Probabilistically defining environmentally relevant concentrations in ecotoxicology.
A long-standing issue in ecotoxicology is the arbitrarily chosen and ambiguous definition of "environmentally relevant" concentrations, which undermines research comparability and hampers risk characterization. Here, we propose a probabilistic framework that anchors exposure levels to percentiles of environmental concentration distributions, defining low (<5th percentile), typical (5th to 95th percentiles), and high (>95th percentile) concentrations, as well as worst-case scenarios (e.g.,
A long-standing issue in ecotoxicology is the arbitrarily chosen and ambiguous definition of "environmentally relevant" concentrations, which undermines research comparability and hampers risk characterization. Here, we propose a probabilistic framework that anchors exposure levels to percentiles of environmental concentration distributions, defining low (<5th percentile), typical (5th to 95th percentiles), and high (>95th percentile) concentrations, as well as worst-case scenarios (e.g., the 99th percentile) for specific contexts. This framework transforms test concentration selection from a subjective assertion into a statistically justified practice, where each concentration corresponds to an explicit occurrence probability. Using global monitoring data, we demonstrate that environmental concentrations reliably follow cumulative probability distributions, validating the fundamental assumption. Crucially, the framework offers a statistical solution for designing proof-of-relevance and proof-of-concept studies. Collectively, this work provides an empirically grounded, immediately usable template for designing ecotoxicological experiments that bridge environmental monitoring, laboratory testing, and regulatory decision-making.
