Background: Broad spectrum of activity and low risk of groundwater contamination are important arguments for the application of the herbicide glyphosate on railway tracks. Nevertheless, certain weeds are insufficiently controlled or develop resistance, and there is also an ongoing controversial discussion about possible carcinogenicity of glyphosate. Alternatives are thus strongly desired. 2,4-D and quizalofop-P-ethyl (QE) are two selective herbicides with a complementary spectrum of activity. When used in agriculture, the compounds and their metabolites pose low risk to groundwater. Uses on railway tracks are, however, more critical since degradation likely is slower and mobility higher than in agricultural soils. In this study, we investigated degradation and sorption of the two active substances as well as five metabolites in three soils collected from railway tracks and in a crushed sand, used for construction works.Results: In these railway materials, the compounds were indeed degraded slower than in agricultural soils (mean half-lives differed by a factor of 2-26) and mobility was higher (mean sorption coefficients differed by a factor of 3-19). Half-lives and sorption coefficients were also estimated by extrapolation of data from agricultural soils, considering the organic carbon content of railway soils and agricultural soils. Estimated values were more conservative than measured values, i.e., for herbicides, for which no measured parameters are available for railway soils, estimated values may be a viable alternative for a first tier groundwater assessment.Conclusions: Based on our experimental data, possible leaching to groundwater is predicted to be highest for 2,4-D and quizalofop-acid, the primary metabolite of QE, moderate for 2,4-dichlorophenol, the primary metabolite of 2,4-D, but low for QE. Secondary and tertiary metabolites were formed in only low quantities. For an advanced groundwater risk assessment, leaching studies, e.g., with lysimeters, would be a reasonable next step.