The use of modelling to predict levels of estrogens in a river catchment: how does modelled data compare with chemical analysis and in vitro yeast assay results?

Abstract

Effluent discharges at Rodbourne sewage treatment works (STWs) were assessed using chemical and in vitro biological analysis as well as modelling predictions. Results showed that Rodbourne STW discharged less estrone (E1) than expected, but similar 17beta-estradiol (E2) and 17alpha-ethinyl estradiol (EE2) to those predicted by a widely cited effluent prediction model. The Exposure Analysis Modelling System (EXAMS) model was set up using measured effluent concentrations as its starting point to predict estrogen concentrations along a 10 km length of the receiving water of the River Ray. The model adequately simulated estrogen concentrations along the river when compared to July 2007 measured data. The model predicted combined estrogen equivalents in reasonable agreement with estrogenicity as measured by passive sampler (POCIS) extracts using the yeast estrogen screen. Using gauged mean flow values for 2007 the model indicated that the most important determinand for estrogen exposure in the Ray was not season, but proximity to the Rodbourne effluent. Thus, fish in the first 3 km downstream of Rodbourne were typically exposed to two or even three times more estrogens than those living 7-10 km further downstream. The modelling indicated that, assuming the effluent estrogen concentrations measured in February 2008 were typical, throughout the year the whole length of the Ray downstream of Rodbourne would be estrogenic, i.e. exceeding the 1 ng/L E2 equivalent threshold for endocrine disruption.