Ensuring robust radiological risk assessment for wildlife: insights from the International Atomic Energy Agency EMRAS and MODARIA programmes.

In response to changing international recommendations and national requirements, a number of assessment approaches, and associated tools and models, have been developed over the last circa 20 years to assess radiological risk to wildlife. In this paper, we summarise international intercomparison exercises and scenario applications of available radiological assessment models for wildlife to aid future model users and those such as regulators who interpret assessments. Through our studies, we have assessed the fitness for purpose of various models and tools, identified the major sources of uncertainty and made recommendations on how the models and tools can best be applied to suit the purposes of an assessment. We conclude that the commonly used tiered or graded assessment tools are generally fit for purpose for conducting screening-level assessments of radiological impacts to wildlife. Radiological protection of the environment (or wildlife) is still a relatively new development within the overall system of radiation protection and environmental assessment approaches are continuing to develop. Given that some new/developing approaches differ considerably from the more established models/tools and there is an increasing international interest in developing approaches that support the effective regulation of multiple stressors (including radiation), we recommend the continuation of coordinated international programmes for model development, intercomparison and scenario testing.

Protection of the environment.

The International Commission on Radiological Protection's (ICRP) system to protect the living components of the environment is designed to provide a broad and practical framework across different exposure situations. The framework recognises the need to be able to demonstrate an adequate level of protection in relation to planned exposure situations, whilst also providing an ability to manage existing and emergency situations in an appropriate way. In all three exposure situations, the release of radionuclides into the natural environment leads to exposures of non-human biota (wildlife), as well as having the potential for exposures of the public. How the key principles of the ICRP system of radiological protection apply in each of these exposure situations will be discussed. Using examples, we will demonstrate how the overall approach provides a mechanism for industry to assess and demonstrate compliance with the environmental protection objectives of relevant (national) legislation, and to meet stakeholder expectations that radiological protection of the environment is taken into consideration in accordance with international best practice. However, several challenges remain, and these will be discussed in the context of the need for additional guidance on the protection of the environment.

Plutonium and other radionuclides persist across marine-to-terrestrial ecotopes in the Montebello Islands sixty years after nuclear tests.

Since the 1956 completion of nuclear testing at the Montebello Islands, Western Australia, this remote uninhabited island group has been relatively undisturbed (no major remediations) and currently functions as high-value marine and terrestrial habitat within the Montebello/Barrow Islands Marine Conservation Reserves. The former weapons testing sites, therefore, provide a unique opportunity for assessing the fate and behaviour of Anthropocene radionuclides subjected to natural processes across a range of shallow-marine to island-terrestrial ecological units (ecotopes). We collected soil, sediment and biota samples and analysed their radionuclide content using gamma and alpha spectrometry, photostimulated luminescence autoradiography and accelerator mass spectrometry. We found the activity levels of the fission and neutron-activation products have decreased by ~hundred-fold near the ground zero locations. However, Pu concentrations remain elevated, some of which are high relative to most other Australian and international sites (up to 25,050 Bq kg-1 of 239+240+241Pu). Across ecotopes, Pu ranked from highest to lowest in the following order: island soils > dunes > foredunes > marine sediments > and beach intertidal zone. Low values of Pu and other radionuclides were detected in all local wildlife tested including endangered species. Activity concentrations ranked (highest to lowest) terrestrial arthropods > terrestrial mammal and reptile bones > algae > oyster flesh > whole crab > sea turtle bone > stingray and teleost fish livers > sea cucumber flesh > sea turtle skin > teleost fish muscle. The three detonations (one from within a ship and two from 30 m towers) resulted in differing contaminant forms, with the ship detonation producing the highest activity concentrations and finer more inhalable particulate forms. The three sites are distinct in their 240/239Pu and 241/239Pu atom ratios, including the Pu transported by natural process or within migratory living organisms.

Measurement of fallout radionuclides, (239)(,240)Pu and (137)Cs, in soil and creek sediment: Sydney Basin, Australia.

