Impacts on the marine environment in the case of a hypothetical accident involving the recovery of the dumped Russian submarine K-27, based on dispersion of 137Cs

Journal of Environmental Radioactivity 167 (2017) 170e179

Contents lists available at ScienceDirect

Journal of Environmental Radioactivity journal homepage: www.elsevier.com/locate/jenvrad

Impacts on the marine environment in the case of a hypothetical accident involving the recovery of the dumped Russian submarine K27, based on dispersion of 137Cs A. Hosseini a, *, I. Amundsen a, J. Brown a, M. Dowdall a, M. Karcher b, F. Kauker b, R. Schnur b a b

Norwegian Radiation Protection Authority, Po Box 55, N-1332, Østerås, Norway O.A.Sys - Ocean Atmosphere Systems GmbH, Tewessteg 4, 20249 Hamburg, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 16 August 2016 Received in revised form 28 November 2016 Accepted 28 November 2016 Available online 5 December 2016

There is increasing concern regarding the issue of dumped nuclear waste in the Arctic Seas and in particular dumped objects with Spent Nuclear Fuel (SNF). Amongst dumped objects in the Arctic, the dumped Russian submarine K-27 has received great attention as it contains two reactors with highly enriched fuel and lies at a depth of about 30 m under water. To address these concerns a health and environmental impact assessment has been undertaken. Marine dispersion of potentially released radionuclides as a consequence of different hypothetical accident scenarios was modelled using the model NAOSIM. The outputs from the dispersion modelling have been used as inputs to food-chain transfer and environmental dosimetry models. The annual effective doses for subsistence fishing communities of the Barents-Kara seas region do not exceed 0.6 mSv for hypothetical accidents located at Stepovogo fjord or the Barents Sea. For high rate consumers of fish in Norway, following a potential accident at the Gremikha Bay, annual effects doses would be at around 0.15 mSv. Accumulated doses (over 90 days) for various organisms and for all release scenarios considered were never in excess of 150 mGy. The levels of 137 Cs derived for marine organism in areas close to Norway were not values that would likely cause concern from a regulatory perspective although for subsistence fishing communities close to the considered accident locations, it is not inconceivable that some restrictions on fishing etc. would need to be introduced. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction There has been and continues to be concern over potential radioactive contamination of the Arctic due to the presence of a wide range of nuclear sources within this region. Dumped radioactive waste is the greatest contributor to the total activity found in the Arctic, inputs from Sellafield and global fallout comprising the next most significant contributions. Of dumped radioactive wastes, those objects containing Spent Nuclear Fuel (SNF) are of special importance. Dumping of liquid and solid radioactive waste in Arctic waters was a disposal practice conducted by the former USSR and later by Russia from the early 1960s through to the early 1990s. Oceanic

* Corresponding author. E-mail address: [email protected] (A. Hosseini). http://dx.doi.org/10.1016/j.jenvrad.2016.11.032 0265-931X/© 2016 Elsevier Ltd. All rights reserved.

dumping of radioactive wastes, carried out by thirteen countries, occurred in the Atlantic and Pacific oceans during the same period. Assessments of the total activity of the liquid and solid radioactive waste dumped into the Barents and Kara Seas were first highlighted in a 1993 report published by the Russian Governmental Commission on Issues of At-sea Disposal. Commonly referred to as the White Book 1993 (WB, 1993), the report was subsequently revised in 1993e1996 by the International Arctic Seas Assessment Project (IASAP) and then summarized by the IAEA in their technical document ‘Inventory of radioactive waste disposal at sea’ (IAEA, 1999). The White Book 2000 (Sivintsev et al., 2005) contained a reassessment of the material originally published in 1993 identifying a number of inaccuracies and omissions. The November class nuclear submarine K-27 was laid down in June 1958, and launched four years later in April 1962. Powered by two liquid metal (Pb-Bi) cooled reactors (LMRs) of 70 MW(t) each, both reactors were refuelled in September 1967 (with 91.5 kg U-

A. Hosseini et al. / Journal of Environmental Radioactivity 167 (2017) 170e179

235 each) and remained operational until May 24, 1968. On this date, blockage of some of the portside reactor core channels initiated a reactor event damaging approximately 20% of the portside reactor's fuel (IAEA, 1997). After attempts to repair the submarine, a decision was made to decommission it and the submarine was ultimately scuttled in September 1981in the shallow waters of Stepovogo Fjord (Fig. 1) at an estimated depth of 30 m. More than three decades later, the K-27 has become a topic of discussion as part of plans for remediation of the Arctic Seas (NES, 2013). In this context, Russian experts have elaborated upon three possible handling scenarios related to the submarine. These are, in essence: taking no action, in-situ isolation of the submarine, and raising and transporting it to land for dismantling. With regards to the last scenario, different recovery options have been evaluated. A detailed description of these options can be found in NES (2013). To address the abovementioned concern, a health and environmental impact assessment of the submarine K-27 was initiated in which a number of hypothetical accident scenarios and the evaluation of possible consequences of these were considered (Hosseini et al., 2015). The study included the impacts on both aquatic and terrestrial ecosystems. The aim of the work presented in this article was to focus on the marine part of the study and provide a comprehensive human health and environmental impact assessment for potential marine releases of radionuclides that could occur from the dumped nuclear submarine K-27. Making use of the most up-to-date information available as well as state of the art 3-dimensional hydrodynamic dispersion models, attempts were made to elucidate the transport, distribution and fate of relevant radionuclides in marine ecosystems following hypothetical accidents. Focus was placed upon the assessment of impact in the near field (local) and intermediate field (regional) over various timescales. The issue of an uncontrolled chain reaction occurring as a result of various management options was also considered in the assessment. 2. Methods To evaluate the marine dispersion of potentially released radionuclides as a consequence of a prolonged stay under water and/ or as a result of a possible recovery of K-27, the large scale numerical model NAOSIM (North Atlantic/Arctic coupled Ocean Sea € berle and Gerdes (2003)) was Ice Model) (Karcher et al., 2003; Ko employed (Hosseini et al., 2016). The model, derived from the Geophysical Fluid Dynamics Laboratory modular ocean model MOM-2 (Pacanowski, 1995), is coupled to a dynamicthermodynamic sea ice model with a viscous-plastic rheology (Hibler, 1979). The version utilised in this work had 30 unevenly spaced vertical levels, starting from 20 m down to 100 m depth with the thickness gradually increasing with depth. The model domain covers the Nordic Seas, the Arctic Ocean and the northern North Atlantic to 50
N and the Canadian Archipelago, allowing throughflow between the central Arctic and the Labrador Sea. NAOSIM has previously been employed and validated successfully in a number of applications involving Nordic Sea and Arctic Ocean circulation (Karcher et al., 2008; Gerdes et al., 2005) and tracer dispersion (Gerdes et al., 2001; Karcher et al., 2004, 2012; Kauker et al., 2016). A limitation of the model is that the interactions of radionuclides with the solid phase are not accounted for and therefore simulations are most applicable for those radionuclides behaving similarly to passive tracers but not those radionuclides where substantial adsorption to particles occurs. Further details regarding the model and its application to this assessment can be found in Karcher et al. (submitted) and Hosseini et al. (2016). With respect to potential releases to the marine environment, the following accident locations were considered:

