pH-dependent zinc and cadmium dynamics: farmyard manure as a double-edged sword in alluvial soils

Heavy metal contamination of agricultural soils threatens food safety across South and Southeast Asia, yet the mechanistic basis for organic amendment efficacy in metal immobilization remains contested. We coupled vertical-column leaching with batch sorption-release kinetics to disentangle the roles of soil pH, clay content, and farmyard manure (FYM) in governing zinc (Zn) and cadmium (Cd) behaviour in two contrasting alluvial soils of West Bengal, India: an acidic sandy-loam from Pundibari (pH 5.48) and a near-neutral clay-loam from Malda (pH 7.51). The specific gap we address is whether FYM efficacy is uniform or instead conditional on soil reaction and texture, and whether the coupled kinetics can be reduced to an operational, property-based decision rule; these questions have not previously been resolved within a single soil-paired design. Breakthrough curves shifted markedly leftward upon FYM addition, with breakthrough at 0.38 pore volumes (Pv) for Zn and 0.35 Pv for Cd in the FYM-amended Malda soil versus 0.95 Pv in the untreated Pundibari soil. Modified Freundlich modelling showed sorption capacity (kaC0) and energy (n) rising with clay content and temperature, with kaC0 for Zn at 35 °C reaching 14.1 in FYM-amended Malda versus 1.38 in untreated Pundibari, consistent with endothermic chemisorption (R² = 0.90–0.98). The Modified Elovich model revealed a striking pH-dependent paradox: FYM accelerated Zn release in the acidic soil (α rose from 4.88 to 16.3; β fell from 0.542 to 0.172) while suppressing release in the near-neutral soil (α fell from 4.77 to 3.73; β rose from 0.46 to 0.60); Cd showed the same pattern (R² = 0.88–0.99). Artificial neural networks, used as a complementary non-linear feature-attribution tool because the controlling properties co-vary and act non-additively, ranked clay and pH as dominant predictors, explaining up to 97.6% of the variance in kinetic parameters. We infer that organic-amendment efficacy may be associated with the protonation state of the soil exchange complex; because the dissolved organic carbon content of the FYM was not measured, this ligand-based interpretation is offered as a hypothesis consistent with the kinetic and statistical evidence rather than as a directly verified mechanism: in acidic, low-clay matrices, protonated humic groups contribute to the labile metal pool, whereas in near-neutral, clay-rich matrices, deprotonated organic surfaces synergize with permanent charge to form stable organo-mineral complexes. These results caution against one-size-fits-all amendment recommendations and support property-targeted, pH- and texture-targeted remediation in contaminated rice-growing landscapes, where amendment decisions must minimize metal transfer to water, crops, and ultimately human diets.