The accidental sinking of the nuclear submarine, the Kursk:monitoring of radioactivity and the preliminary assessment

of the potential impact of radioactive releases

I. Amundsen, M. Iosjpe *, O. Reistad, B. Lind, K. Gussgaard, P. Strand,S. Borghuis, M. Sickel, M. Dowdall

Norwegian Radiation Protection Authority, P.O. Box 55, Grini naeringspark 13, 1332 Oesteraas, Norway

Abstract

Measurements of samples taken from the close vicinity of the Kursk during two expeditions to the site in August and October

2000, indicate that no leakage of radionuclides from the reactors has been observed. Only background levels in the range

0:0–0:1 lSv=h have been measured by use of the remote operating vehicle (ROV) or by the divers working on and inside thesubmarine. Preliminary model calculations based on two different scenarios, representing short- and long-term releases of 100% of

the reactors radionuclide inventory, show that the impact on man and the environment from the Kursk should not be deemed very

serious. The conservative estimates indicate a maximum 137Cs activity concentration in fish in the order of about 80–100 Bq/kg and

a total collective dose of 97 manSv. � 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Radioactive contamination; Kursk; Impact assessment; Inventory; Source term

1. Introduction

On the morning of the 12th of August 2000, anaccident involving a Russian nuclear submarine, theKursk, occurred in international waters east of the Ry-batschi Peninsula in the Barents Sea, approximately 250km off the coast of Norway. The accident occurredduring a large Russian military exercise in which theKursk was taking part. The submarine, a Russian Oscarclass II attack submarine, sank to a depth of 116 m atthe position 69� 36.99 N, 37� 34.50 E (Fig. 1). The Kurskis 154 m long with a submerged displacement of 24 000ton. It is equipped with two pressurized water reactorseach having a thermal effect of 190 MW, or less than10% of that of a typical nuclear power plant reactor. Thetotal amount of radionuclides contained in the reactorsand technical details of the reactor’s construction arenot well known. All reactor and fuel configuration dataconcerning Russian naval, military vessels are classifieddue to military restrictions.

The Kursk is of double hull construction with ninewatertight compartments. The outer hydrodynamic hullis made of 8 mm steel plates covered with up to 80 mmof rubber, the purpose of which is to prevent detectionby other submarines or surface vessels via the elimina-tion of sonar echoes. The inner pressure hull is made of50 mm steel plates, and the distance between the twohulls varies between approximately 1 and 2 m.Additional information is available from Amundsen

et al. (2001).

1.1. Expeditions to Kursk

A rescue operation took place during the period of17–22 August 2000 with the vessel ‘‘Seaway Eagle’’,owned by the Norwegian company Stolt Offshore. Themain purpose of the expedition was to open the rescuehatch in compartment no. IX in an attempt to rescuemembers of the crew that may still have been alive. Thiswork was performed by specialist Norwegian and Brit-ish deep-water divers. The Russian participation in theoperation was administered by the Northern Fleet of theRussian Navy. During the night of the 20th and 21st ofAugust, the divers managed to open the rescue hatch.

*Corresponding author. Tel.: +47-67-16-25-39; fax: +47-67-147407.

E-mail addresses: ingar.amundsen@nrpa.no (I. Amundsen),

mikhail.iosjpe@nrpa.no (M. Iosjpe).

0025-326X/02/$ - see front matter � 2002 Elsevier Science Ltd. All rights reserved.

PII: S0025-326X(01 )00258-2

www.elsevier.com/locate/marpolbul

Marine Pollution Bulletin 44 (2002) 459–468

Some air was expelled upon opening the hatch butcompartment IX proved to be flooded with water.After the expedition with the Seaway Eagle, the

Russian officials commenced planning a second expe-dition to gain access for the divers to the interior of theKursk. The main purpose of this operation was to re-cover the bodies of the casualties. This expedition wouldprovide opportunities for the location of documents andinstrumentation of special interest and which may pro-vide insight into the reason for the catastrophe. Theexpedition took place during the period from the 20th ofOctober to the 7th of November 2000. The Americanfirm ‘‘Halliburton’’ and their vessel, the MSV ‘‘Regalia’’,took part in the operations conducted during the courseof this expedition. A total of 12 casualties were removedfrom the submarine (from compartment no. IX) andwere brought onboard the Regalia. Debris from theseabed and documents from the control room were alsobrought up. During both expeditions, the NRPA wasasked to assist the operators in relation to radiationsafety for the divers and the general crew and to performenvironmental monitoring (Amundsen et al., 2001).Onboard both the Seaway Eagle and the Regalia, a

mobile radiation-monitoring laboratory was set up. Thelaboratory was equipped with two different types ofdose-rate meters and two types of gamma spectrometricdevices, a high resolution (2.0 keV for 137Cs) germanium

detector (HPGe) and a sodium iodide detector (NaI)with a lower resolution (58 keV for 137Cs) but higherefficiency. Two types of NaI equipment were used: a2 in:� 2 in: detector with an EasySpec multi-channelanalyser and a 3 in:� 3 in: detector with a Canberraseries 10 multi-channel analyser.

