Prediction of the mobility and persistence of eight antibiotics based on soil characteristics

Antibiotics are widely used in intensive animal husbandry in the Netherlands and are subsequently emitted to soil via manure. To predict degradation and mobility in soil, generic sorption models have been derived. However, most of the coefficients used in generic models are based on a limited range of soils and have not been validated for agricultural soils in the Netherlands. To improve model predictions and assess to what extent differences among soils affect sorption and degradation, an experimental study has been performed. Using a recently developed experimental approach, both the degradation (DT50) and mobility (K_d) of eight selected commonly used antibiotics were determined in 29 typical Dutch agricultural soils. Median DT50 values range from 5.3 days for Sulfadiazine to 120 days for Trimethoprim but are affected by soil type. The ratio of the lowest and highest DT50 for a given antibiotic among soils can be as large as 151, for Tylosin. Measured values of the logK_d also range from 0.19 for Sulfadiazine to more than 2 for Doxycycline, Flumequine, Trimethoprim, Tylosin and Enrofloxacine. The impact of soil on K_d is large, especially for more mobile antibiotics such as Sulfadoxine and Sulfadiazine. Both the range in DT50 and K_d can be predicted reasonably well using a Freundlich type regression model that accounts for the variation in soil type and sampling depth. Organic matter, iron oxides, pH and clay content appear to be the main constituents and explain between 29 % (Trimethoprim) and 77 % of the variation in DT50 and between 64 % (Lincomycin) and 87 % (Sulfadoxine and Sulfadiazine) of the variation of K_d. The effect of depth on DT50 and K_d is however limited. The information thus obtained in combination with local data on soil type can be used to more accurately predict the potential risk of relevant antibiotics in soil and transport to ground- and nearby surface waters.