Phytotoxicity testing for herbicide regulation: shortcomings in relation to biodiversity and ecosystem services in agrarian systems.

The purpose of this paper is to present current knowledge on methods employed to perform phytotoxicity tests and risk assessments and to highlight shortcomings in relation to biodiversity and ecosystem services. Ecosystem services are benefits provided to humankind by a multitude of organisms present in natural ecosystems. Several studies were conducted between 2001 and 2010 aimed at investigating some of the deficiencies in phytotoxicity testing (new and existing data are presented). Herbicide toxicity responses were similar when comparing a suite of crop versus wild species. However, the validity of the evaluation was limited because of the narrow types of species tested. The number of species tested, currently set between six and ten, appears insufficient. The trait-based approach (i.e. the use of plant attributes to predict species sensitivity to toxicants) can be used to improve species selection. This approach puts more emphasis on the shared biological characteristics that affect plant function within ecological communities rather than on plant phylogeny. Results presented showed that further studies are needed. In test guidelines, protocols require that crop species be sprayed as young vegetative plants, which is assumed to be the most sensitive growth stage to herbicides. In contrast, during herbicide spray, herbicides may reach non-target plants that are at various phenological stages. Several studies demonstrated that plants may be at greater risk when contamination occurs at the reproductive stage. No data on long-term effects, plant recovery or on effects on reproductive stages are requested in current guidelines. Preliminary evidence suggests that this may be an important aspect to consider in risk assessment. In addition, herbicide impacts on plant community diversity as well as biodiversity at other trophic levels have been demonstrated in only a limited number of studies and therefore should warrant more attention in risk assessment. Crown