78
Chapter 7
Potential Accident Scenarios
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
using assumptions relating radiation doses to health
7.1. Introduction
impacts. Such information is essential for risk manage-
This chapter considers the nature and scale of conse-
ment.
quences arising from potential accidental releases of ra-
dioactivity into the Arctic environment from sources
7.1.1. Risk management
under human control. The sources considered to war-
rant accident assessment are those described in Chapter
Risk management includes the analysis of accident sce-
2. All scenarios are described in more detail in the rele-
narios and consequences and, where possible, an assess-
vant literature. Relevant accidents and subsequent im-
ment of the probabilities of accidents and their conse-
pact assessments are reviewed and, where possible, ex-
quences. Sources may occur within the Arctic, in which
trapolated to provide perspectives on the consequences
case these `point sources' require analysis, and outside
of accidents associated with other sources. The chapter
the Arctic. The potential for accidents occurring outside
concludes with recommendations for further impact as-
the Arctic to contaminate the Arctic environment de-
sessments.
pends on the dispersal characteristics.
Scenario analysis begins with a consideration of the
Generally, the larger the inventory of radionuclides,
different possibilities for sequences of events and pro-
the greater the hazard. In most cases, the inventory in
cesses (such as containment failure mechanisms) that
Becquerels (Bq) is well established owing to the applica-
can lead to radionuclide release. These depend upon the
tion of well-proven technology and associated regulatory
specific management and engineering features of the fa-
requirements. However, in some cases, the information
cilities under consideration. The release mechanisms and
may be less complete, as for example, in the case of old
characteristics are important determinants of the envi-
waste storage facilities for which information is limited.
ronmental and human health impacts. Relevant vari-
The risks associated with hazardous sources may
ables might include: isotopic composition; amounts of
also be modified by measures to control the source term.
each isotope released; physical-chemical form of release
That is, consideration of the risks must address both the
(gas, solution, aerosol, etc.); time development of the re-
scale of the consequences and the likelihood of their oc-
lease; release point and plume height; and the energy
currence. Thus, while a very large source term may pres-
content of the release.
ent the greatest hazard (potential for harm), measures to
The scenarios discussed here were not necessarily de-
reduce the chances of release may reduce the risks to a
veloped for the same purposes as the AMAP assessment.
tolerable level (HSE, 1988). Nuclear safety initiatives to
They are illustrative of selected aspects of the possible
reduce the likelihood of accidents are discussed in Chap-
consequences of radionuclide release rather than repre-
ter 6. Other ways of reducing risk include measures to
sentative of comprehensive risk or scenario analyses for
reduce the consequences of potential accidents. Risk
the Arctic.
management must account for both. A simple example is
The radionuclide release information provides in-
the case of spent nuclear fuel, which is a significant ra-
put to a radionuclide transport model that is used to
dioactive source term. Left on the surface with limited
predict the subsequent environmental distribution of
containment, the chance of releases into the environ-
contamination. For accident scenarios resulting in re-
ment, before radioactive decay has reduced the hazard
leases to the atmosphere, the assumed (or actual, if as-
significantly, whether as a result of waste container de-
sessing the consequences of past accidents) meteoro-
gradation or by human sabotage, is relatively high. Deep
logical data are very important as they can radically
disposal is considered to reduce the risks by reducing the
affect the degree of atmospheric dispersion. Similarly,
likelihood of gross and acute environmental releases.
for releases to the aquatic environment, the hydrody-
See, for example, discussions concerning high level waste
namic characteristics of the receiving environment are
disposal in Japan (JNC, 2000).
equally important. Distribution following releases to the
Another aspect of risk management is the introduc-
ground are strongly dependent upon surface geology
tion of measures to mitigate the impact of accidental re-
and hydrology. The radiation characteristics of the ra-
lease. International guidance concerning countermea-
dionuclides and their environmental mobility are also
sures is provided by the International Commission on
important determinants of the magnitude of the conse-
Radiological Protection (ICRP, 1992). Examples include
quences following release. The receiving environments
evacuation, advice to remain indoors, and the distribu-
themselves also influence the scale of the consequences,
tion of iodine tablets. Interventions should be based
since some are more susceptible to incorporating ra-
upon evaluations of their benefit, expressed as averted
dionuclides into human exposure chains than others, as
doses, and disbenefits, especially those of an economic
discussed in Chapter 4 (NRPA, 1999; Skuterud et al.,
or social nature. Assessments of the consequences of ac-
1999). Radiation doses to humans and other biota are
cidents should take account of the planned emergency
assessed using assumptions about the ways in which
response in the early and latter phases of the accident; in
they interact with contaminated media. Finally, the im-
the long-term the effects of clean-up measures may be
pact on human and environmental health is assessed
important, see Brown et al. (2000).

