Geochemical Analysis and Machine Learning Approach in Assessing the Suitability of Surface Water for Domestic and Agricultural Uses Around an Abandoned Limestone Quarry Region
摘要
Surface water around limestone quarry activities tested for physicochemical and toxic elements concentrations and its suitability for drinking, domestic and irrigation uses. Mean levels of pH, dissolved oxygen, lead and zinc exceeded permissible levels by WHO for potable water. Multivariate statistical analysis shows rock and silicate minerals weathering and anthropogenic activities are sources of chemical species in the water. Durov graphical plot shows that the water is Na + K, HCO3 + CO3 hydrochemical facie type. Computed water quality index ranged from 78.06 to 91.54 showing poor water quality status that is unsuitable for drinking, but useful for irrigation only. Geochemical speciation shows toxic metals occur dominantly as free ionic metal species at a percentage speciation range of 89–99%. Only Cr2+ exist predominantly as a complex of CrOH2+at a percentage speciation of 76–84%. The minerals calcite, dolomite, larnakite, vateirte, smithsonite, zincite occur mostly as under saturation at a saturation index less than 1 controlling chemical composition of water such as Pb, Zn, Ca, and Mg. The minerals Pb (OH)2, cerrusite and hydrocerrusite were oversaturated and precipitated from water. Findings using random forest, sector vector machine, and extreme gradient boosting machine learning algorithms are in agreement with water quality index calculations that the water is of poor quality. The study recommend water treatment using ion exchange, adsorption, or membrane filtration to remove Pb from water. The study is useful to water supply managers to improve on their services in delivering potable water in line with the sustainable development goals.