2. Sampling and monitoring

A variety of sampling and monitoring activities wereperformed during the two expeditions which can becategorised as follows: (i) dose-rate measurements, (ii)sediment sampling, (iii) water sampling and (iv) air andair filter measurements.During both expeditions, an underwater remote op-

erating vehicle (ROV) performed an initial surveyaround the Kursk with a dose-rate meter mounted in awaterproof and pressure resistant transparent box. Thepurpose of this survey was to monitor the radiationlevels at any time to ensure that the working conditionswere safe for the divers. During the first expedition, theROV was only allowed to survey the environs of thesubmarines’ stern, at the position of the reactor com-partment (compartment no. VI) and backwards fromthis point. During the October recovery expedition, theROV travelled all around and on top of the submarine.

Fig. 1. Location of the Kursk submarine.

460 I. Amundsen et al. / Marine Pollution Bulletin 44 (2002) 459–468

The distance from the dose-rate meter to the submarinehull was estimated to be 0.5–1 m.During the August expedition, a dose-rate reading

was performed after the opening of the rescue hatch,while during the recovery operation in October, dose-rate readings were performed by the divers each time apiece of the outer or inner hull was cut out to gain accessto the interior of the Kursk. When pieces of the innerhull (the pressure hull) were removed, measurementswere performed at the edge of the resulting hole and alsoinside the submarine by mounting the dose-rate meterand a camera on a pole. Measurements were performedon pieces of debris from the Kursk, which were broughtonto the main deck of the Regalia (Fig. 2). Such piecesconsisted of debris from the seabed, pieces of the hull,pipe-work from between the two hulls and pieces fromthe interior of the Kursk.All dose-rate measurements taken by the ROVs,

during both expeditions, indicated normal backgroundlevels in the range 0:0–0:1 lSv=h. The readings did notprovide any indication of leakage from the submarine.Readings taken outside the reactor compartment andclose to visible cracks in the outer hull of the submarinedid not show any sign of increased levels. Due to theshielding provided by the hull and the distance at whichthe readings were taken, it is estimated that the waterinside the reactor section would have to exceed an ac-tivity concentration of about 37 kBq/l (1 lCi=l) before itwould be possible to detect enhanced dose-rate levelsoutside the hull. Dose-rate measurements conducted bythe divers, and also measurements of debris brought ontothe main deck, showed no elevated levels of radiation.A total of 17 sediment samples were taken, three on

the first expedition and 14 on the second (Fig. 3). On thefirst expedition, in August 2000, Russian officials per-mitted sediment sampling to be conducted only next tothe rear part of the Kursk, while in October sampleswere taken from all around the submarine. Twelvesamples were taken, six on a straight line on either sideof the Kursk. Sample separation distance was approxi-

mately 30 m, the distance to the submarine hull beingabout 8 m. A special steel corer device was made on theRegalia, which made it possible for the ROV to pick upeach corer from a rack of six. The diameter of the steelcorer was 70 mm with a length of 400 mm.Several water samples were taken on both expeditions

to the Kursk. These included samples of surface water,near-surface water (16 m deep), bottom water and sam-ples from inside the submarine. Surface and near-surfacesampling were performed utilising the water intakes ofthe ships. A Nansen water sampler device (5 l) was usedfor bottom water sampling. Water from inside the sub-marine was taken from the bottom of the rescue hatchand from inside compartments III, IV and VIII.Selected samples of water and sediments from the

