Kara Sea radioactivity assessment

The Science of the Total Environment 237r238 Ž1999. 167]179

Kara Sea radioactivity assessment Iolanda OsvathU , Pavel P. Povinec, Murdoch S. Baxter Marine En¨ ironment Laboratory, International Atomic Energy Agency, B.P. 800, MC-98012, Monaco

Abstract Investigations following five international expeditions to the Kara Sea have shown that no radiologically significant contamination has occurred outside of the dumping sites in Novaya Zemlya bays. Increased levels of radionuclides in sediment have only been observed in Abrosimov and Stepovoy Bays very close to dumped containers. Evaluations of radionuclide inventories in water and sediment of the open Kara Sea and Novaya Zemlya bays as well as soil from the shore of Abrosimov bay have shown that radionuclide contamination of the open Kara Sea is mainly due to global fallout, with smaller contributions from the Sellafield reprocessing plant, the Chernobyl accident run-off from the Ob and Yenisey rivers and local fallout. Computer modelling results have shown that maximum annual doses of approximately 1 mSv are expected for a hypothetical critical group subsisting on fish caught in the Novaya Zemlya bays whereas populations living on the mainland can be expected to receive doses at least three orders of magnitude lower. Q 1999 Elsevier Science B.V. All rights reserved. Keywords: Arctic; Kara Sea; Marine radioactivity; Radionuclide inventories; Marine modelling; Radiological assessment

1. Introduction The information that high-level radioactive wastes had been dumped in the shallow waters of the Kara Sea ŽYablokov et al., 1993. has triggered several assessment and research programmes. The latest estimates ŽIAEA, 1997. indicate that the

U

Corresponding author. Tel.: q377-97-97-7272; fax: q37797-97-7273. E-mail address: [email protected] ŽI. Osvath.

present-day inventory of waste dumped in the Kara Sea is 4.7 PBq, most of it being high-level waste in naval reactor assemblies scuttled in the shallow bays of Novaya Zemlya and in the Novaya Zemlya Trough, at depths between 12 and 380 m. Although the Kara Sea dumpsites are remote from inhabited areas and fishing grounds, it was necessary to quantify the hazard to human and environmental health associated with dispersal and transfer of potentially released radionuclides ŽIAEA, 1998.. The main pathway of export of radionuclides from the Kara Sea is related to

0048-9697r99r$ - see front matter Q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 8 - 9 6 9 7 Ž 9 9 . 0 0 1 3 3 - 3

168

I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179

outflow of water towards the Laptev Sea and the Arctic Ocean ŽHarms, 1997.. The export of radionuclides from the Kara Sea by ice-entrained sediments is relatively insignificant ŽCooper et al., 1998; IAEA, 1999a,b.. The IAEA Marine Environment Laboratory’s Arctic programme, most activities of which were coordinated in the framework of the IAEA’s Arctic Seas Assessment Project ŽIASAP., included: 1. participation in five expeditions to the open Kara Sea and to the Novaya Zemlya dumpsites organized between 1992 and 1995 by a joint Russian]Norwegian expert group, the M urm ansk M arine Biology Institute ŽMMBIRAS. and a US]Norwegian group ŽFig. 1.; 2. radiometric analysis of water, sediment and biota samples collected during these expeditions, as well as in situ underwater investigations of the sea-bed at the dump sites using

3.

4.

5.

6.

HPGe and NaI ŽTl. spectrometers ŽHamilton et al., 1994; Povinec et al., 1996, 1997a.; organization of analytical quality assurance intercomparison exercises amongst the participating laboratories ŽPovinec et al., 1997a.; development of a Global Marine Radioactivity Database ŽGLOMARD. which includes Arctic marine radioactivity measurements reported since the early 1960s up to present ŽPovinec et al., 1997b.; theoretical, laboratory and field investigations of distribution coefficients Ž K d s. and concentration factors ŽCFs. in the Arctic marine environment to document whether they differ significantly from those found in temperate regions ŽBoisson et al., 1997; Carroll et al., 1997.; and development of computer models on local, regional and global scales to study dispersion and transfer of radionuclides which could potentially be released from dumped radioactive

Fig. 1. Expeditions to the Kara Sea that IAEA-MEL has participated in.

I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179

wastes ŽOsvath et al., 1995; Baxter et al., 1998.. In this publication we summarize the radioanalytical results, present new data on radionuclide inventories in the open Kara Sea and Novaya Zemlya bays estimated from water, sediment and soil data obtained from our own as well as published results which are included in the GLOMARD data base ŽIAEA, 1999a., discuss contributions from different sources to the radioactive contamination of the Kara Sea and radiological doses to humans on local, regional and global scales.