171

a) Stepovogo Fjord b) Barents Sea (under transportation) c) Gremikha Bay These were chosen based upon the current position of K-27, and the projected route to and destination of the submarine in the event of a potential salvaging operation (Fig. 1). For each of the three, three separate large-scale atmospheric and oceanic circulation regimes were considered, these being a weak and strong flushing of the Barents and Kara Seas and a special case exhibiting reverse flow through Kara Gate. For each combination of accident scenario and circulation regime, hypothetical release situations of both an instantaneous (1 PBq) and continuous nature (1 TBq/yr) were studied. The focus of this work was upon the potential consequences of instantaneous release scenarios reflecting the fact that they would result in potentially greater radiological impacts than those associated with continuous release scenarios. The results from the advection-dispersion simulations were subsequently applied to more realistic situations by simply scaling to the actual release simulated for a given scenario. This was achieved by conducting detailed analyses of source term and release fractions, the basis of which were the decay-corrected inventories associated with the K-27 and calculations related to the potential for and resulting inventories following a Spontaneous Chain Reaction (SCR), i.e. criticality which could be initiated following the ingress of a small volume of water into the reactor compartment of the submarine. The details around this analyses are considered beyond the scope of this paper but further information can be found in Hosseini et al. (2015) and the supplementary material, S1 for this article. The simulation of a passive tracer necessitated substantial simplification in relation to the source term for modelling and assessment purposes, facilitated by the assumption that the entire inventory was present as 137Cs, a radionuclide that behaves relatively conservatively in seawater (see Jefferies et al., 1973). This assumption could be justified as 137Cs constitutes over 80% of the long-lived fission products released to water following a hypothetical accident (as shown in Supplementary material, S1 Table S1.10). Subsequent determinations refer simply to a total activity in units of Bq, which is assumed to be entirely attributable to 137 Cs assessment purposes. Regarding releases to the marine environment, the following accident scenarios were considered: I) SCR in both reactors, entire inventory released, II) The entire inventory of the starboard reactor released during an SCR with fractional releases from the portside reactor, III) SCR in the starboard reactor, entire inventory released, no release from the portside reactor, IV) SCR in the starboard reactor with a fractional release only. The first three represent very conservative assumptions, the last scenario might be considered to be the most plausible based on estimates and prognoses provided by the Russian Energy Safety Analysis Centre of IBRAE (Hosseini et al., 2015). A compromise was necessary between conservatism and realism such that the considered release scenario, while conservative, retained plausibility. These considerations led to adoption of the third scenario and an assumed maximum total release of 0.4 PBq for all subsequent calculations. The assessment approach involved several steps:  Analysis of potential exposure pathways  Derivation of doses (effective doses for people, weighted absorbed doses for non-human biota)

172

A. Hosseini et al. / Journal of Environmental Radioactivity 167 (2017) 170e179

Fig. 1. Main dumping areas in the Barents and Kara Seas as reported in the White Book 2000 (Sivintsev et al., 2005). Adapted from (JNREG, 2014). Location of the dump site of K27 at Stepovogo Fjord (Nr. 5), and the probable post-recovery location of the submarine at Gremikha Bay is indicated on the map. Dotted line indicates envisioned transport route after eventual recovery. Revised estimates of the maximum total activity (TBq) of the dumped solid waste at the time of dumping as reported in the White Book 2000 (Sivintsev et al., 2005) are given in parentheses. Adapted from (JNREG, 2014).

 Contextualisation of the doses in terms of appropriate benchmarks The concepts of ‘Representative Person’ (ICRP, 2007) and Reference Animals and Plants (ICRP, 2008) were used in the assessment. Much effort has been expended on the selection of suitable representative organisms1 for Arctic ecosystems (Brown et al., 2003) and drawing on this information, the final list for the marine ecosystem adopted included fish, seals and sea birds. 2.1. Modelling transfer and estimation of doses The transfer of contaminants to marine plants and animals is often estimated using concentration ratios (or factors) (see IAEA, 2004; IAEA, 2010), the applicability of such parameters being more suited to situations when steady-state conditions are prevalent. Where environmental activity concentrations are changing rapidly with time, dynamic models are more appropriate (Vives i Batlle et al., 2008). For this reason, marine food-chain models

1

Organisms which are representative or typical of a contaminated environment.

were applied (supplementary material, S2) to estimate activity concentrations in representative aquatic species, focus being placed primarily upon 137Cs as discussed earlier. The modelling platform software ECOLEGO-6 (Avila et al., 2005) was employed for this purpose. 2.2. Environmental impact assessment The environmental impacts were estimated using the ERICA Integrated Approach (Larsson, 2008). Once activity concentrations in environmental media and biota were derived, dose-rates were estimated through application of the ERICA Tool (Brown et al., 2008, 2016a). Activity concentration data were used as the basis for deriving internal (Dint) and external (Dext) absorbed dose-rates (in units of mGy h1), the total absorbed doserate being the sum of these components, calculated through the application of dose conversion coefficients (DCCs) (ICRP, 2008). Surface and depth-averaged activity concentrations of 137Cs in seawater and biota were selected for an area encompassing the most elevated levels associated with the main plume of contamination for the release scenarios. While these data might correspond to a single grid cell in the modelling domain, the spatial averaging