Kursk have been analysed by gamma spectrometry andalso analysed for plutonium isotopes. The results ofthese measurements did not indicate the presence ofradionuclides that may have leaked from the submarineand did not indicate activity levels above normal. Table1 shows the concentrations of radionuclides in sedi-ments, seawater and air filters. Concentrations in therange 0.7–1.5 Bq/kg of 137Cs were found in the sedimentsamples after measurements by HPGe detectors at thelaboratories of the NRPA. These levels are ‘‘normallevels’’ in the Barents Sea and are attributed to releasesfrom the Sellafield reprocessing facility in UK, falloutfrom the Chernobyl accident and from the bomb tests inthe 1950s and 1960s (AMAP, 1998; Grøttheim, 2000).Activity levels of 131I, 134Cs or 60Co were not detectedin any of the samples.Three samples of air from inside the submarine were

collected. A sample was obtained when the rescue hatchwas opened in August, while two air samples were ob-tained from inside compartments no. VIII and IV inOctober. An air sampling device, drawing 140 m3=hthrough a Whatman GF/A glass fibre filter (diameter22 cm), was used on both expeditions. The filter wasanalysed by use of the HPGe detector. Again, elevatedlevels of radioactivity were not detected.

Fig. 2. A dose-rate measurement of a piece of the inner hull (compartment IV) from the Kursk (a) being performed on the deck of the Regalia. A NaI

spectrum from a piece of the inner hull of the submarine (compartment VIII) is obtained (b).

I. Amundsen et al. / Marine Pollution Bulletin 44 (2002) 459–468 461

Six sediment samples from near the front part of thesubmarine have been measured for 238Pu and 239;240Puactivity. Activity concentrations in the range 0.006–0.015 and 0.03–0.07 Bq/kg were detected for 238Pu and239;240Pu, respectively. These concentrations are typicalfor the Barents Sea. A 238Pu to 239;240Pu ratio in therange 0.03–0.07 indicates that the plutonium originatesmainly from the global fallout, having a reported ratioof about 0.04 (UNSCEAR, 1982).

3. Assessment of the integrity of the reactor section of the

Kursk

The integrity of the reactor section of the Kursksubmarine and subsequent barriers against releases ofradioactivity have been considered by the joint Norwe-gian–Russian working group as part of the impact as-sessment for the Kursk accident (Reistad, 2001). Someof the information obtained from the Russian partici-pants is briefly described below.

• No additional amounts of radionuclides have beendetected outside the reactor compartment.

• The reactors have been shut down by the emergencyshut down system as a result of the explosions.

• All relevant safety functions were initiated. Fuel fail-ures as a result of the shut down may not be excluded.

• Natural water circulation should be adequate to dis-sipate the decay heat. Russian nuclear safety author-

ities do not consider criticality to be of significantrelevance when considering accident scenarios.

• The most significant risk contributions in the sal-vage operation are considered to be potentialhuman errors in the operation of removing the fuelfrom the reactors when the Kursk has been takento harbour.

4. Radionuclide inventory for the Kursk submarine

Due to the inherent secrecy surrounding the con-struction and operational data of Russian submarinereactors, NRPA has calculated the inventory for theKursk submarine on the basis of a computer reactormodel of the Kursk reactors. A set of assumed opera-tional, construction and fuel data for the submarinehave been used. The modeling of the reactor is based onthe computer tool HELIOS (Stamm‘ler et al., 1996)developed and supported by Studsvik Scandpower.HELIOS has been extensively qualified by comparisonswith experimental data and international benchmarkproblems for reactor physics codes as well as throughfeedback from applications concerning fuel enrichmentsup to 90% and Russian naval reactors (Stamm‘ler et al.,1996; SSP, 1998).There has been some speculation as to whether or

not the Kursk was carrying nuclear warheads. Russianauthorities have stated that the submarine was equippedwith conventional warheads at the time of the accident.

Fig. 3. Sediment sampling locations in August and October 2000 is indicated. The activity levels of 137Cs (Bq/kg) are also shown, indicating only

normal background levels.

462 I. Amundsen et al. / Marine Pollution Bulletin 44 (2002) 459–468

Therefore, this question is not discussed further in thispaper. It should be noted that the reactors would posea greater environmental impact than the nuclear war-heads, due to a higher radionuclide inventory. This isbased on an assessment of the environmental impact ofthe sinking of the Komsomolets in 1989 with two ura-nium/plutonium warheads onboard (AMAP, 1998).