---

79
in control systems (e.g., control rod ejection), and loss
7.1.2. First AMAP assessment
of flow.
The first AMAP assessment concluded that of greatest
CCIs (e.g., power transients and earthquakes) lead to
concern were the possible accidents associated with: nu-
multiple failures and may affect several components in
clear power plant (NPP) operation; nuclear weapons
the system.
handling and storage; decommissioning of nuclear sub-
According to Stokke (1997), loss of coolant in pres-
marines; and the management of spent fuel from nu-
surized water reactors does not immediately signify a
clear-powered vessels.
large radioactive release. The vessel should be able to
Consideration is given to the consequences of each of
contain an overheated core for a period that may be suf-
these types of accident. Also, to additional or modified
ficient to allow restoration of adequate core cooling. If
potential accident sources, particularly reactors in sunken
there is extensive core damage, it is unavoidable that ra-
submarines such as the Kursk (Amundsen et al., 2002a)
dioactivity leaks occur. A core melt by itself, however,
and the management of damaged spent fuel, as stored on
does not create an explosive situation unless reactor
the Lepse (NRPA, 2001).
containment fails. A source term for the Kola NPP hav-
ing a very high radioactive plume rise and thus exposure
of core and fuel to the open air has therefore a low prob-
7.2. Land-based nuclear power plants
ability. Nevertheless, releases of noble gases and volatile
Operational land-based NPPs in the Arctic include the
radioactive compounds should be expected in a severe
Kola and Bilibino NPPs. The Kola plant comprises four
core damage accident.
VVER-440 pressurized water reactors each with a de-
There are differences between the two older and the
sign output of 1375 MW(th) and 411 MW(e). The Bili-
two newer plants. For the older Model 230 reactors, the
bino NPP is located in the Chukotka region in eastern
effectiveness of the confinement structure in containing
Russia and comprises four light-water cooled, graphite-
the radioactive steamญgas mixture after a LOCA is un-
moderated reactors each of output 62 MW(th) and 12
certain. The airtightness of the confining structure is not
MW(e).
assured and there may be considerable leakage even
Owing to design differences, a direct comparison of
without open valves or other penetrations. Breaks in the
the risks posed by the Kola and Bilibino NPPs is not
largest coolant pipes may generate a steam pressure that
straightforward. Risk assessments need to include a con-
could crack or rupture the confinement structure and
sideration of the engineered features and management at
create an open passage from the core to the environ-
the respective plants, but for assessing significant re-
ment, although the reactor vessel would still be intact
leases, there are obvious differences owing to the reac-
(Stokke, 1997).
tors at the Kola NPP being more than 20 times larger.
The power plants on the Kola Peninsula clearly rep-
7.2.1.2. Probabilities
resent the major potential reactor accident source within
the Arctic.
The probability that a severe accident may occur is de-
pendent on many factors such as design features, con-
struction quality, and human performance. The proba-
7.2.1. Accident scenarios and consequences
bility that an event may lead to an unintentional core
for the Kola NPP
melt can be assessed on the basis of engineering judg-
The Kola NPP is located in Murmansk Oblast in north-
ment or by performing a Probabilistic Safety Assessment
west Russia and severe accidents at the site have the po-
where, in principle, all realistic chains of events leading
tential to substantially contaminate both northwest Rus-
to core melt are analyzed and their probabilities of oc-
sia and northern Fennoscandia. Studies by Stokke (1997),
currence calculated. The sum of all probabilities for all
including a review of the Kola reactor safety systems,
possible initiating events to cause a core melt is the Core
have provided detailed information on the Kola plant
Melt Frequency (CMF), given as the probability per re-
and its reactor inventories.
actor operating year. The CMF does not include the
probability of human failure, sabotage, or terrorist at-
tack (Stokke, 1997). For modern NPPs, the CMF is con-
7.2.1.1. Initiating events
sidered to be within the range 10ญ4 to 10ญ5. At present,
Initiating events that may lead to core melt sequences in
there is no CMF for the Kola NPP for use in accident
pressurized water reactors are generally grouped into
consequence analysis. However, the International
three classes: loss of coolant accidents (LOCAs), tran-
Atomic Energy Agency (IAEA) has reported a prelimi-
sients, and common cause initiators (CCIs).
nary estimate of 5.5 10ญ3 per year for the oldest Kola
LOCAs may be initiated by large leaks or breaks in
reactors (Stokke, 1997).
the primary circuit, which in turn may be caused by me-
chanical failure (such as pipe breaks, fire, and corrosion)
7.2.1.3. Accident source terms
resulting from poor maintenance. The loss of cooling
may take place early in the sequence or at a later stage.
