RESUMO
Three different risk assessment procedures are described that aim to protect freshwater habitats from risks of the photosynthesis-inhibiting herbicides metribuzin and metamitron. These procedures are (1) the first-tier approach, based on standard toxicity tests and the application of an assessment factor, (2) the Species Sensitivity Distribution (SSD) approach, based on laboratory tests with a wider array of species and the application of a statistical model to calculate the HCx (the Hazardous Concentration for x% of the species), and (3) the model ecosystem approach, based on the evaluation of treatment-related effects in field enclosures. A comparison of the risk assessment procedures reveals that the first-tier approach is the most conservative for metamitron and metribuzin, and that HC5 values (and even HC10 values) based on acute EC50 values of algae and aquatic vascular plants may be used to derive maximum permissible concentrations for single applications. For both compounds these HC5 values were very similar to the ecological threshold concentrations in the enclosure studies. In contrast to model ecosystem experiments, however, HCx values based on lab toxicity tests do not provide information on the recovery potential of sensitive endpoints and on indirect effects, which may be important for regulatory decision-making. In the enclosure study, indirect effects of metribuzin on invertebrate populations were observed at an exposure concentration that was approximately 20 times lower than the corresponding HC5 value based on lab toxicity data for aquatic invertebrates.