Natural variations in flow are critical in determining concentrations of point source contaminants in rivers: an...

Daily steroid estrogen concentrations as 17β-estradiol equivalents (E2 equiv.) were modeled from 1992 to 2008 for single locations on the well populated Thames and Soar rivers in England. The historic daily mean flow values which were the basis of this exercise came from a selected gauging site on each river. The natural variation in flow from winter to summer typically produced a 20- to 30-fold difference in predicted estrogen concentration over the course of a year. Based on all the predicted values from minimum to maximum over the 1992 to 2008 period there was a 98-fold difference in estrogen concentrations on the basis of flow alone for the Thames (0.1-12.7 ng/L E2 equiv.) and 67-fold for the Soar (0.2-13.3 ng/L E2 equiv.). This compares to a predicted 0.5-fold difference that could arise from differences in sewage treatment and 0.1-fold difference due to differences in in-stream biodegradation. The seasonal variation in flow generated a repeating "roller coaster" in predicted estrogen concentrations. Regularly measured phosphate data for the river Avon over the period 1993 to 1996, where point sources also dominate, was compared against flow and predicted estrogen concentrations. The pattern of predicted estrogen and measured total phosphate concentration were very closely related. This dramatic variation in contaminant concentration over the year due to flow poses questions over what we mean by environmental relevance and the representation of the real environment in aquatic ecotoxicity tests.