Accident source terms depend on the initiating events,
Early loss of cooling is potentially the most dangerous as
which may result in different accident scenarios. Ex-
it gives little time to re-establish cooling and because sig-
amples of scenario development are provided by Stokke
nificant decay of short-lived radioisotopes will not have
(1997). Worst case scenarios concern situations in which
occurred.
the reactor core contains the maximum number of prod-
Transients can be failures in power supply, reactiv-
ucts and maximum activity concentrations of radio-
ity transients (sudden increases in reactivity), failures
nuclides at the end of the normal fuel burn-up cycle.

---

80
AMAP Assessment 2002: Radioactivity in the Arctic
Table 7ท1. Combined inventory of radionuclide groups with release potential in a VVER-440 Kola NPP (Stokke, 1997).
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
Total activity in
core, Bq
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
Noble gases
85mKr, 87Kr, 88Kr, 89Kr, 90Kr, 133Xe, 135Xe, 138Xe
1.21
1019
Halogens
84Br, 87Br, 131I, 133I, 134I, 135I
1.48
1019
Alkaline metals
86Rb, 88Rb, 89Rb, 90Rb, 134Cs, 136Cs, 137Cs, 138Cs
7.53
1018
Tellurium group
127mTe, 127Te, 129Te, 129mTe, 131mTe, 132Te, 127Sb, 129Sb
3.90
1018
Alkaline earth metals
89Sr, 90Sr, 91Sr, 140Ba
6.42
1018
Transition metals
90Y, 91Y, 95Zr, 97Zr, 95Nb, 99Mo, 99mTc, 103Ru, 105Ru, 106Ru, 105Rh
2.14
1019
Lanthanides
140La, 141Ce, 143Ce, 144Ce
9.57
1018
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
Table 7ท1 summarizes information on the relatively mo-
Releases of noble gases, radioiodine, and radioce-
bile radionuclides present at the end of the fuel burn-up
sium are the most important, as these are assigned the
cycle that are most likely to be released in the event of an
highest release level (level 7) on the International Nu-
accident.
clear Event Scale (INES) for this reactor type. Releases
There is currently no information on the inventory of
of radioiodine, radiocesium, and radiostrontium are im-
actinides for the Kola NPP and so the consequences of
portant from a radiological hazard point of view, while
actinide release have not been assessed. Stokke (1997)
the long-term consequences of much smaller releases of
has estimated the highest release fractions for various
actinides are also significant. The inventory estimates in
potential accidents from the Kola reactors (Table 7.2).
Table 7ท1 and fractions released in Table 7ท2 are assumed
to be valid for both VVER reactor types. Various acci-
Table 7.2. Highest release fractions (%) from the core inventory for
dental release scenarios have been considered, examples
different initiating events (Stokke, 1997).
of which are given in Sections 7.2.1.4 to 7.2.1.7, for an
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
VVER-440/213
VVER-440/230
unintentional `worst case' scenario (i.e., where the acci-
---------------------------------------------------------------------------------------
dent that is not a result of malicious intent, e.g. terror-
Small Large
Large
ism) with a large LOCA and a less severe transient sce-
Transient
LOCA
LOCA
LOCA
nario. Table 7ท3 shows the inventory, fraction released,
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
and consequent activity emitted to atmosphere for the two
Noble gases
100
20
50
100
Iodine
2.5
0.05
1
15
scenarios. A major fraction of the radionuclides, including
Cesium
2.5
0.05
1
12
Cs- and Sr-isotopes will be present as particles. Since the
Tellurium
0.1
0.05
0.2
10
air dispersion and transfer model does not currently in-
Strontium
1
0.1
1
2
clude radioactive particles, these are not considered in
Barium
0.5
0.05
0.5
2
the estimated ecosystem transfer and doses to humans.
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
For a given event, the estimated release for a single nu-
7.2.1.4. Initial dispersion
clide is calculated by multiplying the amount of the nu-
clide in the core by the release fraction. The source term
The dispersion of radionuclides from a source depends
suggested for the VVER-440 230 model (Stokke, 1997)
on the release height and meteorological conditions at
is based on source terms applied in earlier consequence
the release site and along the transport route, in addition
assessments of accidents at the Kola NPP. The source
to the properties of the released material such as size dis-
term for the VVER-440 213 model is based on the IAEA
tribution and the degree of volatilization. Buildings and
Technical Co-operation Project on Evaluation of Safety
other structures near the release point can also affect the
Aspects for VVER-440 model 213. Because all aspects
initial dispersion, especially in the case of releases at low
of a potential accident are not yet completely under-
height. This effect, however, becomes insignificant at
stood, a conservative approach should be taken so as
distances over a few kilometers from the source, and is
not to underestimate the risk. The source terms for
negligible for releases with a high effective release height.
model 230 reflect a conservative approach that results in
Scenarios for a hypothetical release from the Kola NPP
source terms that are larger than most other source
are based on data from the Norwegian and Danish Me-
terms previously applied for VVER-440 reactors.
teorological Institutes.