4.1. Technical data

The basic source for the reactor model used in thecalculations below is the Safety Report for the Russianicebreaker, the Sevmorput. This report, submitted bythe Russian authorities before the icebreaker visited theNorwegian city of Tromsø in 1988, contains all thenecessary data to establish a usable reactor model forthe HELIOS reactor calculation code (Sevmorput,1988). This gives a flexible model based on a Russiannaval reactor while maintaining the option to include allother data available on these reactors. Based on earlier

efforts to model the fuel behaviour in Russian navalreactors, a reactor model with the hexagonal shape andthe Sevmorput fuel assembly geometry, as sketched inFig. 4, was chosen as the basis for the present work. Thereactor and fuel data, which are included in the model,are summarized in Table 2.With the former Norwegian nuclear regulatory au-

thority as initiator, and official Russian data for selectedfission products (caesium, strontium) and plutonium,Scandpower AS performed calculations of probableconfiguration and fuel enrichment in the sunken Kom-somolets submarine (SSP, 1991). One of the results wasthat the Russian data was consistent with a possible fuelenrichment of 30%. This is also consistent with otherstudies referring to the fuel enrichment in third genera-tion Russian submarines to vary in the range 21–45%(Chenyal and Hippel, personal communication, 2000).Few independent sources exist. However, the amountof 235U used in the model, and the figure used in thispaper (150.7 kg) is consistent with the content in the

Table 1

Activity concentrations in environmental samples taken from the close vicinity of the Kursk

Sampling location Period of measuring,

date in 2000

Concentrations in air (Bq=m3)

I-131 Cs-137 Cs-134 Co-60

Air filters

Seaway Eagle; SE

Regalia; REG

20.08–21.08 < 0:0109� 10�3 < 0:0109� 10�3 < 0:0109� 10�3 < 0:0109� 10�3

Sample no. Sampling date Concentrations in sediments (Bq/kg) d.w.

I-131 Cs-137 Cs-134 Pu-238 Pu-239,240

Sediments

Sed-1SE 20.08 <0.7 0.7� 38% <0.6 n.a. n.a.

Sed-2SE 20.08 <0.6 0.7� 25% <0.6 n.a. n.a.

Sed-3SE 22.08 <0.3 0.7� 11% <0.3 n.a. n.a.

Sed-1REG 20.10 <0.7 1.3� 10% <0.6 0.006� 67% 0.04� 61%Sed-2REG 20.10 <0.7 1.0� 10% <0.6 0.013� 38% 0.04� 40%Sed-3REG 20.10 <0.7 1.2� 12% <0.6 0.015� 47% 0.07� 42%Sed-4REG 20.10 <0.7 1.0� 20% <0.6 n.a. n.a.

Sed-5REG 20.10 <0.7 1.2� 8% <0.6 n.a. n.a.

Sed-6REG 20.10 <0.7 0.9� 11% <0.6 n.a. n.a.

Sed-7REG 20.10 <0.7 1.2� 8% <0.6 n.a. n.a.

Sed-8REG 20.10 <0.7 1.2� 9% <0.6 n.a. n.a.

Sed-9REG 20.10 <0.7 0.7� 11% <0.6 n.a. n.a.

Sed-10REG 20.10 <0.7 1.5� 7% <0.6 0.014� 50% 0.03� 52%Sed-11REG 20.10 <0.7 1.4� 11% <0.6 0.015� 40% 0.04� 36%Sed-12REG 20.10 <0.7 0.9� 17% <0.6 0.008� 63% 0.03� 61%Sed-13REG 07.11 <0.7 1.2� 9% <0.6 n.a. n.a.

Sample no. Sampling date Concentrations in water (Bq/l)

I-131 Cs-137 Cs-134 Co-60

Water samples

Seawater-1-5SE 20.08 <0.5 <0.5 <0.5 <0.5

Seawater 4 REG 27.10 3:4� 10�3 � 4%Concentrations in water (Bq=m3)

Pu-238 Pu-239,240

Seawater 1A+5 SE 20–22.08 0.0004 0.0034� 0.0001Seawater 5 REG 28.10 <0.0005 0.0050� 0.0001

I. Amundsen et al. / Marine Pollution Bulletin 44 (2002) 459–468 463

icebreaker, the Sevmorput. The basic fuel geometry hasbeen taken from the Sevmorput Safety Report.

4.2. Operational data

With respect to fuel burn-up, two different versions ofthe operational history have been considered (Tables 3

and 4). Both versions are based on the submarine beingoperative for an average of 50 days per year, for eachyear since commissioning at the end of 1994. Howeverversion 2 includes extensive operation of the reactors forproduction of electrical power when in port as reportedin several sources during recent years.

