Enhancing Sustainable Agriculture and Food Security in Developing Economies: Assessing and Predicting Irrigation Water Quality Indices of Urban Groundwater Systems Using an Integrated Approach

Abstract Rapid urbanization intensifies pressure on groundwater resources used for peri-urban agriculture. This study evaluates and models groundwater suitability for irrigation in Awka and Nnewi, Nigeria. Seven irrigation water quality indices were calculated: sodium adsorption ratio (SAR), soluble sodium percentage (SSP), permeability index (PI), Kelly’s ratio (KR), magnesium hazard (MH), residual sodium carbonate, and potential salinity (PS). Analysis of sample distribution revealed distinct, aquifer-specific risk patterns. Nnewi exhibited sodium-dominated hazards (40% of samples in the “doubtful” class for KR and 50% in the “permissible” range for SSP), linked to sandy geology. Awka showed magnesium influence, with 50% of samples exceeding the MH threshold and 30% rated “unsuitable” by PI, associated with mudrock aquifers. The integration of diagnostic analysis with predictive modeling using multiple linear regression and artificial neural networks was implemented. City-specific models achieved high accuracy [ R 2 : 0.806–1.000; low standard error of estimate (SEE) and sum of squared errors (SOSE)]. Combined regional models showed variable, reduced performance ( R 2 : 0.558–1.000; higher SEE and SOSE), emphasizing the necessity for localized models. Sensitivity analysis identified sodium as controlling SAR, SSP, and KR; calcium and magnesium governing MH and PI; and chloride governing PS. This approach offers a framework for sustainable groundwater management in urbanizing regions.