Risk assessments for herbicides applied to roads are limited by the lack of knowledge on the fate and behaviour of the compounds in the urban environment. This study was designed to address this deficiency by quantifying the percentage loss of six herbicides following application to a roadside. Herbicides were applied on two occasions to a 16-m length of roadside and kerb edge. An automatic water sampler was used to collect run-off, draining to a single gulley pot, until 25 mm of rain had fallen. Samples were analysed for glyphosate, atrazine, diuron, oxadiazon and oryzalin, and peak concentrations were 650, 2210, 1810, 390 and 70 micrograms litre-1 respectively. Isoxaben was also applied, but concentrations in run-off were below the limit of detection (10 micrograms litre-1). Herbicide concentrations all followed a similar pattern of rapid decline throughout the first rain event following application, with the majority of loss occurring within the first 10 mm of accumulated rainfall, but compounds of high solubility and low Koc produced the highest peak concentrations. For those compounds of relatively low solubility and moderate Koc, application rate may be an influential factor in determining herbicide loss for these compounds. The percentage loss of the active substances applied differed between compounds, ranging from < 10% to 73%. The ecotoxicological significance of the results is discussed.

Railways have been identified as a potential source of herbicides detected in surface and groundwaters, but there are few data to support this theory. Two studies were undertaken to investigate the fate of herbicides applied to railway trackbeds: a pilot study in a section of a disused, but intact, cutting where runoff and throughflow were sampled from trenches adjacent to the treated area, and a larger scale study on 0.75 km of embankment where surface water from the drainage ditch at the base of the embankment and groundwater were sampled. In the pilot study, peak concentrations of atrazine, diuron and glyphosate (1280, 210 and 15 microg litre(-1) respectively) were detected 6days after treatment (DAT). Oxadiazon, oryzalin and isoxaben were not detected above their limits of quantification. Lower concentrations were detected 81 DAT (10 and 0.8 microg litre(-1) of atrazine and glyphosate respectively). In the larger scale study, herbicides were not detected, in either the surface water or groundwater, at concentrations above the limit of detection that could be attributed to application to the railway. Rainfall volume and depth to sampling point may partly explain the different results obtained from the two studies. The findings are compared with herbicide losses from other 'hard surfaces'.