4.3. Source term – rate of release from the reactors

According to the Russian authorities, the reactorswere shut down with the emergency shut down mecha-nism as a result of the explosion. This mechanism isactivated when the hull is exposed to heavy vibrations.The activity dynamic of radionuclides in the subma-

rine reactors, having an activity more than 1 TBq after100 years from accident time is shown in Fig. 5.At the moment, it is not possible to calculate radio-

activity release from the submarine on the basis of athorough safety analysis report due to the consistentsecrecy surrounding these submarine reactors. However,if the Kursk remains on the seabed, fission productsand activation products from the reactor fuel, as well asactivation products from the reactor pressure vessel andother parts of the reactor will eventually be released tothe sea. The hypothetical release rate of radionuclidesdepends heavily on the release conditions. These con-ditions may range from instantaneous release as a resultof explosions of torpedoes or cruise missiles left in thesubmarine, to the slow long-term corrosion of the fuelmaterial. The latter may occur when seawater has pen-etrated the fuel cladding. If the cladding is zirconium,penetration may take several hundred years or more.However, if conditions for galvanic corrosion are pre-sent, the cladding may perhaps be corroded through inless than a year.The following two scenarios for releases of radio-

nuclides have been selected for the wide range of pos-sibilities in the present situation:Scenario 1, corresponding to an abnormal event one

year after the accident, i.e. during the salvage operation,100% of inventory in both reactors is released instan-taneously.Scenario 2, corresponding to the assumption that all

barriers, for all practical purposes, have been removedafter 100 years, and 100% of the inventory in bothreactors are released at this point.It should be noted that an estimated release of 100%

of the inventory is considered to be a very conservativeapproach.

5. Assessment of potential impact of radioactive releases

Estimations of the radiological consequences afterpotential releases of radionuclides from the Kursk sub-marine have been performed. The elements for the

Fig. 4. Schematic diagram of one-sixth of a Kursk reactor assembly as

used in the reactor model. U–Al alloy, 30% enriched, 150.7 kg 235U,

241 assemblies, 6 Gd pins per assembly.

Table 2

General reactor core and fuel assembly dimension data as a basis for

inventory calculations for the Kursk

Used in model of Kursk

Generation Third

Max thermal power (MWt) 200 MW

U-235 (kg) Basic: 150.7 kga

Range: 75–200 kg

Enrichment Basic: 30%

Range: 20–90%

#Fuel assemblies 241a

Fuel composition (1) U–Al alloy foil cladded in Zr

tubes

(2) U–Al alloy dispersed in a matrix

cladded in Zr tubes

Fuel geometry Circular pins in hexagonal latticea

U-235 pr. fuel assembly (kg) Basic: 0.625 kg

Core diameter 121.2 cma

Assembly Outer diameter 6 cma

Outer clad Thickness: 0.06 cm

Material: Zra

Inner clad Thickness: 0.06–0.006 cm

Material: Zra

Number of pins/assembly 55a

Active core height 100 cma

Coolant flow area 0.26 m2a

Reactor burn 12 000/24 000 MWd

aData taken directly from the Sevmorput safety report (Sevmorput,

1988).

464 I. Amundsen et al. / Marine Pollution Bulletin 44 (2002) 459–468

modelling work are the two given scenarios, and a boxmodel, which estimates radionuclide transport overlarge distances (> 1000 km) and over long time-scales(up to centuries or millennia). The box modelling hasbeen recommended by the European Commission forradiological assessment and has been used for manyprevious investigations (see, for example, Nielsen et al.,1997; EC, 1997, 1999; IAEA, 1999).The present model is based on the box modelling

approach, which includes terms that describe the dis-persion of radionuclides into oceanic space with time(Iosjpe et al., 1997, 2001; Iosjpe and Strand, 1998). Thepresent model is a revised version of the box modelcovering the European coastal waters, Arctic and theNorth Atlantic Oceans, which was described by Nielsenet al. (1995, 1997).

6. Model description

The equations describing the transfer of radionuclidesbetween the boxes are of the form:

dAi

dt¼

Xn

j¼1kjiAj �

Xn

j¼1kijAic ðtP TjÞ � kiAi þ Qi; tP Ti;

Ai ¼ 0; t < Ti; ð1Þ

where kii ¼ 0 for all i, Ai and Aj are activities (Bq) at timet in boxes i and j; kij and kji are rates of transfer (y�1)between boxes i and j; ki is an effective rate of transfer ofactivity (y�1) from box i taking into account loss ofmaterial from the compartment without transfer to an-other, for example radioactive decay; Qi is a source ofinput into box i (Bq/y); n is the number of boxes in thesystem. Ti is the time of availability for box i (the firsttime when box i is open for dispersion of radionuclides)and c is a unit function:

cðtP TiÞ ¼1; tP Ti0; t < Ti:

�

The time of availability Ti

Ti ¼ minlmðv0;viÞ2Mi

Xj;k

wjk

are calculated as a minimized sum of the weights for allpaths lmðv0; . . . ; viÞ from the initial box (v0) with dis-charge of radionuclides to the box i on the orientedgraph G ¼ ðV ;EÞ with a set V of nodes, vj, correspond-ing to boxes and a set, E, of arcs, ejk, corresponding tothe transfer possibility between the boxes, j and k.Every arc ejk has a weight wjk which is defined as the

Table 4

Operational data

Year

1995 1996 1997 1998 1999 2000

Power MW 40 0 0 40 40 0 15 15 0 40 40 0 15 15 40

Days 50 315 315 50 50 115 200 200 115 50 70 75 220 180 30PMWd 2000 4000 9000 14000 20100 24000

The estimated total reactor burn is 24 000 MWd. The basic time of operation is 50 days per year, but it includes an extensive operation of reactors in

port to produce electric power (Case 2).

Table 3

Operational data

Year

1995 1996 1997 1998 1999 2000

Power MW 40 0 40 0 40 0 40 0 40 0 0 40

Days 50 315 50 315 50 315 50 315 50 295 180 30PMWd 2000 4000 6000 8000 10 800 12 000

The estimated total reactor burn is 12 000 MWd with a basic time of operation of 50 days per year (Case 1).

Fig. 5. The estimated amount of radionuclides (Bq) in one of the two

submarine reactors in Kursk as a function of time (years) after shut-

down. All radionuclides in the reactor having an activity of more than

1 TBq after 100 years are shown.

I. Amundsen et al. / Marine Pollution Bulletin 44 (2002) 459–468 465

time required before the transfer of radionuclides frombox j to box k can begin (without any path throughother boxes).The contamination of fish, crustaceans and molluscs

is further calculated from the radionuclide concentra-tions in filtered seawater in the different water regions.Doses to man are calculated on the basis of data for theamount of seafood caught and statistics pertaining tohuman diet.

7. Results of calculation

The model calculation of transport, transfer to fishand collective doses to humans following the givenscenarios is performed for a range of different radio-nuclides, which are present in the reactors. However,most attention is focused on 137Cs release. This is due tothe fact that 137Cs has a relatively long physical half-life

(30 years), is readily dissolved in the water phase andaccumulates readily in edible parts of fish and shellfish.Dispersion of 137Cs in the oceanic surface water,

corresponding to the worst case of the potential acci-dental releases (version 2 of Scenario 1) from the sub-marine, the Kursk, is shown in Fig. 6. The boxescorrespond to seafood catchment areas. Calculationsshow that 0.5 years after a hypothetical accidental re-lease of 100% of the inventory, the average water con-centration in the Barents Sea will be in the range160–210 Bq=m3 for areas in close vicinity to the sub-marine. 137Cs activities will decrease rapidly, and after10 years it is estimated that the average water concen-tration in the Barents Sea will be in the range 0:1–2:8 Bq=m3. Notice that 137Cs distribution presented inFig. 6 is in reasonable agreement with concentration of apassive, conservative tracer after release of radioactivityfrom the Kursk submarine evaluated by Gerdes et al.(2001).

Fig. 6. Dispersion of 137Cs (Bq=m3) after a potential release from the submarine, the Kursk into the Barents Sea.

466 I. Amundsen et al. / Marine Pollution Bulletin 44 (2002) 459–468

The dynamics for the 137Cs concentration in fish forthe Barents Sea region (corresponding to the same sce-nario) is shown in Fig. 7. The maximum, minimum andaverage activity concentrations in fish correspond toareas at different distances from the Kursk. The plotsdisplayed in the figure indicate that during the first yearsof potential dispersion, the 137Cs activity concentrationin fish varies widely depending on the habitat of fish.During the early stages of the dispersion, the BarentsSea contains regions with relatively high contaminationand regions that are not affected by the release of ra-dionuclides. The model calculation of transfer to fish issubject to large uncertainties, and other more hypo-thetical transfer pathways (e.g. ingestion of particles)have not been considered. Calculations show a maxi-mum activity concentration of 137Cs in fish, duringthe first year following a hypothetical leakage from theKursk, in the range 0–100 Bq/kg. For comparison,the intervention level for concentrations of 137Cs in basicfoodstuffs (as recommended by EC and adopted byseveral countries, e.g. Norway) is defined as 600 Bq/kg.The preliminary calculation of the collective dose to

man is shown in Table 5. These results correspond toreleases of all radionuclides from both reactors accord-ing to the aforementioned scenarios. The collective doseto man is dominated by the contribution from 137Cs forScenario 1. The calculations showed that a collective