2. Investigation of radionuclide distributions in the Kara Sea 2.1. The open Kara Sea The first international expedition involving IAEA-MEL was organized by the Joint Russian]Norwegian Expert Group in 1992. The

169

radioanalytical results obtained on samples collected at 11 locations covering the Western Kara Sea ŽHamilton et al., 1994; Strand et al., 1994. showed unambiguously that no contamination attributable to leakage from dumped radioactive wastes could be detected in the investigated areas. A time-trend analysis of the data contained in the GLOMARD database shows that 90 Sr and 137 Cs concentrations in open Kara Sea waters have actually decreased ŽIAEA, 1999a.: average 90 Sr concentrations in surface water ranged from 40 Bq my3 in 1965 to 15 Bq my3 in 1982 and 5 Bq my3 at present. Average 137 Cs concentrations of 20 Bq my3 were recorded in the mid 1960s and mid 1980s, present values being on the average four times lower. The ventilation rate of the Kara Sea is approximately 3.5 years ŽPavlov and Pfirman, 1995. and winter convection results in effective mixing of the top 50]70-m layer ŽPavlov et al., 1993.. Reduced ventilation is observed nearbottom in the Novaya Zemlya Trough ŽPavlov and Pfirman, 1995.. Because of the temporal changes in the sources inputting radionuclides to the Kara Sea, coupled with the various origins and resi-

Fig. 2. Caesium-137 in Barents and Kara Sea bottom waters Žbelow 50 m..

170

I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179

dence times of different water masses, deeper waters in the Kara Sea show higher concentrations of 137Cs and 239,240 Pu than surface waters by approximately a factor of 2]3. The spatial distribution of 137Cs concentrations in the Barents and Kara Sea bottom waters in 5-year time intervals covering the period 1981]1995 is shown in Fig. 2. The successive snap-shots clearly show the progression of Sellafield 137 Cs towards the Barents Sea. The Ob and Yenisey rivers are important suppliers of fresh water and contaminants Žincluding, e.g. 137 Cs, Fig. 2, and 90 Sr. from land-based sources to the Kara Sea. Data obtained on samples collected during the 1993 MMBIRAS expedition ŽFig. 1. indicate, as expected, a fairly good correlation Ž r 2 s 0.74. between 137 Cs concentrations in surface water and salinity in the area of the estuaries. On the other hand, 90 Sr, which is less particle-reactive, showed no obvious trend within the investigated area. As radionuclide removal from the water column is enhanced in the estuary mixing zone, relatively high sediment inventories of 137 Cs and 239,240 Pu were observed here ŽPovinec et al., 1997a. as compared to the open Kara Sea. Clearly related to fresh water input from the Ob and Yenisey rivers, the 137 Csr 90 Sr ratios in surface water vary across the western Kara Sea, as do the concentrations of 90 Sr, while those of 137 Cs and 239,240 Pu are more homogeneous. Data obtained at IAEA-MEL ŽHamilton et al., 1994; Povinec et al., 1997a. and by the Joint Russian]Norwegian Expert Group ŽJoint Russian]Norwegian Expert Group, 1993, 1994; Strand et al., 1994. on samples collected at stations occupied in the open Kara Sea and the estuaries’ area in 1992]1993 ŽFig. 1. show that 137 Csr 90 Sr ratios were in the range 2.1]2.8 in the open sea and dropped to 0.4 at the mouths of the Ob and Yenisey. This gives a valuable indicator for the mixing of river waters in the Kara Sea. It can be noted also that the open Kara Sea 137 Csr 90 Sr ratios were higher than those typical of global fallout Žnear 1.6 at present cf. UNSCEAR, 1993., while at the mouth of the rivers they were considerably lower. Radionuclide data for sediments in the open Kara Sea are very sparse, only two significant data

sets being available up to the 1990s ŽNOO, 1969; Vakulovsky, 1993.. The highest 137 Cs levels in sediment Žapprox. 200 Bq kgy1 dry wt.. were observed in the 1960s in the central and eastern Kara Sea ŽNOO, 1969., probably due to discharges from the Ob river. Data from the 1980s show Ž4]20. Bq kgy1 dry wt. of 137 Cs in surface sediment ŽVakulovsky, 1993., which is similar to the present values Ž18]32. Bq kgy1 dry wt. ŽHamilton et al., 1994; Strand et al., 1994.. No local contamination of sediment has been observed in the open Kara Sea. The 238 Pur 239,240 Pu isotopic ratio is a good indicator of the Pu’s origin. Observed values ŽHamilton et al., 1994; Joint Russian]Norwegian Expert Group, 1994. are generally consistent with that typical of global fallout Ž0.02]0.04. resulting from nuclear weapons testing and debris from the 1964 satellite re-entry ŽUNSCEAR, 1993; Aarkrog, 1988.. Lead-210 dating of sediments from the Novaya Zemlya Trough Žin the 250]350 m water depth range. indicates sedimentation rates of 0.9]2.7 mm yeary1 ŽHamilton et al., 1994.. 2.2. No¨ aya Zemlya bays Data on radionuclide levels in the bays were first published following the Joint Russian] Norwegian cruises organized in 1993 and 1994. In Abrosimov Bay the concentrations of radionuclides in sediment cores within one of the container dumpsites were higher by a factor of 10]1000 than those at uncontaminated sites inside the same bay, e.g. up to 30 kBq kgy1 dry wt. of 137 Cs, ; 10 kBq kgy1 dry wt. of 90 Sr, ; 200 Bq kgy1 dry wt. of 60 Co and ; 15 Bq kgy1 dry wt. of 239,240 Pu were measured ŽJoint Russian]Norwegian Expert Group, 1994, 1996; Povinec et al., 1997a.. The 238 Pur 239,240 Pu activity ratio ranged from 0.3 to 0.7. This is over one order of magnitude higher than the global fallout ratio, confirming that the contamination is derived from the waste. The sites of highest contamination were clearly associated with dumped waste containers. The concentrations observed in sediment at the outlet of the bay were typical for the open Kara Sea ŽFig. 3.. The maximum activity levels measured in water Ž1 kBq my3 3 H, 2.5 Bq my3

