In safety assessments of underground radioactive waste repositories, understanding radionuclide fate in ecosystems is necessary to determine the impacts of potential releases. Here, the reliability of two mechanistic models (the compartmental K-model and the 3D dynamic D-model) in describing the fate of radionuclides released into a Baltic Sea bay is tested. Both are based on ecosystem models that simulate the cycling of organic matter (carbon). Radionuclide transfer is linked to adsorption and flows of carbon in food chains. Accumulation of Th-230, Cs-135, and Ni-59 in biological compartments was comparable between the models and site measurements despite differences in temporal resolution, biological state variables, and partition coefficients. Both models provided confidence limits for their modeled concentration ratios, an improvement over models that only estimate means. The D-model enables estimates at high spatio-temporal resolution. The K-model, being coarser but faster, allows estimates centuries ahead. Future developments could integrate the two models to take advantage of their respective strengths.Electronic supplementary materialThe online version of this article (doi:10.1007/s13280-013-0398-2) contains supplementary material, which is available to authorized users.

As a result of nuclear accidents and weapons tests, the radionuclides Cs-137 and Sr-90 are common contaminants in aquatic ecosystems. Concentration ratios (CR) based on concentrations of stable Cs and Sr in biota and media are used for the estimation of transfer of their radioisotopes for radiation dose calculations in environmental and human safety assessments. Available element-specific CRs vary by over an order of magnitude for similar organisms, thus affecting the dose estimates proportionally. The variation could be reduced if they were based on a better understanding of the influence of the underlying data and how that affects accumulation and potential biomagnification of stable Cs and Sr in aquatic organisms. For fish, relationships have been identified between water concentrations of K and CR of Cs-137, and between water concentrations of Ca and CR of Sr-90. This has not been confirmed for stable Cs and Sr in European waters. In this study, we analysed an existing dataset for stable Cs and Sr, as well as K and Ca, in four Swedish lakes and three Baltic Sea coastal areas, in order to understand the behaviour of these elements and their radioisotopes in these ecosystems. We found significant seasonal variations in the water concentrations of Cs, Sr, K and Ca, and in electrical conductivity (EC), especially in the lakes. CR values based on measurements taken at single or few time points may, therefore, be inaccurate or introduce unnecessarily large variation into risk assessments. Instead, we recommend incorporating information about the underlying variation in water concentrations into the CR calculations, for example by using the variation of the mean. The inverse relationships between fish CR(Cs)-[K]water and fish CR(Sr)-[Ca]water, confirmed that stable Cs and Sr follow the same trends as their radioisotopes. Thus, they can be used as proxies when radioisotope data are lacking. EC was also strongly correlated with K and Ca concentrations in the water and could potentially be used as a quick and cost-effective method to estimate water chemistry to obtain less variable CR. We also recommend some simple improvements to data collection that would greatly enhance our ability to understand Cs and Sr uptake by fish.

No abstract

In environmental risk assessments of nuclear waste, there is need to estimate the potential risks of a large number of radionuclides over a long time period during which the environment is likely to change. Usually concentration ratios (CRs) are used to calculate the activity concentrations in organisms. However, CRs are not available for all radionuclides and they are not easily scalable to the varying environment. Here, an ecosystem transport model of elements, which estimates concentrations in organisms using carbon flows and food transfer instead of CR is presented. It is a stochastic compartment model developed for Lake Eckarfjärden at Forsmark in Sweden. The model was based on available data on carbon circulation, physical and biological processes from the site and identifies 11 functional groups of organisms. The ecosystem model was used to estimate the environmental transfer of 13 elements (Al, Ca, Cd, Cl, Cs, I, Ni, Nb, Pb, Se, Sr, Th, U) to various aquatic organisms, using element-specific distribution coefficients for suspended particles (K) and upper sediment (K), and subsequent transfer in the foodweb. The modelled CRs for different organism groups were compared with measured CRs from the lake and literature data, and showed good agreement for many elements and organisms, particularly for lower trophic levels. The model is, therefore, proposed as an alternative to measured CR, though it is suggested to further explore active uptake, assimilation and elimination processes to get better correspondence for some of the elements. The benthic organisms (i.e. bacteria, microphytobenthos and macroalgae) were identified as more important than pelagic organisms for transfer of elements to top predators. The element transfer model revealed that most of the radionuclides were channelled through the microbial loop, despite the fact that macroalgae dominated the carbon fluxes in this lake. Thus, element-specific adsorption of elements to the surface of aquatic species, that may be food sources for organisms at higher trophic levels, needs to be considered in combination with generic processes described by carbon fluxes.