Dynamics of biocide emissions from buildings in a suburban stormwater catchment - Concentrations, mass loads and emission processes

Biocides such as isothiazolinones, carbamates, triazines, phenylureas, azoles and others are used to protect the surfaces of buildings, e.g. painted or unpainted render or wood. These biocides can be mobilized from the materials if rainwater gets into contact with these buildings. Hence, these biocides will be found in rainwater runoff (stormwater) from buildings that is traditionally managed as "clean water" in stormwater sewer systems and often directly discharged into surface waters without further treatment. By means of a 9 month event-based high resolution sampling campaign the biocide emissions in a small suburban stormwater catchment were analysed and the emission dynamics throughout the single rain events were investigated. Five out of twelve of the rain events (peak events) proved significantly higher concentrations than the rest (average) for at least one compound. Highest median concentrations of 0.045 and 0.052μg L^-1 were found for terbutryn and carbendazim, while the concentrations for isoproturon, diuron, N-octylisothiazolinone, benzoisothiazolinone, cybutryn, propiconazole, tebuconazole, and mecoprop were one order of magnitude lower. However, during the peak events the concentrations reached up to 1.8 and 0.3μgL^-1 for terbutryn and carbendazim, respectively. Emissions of an averaged single family house into the stormwater sewer turned out to be 59 and 50μgevent^-1house^-1 terbutryn and carbendazim, respectively. Emissions for the other biocides ranged from 0.1 to 11μgevent^-1house^-1. Mass load analysis revealed that peak events contributed in single events as much to the emissions as 11 average events. However, the mass loads were highly dependent on the amounts of rainwater, i.e. the hydraulic flow in the receiving sewer pipe.The analysis of the emission dynamics showed first flush emissions only for single parameters in three events out of twelve. Generally biocides seemed to be introduced into the stormwater system rather continuously during the respective events than in the beginning of them. Mass flows during the events did correlate to driving rain, whereas mass loads neither correlated to the length or the intensity of rainfall nor the length of dry period.