I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179

171

Fig. 3. Caesium-137 in Abrosimov Bay surface sediments in 1994.

Sr, 7 Bq my3 137 Cs, 5 mBq my3 239,240 Pu. were similar to those in the open Kara Sea, implying that any present leakage from radioactive wastes is negligible. In Stepovoy Bay, 137 Cs levels of up to 110 kBq kgy1 dry wt., 60 Co levels of up to 3 kBq kgy1 dry wt., 90 Sr levels of up to 0.3 kBq kgy1 dry wt. and 239,240 Pu of up to 15 Bq kgy1 dry wt. were measured only in very localized areas around dumped containers. As in Abrosimov Bay, the leakage has not led to a measurable increase of sediment contamination in the outer part of the bay. In the inner part of the bay, however, in the close vicinity of a containered waste dumpsite, measured concentrations of 137 Cs and 90 Sr were significantly higher in the near bottom water Ž30 Bq my3 and 25 Bq my3 , respectively. than in the surface water Ž9 Bq my3 and 7 Bq my3 , respectively., where levels were those typical for the open Kara Sea. This indicates that a release was occurring at that time from the waste, leading to the observed increased concentrations in water due to the lower flushing rates of the inner Stepovoy Bay. This observation has lead to detailed investigations of the hydrodynamics and radionuclide transport and transfer in the bay 90

being undertaken at IAEA-MEL, the results of which are presented in Harms Ž1997. and Baxter et al. Ž1998.. In Tsivolki Bay the 137 Cs and 239,240 Pu levels in sediments were similar to those of the open Kara Sea. The presence of traces of 60 Co Ž1]2 Bq kgy1 dry wt.., however, suggested a local source of contamination. The observed 137 Cs and 239,240 Pu profiles in conjunction with 210 Pb dating ŽBaxter et al., 1995. indicate fast mixing of the surface sediment layer. The spent nuclear fuel of the Lenin reactor has not been localized yet and neither has any leakage associated with this waste been detected. Radionuclide concentrations in water are typical of the open Kara Sea.

3. Radionuclide inventories in the Kara Sea 3.1. The open Kara Sea Caesium-137, 90 Sr and 239,240 Pu inventories in the water column in the Kara Sea ŽFig. 4. show a good correlation with water depth Žcorrelation coefficient r 2 s 0.92, 0.82 and 0.89, respectively, excluding the encircled points in Fig. 4, which

172 I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179

Fig. 4. Caesium-137, occurred.

90

Sr and

239,240

Pu inventories in the Kara Sea water column and sediment decay-corrected to 1993. Circles single out sites where leakage has