A. Hosseini et al. / Journal of Environmental Radioactivity 167 (2017) 170e179

involved was nonetheless quite substantial covering an area of approximately 900 km2. A spatial averaging of this magnitude might over-compensate for the migratory/peripatetic nature of the animal and fish species covered (Sazykina, 1998; Wienerroither et al., 2011), but no higher resolution was available for the model. This observation provided the rationale for additionally considering the surface water only (to a depth of 10 m) in the model output, and for which activity concentrations were at a maximum, to more robustly characterise the potential levels in the environment that could be attained for the release. For consistency, and to address similar concerns regarding spatial averaging, surface water data were also additionally used for the human assessment. 2.3. Human dose assessment For the estimation of human exposure (see e.g. IAEA, 2003) the ingestion of contaminated seafood pathway was considered. The total annual effective dose from the ingestion of food, Eing, food, public (in Sv/a), was estimated using Equation (1):

Eing;food;public ¼

X k

HB ðkÞ

X

CB ðj; kÞDCing;j ðjÞ



[1]

j

where:HB(k) is the rate of human consumption of foodstuff k (in kg/ a);DCing(j) is the dose coefficient for ingestion of radionuclide j (in Sv/Bq); (ICRP, 1995; ICRP, 2012)CB(j,k) is the concentration of radionuclide j in the edible fraction of foodstuff k (in Bq/kg, fresh weight). The human dose assessment was split into 2 main parts e the first relating to the local area and region surrounding the dumping site and the second for the Norwegian representative person. For the region surrounding the dumping site, the subsistence fishing communities of the Kara and Barents seas was considered. It was assumed that ingestion of seafood was the dominant exposure pathway for these groups. The IASAP (International Arctic Seas Assessment Project) postulated relevant consumption rates for these populations as being 500 g/d of sea fish, 80 g/d of sea mammals, 20 g/d of sea birds and 20 g/d of seabird eggs. These values were used in further calculations in this work. In view of its proximity to the subsistence fishing communities being considered, a hypothetical release point in Stepovogo Fjord was selected for further analysis. The second group considered was a high rate consumer of seafood in the Norwegian adult population. The EFSA (European Food Safety Authority) Comprehensive European Food Consumption Database (EFSA, 2011) was used as a source of dietary information. Consumption rates were provided for fish and seafood of 63 g/d and 250 g/d for an average person and person in the 99th percentile, respectively. Some consideration was given to the appropriate area for spatial averaging of the activity concentrations in seafood being used. Cod (Gadus morhua L.), haddock (Melanogrammus aeglefinus) and capelin (Mallotus villosus) were focused upon as these are important species for fisheries in the Barents Sea (Wienerroither et al., 2011). In view of uncertainties as to where fish have been prior to capture, a conservative assumption was adopted that the fish considered were distributed over a small area commensurate with the resolution of the marine model used (28  28 km) and in close proximity to the hypothetical accident site. The time series data for activity concentrations in seawater for this area from the NAOSIM model were utilised as input to the food-chain transfer models. 2.4. Contextualizing impacts to humans and the environment One way of contextualizing potential impacts to humans is via

173

the direct comparison of estimated foodstuff activity concentrations with corresponding levels derived from various pertinent, e.g. radiological, criteria. As an example, Heldal et al. (2013) used the intervention level of 600 Bq kg1 f.w., as currently applied by Norwegian authorities for basic foodstuffs (Liland et al., 2009), to contextualize activity concentration data derived for commercial fish species in relation to other sunken submarines. Similarly, by considering cod and capelin abundance data, Heldal et al. (2013) were also able to define the percentage of a given Barents Sea population that exceeded a level of 0.2 Bq kg1 137Cs commensurate with the current contamination level in cod within the Norwegian and Barents Seas. While such approaches have clear, practical merits, a decision was made not to focus on foodstuff based criteria as a sole indication of radiological significance. As the setting of intervention levels is arguably ambiguous (compare the 100 Bq kg1 f.w. applied by Japanese authorities to foodstuffs such as fish following the Fukushima accident (Buesseler, 2012) with the 600 Bq kg1 f.w. noted above). Kocher (1987) introduced the concept of a de minimis dose defining a level below which control of radiation exposures would be deliberately and specifically curtailed. In the context of setting criteria to allow or forbid the dumping of radioactive material at sea, the IAEA (1999, 2015) have used a de minimis dose of 10 mSv per annum. Although not strictly applicable, because of its development specifically for the London Convention, the de minimis level specified above provides a robust indication of what might be considered to be a trivial radiation dose and provides an appropriate benchmark with which the doses calculated in this assessment might be compared. For emergency (and existing) exposure situations, the sourcerelated restriction recommended by the ICRP is termed a “Reference level” (ICRP, 2007). In emergency or existing controllable exposure situations, the Reference levels represent the level of dose or risk, above which it is judged to be inappropriate to plan to allow exposures to occur, and for which therefore protective actions should be planned and optimised. At doses higher than 100 mSv, there is an increased likelihood of deterministic effects and a significant risk of cancer. For these reasons, the ICRP considers that the maximum value for a Reference level is 100 mSv incurred either acutely or in a year. The ICRP set a Reference level band of 20e100 mSv set for the highest planned residual dose from a radiological emergency. In relation to environmental impacts, there are activity concentration based criteria available to contextualize the impact of given levels of radionuclides in seawater or sediment, in the form of Environmental Media Concentration Limits (see Brown et al., 2008), although the standard methodology for making inferences about potential environmental effects involves the application of dose-based criteria. The ICRP (ICRP, 2008) recommend the application of a set of Derived Consideration Reference Levels (DCRLs) for particular categories of Reference Animals and Plants (RAPs) (ICRP, 2009). For environmental protection the ICRP recommend the following DCRL bands for the marine RAPs: 1e10 mGy d1, 1e10 mGy d1 and 10e100 mGy d1 for Seaweed, Flatfish and Crab, respectively (ICRP, 2014). In ICRP-124 (ICRP, 2014), the Commission has elaborated on the application of these criteria under cases other than planned exposure situations. For existing emergency exposure situations where control of the source has not been attained, if the dose rates are above the relevant DCRL band, the ICRP recommends that the aim should be to reduce exposures to levels that are within the DCRL bands for the relevant populations, with full consideration of the radiological and non-radiological consequences of so doing. If dose rates are within the bands, the ICRP recommends that consideration should be given to reduce exposures, assuming that the costs and benefits are such that further efforts are warranted.