Soil and sediment samples from the Sydney basin were measured to ascertain fallout radionuclide activity concentrations and atom ratios. Caesium-137 ((137)Cs) was measured using gamma spectroscopy, and plutonium isotopes ((239)Pu and (240)Pu) were quantified using accelerator mass spectrometry (AMS). Fallout radionuclide activity concentrations were variable ranging from 0.6 to 26.1 Bq/kg for (137)Cs and 0.02-0.52 Bq/kg for (239+240)Pu. Radionuclides in creek sediment samples were an order of magnitude lower than in soils. (137)Cs and (239+240)Pu activity concentration in soils were well correlated (r(2) = 0.80) although some deviation was observed in samples collected at higher elevations. Soil ratios of (137)Cs/(239+240)Pu (decay corrected to 1/1/2014) ranged from 11.5 to 52.1 (average = 37.0 ± 12.4) and showed more variability than previous studies. (240)Pu/(239)Pu atom ratios ranged from 0.117 to 0.165 with an average of 0.146 (±0.013) and an error weighted mean of 0.138 (±0.001). These ratios are lower than a previously reported ratio for Sydney, and lower than the global average. However, these ratios are similar to those reported for other sites within Australia that are located away from former weapons testing sites and indicate that atom ratio measurements from other parts of the world are unlikely to be applicable to the Australian context.

Voxel modeling of rabbits for use in radiological dose rate calculations.

Radiation dose to biota is generally calculated using Monte Carlo simulations of whole body ellipsoids with homogeneously distributed radioactivity throughout. More complex anatomical phantoms, termed voxel phantoms, have been developed to test the validity of these simplistic geometric models. In most voxel models created to date, human tissue composition and density values have been used in lieu of biologically accurate values for non-human biota. This has raised questions regarding variable tissue composition and density effects on the fraction of radioactive emission energy absorbed within tissues (e.g. the absorbed fraction - AF), along with implications for age-dependent dose rates as organisms mature. The results of this study on rabbits indicates that the variation in composition between two mammalian tissue types (e.g. human vs rabbit bones) made little difference in self-AF (SAF) values (within 5% over most energy ranges). However, variable tissue density (e.g. bone vs liver) can significantly impact SAF values. An examination of differences across life-stages revealed increasing SAF with testis and ovary size of over an order of magnitude for photons and several factors for electrons, indicating the potential for increasing dose rates to these sensitive organs as animals mature. AFs for electron energies of 0.1, 0.2, 0.4, 0.5, 0.7, 1.0, 1.5, 2.0, and 4.0 MeV and photon energies of 0.01, 0.015, 0.02, 0.03, 0.05, 0.1, 0.2, 0.5, 1.0, 1.5, 2.0, and 4.0 MeV are provided for eleven rabbit tissues. The data presented in this study can be used to calculate accurate organ dose rates for rabbits and other small rodents; to aide in extending dose results among different mammal species; and to validate the use of ellipsoidal models for regulatory purposes.

Accumulation of plutonium in mammalian wildlife tissues following dispersal by accidental-release tests.

We examined the distribution of plutonium (Pu) in the tissues of mammalian wildlife inhabiting the relatively undisturbed, semi-arid former Taranaki weapons test site, Maralinga, Australia. The accumulation of absorbed Pu was highest in the skeleton (83% ± 6%), followed by muscle (10% ± 9%), liver (6% ± 6%), kidneys (0.6% ± 0.4%), and blood (0.2%). Pu activity concentrations in lung tissues were elevated relative to the body average. Foetal transfer was higher in the wildlife data than in previous laboratory studies. The amount of Pu in the gastrointestinal tract was highly elevated relative to that absorbed within the body, potentially increasing transfer of Pu to wildlife and human consumers that may ingest gastrointestinal tract organs. The Pu distribution in the Maralinga mammalian wildlife generally aligns with previous studies related to environmental exposure (e.g. Pu in humans from worldwide fallout), but contrasts with the partitioning models that have traditionally been used for human worker-protection purposes (approximately equal deposition in bone and liver) which appear to under-predict the skeletal accumulation in environmental exposure conditions.

Dietary ingestion of fine sediments and microalgae represent the dominant route of exposure and metal accumulation for Sydney rock oyster (Saccostrea glomerata): A biokinetic model for zinc.