I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179

correspond to the leaking containers site in inner Stepovoy Bay.. This confirms that significant local sources of contamination are absent in general. The points corresponding to depths shallower than 80 m in Fig. 4 are, with the exception of two shallow open sea sites, situated inside or at the mouths of the Novaya Zemlya bays investigated in 1993]1994. All other points correspond to stations in the open Kara Sea occupied during 1992]1993 cruises ŽFig. 1.. Inventories in the open Kara Sea water column range between 0.3 and 3.7 kBq my2 for 137Cs, 0.1]1.7 kBq my2 for 90 Sr and 0.2]3.5 Bq my2 for 239,240 Pu. The scatter of the values of 90 Sr inventories corresponding to the open Kara Sea, obvious in Fig. 4, can be traced back to the location of the respective sites relative to the Ob and Yenisey plumes. Caesium-137, 90 Sr and 239,240 Pu inventories in sediments in the open Kara Sea are also shown in Fig. 4. With the exclusion of dumpsites Žthose appearing in Fig. 4 are singled out., the inventories are between 0.1 and 1 kBq my2 for 137 Cs, 0.01 and 0.2 kBq my2 for 90 Sr and 8 and 25 Bq my2 for 239,240 Pu. While 90 Sr and 137 Cs have dominant inventories in the water column, 239,240 Pu, due to its more particle reactive nature, is predominantly found in sediment. From the combined water and sediment data, depth averaged 90 Sr, 137 Cs and 239,240 Pu inventories of 0.5, 1.5 and 0.02 kBq my2 can be obtained for the Kara Sea. The average 137 Cs global fallout cumulative inventory in the Kara Sea is estimated at 0.7 kBq my2 in 1993 ŽUNSCEAR, 1993; Aarkrog, 1994. which is lower than the average value weighted over the mean water depth in the Kara Sea Ž1.5 kBq my2 . of inventories calculated on the basis of measured environmental concentrations. The difference may reflect contributions from Sellafield, river input, local fallout and the Chernobyl accident. Due to proximity to the underwater test site in Guba Chernaya, the observed 137 Cs concentrations in the Pechora Sea are much higher than in the Kara Sea ŽHamilton et al., 1994; Smith et al., 1995.. At the mouth of the Yenisey river, higher 137 Cs, 238 Pu and 239,240 Pu inventories were found in sediment Ž5.3 kBq my2 , 1.6 and 45 Bq my2 , respectively. due to enhanced

173

scavenging of radionuclides from the water column. The total 90 Sr, 137 Cs and 239,240 Pu inventories in the open Kara Sea can be roughly estimated to be 0.45, 1.3 and 0.02 PBq, respectively. 3.2. No¨ aya Zemlya bays In Abrosimov Bay radionuclide inventories in sediment varied for 137Cs from 0.3 to 3000 kBq my2 at locations close to dumped containers, where leakages occurred. Similarly, 60 Co, 90 Sr and 239,240 Pu inventories up to approximately 3, 0.2 and 0.2 kBq my2 , respectively, were calculated. The results corresponding to the area confined in the vicinity of the leaking containers inside Abrosimov Bay, 2]3 orders of magnitude higher than those outside this area, were not represented in Fig. 4. In Stepovoy Bay 137 Cs inventories in sediment varied from below 1 kBq my2 in uncontaminated locations to 110 kBq my2 where leakage around dumped containers was observed. Cobalt-60 and 239,240 Pu inventories were up to 26 and 0.2 kBq my2 , respectively. In Tsivolki Bay, 137 Cs inventories in sediment ranged from 0.3 to 3 kBq my2 and for 60 Co and 239,240 Pu, up to approximately 0.3 and 0.02 kBq my2 , respectively. For a better estimation of local fallout from the atmospheric nuclear bomb explosions carried out at Novaya Zemlya test sites, a series of soil cores form the eastern coast of Novaya Zemlya was collected. Results are given here for a core collected nearshore at 718569 N in 1994. The 137 Cs profile ŽFig. 5. peaks at a depth of approximately 5 cm. The profile is very sharp, most likely in relation to limited downward transfer of 137 Cs due to permafrost, therefore the maximum measured concentration Ž300 Bq kgy1 dry wt.. is quite high. However, the total 137 Cs inventory is calculated to be 1.3 kBq my2 , a value similar to the average one for the open Kara Sea Ž1.5 kBq my2 .. As already mentioned, the expected value from global fallout is approximately 0.7 kBq my2 . The Chernobyl accident contribution could be estimated at approximately 0.2 kBq my2 and the

174

I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179

Fig. 5. Caesium-137 concentration and cumulative inventory in a soil core from the shore of Abrosimov Bay, 1994.

remainder Ž0.4 kBq my2 . may be a contribution from local fallout. The difference between the open Kara Sea value and the soil value Ž0.2 kBq my2 . may be attributed to the contribution from Sellafield and the Ob and Yenisey rivers. Obviously, this is a rough estimate, as the relative importance of all the sources mentioned above varies across the Kara Sea. To clarify the contributions from local sources the analysis of the soil cores is continued, including determinations of 240 Pur 239 Pu isotopic ratios. 4. Modelling and radiological assessment As shown, past and present radionuclide concentrations in the Kara Sea water and sediments have been and still are low, except for areas in the immediate vicinity of low-level waste containers inside the Novaya Zemlya bays. No evidence of significant leakage from the high-level waste contained in naval reactors has been found.