174

A. Hosseini et al. / Journal of Environmental Radioactivity 167 (2017) 170e179

3. Results and discussion 3.1. Derived activity concentrations Simulated surface activity concentrations of 137Cs in seawater and concomitant levels in marine biota are presented in Fig. 2. The pulsed nature of the activity concentrations in seawater are quite evident with maximum surface levels (Table 1) occurring within the first 7 days of simulation (Fig. 2). The activity concentrations are then predicted to decrease rather rapidly to levels below 1 Bq/l once a period of 50e60 days has elapsed. To place the simulated and predicted activity levels (both in seawater and in biota) into context, the measured 137Cs activity concentrations at the areas of interest are summarized in Table 2. Simulations for the case of Stepovogo yielded the highest peak 137 Cs activity concentrations in biota with levels of approximately 150 Bq/kg f.w. and 340 Bq/kg f.w. derived for fish and seal respectively (Table 1). These elevations beyond those predicted for Gremikha and the Barents Sea, undoubtedly reflect the protracted, relatively high, concentrations of 137Cs in seawater for the Stepovogo scenario. The maximum activity concentration of 137Cs predicted for seabird fell slightly below 1000 Bq/kg f.w. Although the 137Cs activity concentrations predicted for fish fall, strictly speaking, below the aforementioned 600 Bq/kg intervention limit by some margin, the levels are close enough to raise some concerns. In view of the uncertainties involved, not least those associated with the imposed degree of spatial averaging (essentially over an area of 900 km2), the levels in fish for an actual pulsed release of 0.4 PBq 137Cs could conceivably exceed the intervention limit. Such an outcome would potentially require the imposition of fishing restrictions/fishery closures. Nonetheless, this expressed view is tempered by the knowledge that considering the total inventory of 137Cs and a more realistic release fraction (Hosseini et al., 2015), the most plausible simulated release of 137Cs could be a factor of 10 below the total release of 0.4 PBq considered here as shown in Supplementary material, S1. The predicted 137Cs contamination levels in fish for the given scenarios would far exceed existing “background” contamination levels (Table 2). According to Heldal et al. (2013), the levels of 137Cs in cod (Gadus morhua L.) muscle in the Barents Sea in the period around 2010 were approximately 0.1e0.2 Bq/kg f.w. Results of simulations of depth averaged activity concentrations of 137Cs in seawater along with resultant levels of 137Cs in marine biota are displayed in Fig. 3. The levels of contamination are substantially lower than those associated with surface activity concentrations reflecting the initial presence of 137Cs near the top of the water column (following the release of radioactivity from K-27 assumed to be at the sea surface) prior to the occurrence of substantial mixing. The maximum 137Cs activity concentrations in seawater, forming a similar pulsed time profile to that seen for the surface data series, attain levels of up to 3 Bq/l within the first week of the release. The Gremikha scenario differed from the other two release locations within the Barents Sea and Stepovogo Fjord as seabirds were not included in the analysis of the former. This accounts for the fact that seabirds and their eggs form a potential ingestion exposure pathway for humans for the Barents Sea and Stepovogo but not at Gremikha. The activity concentrations of 137Cs in seabirds actually exhibited the highest levels among all biota groups considered in the scenarios. The prediction of elevated transfer to seabirds relative to other organism types is in accordance with patterns observed within collations of empirically based concentration ratio data (IAEA, 2014). Nonetheless, the uncertainties associated with this prediction are significant and reflect the challenges in generating experimentally-based transfer parameters

Fig. 2. Sea water activity concentration of 137Cs at surface (Bq/l) along with the associated activities for fish, seal and seabird (Bq kg1 f.w.) based on releases to the marine environment for (a) the Gremikha scenario, (b) the Barents Sea scenario and (c) the Stepovogo scenario.

A. Hosseini et al. / Journal of Environmental Radioactivity 167 (2017) 170e179

175

for wild birds. The prognosis provides only an approximate indication of activity concentrations that might be observed but the fact that ingestion of seabirds/seabird eggs is unlikely to form an important (human) exposure pathway renders this putatively high uncertainty in prediction less critical than it might otherwise have been. Comparing predicted activity concentrations of all release scenarios highlights a similar pattern; the highest depth-averaged activity concentrations for the Stepovogo release scenario were noticeably lower than those predicted for Barents Sea and Gremikha. On the other hand, the trend observed for the highest surface activity concentrations provided predictions where the activity concentrations for the Barents Sea release scenario were markedly lower than those predicted for Gremikha and Stepovogo. A probable explanation lies in the differences in the nature of flow regimes considered for these scenarios. Two maximally different flow regimes were considered for these scenarios (Hosseini et al., 2016) - a weak and strong through-flow through the regional seas for Gremikha and the Barents Sea, respectively. Regionally, this translates into a longer flushing time for the south-eastern Barents Sea (where Gremikha is situated) corresponding to a weak flow regime and hence higher concentrations, compared to the stronger flow regime employed for open Barents Sea. Another evident general trend is the delayed occurrence of the maximum values for biota compared to those for water. Such delays in exhibiting peak activities are to be expected from an understanding of the general behaviour for the types of (bio)kinetic models applied and with regards to how transfer parameters have been selected. Differences in when maxima occur are primarily dictated by uptake and turnover rates, i.e. biological half-lives, the specifics of which are provided in the Supplementary material, S2. In contrast, the application of concentration ratios to the time

Table 1 Maximum activity concentration in water and marine biota at different accident locations considered in the study along with the estimated associated doses to biota. Accident location

Maximum dose (mGy/h)

Maximum activity concentration, Cs-137 Water (Bq/l)

Biota (Bq/kg f.w.)