dose of 61 manSv was attributable to the intake of 137Csfrom the Barents Sea alone for the ‘‘worst case scenario’’(Scenario 1, version 2), while the total collective dosefrom all radionuclides from the whole marine area wasestimated to be 97 manSv. In this case, contributionsfrom 137Cs and 239Pu correspond to a total dose of 69and 5.5 manSv, respectively. For comparison, collectivedoses from other radionuclides for the ‘‘worst case’’scenario are estimated to be 6.5, 4.3, 2.2, 0.37 and 0.27manSv for 90Sr, 134Cs, 241Am, 147Pm and 106Ru, re-spectively. Considering scenario 2, with corrosion lead-ing to release after 100 years, the total collective dosewas estimated to be 8.4 manSv for an operational periodof 12 000 MWd. Furthermore, approximately 80% ofthe collective dose from the Barents Sea is attributableto 137Cs exposure. 239Pu does not contribute significantlyto the collective dose for Scenario 1. For Scenario 2,however, the contribution of 239Pu to the total collectivedose is comparable to that of 137Cs. This is mainly dueto the radioactive half-life of 137Cs, which is 30 years.

8. Conclusions

No indications of leakage from the submarine havebeen observed during the two expeditions to the Kurskin August and October 2000. Elevated levels of radio-activity have not been observed neither in any dose-ratereadings nor any measurements of environmental sam-ples taken from close to or even from inside the Kursksubmarine.A model calculation based on a scenario representing

an abnormal event occurring one year after shutdown ofthe reactors, e.g. during the lifting operation, was con-ducted. Using the conservative assumption of a releaseof 100% of the reactors radioactivity inventory to thesurrounding seawater, a potential activity concentrationin fish of about 80–100 Bq/kg of 137Cs was calculated.Following the same scenario, the total collective dosefrom all radionuclides from the whole marine area wasestimated to 97 manSv. 137Cs has been shown to be themost important radionuclide. These estimates are ofcourse subject to large uncertainties i.e. due to militarysecrecy surrounding information pertaining to the

Table 5

Collective doses to man (manSv) calculated to 1000 years for two release scenarios (Scenario 1: release after 1 year; Scenario 2: corrosion after 100

years) with two different operational periods for the reactors (Case 1: 12 000 MWd; Case 2: 24 000 MWd)

Scenarios/operational

period

Collective dose in the Barents Sea (manSv) Total collective dose (manSv)

137Cs 239Pu All relevant radionuclides 137Cs 239Pu All relevant radionuclides

Scenario 1 12 000 MWd 29 0.25 34 33 3.2 43

Scenario 1 24 000 MWd 61 0.42 73 69 5.5 97

Scenario 2 12 000 MWd 3.1 0.25 3.8 3.6 3.2 8.4

Scenario 2 24 000 MWd 6.1 0.42 7.4 6.9 5.5 19

Fig. 7. Dynamics of the 137Cs concentration in fish (Bq/kg) from the

Barents Sea regions following a potential release of 100% of the ac-

tivity in the two reactors (scenario 1, case 2).

I. Amundsen et al. / Marine Pollution Bulletin 44 (2002) 459–468 467

reactor inventory, the damage to the hull and reactorcompartment and the possibility of releases of radioac-tivity to seawater following an accidental event. How-ever, the conservative calculation indicates that nohealth impact on the population is expected resultingfrom the given worst-case scenario. However, the eco-nomic consequences following a serious leakage fromthe Kursk are difficult to estimate.If the submarine fails to be lifted in mid-September

2001, it is of interest to improve the modelling of thepotential long-term leakage from the Kursk. Better in-formation regarding the radionuclide inventory and thesource term will then be needed.

Acknowledgements

We would like to thank the leading personnel andcrew of Stolt Offshore and the Haliburton for their co-operation and the excellent facilities on the expeditionsof August and October 2000, respectively. Furthermore,we are grateful to Lars Føyn of the Norwegian Instituteof Marine Research for provision of sampling equip-ment on both expeditions and to Tone Bergan, Anne-Liv Rudjord and Finn Ugletveit for their support. Thepersonnel working at the laboratory of NRPA, AnneLene Brungot, John Cobb and Rikke Engstrøm whoperformed an outstanding job in preparing and mea-suring the environmental samples.

References

AMAP, 1998. Radioactivity. In: Strand, P. et al. (Eds.), AMAP

Assessment Report: Arctic Pollution Issues, Arctic Monitoring and

Assessment Programme (AMAP), Oslo, Norway, pp. 525–619

(Chapter 8).