Therefore the issue of interest concerning this waste is its future environmental and radiological impact. Six of the 16 dumped naval reactors contain the spent nuclear fuel. These have been decommissioned prior to dumping. Approximately 60% of the fuel from one of the three reactors of the Lenin ice-breaker is also dumped in a concrete and stainless steel containment. In general the high level waste was encased in a hardening resin called furfurol, which inhibits the release of radionuclides for a period of 100]500 years ŽYablokov et al., 1993; IAEA, 1997.. A detailed study of the radionuclide composition, inventories and release rates has been undertaken within IASAP for each individual dumped object containing high-level waste ŽIAEA, 1997. on timescales up to 4500 years in the future. Three release scenarios were developed for the assessment ŽIAEA, 1999b.: 1. a best estimate release scenario, whereby ra-

I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179

dionuclides are released due to gradual corrosion of the containment and the spent nuclear fuel; 2. a plausible worst case scenario which superimposes on Scenario A a disruptive event involving the ‘Lenin’ ice-breaker waste taking place in the year 2050; and 3. a climate change scenario, identical to Scenario A up to the year 3000, followed by instantaneous release of the whole remaining inventory. Of the estimated 1994 total inventory of 4.7 PBq, 86% are fission products Žmainly 90 Sr and 137 Cs., 12% activation products and the remaining 2% actinides ŽIAEA, 1997.. Approximately 46% of the total inventory is in the ‘Lenin’ ice-breaker waste in Tsivolka Bay, releases from which are estimated to peak at around 3 TBq yeary1 for Scenario A. This is the order of magnitude estimated for maximum release rates anywhere in the Kara Sea. Local, regional and global dispersion models were developed at MEL ŽOsvath et al., 1995; Baxter et al., 1998. and were applied to estimate the environmental and radiological impact of possible releases from the waste dumped in the Kara Sea for several scenarios. Predictions were given for radionuclide concentrations in seawater, superficial bottom sediment and marine edible biota and for doses to critical groups and global collective committed doses up to 1000 years in the future. The principal pathways of radionuclides out of the Kara Sea include transport by water and ice ŽIAEA, 1998, 1999a,b.. The current knowledge of water circulation in the Kara Sea indicates that the main water outflow, of 18 000]22 000 km3 yeary1 , is between Franz Josef Land and Severnaya Zemlya, towards the Arctic Ocean. A one order of magnitude smaller outflow occurs towards the Barents Sea through the Kara Gate. An outflow of 5000]10 000 km3 yeary1 is reported towards the Laptev Sea. The uncertainties in these flows are considerable and the seasonality related to the major fresh water input Žover 1100 km3 yeary1 . by the rivers Ob and Yenisey is

175

essential to local and regional scale dispersion modelling. Export of radionuclides by ice is negligible in terms of inventory when compared to that through water. Mechanisms through which contaminated sediment trapped in ice can be exported from the Novaya Zemlya bays or the Siberian rivers have been discussed ŽStrand et al., 1996.. Recent work indicates that the Kara Sea is unlikely to be a major source of ice-entrained contaminated sediments to the Arctic Ocean ŽCooper et al., 1998., still leaving open the question of possible export towards the neighboring Barents and Laptev seas. 4.1. Local scale Modeling the dispersion of the released radionuclides within Novaya Zemlya bays has shown ŽBaxter et al., 1998. that for release rates of 1 TBq yeary1 of 137 Cs and 239 Pu from the inner part of Abrosimov Bay for instance, concentrations of approximately 1 kBq my3 and of below 0.5 kBq my3 , respectively, can be expected for the water leaving the bay. Caesium-137 and 239 Pu concentrations in bay sediments would not exceed 0.1 kBq kgy1 and 1 kBq kgy1 , respectively. An assessment of the radiological impact at bay scale indicated that for the ingestion pathway the dominant radionuclide is 137 Cs. Maximum individual doses for a hypothetical group of persons subsisting on fish from the bay calculated for a release of 1 TBq yeary1 137 Cs wwhich is of the order of magnitude of the maximum plausible release rate according to ŽIAEA, 1997.x are below 1 mSv yeary1 . Since the bays are uninhabited and only military personnel Žsubsisting on imported food. are allowed access, this exposure pathway does not present a significant problem. We also calculated exposure to military personnel patrolling the shores of the bays for several release scenarios ŽIAEA, 1999b.. For the plausible worst case scenario this critical group would receive doses up to 3 mSv yeary1 , comparable to those received from natural radiation. Practically 100% of this estimated maximum dose-rate is given by actinides in inhaled sea-spray and resuspended shore sediment.

