Runoff of veterinary pharmaceuticals from arable and grassland-A comparison between predictions from model simulations and experimental studies

Generally, substantial variations between the predicted and measured concentrations of VPs in the runoff were uncovered. Although the FOCUS prediction suggested higher concentrations than were actually measured in 65% of the cases, the runoff concentrations of VPs were underestimated in the remaining 35%. This frequent underestimation of runoff concentrations was primarily observed for the grassland plots (85% of the underestimated situations), whereas the FOCUS predictions largely overestimated the measured concentrations for the arable plots. More strikingly, when involving a difference between the measured and predicted concentrations of 10% as a validity criterion for the model, only one (out of a total of 17) runoff concentration among the five VPs (i.e., sulfadimidine) assessed in the seven scenarios fell within this tolerance margin. Thus, these results demonstrate the substantial uncertainties related to the use of the FOCUS surface water modeling approach for the prediction of VPs introduced with manure and their environmental risk.