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 (mu S/cm), turbidity 0.94-10.86 (NTU), pH 6-8.4, TDS 40-830 (mg/L), hardness 30-305 (mg/L), Ca2+ 35-135 (mg/L), K+ 2-12.2 (mg/L), Mg2+ 0-17.9 (mg/L), Na+ 2-27.6 (mg/L), Zn2+ 0-3.5 (mg/L), Cu2+ 0.27-3.51 (mg/L), Cl- 25-162 (mg/L), SO42- 0-83 (mg/L), NO3- 0-0.94 (mg/L), and HCO3- 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.