Stepovogo Fjord

18

Barents Sea

13

Gremikha Bay

21

Fish Seal Seabird Fish Seal Seabird Fish Seal

153 342 987 72 164 470 129 285

0.03 0.12 0.19 0.013 0.054 0.09 0.024 0.094

Table 2 Measured activity concentrations of 137Cs in seawater and biota based on samples collected in the Stepovogo Fjord, Barents Sea and Norwegian Sea in recent years. Location

Measured activity concentration, Cs-137 Water (Bq/l)

Biota (Bq/kg f.w.)

Stepovogo Fjord

1.5E-03 e 1.8E-03

Barents Sea Norwegian Sea

1.6E-03 e 2.0E-03 1.1E-03 e 5.9E-03

Fish Seal Fish Fisha

a

<0.3 <0.2 <0.3 <0.5

Caught in coastal waters of Finnmark and Troms.

Reference

JNREG 2014 (Gwynn et al., 2012) NRPA (2011) & NRPA (2015)

Fig. 3. Depth-averaged activity concentrations of 137Cs in sea water (Bq/l) along with the associated activities for fish, seal and seabird (Bq kg1 f.w.) based on releases to the marine environment for (a) the Gremikha scenario, (b) the Barents Sea scenario and (c) the Stepovogo scenario.

176

A. Hosseini et al. / Journal of Environmental Radioactivity 167 (2017) 170e179

profile of 137Cs activity concentrations in seawater would have led to highly elevated biota concentrations in the initial phase followed by relatively low values in the proceeding phase (weeks and months later). Application of such methods would arguably lead to a poor characterisation of the 137Cs levels in the marine food-chain and provide misleading insights into the protracted nature of food contamination following an accidental release. A comparison can be made between the Barents Sea release in the current analysis and the modelling work of Heldal et al. (2013) for a nearby source in the Barents Sea. Heldal et al. (2013) used a 3D numerical ocean model and predicted maximum (essentially depth averaged) levels in seawater of around 0.5 Bq/l137Cs for a 5.2 PBq instantaneous input (although for a nominally different source, i.e. K-159). The maximum (depth averaged) seawater concentrations obtained in the present study were approximately 2 Bq/l137Cs for a 0.4 PBq instantaneous input. Furthermore, in the study of Heldal et al. (2013), levels in fish (cod) were predicted to be at around 63 and 123 Bq/kg f.w. in the near-surface and near-bottom layer, respectively compared to our values of around 70 Bq/kg f.w.. These values are apparently similar but this is misleading because of the above-noted differences in the source term magnitudes applied. This discrepancy can only be partly explained by the release conditions, i.e. near surface for the present study versus near-bottom for the study of Heldal et al. (2013). The difference would have been even more pronounced had the option been taken, as it was in Heldal et al. (2013), to use seawater-to-fish concentration ratios as opposed to the kinetic models applied in the present work. It is apparent that the main cause of the discrepancy lies in the fact that the numerical advection dispersion model applied by Heldal et al. (2013) predicts much lower activity concentrations of radionuclides in seawater, for a given release of 137Cs, compared to the corresponding model used in the present study. 3.2. Estimated environmental dose rates The dose rates for fish and seals predicted to arise from the Gremikha release scenario using the surface activity concentration in seawater (for a given grid cell) are presented in Fig. 4a. The dose rates were of the same order of magnitude as those associated with exposure from naturally occurring radionuclides (cf. 0.15 mGy/h for flatfish (see Hosseini et al., 2010) with the maximum of 0.024 mGy/h for fish (present study) and 0.1 mGy/h for marine mammals (Brown et al., 2004) with the maximum of 0.094 mGy h1 for seal determined here). The accumulated doses (over 90 days) of 43 mGy and 130 mGy were determined for fish and seals respectively. For the case of Barents Sea, the accumulated doses (over 90 days) of 24, 74 and 132 mGy were determined for fish, seals and sea bird, respectively (see Fig. 4b). Dose-rates pertaining to surface activity concentrations of 137Cs for the Stepovogo scenario are depicted in Fig. 4c. Seabirds were predicted to experience the most elevated exposures with dose rates of around 0.19 mGy/h. The dose rates for fish and seal were of a similar order of magnitude to the dose rates associated with exposures for corresponding groups of organisms from naturally occurring radionuclides (Brown et al., 2004; Hosseini et al., 2010). Generally, even at maximum dose-rates (see Table 1), the exposures are orders of magnitude below the 0.1e1 mGy/d DCRL band recommended for application to mammals (strictly speaking Reference Deer and Rat but mammals generally are known to exhibit similar radiosensitivity) by the ICRP (2008). This band is considered to correspond to dose-rates where the probability of radiation-induced effects occurring is low. Using the depth averaged seawater activity concentrations as input to the kinetic and dose rate model, the resultant dose-rates were lower by a factor of 4e5. The total dose rates essentially

Fig. 4. Dose-rates (mGy/h) derived from surface activity concentrations of 137Cs in sea water for fish, seal and seabird based on releases to the marine environment for (a) the Gremikha scenario, (b) the Barents Sea scenario and (c) the Stepovogo scenario.

reflect the activity concentrations of 137Cs in the organisms, per se, as opposed to ambient seawater concentrations with >95% doserate attributable to internal body burdens of radioaceasium at