Amundsen, I., Lind, B., Reistad, O., Gussgaard, K., Iosjpe, M., Sickel,

M., 2001. Str�aalevern Rapport: 5 The Kursk Accident, Norwegian

Radiation ProtectionAuthority, Øster�aas, Norway. ISSN 0804-4910.Chenyal, Ma., vol Hippel, F., Ending the production of highly-

enriched uranium for naval reactors, October 2000, p. 9 (submitted

by the author, draft 30.11.00).

EC, 1997. Evolution of the radiological situation around the nuclear

reactors with spent fuel, which have been scuttled in the Kara Sea,

European Commission, Luxembourg, EUR 17634 EN.

EC, 1999. The radiological exposure of the population of the

European Community to radioactivity in the Baltic Sea, Marina-

Balt Project. In: Nielsen, S.P. (Ed.), Proceedings of a seminar at

Hasseludden Conference Centre, Stockholm, 9–11 June 1998,

Radiation Protection 110. EUR 19200. Office for Official Publica-

tions of the European Communities. Luxemborg, 2000.

Gerdes, R., Karcher, M., Kauker, F., Koeberle, C., 2001. Predicting

the spread of radioactive substances from the Kursk. Eos, Trans-

actions, American Geophysical Union vol. 82 (23, June 5), 256–257

(see also p. 253).

Grøttheim, S., 2000. Artificial radionuclides in the Northern European

Marine Environment. S. NRPA-report 2000:1 Norwegian Radia-

tion Protection Authority, Øster�aas, Norway.

IAEA, 1999. Radioactivity in the Arctic Seas, Report for the

International Arctic Seas Assessment Project (IASAP), Interna-

tional Atomic Energy Agency, Vienna, IAEA-TECDOC-1075.

Iosjpe, M., Strand, P., Salbu, B., 1997. Estimations of significance of

some processes for modelling of consequences from releases in the

Arctic Ocean. In: The third International Conference on Environ-

mental Radioactivity in the Arctic. Extended Abstract, Tromsø,

Norway, June 1–5, 1997, pp. 74–75.

Iosjpe, M., Strand, P., 1998. Influence of time and space resolution on

dispersion of radionuclides and radioecological assessment in the

marine environment. In: International Symposium on Marine

Pollution, Monaco, Extended Synopses, pp. 329–330.

Iosjpe, M., Brown, J., Strand, P., 2001. Modified approach for box

modelling of radiological consequences from releases into marine

environment. Journal of Environmental Radioactivity, in press.

Nielsen, S.P., Iosjpe, M., Strand, P., 1995. A preliminary assessment of

potential doses to man from radioactive waste dumped in the

Arctic Sea. NRPA report 1995:8, ISSN 0804-4910, Øester�aas,

Norwegian Radiation Protection Authority.

Nielsen, S.P., Iosjpe, M., Strand, P., 1997. Collective doses to man

from dumping of radioactive waste in the Arctic Seas. The Science

of the Total Environment 202, 135–146.

Reistad, O., 2001. Minutes from meeting in Joint Russian–Norwegian

Expert Group on Impact Assessment of the Kursk Accident,

Norwegian Radiation Protection Authority, Øster�aas, June 2001

(not published).

Sevmorput, 1988. Information regarding Safety of Icebreaker –

Transport lighter/container-ship with nuclear propulsion Sevmor-

put. Approved by the Register of Shipping USSR. Submitted by

the Russian authorities before a visit of the icebreaker in the

Norwegian city Tromsoe in 1988.

Stamm�ller, R.J.J., Boerresen, S., Casal, J.J., Forslund, P., 1996.‘‘Helios’’ – Verification against Kritz and other critical experi-

ments. In: International Conference on Physics of Reactors

PHYSOR 96, Mito, Ibaraki, Japan, September 16–20, 1996, vol.

3, pp. F58–F65.

SSP, 1991. Scandpower. Calculation of the Isotopic Content in a

Submarine Reactor (Technical Note) – 1991 (Norwegian).

SSP, 1998. Criticality Considerations of Russian Ship Reactors and

Spent Nuclear Fuel, Report no: TR1/41.20.03, Studsvik Scand-

power; Kjeller Norway 1998.

UNSCEAR, 1982. Ionising radiation: Sources and biological effects.

United Nations Scientific Committee on the Effects of Atomic

Radiation, 1982 Report to the General Assembly, with annexes,

United Nations, New York, 1982.

468 I. Amundsen et al. / Marine Pollution Bulletin 44 (2002) 459–468