Quantitative models for predicting adsorption of oxytetracycline, ciprofloxacin and sulfamerazine to swine manures with contrasting properties

Understanding antibiotic adsorption in livestock manures is crucial to assess the fate and risk of antibiotics in the environment. In this study, three quantitative models developed with swine manure-water distribution coefficients (LgKd) for oxytetracycline (OTC), ciprofloxacin (CIP) and sulfamerazine (SM1) in swine manures. Physicochemical parameters (n = 12) of the swine manure were used as independent variables using partial least-squares (PLSs) analysis. The cumulative cross-validated regression coefficients (Q2cum) values, standard deviations (SDs) and external validation coefficient (Q2ext) ranged from 0.761 to 0.868, 0.027 to 0.064, and 0.743 to 0.827 for the three models; as such, internal and external predictability of the models were strong. The pH, soluble organic carbon (SOC) and nitrogen (SON), and Ca were important explanatory variables for the OTC-Model, pH, SOC, and SON for the CIP-model, and pH, total organic nitrogen (TON), and SOC for the SM1-model. The high VIPs (variable importance in the projections) of pH (1.178–1.396), SOC (0.968–1.034), and SON (0.822 and 0.865) established these physicochemical parameters as likely being dominant (associatively) in affecting transport of antibiotics in swine manures.