Table 7ท3. Calculated release fractions for selected radionuclides under a large LOCA
and a transient scenario (Larsen et al., 1999; Stokke, 1997).
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
Inventory,
Fraction released, %
Activity released, PBq
PBq
LOCA
Transient
LOCA
Transient
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
137Cs
117
12
2.5
14.0
2.9
134Cs
156
12
2.5
18.7
3.9
90Sr
85
2
1
1.7
0.9
132Te
2240
10
0.1
224
2.2
132I
2330
15
2.5
233
57.5
131I
1570
15
2.5
236
39.3
103Ru
2350
1
0.1
23.5
2.4
140Ba
2790
2
0.5
55.8
55.8
140La
2860
0.2
0.1
5.7
2.9
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ

---

Chapter 7 ท Potential Accident Scenarios
81
The information was supplemented by a study of the
culated effective doses are 3.5 mSv and 5.1 mSv for Sce-
meteorology and transport of radioactive contamination
narios A and B, respectively, with 134Cs and 137Cs the
from the Kola NPP (Bartnicki and Saltbones, 1997; Salt-
most important contributors (60 to 70%). High activity
bones et al., 1997). The Norwegian Meteorological In-
of deposited radiocesium is the main reason for the higher
stitute defined the meteorology in the area, calculated
annual dose in Scenario B, the precipitation scenario.
the transport times, and investigated the likelihood of
External irradiation from airborne radionuclides is a
nuclear contamination at specific locations. Three sce-
relatively insignificant exposure pathway a week or
narios were selected as initial conditions for the disper-
more after the hypothetical accident. However, during
sion model (Saltbones et al., 1995). The scenarios used
cloud passage, airborne radionuclides, such as noble
three sets of weather situations that would provide par-
gases with very short half-lives (e.g., 133Xe), could be of
ticularly unfavorable consequences for Norway. Two of
some significance for the dose rate.
the three situations were relevant to Arctic Norway.
Owing to large quantities of short-lived radionu-
Scenario A, with a rapid transit time to northern
clides emitted during the hypothetical accident, a consid-
Norway, where one-eighth of the released material was
erable fraction of the external dose will occur during the
deposited within 72 hr.
first few days after contamination. In Scenario A, where
Scenario B, with precipitation during the passage of
the atmospheric transport of radioactivity occurs rap-
the plume over northern Norway, with wet deposition
idly, about 20% of the effective external dose will result
such that nine-tenths of the release was deposited over
from irradiation during the first week after deposition.
Norwegian territory.
In Scenario B, where the transport is considerably
slower, the corresponding value was estimated at about
10%. Furthermore, if radionuclide migration through
7.2.1.5. Consequences
soil profiles, and the subsequent radioactive shielding by
There have been recent assessments of both the short-
overlying soil, is considered, then the external dose re-
term and long-term doses from hypothetical accidents at
ceived from the first few days may be of even greater rel-
the Kola NPP. The analysis of short-term doses was con-
ative importance than that estimated.
fined to Norway and to external and inhalation doses.
Of the 50 radionuclides considered, only a few are
The short-term analysis did not include ingestion doses
significant contributors to the total external dose. Over
based on the assumption that the accident occurred out-
a short time-scale (days or weeks) the dominant nuclides
side the short growing period, when dairy animals
are 132Te/132I, 131I, 103Ru, and 140Ba/140La. After these
would be housed; a situation which prevails for most of
decay the external dose is dominated by 134Cs and 137Cs.
the year in the Arctic. The long-term doses were esti-
The effective dose from inhaled radionuclides is pre-
mated for northern Norway and for various regions of
dicted to be <1 mSv for both scenarios and is generally
northwest Russia and considered external and ingestion
highest for young individuals. The dominant nuclide is
doses arising from the mobile, long-lived radionuclides,
131I, which contributes 50 to 70% to the total inhaled
radiocesium and radiostrontium only. For both assess-
dose, depending on the scenario and age group. Equiva-
ments, the unintentional worst-case accident was se-
lent doses to the thyroid gland were calculated at 4.6 to
lected for an assessment of radiation levels and doses.
10.6 mSv (Scenario A) and 1.6 to 3.6 mSv (Scenario B).