A. Hosseini et al. / Journal of Environmental Radioactivity 167 (2017) 170e179

times after 1 month. The maximum dose rates calculated for the Gremikha scenario were slightly above 5.0E-03 mGy/h for fish and slightly above 2.0E-02 mGy/h for seal. The accumulated doses (90 days) of approximately 9 and 27 mGy were derived for fish and seals, respectively. The maximum dose rates predicted for the Barents Sea release scenario using the depth averaged activity concentration in seawater were 3.0E-03, 1.3E-02 and 2.1E-02 mGy/h for fish, seal and sea bird, respectively. The accumulated doses (90 days) of approximately 5.4, 15 and 28 mGy were derived for fish, seals and sea bird, respectively. As for Gremikha, internal doserates for the Stepovogo release scenario dominate following the initial 30 day period reflecting the gradual uptake and transfer through the marine food-chain. Maximum dose rates at around 100 days for seal and seabirds with an absolute maximum of just under 0.02 mGy/h being derived for the latter. Accumulated doses (30 days) of approximately 14 mGy for both fish and mammal and approximately 28 mGy for seabirds were derived. All the predicted dose rates are extremely low doses falling orders of magnitude below levels where any types of effect on organisms might be expected. 3.3. Estimation of doses to humans The doses for subsistence fishing communities of the BarentsKara seas region are provided in Table 3 and fall far below the reference level of 100 mSv (ICRP, 2007) applicable to emergency exposure situations but certainly well above a level which might be considered trivial. As considered above, the environmental management of this situation might be driven by restrictions placed on foodstuffs because of applied intervention levels. In view of the activity concentrations that might hypothetically be attributed to foodstuffs for the pessimistic scenarios considered, it is not inconceivable that some restrictions on fishing etc. would need to be introduced. For the Gremikha scenario, in contrast to the other two scenarios, it has been assumed that ingestion of contaminated fish would be the main exposure pathway for the considered population group in Norway (high rate consumers of fish) following a potential accident at Gremikha Bay. The results shown in Table 3 provide a conservative measure of the potential exposures of human critical groups as a result of an accident at Stepovogo, Barents Sea and Gremikha. It can be seen that, of the accident scenarios considered, the accident at Stepovogo Fjord gives the highest dose. However, there is a caveat on direct comparison of the consequences of these accident scenarios as they consider different types of accident and also different target population groups. 3.4. Dealing with uncertainties It must borne in mind that for the case of K-27 (or any other dumped or sunken object for which salvage is an option) there are factors which compound the existing uncertainty. Such factors might include, among other things, the time and circumstances of lifting and the characteristics of the scenarios considered. When, in

177

the future, a salvage operation would take place and which technology would be used has an impact on the applicability of the results arising from any concomitant impact assessment (NES, 2013). The use of hypothetical scenarios implies making assumptions that often cannot be verified. Uncertainties of this kind most probably mask much of the effort in characterising other types of uncertainties that could be described statistically. The choice of grid cell dimensions for the advection-dispersion modelling of radionuclides may clearly also influence the maximum radionuclide levels in sea water that are simulated but our adoption of 28  28 km grid cells leads to an acceptable spatial averaging commensurate with our subsequent purpose of characterising activity concentrations over populations of organisms or over entire fishing areas. The issue of spatial averaging, a critical feature relating to the spatial extent of various biota populations, has been raised earlier by Hosseini et al. (2010). With regards to interpreting doses to biota, chronic effects of ionising radiation in marine animal species are scarcely documented in radiobiology literature for species besides fish (see Copplestone et al., 2008). Conscious of these sources of great uncertainty, spending resources to characterise model and parameter uncertainties, measurement and sampling errors and other location relevant uncertainties would not be fully justified. Hence, a resource efficient and pragmatic approach to deal with uncertainty, and consequently lend credibility to the outcomes of a study of the kind conducted, would be to apply conservative assumptions, consider extreme accident scenarios and employ high-end input values (IPCS, 2014). Such an approach has been adopted in this study. Conservatism has been introduced at various points in the assessment. This has been done by assessing the worst case scenarios which represent extreme situations (e.g. different flow regimes in marine dispersion modelling), considering various accident scenarios, employing conservative parameters and assumptions (spontaneous release of total inventory, highest possible SCR), focusing on higher end input values (considering critical groups, using 99th percentiles). In addition, to reduce uncertainty, further attempts have been made to use best available knowledge/ information through consulting the most relevant information sources and employing state of the art models.

4. Conclusions The model NAOSIM was employed to simulate dispersion of radionuclides potentially released as a consequence of a possible recovery of K-27 or with regards a prolonged stay under water. The latter scenario reflected the modelling work conducted by IASAP (IAEA, 1997) where different corrosion rates of containment and barrier materials were considered as well as possible associated releases in the future. Accident scenarios at three locations were considered: Gremikha Bay, Barents Sea and Stepovogo Fjord. The doses for humans derived for the ingestion of contaminated marine foodstuffs provide a conservative measure of the potential exposures of human critical groups as a result of an accident at Stepovogo, Barents Sea and Gremikha. These are low relative to potential

Table 3 Estimated Cs-137 annual effective doses from ingestion of marine food for release scenarios at Stepovogo Fjord, Barents Sea and Gremikha Bay based on maximum (surface) activity concentrations. Scenario

Stepovogo Barents Sea Gremikha

Cs-137 (Bq/l)

Dose contribution (mSv)

Water

Fish

Seal

Sea bird

Bird egg

Annual effective dose (mSv)