176

I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179

4.2. Regional scale On regional scale, the modeling results obtained by Harms Ž1997. and Baxter et al. Ž1998. have shown that the dispersion of 137 Cs following an instantaneous release of the whole present-day inventory of 1 PBq from dumpsites in Abrosimov, Stepovoy and Tsivolka Bays and the Novaya Zemlya Trough results in maximum concentrations in open Kara Sea water ranging from a few kBq my3 close to the dumpsites to 50 Bq my3 in the central Kara Sea and 10 Bq my3 at the mainland coast ŽBaxter et al., 1998.. The maximum doses due to fish ingestion to individuals on the Yamal and Taymir coasts were estimated to be below 5 mSv yeary1 . For the above release points the major export of 137 Cs occurs towards the Laptev Sea, but expected concentrations in water leaving the Kara Sea do not exceed 20 Bq my3 , even for most conservative release scenarios considered. A rapid flushing of the central Kara Sea on time-scales of 5]6 years is predicted by the model ŽHarms, 1997.. The calculated increases of 137 Cs concentrations in water close to the inhabited coasts are only slightly in excess of those presently measured. Doses to fish eaters on the coasts of the Yamal and Taymir Peninsulas were estimated for a more realistic release rate of 1 TBq yeary1 of 137Cs from Abrosimov Bay to be in the order of 1 nSv yeary1 . Similar releases from the other dumpsites result in even lower doses. In the region, the Kola Peninsula Žon the mainland coast of the southern Barents Sea. is the coastal area which is nearest to the dump sites having a numerous population whose diet includes sea food harvested in the Arctic marginal seas. For this site a detailed study was carried out with the purpose of evaluating the reliability of the estimated doses to the local public. Radionuclide concentrations in water, sediment and biota in the southern Barents Sea were calculated using the compartmental model presented in Section 4.3. For the best estimate ‘normal’ release scenario ŽScenario A above., the critical exposure pathway is seafood ingestion and over 95% of the maximum dose-rate is delivered by 137 Cs. The maximum individual ingestion dose-rate was esti-

mated based on a seafood diet including fish, molluscs and crustaceans. Obviously the reliability of predictions for radionuclide concentrations and further on, for doses to humans, will depend on the choice of parameters for both the dispersion and radiological models. Model sensitivity and uncertainty analyses ŽIAEA, 1999b. and a probabilistic assessment of individual doses were carried out. The probabilistic approach to dose calculation allows a better quantification of uncertainties in the calculated doses related to the uncertainties in the model parameters. A number of other sources of uncertainty also affect the final dose estimates, such as the specification of the scenario, the formulation of the conceptual and numerical models and the interpretation of results, but these cannot be readily quantified ŽBIOMOVS II, 1993.. Uncertainty analysis involves analytical methods ŽCox and Baybutt, 1981; Worley 1987. or statistical methods based on random sampling ŽIAEA, 1989.. The latter approach, using Monte Carlo techniques and Latin hypercube sampling ŽLHS., was adopted for this study. The annual ingestion dose Ž D i . to an individual was calculated using the equation below: D i s C w ? Ž CFf ? A f q CFm ? A m qCFc ? A c . ? DCFi

Sv yeary1

Ž1.

where C w is maximum concentration of 137 Cs Žin Bq my3 . in water offshore the Kola Peninsula, as predicted by the compartmental model for Scenario A; CFf , CFm , CFc are the concentration factors for fish, mollusc and crustacean edible parts: 0.1, 0.03 and 0.03 my3 kgy1 , respectively ŽIAEA, 1985.; A f , A m , A c are the annual intakes of fish, molluscs and crustaceans: 50 kg, 0.5 kg and 1 kg, respectively ŽIAEA, 1999b.; and DCFi is the dose conversion factor for ingestion for adults: 1.3= 10y8 Sv Bqy1 ŽIAEA, 1996.. Probability density functions were constructed for C w CFf , CFm , CFc , A f , A m and A c . It was assumed that these model parameters are distributed log-normally for C w CFf , CFm and CFc and, respectively, normally for A f , A m and A c around the mean values specified above. The

I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179

distribution type, standard deviation, minimum and maximum values for the model parameters were chosen based on a review of published data and expert judgement. No correlations between the input parameters were taken into account. One thousand simulations were run and the mean, standard deviation and Chebyshev confidence bounds were estimated for the output distribution ŽFig. 6.. As expected, the deterministic and probabilistic mean values of the estimated doserate are practically equal. The magnitude of the resulting dose-rate is negligible. It should be noted however that dose-rate values over an order of magnitude higher than the deterministic estimate could be expected to occur, certainly with a very low probability. The standard deviation and Chebyshev confidence bounds which define the uncertainty in the estimated dose-rate are given in Fig. 6. In any case, it is obvious that, depending on the release scenario and population groups considered, the estimated annual individual doses are three to six orders of magnitude lower than the worldwide average dose rate due to natural sources Ž2.4 mSv yeary1 .. 4.3. Global scale For simulating dispersion of radionuclides from the nuclear waste dumpsites in the Kara Sea on the global scale a compartmental model was developed ŽBaxter et al., 1998.. This type of model requires modest computer power to run and is therefore well suited to long-range and long time-scale simulations, moreover allowing straightforward coupling of sediment interaction and radiological dose calculation models. At present the ARCTIC-5 model comprises 48 compartments of water and sediment. It represents in more detail the Arctic seas, in the vicinity of the simulated sources of release, and the UK coastal system, of particular interest for testing the model by using the Sellafield signal. The validation of the model has been achieved using the documented progression of the Sellafield 137 Cs releases through the northern seas and estimated residence and transit times for the Arctic Seas ŽSchlosser et al., 1995..

