Hydrochemical Characterisation, Indexical, and Exploratory Multilinear Regression Appraisal of Groundwater Quality for Irrigation Purposes in South-Eastern Ghana
摘要
With the threats posed by anthropogenic activities and climate change on surface water resources, groundwater is increasingly relied upon for domestic and irrigation purposes, especially in arid and semi-arid regions. Multivariate statistical analysis, groundwater irrigation proxies, and exploratory multilinear regression modelling (MLM) were applied to 47 groundwater samples to characterise groundwater chemistry, assess its suitability for irrigation, and predict the parameters influencing its quality for irrigation. The study revealed the dominant cation and anion trends: Na⁺ > Ca²⁺ > Mg²⁺ > K⁺ and HCO₃⁻ > Cl⁻ > SO₄²⁻ > NO₃⁻ > F⁻, respectively. The results suggest groundwater from the study area is mainly controlled by ion exchange processes and carbonate mineral weathering, dominated by Ca-HCO3 water types in high-elevation areas (recharge zones), which evolve into Na-Cl and Mg-Cl water types as the water becomes more mineralised in the flow regime. The computed irrigation water quality indexes (IWQIs) classified 27% of samples, mainly in the southern and central zones, as having little to no restriction and optimal for most crops. In comparison, 56% of the samples showed moderate to high restrictions, requiring salinity-tolerant crops and careful irrigation management practices. The remaining 17% were categorised as severely restricted for irrigation and unsuitable for purposes of irrigation. The multilinear regression model (MLM) prediction of the IWQI yielded poor performance (R2 = 0.353; SEE = 16.522), likely due to dataset limitations and missing covariates such as temperature. Spatial analysis revealed poorer water quality in the western and northern portions of the study area, emphasising the need for tailored irrigation strategies including leaching, soil amendments, and selective cropping to mitigate salinity and sodicity risks in these areas.