18 13 21

3.6E-01 1.7E-01 1.5E-01

1.3E-01 6.2E-02 1.5E-01

9.4E-02 4.5E-02

1.4E-02 6.6E-03

6.0E-01 2.8E-01

178

A. Hosseini et al. / Journal of Environmental Radioactivity 167 (2017) 170e179

doses from the combined terrestrial exposure pathways (Brown et al., 2016b). The levels of 137Cs derived for marine organism in areas close to Norway were not values that would likely cause concern from a regulatory perspective although for subsistence fishing communities Yamal/Northern Yenisey, it is not inconceivable that some restrictions on fishing etc. would need to be introduced. Despite this consideration, any long-term consequences would not be expected with recovery of the marine system occurring within the first years following any hypothetical release. Bearing in mind the socio-economic impacts of other accidents involving nuclear materials, such as the sinking of the Kursk, as well as public unease evident in relation to the Fukushima accident, there remains a cause for concern. The study indicates the potential for significant and widespread contamination of the Arctic environment with radionuclides in the event of an incident involving a SCR. Such contamination and public perception of the significance of its extent and magnitude is difficult to predict but previous incidents serve to indicate that there would be a potential impact with respect to consumer confidence in marine products. Acknowledgments This work was partly supported by the Research Council of Norway through its Centre's of Excellence funding scheme, project number 223268/F50. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jenvrad.2016.11.032. References Avila, R., Broed, R., Pereira, A., 2005. ECOLEGO - a toolbox for radioecological risk assessment, 6-10 October 2003. STI/PUB/1229. In: Protection of the Environment from the Effects of Ionizing Radiation: Proceedings of an International Conference Stockholm, vol. 2005. IAEA, Vienna, pp. 229e232. Brown, J., Hosseini, A., Børretzen, P., Iosjpe, M., 2003. Environmental impact assessments for the marine environment e transfer and uptake of radionuclides. In: StrålevernRapport, vol. 2003. Norwegian Radiation Protection Authority, Østerås, p. 7, 2003. Brown, J.E., Alfonso, B., Avila, R., Beresford, N.A., Copplestone, D., Prohl, G., Ulanovsky, A., 2008. The ERICA tool. J. Environ. Radioact. 99 (9), 1371e1383. n, R., Thørring, H., Vives i Batlle, J., 2004. Radiation doses Brown, J.E., Jones, S.R., Saxe to aquatic organisms from natural radionuclides. J. Radiol. Prot. 24, A63eA77. Brown, J., Alfonso, B., Avila, R., Beresford, N., Copplestone, D., Hosseini, A., 2016a. A new version of the ERICA tool to facilitate impact assessments of radioactivity on wild plants and animals. J. Environ. Radioact. 153, 141e149. Brown, J., Amundsen, I., Bartnicki, J., Dowdall, M., Dyve, J.E., Hosseini, A., Klein, H., Standring, W., 2016b. Impacts on the terrestrial environment in case of a hypothetical accident involving the recovery of the dumped Russian submarine K27. J. Environ. Radioact. 165, 1e12. Buesseler, K.O., 2012. Fishing for answers off Fukushima. Science 338 (6106), 480e482. Copplestone, D., Hingston, J.L., Real, A., 2008. The development and purpose of the FREDERICA radiation effects database. J. Environ. Radioact. 99 (9), 1456e1463. EFSA (European Food Safety Authority), 2011. Use of the EFSA comprehensive European food consumption database in exposure assessment, 2011 EFSA J. 9 (3), 2097. Gerdes, R., Karcher, M., Kauker, F., Koeberle, C., 2001. Predicting the spread of radioactive substances from the Kursk. EOS Trans. Am. Geophys. Union 82, 253e257. Gerdes, R., Drange, H., Gao, Y., Karcher, M.J., Kauker, F., Bentsen, M., 2005. Ocean general circulation modelling of the Nordic Seas. In: Drange, Furevik, Gerdes, Berger (Eds.), The Nordic Seas: an Integrated Perspective. AGU Monograph Edition, pp. 199e220. Gwynn, J.P., Heldal, H.E., G€ afvert, T., Blinova, O., Eriksson, M., Sværen, I., Brungot, A.L., Strålberg, E., Møller, B., Rudjord, A.L., 2012. Radiological status of the marine environment in the Barents Sea. J. Environ. Radioact. 113, 155e162. Heldal, H.E., Vikebø, F., Johansen, G.O., 2013. Dispersal of the radionuclide caesium137 (137Cs) from point sources in the Barents and Norwegian Seas and its potential contamination of the Arctic marine food chain: coupling numerical ocean models with geographical fish distribution data. Environ. Pollut. 180,

190e198. Hibler III, W.D., 1979. A dynamic thermodynamic sea ice model. J. Phys. Oceanogr. 1979 (9), 815e846. Hosseini, A., Amundsen, I., Brown, J., Dowdall, M., Standring, W., 2015. Inventory and source term evaluation of the dumped nuclear submarine K-27. In: StrålevernRapport, vol. 2015. Norwegian Radiation Protection Authority, p. 6. Hosseini, A., Amundsen, I., Bartnicki, J., Brown, J., Dowdall, M., Dyve, J.E., Harm, I., Karcher, M., Kauker, F., Klein, H., Lind, O.C., Salbu, B., Schnur, R., Standring, W., 2016. Environmental modelling and radiological impact assessment associated with hypothetical accident scenarios for the nuclear submarine K-27. In: StrålevernRapport, vol. 2016. Statens strålevern, Østerås, p. 8, 2016. Hosseini, A., Beresford, N.A., Brown, J.E., Jones, D.G., Phaneuf, M., Thørring, H., Yankovich, T., 2010. (2010). Background dose-rates to reference animals and plants arising from exposure to naturally occurring radionuclides in aquatic environments. J. Radiological Prot. 30 (2), 235e264. IAEA, 1997. Predicted radionuclide release from marine reactors dumped in the Kara sea. In: Report of the Source Term Working Group of the International Arctic Seas Assessment Project (IASAP). IAEA-TECDOC-938, Vienna (1997). IAEA, 1999. Inventory of Radioactive Waste Disposals at Sea. IAEA-TECDOC-1105. IAEA, Vienna, 1999. IAEA, 2003. Determining the Suitability of Materials for Disposal at Sea under the London Convention 1972: a Radiological Assessment Procedure. IAEA-TECDOC1375. IAEA, Vienna, 2003. IAEA, 2004. Sediment Distribution Coefficients and Concentration Factors for Biota in the Marine Environment. Technical reports series 422. IAEA, Vienna, 2004. IAEA, 2010. Handbook of Parameter Values for the Prediction of Radionuclide Transfer in Terrestrial and Freshwater Environments. IAEA technical report series 472. IAEA, Vienna, 2010. IAEA, 2014. Handbook of Parameter Values for the Prediction of Radionuclide Transfer to Wildlife. Technical report series 479. IAEA, Vienna, 2014. IAEA, 2015. Determining the Suitability of Materials for Disposal at Sea under the London Convention 1972 and London Protocol 1996: a Radiological Assessment Procedure. IAEA-TECDOC-1759. IAEA, Vienna, 2015. ICRP, 1995. Age-dependent doses to members of the public from intake of radionuclides - part 3 ingestion dose coefficients. In: Annals of the ICRP 1995, vol. 25. ICRP publication 69 (1). ICRP, 2007. The 2007 recommendations of the international commission on radiological protection. In: Annals of the ICRP 2007, vol. 37. ICRP publication 103 (2-4). ICRP, 2008. Environmental protection: the concept and use of reference animals and plants. In: Annals of the ICRP 2008, vol. 38. ICRP publication 108 (4-6). ICRP, 2009. Transfer parameters for reference animals and plants. International commission on radiological protection. Environmental protection publication 114 (6). In: Annals of the ICRP, vol. 39. Elsevier, Oxford, 2009. ICRP, 2012. Compendium of dose coefficients based on ICRP Publication 60. In: Annals of the ICRP 2012, vol. 41. ICRP publication 119 (Suppl.). ICRP, 2014. Protection of the environment under different exposure situations. In: Annals of the ICRP 2014, vol. 43. ICRP publication 124 (1). IPCS, 2014. Guidance Document on Evaluating and Expressing Uncertainty in Hazard Characterization. The International Programme on Chemical Safety (IPCS). Harmonization Project Document 11. World Health Organization, ISBN 978 92 4 150761 5. Jefferies, D.F., Steele, A.K., Preston, A., 1973. Distribution of Cs-137 in British coastal waters. Mar. Pollut. Bull. 4, 118e122. JNREG, 2014. Investigation into the radioecological status of Stepovogo Fjord. In: Gwynn, Justin P., Nikitin, Alexander I. (Eds.), The Dumping Site of the Nuclear Submarine K-27 and Solid Radioactive Waste. Results from the 2012 Research Cruise. Norwegian Radiation Protection Authority, Østerås. Karcher, M.J., Gerdes, R., Kauker, F., Koeberle, C., 2003. Arctic warming - evolution and spreading of the 1990s warm event in the Nordic seas and the Arctic Ocean. J. Geophys. Res. 108 (C2), 3034. Karcher, M.J., Gerland, S., Harms, I.H., Iosjpe, M., Heldal, H.E., Kershaw, P.J., Sickel, M., 2004. The dispersion of Technetium-99 in the Nordic Seas and the Arctic Ocean: a comparison of model results and observations. J. Environ. Radioact. 74, 185e198. Karcher, M., Gerdes, R., Kauker, F., 2008. Long-term variability of Atlantic water inflow to the Northern Seas: insights from model experiments. In: Dickson, B., Meincke, J., Rhines, P. (Eds.), Arctic-subarctic Ocean Fluxes: Defining the Role of the Northern Seas in Climate. Springer, pp. 111e130. Karcher, M., Smith, J.N., Kauker, F., Gerdes, R., Smethie, W.M., 2012. Recent changes in Arctic Ocean circulation revealed by iodine-129 observations and modeling, 1978e2012 J. Geophys. Res. Oceans 117 (C8). Karcher, M., Hosseini, A., Schnur, R., Kauker, F., Brown, J.E., Dowdall, M., Strand, P., 2016. On the potential dispersal of radioactive contaminants in arctic waters during recovery operations on the dumped submarine K-27. Mar. Pollut. Bull. submitted for publication. Kauker, F., Kaminski, T., Karcher, M., Dowdall, M., Brown, J., Hosseini, A., Strand, P., 2016. Model analysis of worst place scenarios for nuclear accidents in the northern marine environment. Environ. Model. Softw. 77, 13e18. March 2016, Pages 1364e8152. http://dx.doi.org/10.1016/j.envsoft.2015.11.021. €berle, C., Gerdes, R., 2003. Mechanisms determining the variability of Arctic sea Ko ice conditions and export, 2003 J. Clim. 16 (17), 2843e2858. Kocher, D.C., 1987. A Proposal for a Generally Applicable de minimis Dose, 1987 Health Phys. 53 (2), 117e121. Larsson, C.M., 2008. (2008). An overview of the ERICA Integrated Approach to the