177

Fig. 6. Distribution function of maximum individual dose from modelled releases of 137 Cs from nuclear waste dumped in the Kara Sea, calculated for Critical Group on Kola Peninsula.

A first worst case assessment was obtained by assuming the entire inventory in the Kara Sea reactors ŽIAEA, 1997. was instantaneously released. The collective committed effective dose ŽCCED. to the world population integrated over 500 years after release was estimated for the marine food ingestion pathway. A global CCED of approximately 2 man Sv was obtained for the world population. Approximately 84% of this dose is delivered by fission products Ž80% of 137 Cs and 4% of 90 Sr., 2% by actinides Žmainly 239,240 Pu. and 14% by activation products Žmainly 14 C, 63 Ni and 60 Co.. The dominant exposure pathways are through ingestion of fish Žmore than 85% of the CCED. and molluscs Žapprox. 8%.. For the NE Atlantic fishing area the corresponding CCED is 0.7 man Sv.

5. Conclusions A comprehensive programme of Arctic radioactivity studies dedicated to the assessment of consequences of radioactive waste dumping in the Kara Sea was carried out at IAEA’s Marine Environment Laboratory in the period 1992]1998. It involved field, analytical, experimental and modeling work. The results show that at present no

178

I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179

contamination is detectable outside the dumpsites and that the present and future radiological impact on human populations cannot be expected to be significant. Maximum individual dose-rates estimated for the region are at least three orders of magnitude lower than the worldwide average received from natural sources Ž2.4 mSv yeary1 .. Evaluation of radionuclide inventories has shown that radionuclide contamination of the open Kara Sea is mainly due to global fallout with smaller contributions from local fallout, the Sellafield reprocessing plant, the Chernobyl accident, and run-off from the Ob and Yenisey rivers.

Acknowledgements The IAEA-MEL operates under an agreement between the International Atomic Energy Agency and the Government of the Principality of Monaco. Many colleagues from IAEA-MEL’s Radiometric Section contributed to the laboratory’s Arctic programme: L. Huynh-Ngoc, T. Hamilton, J. Carroll, H. Pettersson, S. Ballestra, J. Gastaud, J. Gayol, L. Liong Wee Kwong, P. Parsi. Outside collaborators included German, Norwegian, Russian and US scientists, and we are particularly grateful to these countries’ governments for supporting IAEA-MEL’s participation in Kara Sea investigations. References Aarkrog A. Worldwide data on fluxes of 239,240 Pu and 238 Pu to the oceans. In: Inventories of selected radionuclides in the oceans. Vienna: IAEA-TECDOC-481, 1988. Aarkrog A. Radioactivity in Polar regions } main sources. J Environ Radioact 1994;25:21]35. Baxter MS, Ballestra S, Gastaud J, Hamilton TF, Harms I, Huynh-Ngoc L, Liong Wee Kwong L, Parsi P, Pettersson HBL, Povinec PP, Sanchez A. Marine radioactivity studies in the vicinity of sites with potential radionuclide releases. In: Proc Int Symp, Vienna 1995:125]141. Baxter MS, Harms I, Osvath I, Povinec PP, Scott M. Modelling the potential radiological consequences of radioactive waste dumping in the Kara Sea. J Environ Radioact 1998;39:161]181. BIOMOVS II. Technical Report No.1: guidelines for uncertainty analysis. Swedish Radiation Protection Institute, Stockholm, 1993. Boisson F, Hutchins DA, Fowler SW, Fisher NS, Teyssie J-L.

