ORIGINAL ABSTRACT

Relative contributions of agricultural and urban uses to the glyphosate contamination of surface waters were studied in a small catchment (25 km(2)) in Switzerland. Monitoring in four sub-catchments with differing land use allowed comparing load and input dynamics from different sources. Agricultural as well as urban use was surveyed in all sub-catchments allowing for a detailed interpretation of the monitoring results. Water samples from the river system and from the urban drainage system (combined sewer overflow, storm sewer and outflow of wastewater treatment plant) were investigated. The concentrations at peak discharge during storm events were elevated throughout the year with maximum concentrations of 4.15 μg L(-1). Glyphosate concentrations mostly exceeded those of other commonly used herbicides such as atrazine or mecoprop. Fast runoff from hard surfaces led to a fast increase of the glyphosate concentration shortly after the beginning of rainfall not coinciding with the concentration peak normally observed from agricultural fields. The comparison of the agricultural application and the seasonal concentration and load pattern in the main creek from March to November revealed that the occurrence of glyphosate cannot be explained by agricultural use only. Extrapolations from agricultural loss rates and from concentrations found in the urban drainage system showed that more than half of the load during selected rain events originates from urban areas. The inputs from the effluent of the wastewater treatment plant, the overflow of the combined sewer system and of the separate sewer system summed up to 60% of the total load.