A. Hosseini et al. / Journal of Environmental Radioactivity 167 (2017) 170e179 assessment and management of environmental risks from ionising contaminants. J. Environ. Radioact. 99 (9), 1364e1370. Liland, A., Lochard, J., Skuterud, L., 2009. How long is long-term? reflections based on over 20 years of post-Chernobyl management in Norway. J. Environ. Radioact. 100, 581e584. NES, 2013. Sunken K-27 nuclear submarine recovery: risk overview. Nucl. Environ. Saf. Mag. Nr 1, 2013. NRPA, 2011. Radioactivity in the marine environment 2008 and 2009. Results from the Norwegian National monitoring programme (RAME). In: StrålevernRapport, vol. 2011. Norwegian Radiation Protection Authority, Østerås, p. 4, 2011. NRPA, 2015. Radioactivity in the marine environment 2011. Results from the Norwegian marine monitoring programme (RAME). In: StrålevernRapport, vol. 2015. Norwegian Radiation Protection Authority, Østerås, p. 3, 2015. Pacanowski, R.C., 1995. MOM 2 Documentation, User's Guide and Reference Manual, GFDL Ocean Group Tech. Rep. 3, Geophys. Fluid Dyn. Lab. Princeton Univ., Princeton, N. J. Sazykina, T.G., 1998. Long-distance radionuclide transfer in the Arctic Seas related to fish migrations, 1998 Radiat. Prot. Dosim. 75 (1e4), 219e222.

179

Sivintsev, YuV., Vakulovsky, S.M., Vasiliev, A.P., Vysotsky, V.L., Gubin, A.T., Danilyan, V.A., Kobzev, V.I., Kryshev, I.I., Lavkovsky, S.A., Mazokin, V.A., Nikitin, A.I., Petrov, O.I., Pologikh, B.G., Skorik, YuI., 2005. Technogenic Radionuclides in the Sea Surrounding Russia. Radioecological Consequences of Radioactive Waste Dumping in the Arctic and Far Eastern Seas ‘The White Book e 2000’, Moscow. Vives i Batlle, J., Wilson, R.C., Watts, S.J., Jones, S.R., McDonald, P., Vives-Lynch, S., 2008. Dynamic model for the assessment of radiological exposure to marine biota. J. Environ. Radioact. 99 (11), 1711e1730. WB, 1993. Facts and problems related to radioactive waste disposal in seas adjacent to the territory of the Russian Federation (1993). In: Materials from the Report of the Governmental Commission on Issues of At-sea Disposal Enacted by President's Order No. 613-rr Dated October 24th, 1992President Office, Moscow, 1993 (in Russian). Wienerroither, R., Johannesen, E., Dolgov, A., Byrkjedal, I., Bjelland, O., Drevetnyak, K., Eriksen, K.B., Høines, Å., Langhelle, G., Langøy, H., Prokhorova, T., Prozorkevich, D., Wenneck, T., 2011. Atlas of the Barents Sea Fishes. IMR/PINRO Joint Report Series 1-2011. Institute of Marine Research, Bergen, 2011.