Influence of temperature on the accumulation and retention of eleven radionuclides by the marine alga Fucus ¨ esiculosus ŽL... Mar Pollut Bull 1997;35:313]321. Carroll J, Boisson F, Fowler SW, Teyssie J-L. Radionuclide absorption to sediments from nuclear waste dumping sites in the Kara Sea. Mar Pollut Bull 1997;35:296]304. Cooper LW, Larsen IL, Beasley TM, Dolvin SS, Grebmeier JM, Kelley JM, Scott M, Johnson-Pyrtle A. The distribution of radiocesium and plutonium in sea ice-entrained Arctic sediments in relation to potential sources and sinks. J Environ Radioact 1998;39Ž3.:279]303. Cox DC, Baybutt P. Methods for uncertainty analysis: a comparative survey. Risk Anal 1981;1Ž4.:251]258. Hamilton TH, Ballestra S, Baxter MS, Gastaud J, Osvath I, Parsi P, Povinec PP, Scott EM. Radiometric investigations of Kara Sea sediments and preliminary radiological assessment related to dumping of radioactive wastes in the Arctic Seas. J Environ Radioact 1994;25:113]134. Harms I. Modeling the dispersion of 137 Cs and 239 Pu released from dumped waste in the Kara Sea. J Mar Syst 1997; 13:1]19. IAEA. Sediment K d s and concentration factors for radionuclides in the marine environment. Technical Report Series No. 247. Vienna: IAEA, 1985. IAEA. Evaluating the reliability of predictions made using environmental transfer models. Safety Series No. 100. Vienna: IAEA, 1989. IAEA. International basic safety standards for protection against ionizing radiation and for the safety of radiation sources. Safety Series No. 115. Vienna: IAEA, 1996. IAEA. Predicted radionuclide releases from marine reactors dumped in the Kara Sea. IAEA-TECDOC-938, Vienna: IAEA, 1997. IAEA. Radiological conditions of the western Kara Sea. Vienna: IAEA, 1998. IAEA. Anthropogenic radionuclides in the Arctic Seas. Vienna: IAEA TECDOC, 1999:in print. IAEA. Report of International Arctic Seas assessment project modeling working group. IAEA TECDOC 1999:in print. Joint Russian]Norwegian Expert Group. A survey of artificial radionuclides in the Kara Sea. Results from the Russian]Norwegian 1992 expedition to the Barents and Kara Seas. Osteras: NRPA, 1993. Joint Russian]Norwegian Expert Group. Radioactive contamination at dumping sites for nuclear waste in the Kara Sea. Results from the Russian]Norwegian 1993 expedition to the Kara Sea. Osteras: NRPA, 1994. Joint Russian]Norwegian Expert Group. Dumping of radioactive waste and investigation of radioactive contamination in the Kara Sea. Results from 3 years of investigations Ž1992]1994. in the Kara Sea. Osteras: NRPA, 1996. NOO. Arctic environmental isotope study, Informal report No. 69-4. Washington, D.C.: NOO, 1969. Osvath I, Ballestra S, Baxter MS, Gastaud J, Hamilton TF, Harms I, Liong Wee Kwong L, Parsi P, Povinec PP. IAEAMEL’s contribution to the investigation of the Kara Sea dumping sites. In: Proc Int Conf Oslo, 1995:139]144.

I. Os¨ ath et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 167]179 Pavlov VK, Pfirman SL. Hydrographic structure and variability of the Kara Sea: implications for pollutant distribution. Deep-Sea Res II 1995;42Ž6.:1369]1390. Pavlov VK, Kulakov MYu, Stanovoy VV. Hydrometeorological regime of the Kara and Barents Seas. Report to the International Arctic Seas Assessment Programme. Vienna: IAEA, 1993. Povinec PP, Osvath I, Baxter MS. Underwater gamma-spectrometry with HPGe and NaI ŽTl. detectors. Appl Radiat Isot 1996;47:1127]1133. Povinec PP, Osvath I, Baxter MS, Ballestra S, Carroll J, Gastaud J, Harms I, Huynh-Ngoc L, Liong Wee Kwong L, Pettersson H. Summary of IAEA-MEL’s investigation of Kara Sea radioactivity and radiological assessment. Mar Pollut Bull 1997a;35:235]241. Povinec PP, Baxter MS, Gayol J, Huynh-Ngoc L, Liong Wee Kwong L, Osvath I. Inventories and distributions of radionuclides in the Kara Sea. In: Proc Int Symp Radionuclides in the Ocean. Cherbourg: IPSN, 1997:C2.29]C2.34. Schlosser P, Swift JH, Lewis D, Pfirman S. The role of the large-scale Arctic Ocean circulation in the transport of contaminants. Deep Sea Res 1995;42Ž6.:1341]1367. Smith JN, Ellis KM, Forman S, Polyak L, Ivanov G, Matishov

179

D, Kilius L, Dahle S. Radionculide sources in the Barents and Kara Seas. In: Proc Int Conf Oslo 1995:179]185. Strand P, Nikitin A, Rudjord AL, Salbu B, Christensen G, Foyn L, Kryshev II, Chumichev VB, Dahlgaard H, Holm E. Survey of artificial radionuclides in the Barents Sea and the Kara Sea. J Environ Radioact 1994;25:99]112. Strand P, Sickel M, Aarkrog A, Bewers JM, Tsaturov Y, Magnusson S. Radioactive contamination of the Arctic marine environment. In: Guegueniat P, Germain P, Metivier H, editors. Radionuclides in the oceans. Inputs and inventories. Cherbourg: IPSN, 1996:95]119. UNSCEAR. Sources and effects of ionizing radiation. New York: UNO, 1993. Vakulovsky SM. Radioactive contamination of the Arctic Seas. TYPHOON, Obninsk, Russian Federation, 1993. Worley BA. Deterministic uncertainty analysis. Oak Ridge National Laboratory Report ORNL-6428. Oak Ridge, TN, 1987. Yablokov A, Karasev V, Rumyantsev V, Kokeev M, Petrov O, Lystov V, Yemelyanenkov A. Facts and problems related to radioactive waste disposal in seas adjacent to the territory of the Russian Federation. Office of the President of the Russian Federation, Moscow, 1993.