The doses calculated for the two scenarios are as-
7.2.1.6. Short-term assessment
sumed to represent the worst possible consequences of a
severe nuclear accident at the Kola NPP. Nevertheless,
The assessment included more than 50 radionuclides.
the received doses are much too low to result in any
The dispersion model results were based on the release
acute radiation injuries.
of particles with a given mathematical mass and must
International guidance concerning countermeasures
therefore be combined with information about the ra-
(e.g., ICRP, 1992), such as evacuation, staying indoors,
dioactive emission in order to calculate the atmospheric
or the distribution of iodine tablets is based on evalua-
and ground levels of the various radionuclides. Activity
tions of the benefits and disadvantages of implementa-
concentrations and ground deposited activity were used
tion, expressed as saved (averted) doses. This assessment
as a starting point for calculating doses from the various
indicates that the saving potential is too low to justify
exposure pathways associated with inhalation and ex-
the direct implementation of countermeasures; however,
ternal irradiation from radionuclide deposits on the
this must be investigated further. The uncertainties in the
ground and in the air.
calculations are large, and an evaluation of the pros and
Although foodstuffs are the major contributor to the
cons for an actual situation can result in the use of dif-
total long-term dose, the focus of the short-term conse-
ferent countermeasures, e.g., for special groups.
quence assessment was on external exposure and inhala-
tion. There is a very short growing season in the Arctic
7.2.1.7. Long-term assessment
and the implicit assumption is that the accident occurred
outside this period.
The first AMAP assessment concluded that the vulnera-
For an adult, the effective dose was calculated at
bility of the Arctic (defined as the relationship between
about 1 mSv for Scenarios A and B. In both cases, the
dose and atmospheric deposition of nuclides) is higher
contribution from radionuclides deposited on the
than in most other areas of the world, particularly for
ground predominates, especially where precipitation is
137Cs. This reflects the transfer of radionuclides de-
high and wet deposition considerable.
posited from the atmosphere through terrestrial food
Throughout the first year, external irradiation from
chains to human radiation exposure. The long-term as-
the ground is the most important exposure pathway: cal-
sessment estimated the long-term impact of radioactive

---

82
AMAP Assessment 2002: Radioactivity in the Arctic
contamination from a hypothetical LOCA at the Kola
concentrations were predicted to occur in reindeer meat,
NPP on the two northernmost counties of Norway
sheep meat, mushrooms, and berries, and the highest
(Troms and Finnmark), and on the Murmansk Oblast in
90Sr activity concentrations in berries and potatoes.
Russia. The weather pattern for the Russian scenario
After fifty years, the highest activity concentrations pre-
was based on predicted ground deposition provided by
dicted for foodstuffs were for 137Cs in mushrooms, rein-
the Danish Meteorological Institute, with most deposi-
deer meat, and berries.
tion occurring on the Kola Peninsula.
As for foodstuffs, predicted annual individual inges-
The study considered radionuclide deposition, trans-
tion doses for reindeer herders and other inhabitants
fer to and contamination of locally produced foodstuffs,
vary spatially according to differences in deposition and
and external and ingestion doses for reindeer herders
land cover. Annual ingestion doses for all population
and other inhabitants. A spatial model was developed
groups in the first year after deposition were predicted
within a geographical information system to predict the
to exceed 1 mSv. Annual individual radiocesium inges-
long-term consequences of radionuclide deposition on
tion doses for reindeer herders are significantly greater
northern Norway and northwest Russia. As no site spe-
than for other inhabitants. In the first year after deposi-
cific data were available, general transfer factors were
tion, the most significant contributor to annual individ-
used in the model (JNREG, 2002a,b,c).
ual radiocesium ingestion dose is reindeer meat for all
population groups, with the exception of other Norwe-
External doses
gian inhabitants for whom dairy products and mutton
The highest individual external -doses occur in those
are important contributors. Potatoes and dairy products
areas receiving most accident deposition, but are negligi-
are the largest contributors to the much lower annual in-
ble compared to ingestion doses. Individual external -
dividual 90Sr ingestion doses for all population groups.
doses for reindeer herders are twice those of the other in-
Berries are another important 90Sr contributor to the
habitants owing to the tendency of the latter to occupy
two Russian population groups, while reindeer meat is
areas with higher shielding (i.e., buildings).
also a source of 90Sr for Russian reindeer herders.
Under all accident scenarios, reindeer herder annual
Internal doses
ingestion doses are predicted to exceed 1 mSv for many
Radionuclide transfer to foodstuffs was modelled using
decades after accident deposition (and are much higher
aggregated transfer coefficients (Tag ; Box 3ท2) and effec-
in the first few years); for the other population group,
tive ecological half-lives (Teff ; Box 3ท1). Long-term pre-
ingestion doses exceed 1 mSv for a few years after acci-
dictions were made for the spatial variation in activity
dent deposition in northern Norway and for a decade in
concentrations in foodstuffs, individual external and in-
Murmansk Oblast. Fifty years after accident deposition,
gestion doses for reindeer herders and other inhabitants,
individual 137Cs ingestion doses for reindeer herders are
and radionuclide fluxes (total Bq output from contami-
over two orders of magnitude lower than during the first
nated land areas over specified time periods).
