mg/L). The ion charge balance error was found within the range of ±10% with a mean value of 5.36%. The Gibbs plot and ionic ratios showed that the rock‐weathering, especially carbonate dissolution, was the major hydrogeochemical process that influenced water chemistry. The water samples were classified as excellent (56%), good (40%), and medium (4%) quality based on the entropy water quality index (EWQI). The specificity and sensitivity of the discriminant model were found 100% and 90% respectively with a model‐hit‐ratio of 95.8% which satisfied this analysis.