Management decisions concerning the spraying of herbicides on highway roadsides are evaluated on the basis of their impact on resulting environmental risk. A mathematical transport model was previously applied to the State of California with a Monte Carlo technique, and in this study the results are manipulated to evaluate the risk reduction that results from restricting herbicide application on the basis of site characteristics or changing other application practices. Results show that eliminating herbicide applications where the slope of the grass adjacent to the highway is greater than 30° has little or no effect on risk. Eliminating application where the width of the grass adjacent to the highway is less than 2 m or where soil organic carbon content is less than 0.5% can lead to significant reductions in environmental risk for certain herbicides. Additionally, limiting the width of the spray zone and applying the minimum manufacturer-suggested application rate reduce the risk to aquatic ecosystems. Applying at the minimum rate has the greatest potential to decrease risk. Results of this study show that management decisions can have a significant effect on limiting herbicide runoff risks to aquatic ecosystems. Decision makers would have to weigh costs of alternatives to herbicide spraying for controlling roadside vegetation against the environmental risk reductions.