year; those for the other population group are more than
Data were collated for each study area to derive area-
30 times lower. The largest contributors to annual indi-
specific Tag and Teff values for radiocesium and 90Sr
vidual 137Cs ingestion doses for Norwegian reindeer
(JNREG, 2002a). The biggest difference was the 3-fold
herders fifty years after accident deposition are reindeer
higher Tag value for 137Cs transfer to reindeer meat for
meat, freshwater fish, and dairy products, with dairy
the Murmansk Oblast compared to Norway, and the
products, freshwater fish, mushrooms, and reindeer
longer associated half-life. Together, these were responsi-
meat the most important contributors to the other Nor-
ble for the greater intakes predicted for radiocesium in
wegian population group. Reindeer meat and mush-
reindeer meat per unit deposition, and the greater per-
rooms are the largest contributors to annual individual
sistence in reindeer meat and thus Russian reindeer
137Cs ingestion doses to the Russian population groups
herders. In addition, the Tag value for 137Cs transfer to
50 years after accident deposition.
potato and to a lesser extent berries, was lower for Nor-
Sr-90 is a much less important contributor to inges-
way than Russia. Teff values for freshwater fish were
tion dose and the predicted consequences of the accident
lower for Norway than Russia. The Teff value used for
scenarios are much less certain owing to the paucity of
90Sr in milk was much greater for Russia than that as-
relevant data for the Arctic, in particular for milk. For
sumed for Norway. For 90Sr, Tag values for Russian dairy
reindeer herders, freshwater fish, potatoes, berries, and
products and potatoes were lower than for Norway,
reindeer meat, provide the largest contribution to annual
while those for most other products were higher.
individual 90Sr doses, while potatoes, freshwater fish,
The most obvious difference between the diets of the
and berries, are the most significant contributors for the
Norwegian and Russian inhabitants is in the consump-
other inhabitants.
tion of dairy products; these are important in Norway
The most significant contributor to total doses for all
but much less so in the Murmansk Oblast. Reindeer
population groups is radiocesium ingestion. Vulnerabil-
meat consumption is highest in the male reindeer herd-
ity to 90Sr contamination is much lower than to radioce-
ers in Lovozero in Russia. Sheep and goat meat is only
sium for both reindeer herders and other inhabitants.
consumed in Norway. Potato and freshwater fish con-
There are substantial differences in agricultural pro-
sumption is also greater in Russia.
duction within the various areas of northern Norway.
Under the scenarios considered, high activity concen-
Production of almost all agricultural products in Troms
trations persist in foodstuffs owing to the high Teff val-
is 2- to 5-fold higher than in Finnmark, whereas reindeer
ues. Activity concentrations for 90Sr in foodstuffs are
production is 20-fold higher in Finnmark where most of
much lower than for radiocesium. In the first year after
reindeer herders live. Detailed production data were not
accident deposition, the highest radiocesium activity
available for Murmansk Oblast. Annual radionuclide

---

Chapter 7 ท Potential Accident Scenarios
83
fluxes have been predicted for all locally grown food-
Union (CFSU), and the Taymir Peninsula. Several ap-
stuffs (production of mushrooms, berries, and freshwa-
proaches were used to determine the probability that air
ter fish was estimated by multiplying diet and popula-
would be transported from the Kola NPP to each of
tion). In the first year after deposition, the highest ra-
these regions, transport times, and seasonal variations in
dionuclide fluxes are predicted to coincide with the
atmospheric transport.
areas receiving the highest accident deposition. The
The assessment indicated that Scandinavia would be
largest contributors to radiocesium fluxes are reindeer
affected by a release for 44.5% of days between 1991
meat and dairy products, while dairy products and pota-
and 1995, Europe for 8.1%, the CFSU for 43.2%, and
toes are the largest contributors to annual 90Sr fluxes.
the Taymir Peninsula for 55.5%. The airflow probabil-
The contribution of different foodstuffs to radionuclide
ity field had a similar pattern to that for the one avail-
fluxes changes with time. Fifty years after accident dep-
able assessment of the consequences of hypothetical ac-
osition, the highest radionuclide fluxes do not necessar-
cidents at the Kola NPP, where the released materials
ily occur in those areas receiving the greatest accident
were distributed in an almost circular pattern extending
deposition. High radionuclide fluxes can occur in areas
slightly in a northeasterly direction (Baklanov et al., 2002).
with high food production. In general, reindeer meat
Seasonal variations influenced the transport pattern.
and dairy products remain the significant contributors
Two cases of rapid transport from the Kola NPP to
to 137Cs fluxes in the fiftieth year, while berries, pota-
Scandinavia were selected for more detailed study. In
toes, and freshwater fish are the largest contributors to
both scenarios, 60 PBq of 137Cs were released over a pe-
the lower annual 90Sr fluxes.
riod of 20 hr in a plume rising 400 to 600 m. The areas
This study confirms the outcome of the first AMAP
contaminated by 137Cs to a level exceeding 30 kBq were
assessment, i.e., that Arctic residents are particularly vul-
190 000 and 250 000 km2.
