Concentration ratios in fish, their natural variation, and extrapolation of missing values

In environmental impact assessments the biota-to-water concentration ratio (CR) is needed for calculating the transfer of radionuclides from water to aquatic organisms. Data are frequently missing, especially for rare elements, and various approaches are used for extrapolation of missing values. In this study we have analysed elemental concentrations in fish based on previously published data for brackish and freshwater. We grouped the fish according to their species and used a mixed linear model to calculate site- and ecosystem-specific CR values. These values were compared with values from different extrapolation methods (site, literature, element and ecosystem analogues). We also calculated edible versus whole body conversion factors. The site-specific CR values frequently differed by an order of magnitude or more between the brackish and freshwater ecosystems. For most elements the value was affected by site, locality (within site) and fish type. Also the inclusion of bones affected some CR values. Literature analogues diverged more from the site data than site analogues, but these both performed better than element and ecosystem analogues. The ratio between edible and whole-body concentrations was typically similar in brackish and freshwater environments. We conclude that a joint analysis of different aquatic ecosystems at similar sites allows for robust estimates of systematic and random variation associated with CR. We suggest that there are more possible analogues than the ones considered in the international recommendations, and that the choice of analogue should be based on the available data and the performance of the analogues in the specific assessment at hand.