Past studies disagree on the extent to which dissolved or dietary uptake contribute to metal bioaccumulation in the filter-feeding Sydney rock oyster (Saccostrea glomerata) in urbanized estuaries. Although most data support the assumption that fine sediments are a major route of metal uptake in these bivalves, some studies based in the Sydney estuary, Australia, have indicated a poor correlation. In the present study, seawater, sediment and microalgae were radiolabelled with (65)Zn tracer and exposed to S. glomerata to assess the influence of dissolved and dietary sources to Zn bioaccumulation. Oysters in the dissolved-phase uptake experiment (5, 25 and 50 µg L(-1) (65)Zn for 4 d followed by 21 days of depuration) readily accumulated (65)Zn for all three concentrations with an uptake rate constant of 0.160±0.006 L dry weight g(-1) d(-1). Oysters in the dietary assimilation experiment (1h pulse-feed of either (65)Zn-radiolabelled suspended fine-fraction (<63 µm) sediment or the microalgae Tetraselmis sp.) accumulated (65)Zn, with assimilation efficiencies of 59 and 67% for fine sediment and microalgae, respectively. The efflux rates were low for the three experiments (0.1-0.5% d(-1)). A bioaccumulation kinetic model predicts that uptake of Zn will occur predominantly through the dietary ingestion of contaminated fine sediment particles and microalgae within the water column, with considerably greater metal bioaccumulation predicted if oysters ingested microalgae preferentially to sediments. However, the model predicts that for dissolved Zn concentrations greater than 40 µg L(-1), as observed during precipitation events, the uptake of the dissolved phase may contribute ≥50% to accumulation. Overall, the results of the present study suggest that all three sources may be important exposure routes to S. glomerata under different environmental conditions, but contributions from dietary exposure will often dominate.

Predicting exposure of wildlife in radionuclide contaminated wetland ecosystems.

Many wetlands support high biodiversity and are protected sites, but some are contaminated with radionuclides from routine or accidental releases from nuclear facilities. This radiation exposure needs to be assessed to demonstrate radiological protection of the environment. Existing biota dose models cover generic terrestrial, freshwater, and marine ecosystems, not wetlands specifically. This paper, which was produced under IAEA's Environmental Modelling for Radiation Safety (EMRAS) II programme, describes an evaluation of how models can be applied to radionuclide contaminated wetlands. Participants used combinations of aquatic and terrestrial model parameters to assess exposure. Results show the importance of occupancy factor and food source (aquatic or terrestrial) included. The influence of soil saturation conditions on external dose rates is also apparent. In general, terrestrial parameters provided acceptable predictions for wetland organisms. However, occasionally predictions varied by three orders of magnitude between assessors. Possible further developments for biota dose models and research needs are identified.

Plutonium in wildlife and soils at the Maralinga legacy site: persistence over decadal time scales.

The mobility of plutonium (Pu) in soils, and its uptake into a range of wildlife, were examined using recent and ∼25 year old data from the Taranaki area of the former Maralinga weapons test site, Australia. Since its initial deposition in the early 1960s, the dispersed Pu has been incorporated into the soil profile and food chain through natural processes, allowing for the study of Pu sequestration and dynamics in relatively undisturbed semi-arid conditions. The data indicate downward mobility of Pu in soil at rates of ∼0.2-0.3 cm per year for the most mobile fraction. As a result, while all of the Pu was initially deposited on the ground surface, approximately 93% and 62% remained in the top 0-2 cm depth after 25- and 50-years respectively. No large-scale lateral spreading of the Taranaki plume was observed. Pu activity concentrations in 0-1 cm soils with biotic crusts were not elevated when compared with nearby bare soils, although a small number of individual data suggest retention of Pu-containing particles may be occurring in some biotic crusts. Soil-to-animal transfer, as measured by concentration ratios (CRwo-soil), was 4.1E-04 (Geometric Mean (GM)) in mammals, which aligns well with those from similar species and conditions (such as the Nevada Test Site, US), but are lower than the GM of the international mammal data reported in the Wildlife Transfer Database (WTD). These lower values are likely due to the presence of a low-soluble, particulate form of the Pu in Maralinga soils. Arthropod concentration ratios (3.1E-03 GM), were similar to those from Rocky Flats, US, while values for reptiles (2.0E-02 GM) were higher than the WTD GM value which was dominated by data from Chernobyl. Comparison of uptake data spanning approximately 30 years indicates no decrease over time for mammals, and a potential increase for reptiles. The results confirm the persistence of bioavailable Pu after more than 50 years since deposition, and also the presence of larger-sized particles which currently affect CRwo-soil calculations, and which may serve as an ongoing source of bioavailable Pu as they are subjected to weathering into the future.