nerable to radiocesium contamination and that the vul-
Mean individual doses, collective doses, and collec-
nerability would persist for many years after deposition.
tive risks were calculated for one of the two scenarios,
Reindeer herders are particularly vulnerable due to their
based on assumptions of the relative importance of vari-
higher levels of reindeer meat consumption. Neverthe-
ous nuclides and exposure pathways to the total dose re-
less, other inhabitants of northern Norway and Russia
sulting from the effects of the Chernobyl accident on
would also be potentially exposed to high doses, espe-
Scandinavia. The highest mean individual doses, 1.15
cially if consuming many local products. While reindeer
mSv, occurred in northern Norway. The collective dose
production is the most vulnerable pathway, freshwater
for the area affected was calculated as 1100 manSv, cor-
fish, lamb meat, dairy products, mushrooms, and berries
responding to a collective risk of 54 cases of additional
are also vulnerable foodstuffs. Although game was not
cancer.
included in this study, post-Chernobyl studies show high
and persistent contamination of some game animals.
7.3. Nuclear-powered vessels
The location of communities and their types of agri-
cultural production are important variables determining
The reactors of nuclear-powered vessels located around
vulnerability; if high deposition occurred in the major
the Kola Peninsula represent the greatest density of nu-
reindeer production areas (Finnmark in Norway and
clear reactors in the world. Several types of release have
Lovozero in the Murmansk Oblast) the impact would be
been registered from these vessels, particularly from
much higher than if deposition occurred in areas where
those operating at sea. However, releases have also oc-
other types of agriculture predominated. Conversely, be-
curred at bases on shore, for example in Andreyeva Bay
cause dairy cattle are inside for much of the year, vulner-
and Gremikha Bay. Limited effort has been made re-
ability increases if an accident occurs during the short
garding impact assessments for accident scenarios re-
summer grazing period, especially for 90Sr.
lated to operating vessels, decommissioned vessels, or
Major factors contributing to the uncertainties in the
vessel components after dismantling, owing to the tradi-
estimates of doses and fluxes are the limited number of
tional secrecy surrounding these vessels, their reactors,
nuclides being considered, as well as the use of general
and the composition of their fuel. There is a need to
rather than site specific transfer factors. Also, the scenar-
standardize the existing studies comparing Russian and
ios address releases of gaseous and aerosol components
western efforts and to complete the assessments. Never-
but potential releases of radioactive particles are not taken
theless, some significant contributions have already been
into account. The effects of countermeasures were not
made, such as the IAEA assessment of the risks from
evaluated in this assessment. Doses and fluxes were pre-
the dumped reactors close to Novaya Zemlya (IAEA,
dicted assuming no mitigating actions having been taken.
1998a). Other more recent efforts include the pilot study
However, the results clearly indicate the need for an ef-
by the NATO Committee on the Challenges of the Mod-
fective emergency response, including the application of
ern Society concerning an environmental risk assessment
countermeasures, should an accident of the scale consid-
for decommissioned Russian nuclear submarines still
ered in this assessment ever occur at the Kola NPP.
containing fuel, and an evaluation of the potential im-
pact of large releases from the Kursk at the time of sink-
ing and during subsequent recovery operations (Bak-
7.2.2. Barents region environmental center
study of atmospheric transport
lanov et al., 2003).
The operation, maintenance, decommissioning, and
pathways from the Kola NPP
dismantling of a nuclear vessel fleet is a complex process
An assessment of atmospheric transport pathways from
involving a large number of smaller operations. The ac-
the Kola NPP was undertaken for four geographical re-
tivities include: different modes of operation (training,
gions: Scandinavia, Europe, the central former Soviet
patrolling, tracking, etc.); assignments in port, changes

---

84
AMAP Assessment 2002: Radioactivity in the Arctic
of crew; docking for maintenance and repair; refuelling
7.3.1.1. Kursk
and defuelling; storage onboard of fuelled reactors; on-
and off-loading of fresh and spent fuel from vessels and
The latest accident involving a Russian submarine was
transport ships; mode of fuel transport; and storage of
that of the Kursk in August 2000. The sinking, and sub-
damaged reactors / damaged fuel.
sequent recovery operation, raised considerable concern
To date, some of these operations have been covered
about possible consequences. The accident represented a
by risk assessments. Hopefully, the most serious scen-
significant challenge for the nuclear emergency prepared-
arios involving potential releases have been covered;
ness organization; from the day of the accident until the
however, as this work and international efforts to assist
larger part of the submarine was brought into dock at
Russia in these tasks are reaching new levels of advance-
Roslyakovo in October 2001.
