ORIGINAL ABSTRACT

Glyphosate is a widely used broad-spectrum postemergent herbicide used for weed control in both agricultural and nonagricultural settings. Spray drift of glyphosate can pose a risk to nontarget terrestrial plants and plant communities outside the intended area of application, but the lack of a well-established predicted-no-effect drift rate makes properly assessing such risk difficult. For this reason, a literature review and meta-analysis was carried out with the aim to determine the level of drift that is likely to cause harm to plants and to explore what spray-reducing targets would be sufficiently protective. No-observed-adverse effect rates, lowest-observed-adverse effect rates, and effect rates giving 10, 25, and 50% effects were extracted from a total of 39 different publications. The data were combined per species, and species sensitivity distributions were constructed and fitted with a log-logistic model to assess protectiveness. No systematic differences were detected between the responses of monocotyledons or dicotyledons, but wild plants were found to be generally less sensitive to glyphosate drift than domesticated plants. The results indicate that restricting spray drift to a level below 5 g a.e./ha would protect approximately 95% of all higher plant species against minor adverse effects of glyphosate drift and that rates below 1 to 2 g a.e./ha would be almost completely protective. No studies were encountered that evaluated effects of spray drift against nonvascular plants, and therefore, the conclusions are only valid for vascular plants. Environ Toxicol Chem 2017;36:2879-2886. © 2017 SETAC.