Environmental fate of ciprofloxacin in agricultural soils under realistic irrigation water matrices: kinetic, column transport, thermodynamic, mechanistic and machine learning insights.

The combined influence of soil properties, irrigation water chemistry, and transport processes on antibiotic fate remains poorly understood. Here, ciprofloxacin (CIP) retention, transport, and remobilization were investigated in three agricultural soil types under four different irrigation water matrices using batch, column, thermodynamic, mechanistic, and machine-learning approaches. At elevated concentrations (10 mg/L), all soils exhibited comparable retention capacities, whereas substantial differences emerged at extremely elevated concentrations (150 mg/L), with maximum adsorption capacities following the order black soil (≈17.6 mg/g) > alluvial soil (≈15.6 mg/g) > red soil (≈10.4 mg/g). Transport behaviour differed markedly among soils, with retardation factors following the order black soil (7.9) > alluvial soil (2.24) > red soil (1.4). After 180 pore volumes, relative breakthrough remained below 0.03 in black soil but reached 0.25 and 0.45 in alluvial and red soils, respectively. Rainfall flushing further revealed contrasting environmental outcomes, with black soil strongly immobilizing CIP, red soil promoting rapid transport, and alluvial soil acting as a persistent contaminant leaching reservoir. Thermodynamic analyses indicated predominantly endothermic adsorption with enhanced spontaneity at elevated temperatures. FTIR and zeta potential analyses identified surface complexation, hydrogen bonding, electrostatic interactions, and cation bridging as the primary retention mechanisms. Random Forest modelling identified soil type as the dominant control on CIP retention, accounting for ~92% of total feature importance (R2 = 0.96-0.97). These findings demonstrate that soil properties govern CIP fate and determine contrasting pathways of sequestration, persistence, and groundwater contamination risk.