Whole-organism concentration ratios for plutonium in wildlife from past US nuclear research data.

Whole-organism concentration ratios (CRwo-media) for plutonium (Pu) in wildlife were calculated using data from the broad range of organism types and environmental settings of the US nuclear research program. Original sources included site-specific reports and scientific journal articles typically from 1960s to 80s research. Most of the calculated CRwo-media values are new to existing data sets, and, for some wildlife categories, serve to fill gaps or add to sparse data including those for terrestrial reptile; freshwater bird, crustacean and zooplankton; and marine crustacean and zooplankton. Ratios of Pu concentration in the whole-organism to that in specific tissues and organs are provided here for a range of freshwater and marine fish. The CRwo-media values in fish living in liquid discharge ponds were two orders of magnitude higher than those for similar species living in lakes receiving Pu from atmospheric fallout, suggesting the physico-chemical form of the source Pu can dominate over other factors related to transfer, such as organism size and feeding behavior. Small rodent data indicated one to two order of magnitude increases when carcass, pelt, and gastrointestinal tract were included together in the whole-organism calculation compared to that for carcass alone. Only 4% of Pu resided in the carcass of small rodents compared to 75% in the gastrointestinal tract and 21% in the pelt.

The IAEA handbook on radionuclide transfer to wildlife.

An IAEA handbook presenting transfer parameter values for wildlife has recently been produced. Concentration ratios (CRwo-media) between the whole organism (fresh weight) and either soil (dry weight) or water were collated for a range of wildlife groups (classified taxonomically and by feeding strategy) in terrestrial, freshwater, marine and brackish generic ecosystems. The data have been compiled in an on line database, which will continue to be updated in the future providing the basis for subsequent revision of the Wildlife TRS values. An overview of the compilation and analysis, and discussion of the extent and limitations of the data is presented. Example comparisons of the CRwo-media values are given for polonium across all wildlife groups and ecosystems and for molluscs for all radionuclides. The CRwo-media values have also been compared with those currently used in the ERICA Tool which represented the most complete published database for wildlife transfer values prior to this work. The use of CRwo-media values is a pragmatic approach to predicting radionuclide activity concentrations in wildlife and is similar to that used for screening assessments for the human food chain. The CRwo-media values are most suitable for a screening application where there are several conservative assumptions built into the models which will, to varying extents, compensate for the variable data quality and quantity, and associated uncertainty.

Assessing doses to terrestrial wildlife at a radioactive waste disposal site: inter-comparison of modelling approaches.

Radiological doses to terrestrial wildlife were examined in this model inter-comparison study that emphasised factors causing variability in dose estimation. The study participants used varying modelling approaches and information sources to estimate dose rates and tissue concentrations for a range of biota types exposed to soil contamination at a shallow radionuclide waste burial site in Australia. Results indicated that the dominant factor causing variation in dose rate estimates (up to three orders of magnitude on mean total dose rates) was the soil-to-organism transfer of radionuclides that included variation in transfer parameter values as well as transfer calculation methods. Additional variation was associated with other modelling factors including: how participants conceptualised and modelled the exposure configurations (two orders of magnitude); which progeny to include with the parent radionuclide (typically less than one order of magnitude); and dose calculation parameters, including radiation weighting factors and dose conversion coefficients (typically less than one order of magnitude). Probabilistic approaches to model parameterisation were used to encompass and describe variable model parameters and outcomes. The study confirms the need for continued evaluation of the underlying mechanisms governing soil-to-organism transfer of radionuclides to improve estimation of dose rates to terrestrial wildlife. The exposure pathways and configurations available in most current codes are limited when considering instances where organisms access subsurface contamination through rooting, burrowing, or using different localised waste areas as part of their habitual routines.

