Quantifying sources of variability in fish bioaccumulation factor estimates for perfluoro-n-octane sulfonic acid: study design effects and implications for water quality criteria.

Numeric water quality criteria development approaches for the protection of aquatic life and human health incorporate bioaccumulation factors (BAFs) as key exposure pathways. These are typically derived from field-collected tissue and ambient water concentration data, and are often highly variable within and across species. Existing guidance for conducting BAF field studies is necessarily general to accommodate research objectives, ecosystem characteristics, and the chemicals and species of interest. However, the direction on evaluating study quality and understanding of the relative influence of study design components on BAFs is limited. We used a large, publicly available dataset of perfluoro-n-octane sulfonic acid (PFOS) concentrations in water and fish tissue to (1) evaluate overall patterns and drivers of water and fish tissue PFOS concentrations and derived BAFs, and (2) quantify how biological, environmental, and study design factors regarding water and tissue samples affect derived BAFs. PFOS tissue concentrations and BAFs differed significantly with species and, although variable, there was a significant positive relationship between PFOS water and fish tissue concentrations. Using biological, environmental, and study design variables, boosted regression tree analyses showed that spatial proximity between water and fish tissue sampling locations was the primary driver of BAF variability owing to large differences in water concentrations across hydraulically connected sites. Other variables, including temporal proximity between samples and biological and environmental factors, were less influential relative to spatial proximity. Our findings demonstrate that BAF study design decisions can have significant implications for key water quality criteria derivation components and highlight the need for sampling regimes that accurately characterize exposure to improve the quality of BAFs used in risk assessment and the development of regulatory standards.