Modelling of carbamazepine and diclofenac in a river network : photolytic degradation in Swiss rivers

Abstract: Micropollutants from urban or agricultural sources, for example pharmaceuticals or pesticides, are facing increasing public awareness. In the last years, many studies reported on the occurrence of micropollutions in different water samples, including drinking waters. Moreover, harmful effects could be detected on fish where micropollutants act as endocrine disruptors. Micropollution measurements of a national water-quality inventory in Switzerland (NADUF) indicate a relatively constant load pattern for Carbamazepine (650 g/week ±190 g/week) at chosen river sections throughout the year. This backs up the hypothesis that Carbamazepine is persistent. For Diclofenac the load pattern of the same river was fluctuating to a larger extent (640 g/week ± 400 g/week). Especially during dryer time periods in summer the loads were significant lower than the average. One explanation for the fluctuation in Diclofenac loads is a variation in the input loads to the rivers, another is degradation by photolysis in the river network. To prove this, one method was to analyse the measurements statistically and the other method was to set up a one-dimensional model of the fate of this pharmaceuticals in a pre-alpine river catchment (1750 km2) in Switzerland. The analysis of Diclofenac measurements on two distinctive sites showed the same drop in the mean loads during the summer. This indicates that fluctuations in the input loads occurred. In the model the pharmaceutical input is calculated on the base of yearly sales per capita, a constant yearly consumption and waste water treatment plants as point pollution sources. The model indicated that up to 28 % of the seasonal variation in Diclofenac loads could be explained by direct photolysis. In truth both hypotheses have to be combined to fully explain the drop. Additionally, when consulting hourly simulations we found that the environment is harmed from human pharmaceuticals to a large extent were critical environmental values are exceeded up by to 80 % of the time. In future the model should be extended to non-point source pollution from agriculture for instance.