Movement of a tritium plume in shallow groundwater at a legacy low-level radioactive waste disposal site in eastern Australia.

Between 1960 and 1968 low-level radioactive waste was buried in a series of shallow trenches near the Lucas Heights facility, south of Sydney, Australia. Groundwater monitoring carried out since the mid 1970s indicates that with the exception of tritium, no radioactivity above typical background levels has been detected outside the immediate vicinity of the trenches. The maximum tritium level detected in ground water was 390 kBq/L and the median value was 5400 Bq/L, decay corrected to the time of disposal. Since 1968, a plume of tritiated water has migrated from the disposal trenches and extends at least 100 m from the source area. Tritium in rainfall is negligible, however leachate from an adjacent and fill represents a significant additional tritium source. Study data indicate variation in concentration levels and plume distribution in response to wet and dry climatic periods and have been used to determine pathways for tritium migration through the subsurface.

Radionuclide concentration ratios in Australian terrestrial wildlife and livestock: data compilation and analysis.

Radionuclide concentrations in Australian terrestrial fauna, including indigenous kangaroos and lizards, as well as introduced sheep and water buffalo, are of interest when considering doses to human receptors and doses to the biota itself. Here, concentration ratio (CR) values for a variety of endemic and introduced Australian animals with a focus on wildlife and livestock inhabiting open rangeland are derived and reported. The CR values are based on U- and Th-series concentration data obtained from previous studies at mining sites and (241)Am and (239/240)Pu data from a former weapons testing site. Soil-to-muscle CR values of key natural-series radionuclides for grazing Australian kangaroo and sheep are one to two orders of magnitude higher than those of grazing cattle in North and South America, and for (210)Po, (230)Th, and (238)U are one to two orders of magnitude higher than the ERICA tool reference values. When comparing paired kangaroo and sheep CR values, results are linearly correlated (r = 0.81) for all tissue types. However, kidney and liver CR values for kangaroo are typically higher than those of sheep, particularly for (210)Pb, and (210)Po, with values in kangaroo liver more than an order of magnitude higher than those in sheep liver. Concentration ratios for organs are typically higher than those for muscle including those for (241)Am and (239/240)Pu in cooked kangaroo and rabbit samples. This study provides CR values for Australian terrestrial wildlife and livestock and suggests higher accumulation rates for select radionuclides in semi-arid Australian conditions compared with those associated with temperate conditions.

Hydrologic response and radionuclide transport following fire at semiarid sites.

Infrequent, high-impact events such as wildfires, droughts, biological shifts, floods, and mechanical disturbances can greatly change land surfaces, including vegetative cover and soil characteristics, which in turn can trigger high rates of hydrologic erosion and associated transport of sediments and sediment-sorbed contaminants. Where persistent soil contamination exists, infrequent mobilization of contaminants may dominate in determining long-term risks to human and ecological receptors. Among these infrequent events, fire stands out as having the capacity to cause large increases in sediment transport. This study measured runoff, sediment yield, and mobility of sediment-sorbed contamination (137Cs) on burned and unburned plots at the Waste Isolation Pilot Plant, New Mexico (WIPP), and the Rocky Flats Environmental Technology Site, Colorado (RFETS). Results showed that 137Cs transport from burned plots was up to 22 times greater than that from unburned plots at WIPP and 4 times greater at RFETS. Associated runoff was up to 12 times greater on burned plots at WIPP and sediment yields up to 6 times greater. Further, 137Cs concentrations in transported sediments were enriched compared with parent soils (expressed as enrichment ratio) by a factor of 2.3 at WIPP, and 1.3 at RFETS. However, enrichment ratios were not significantly different in sediments from burned and unburned plots. Our results provide new data on the effects of fire on the transport of sediment-sorbed contaminants, and demonstrate that rare events such as fire can greatly increase contaminant mobility.