ment and maturity, new facts and scenarios are being
Owing to considerable concern in Norway, the Nor-
identified. The most recent and relevant efforts toward
wegian Radiation Protection Authority undertook an
comprehensive impact assessments are presented in the
environmental risk assessment for four scenarios; com-
rest of this section. These have been subdivided into ves-
bining two inventory calculations and two release sce-
sels in operation (Sections 7.3.1. and 7.3.2.), decommis-
narios. The Kursk inventory calculations were based on
sioned vessels still containing spent fuel on board (Sec-
information for the Russian cargo ship Sevmorput with
tion 7.3.3.), and accident scenarios involving spent fuel
some adjustment of the technical input data. The hypo-
and radioactive waste after dismantling of the vessel
thetical release rate for radionuclides depends strongly
(Section 7.3.4.). The focus is on the presence of spent
on release conditions. These range from instantaneous
fuel because 90 to 99% of the radioactivity resides
release owing to the explosion of torpedoes or cruise
within the fuel. However, the reactor compartments, and
missiles within the submarine, to the slow long-term
the solid and liquid high-level, medium-level, and low-
corrosion of fuel material. The latter may occur when
level radioactive waste also constitute formidable prob-
seawater has penetrated the fuel cladding. If the clad-
lems, mainly in the remediation of the bases and sites.
ding is zirconium, penetration may take several hundred
The latter requires further evaluation as remedial work
years. However, if conditions for galvanic corrosion are
involving international participation is to begin shortly.
present, the cladding could be fully corroded in less than
a year.
Two radionuclide release scenarios were considered.
7.3.1. Military vessels
There are around 33 operative nuclear submarines
1. An abnormal event one year after the accident, i.e.,
within the Russian North Fleet. According to ุlgaard
during the salvage operation, in which 100% of the
(2001), these comprise 12 ballistic missile submarines
inventory in both reactors is released instantaneously.
(Typhoon and Delta Classes), 4 cruise missile sub-
2. The assumption that all barriers, for all practical pur-
marines (Oscar Class), 12 attack submarines (Akula,
poses, have been removed after 100 years, and that
Sierra, Yankee, and Victor Classes), 1 cruiser (Kirov
100% of the inventory of both reactors is then re-
Class), and 4 other submarines (Yankee, Uniform, and
leased.
X-ray Classes). These regularly patrol the nearby oceans
as part of their contribution to the Russian defense
Two versions of operational history, resulting in burn-
force. During service, four Russian nuclear submarines
ups of 12 000 (Version 1) and 24 000 (Version 2) MW-
have sunk, 36 accidents have occurred, and there have
days respectively, were considered for each scenario.
been 378 associated fatalities (ุlgaard, 2001).
Both versions were based on the submarine being opera-
The first AMAP assessment made reference to design
tional for an average of 50 days per year for each year
and beyond-design accident scenarios prepared in rela-
since commissioning at the end of 1994. Version 2 in-
tion to Russian nuclear-powered submarine refuelling.
cludes extensive operation of the reactors for electrical
No new assessments of this type were available for the
power in port, as has been reported to occur by several
present assessment and so that in AMAP (1998) remains
sources in recent years. An estimated release of 100% of
the most appropriate. A submarine incident in a ship re-
the inventory, a very pessimistic approach, was chosen
pair yard in Chazhma Bay on the Russian Pacific coast
to demonstrate the consequences of a simple scenario,
on 10 August 1985 (Sarkisov, 1999; Sivintsev et al., 1994)
even if not realistic, to the public concerned. There is a
involved inadvertent criticality in a reactor core. This
lack of comprehensive environmental assessments of ac-
can be used to illustrate the potential circumstances and
cidents involving submarines in operation and the asso-
the nature, scale, and consequences of such accidents.
ciated release mechanisms and source terms. Earlier stud-
The accident claimed ten lives and gave rise to 39 cases
ies concentrated on releases from sunken submarines to
of acute radiation effects. Subsequent on-site observa-
the marine environment (Eriksen, 1990; IAEA, 1997) or
tions and radioecological investigations showed that the
releases from decommissioned non-defuelled submarines
accident did not have a measurable radiological impact
(NATO, 1998) to sea and air. A consideration of sub-
on Vladivostok or the nearby Shkotovo-22 village.
marines in operation, such as the Kursk, might indicate
Residual long-lived radioactive contamination in the
more severe consequences owing to the greater amount
Chazhma Bay region is localized and does not give rise
of short-lived radionuclides present.
to serious radioecological concern.
Estimates of the radiological consequences for the
Risk estimates of criticality events during refuelling
marine environment of potential radionuclide releases
have been performed by NATO (NATO, 1998). The
from the Kursk were performed for Scenarios 1 and 2,
probability of a severe accident in the Russian navy is es-
using a box model to estimate radionuclide transport
timated to be 2 10ญ3 per refuelling.
over large distances (>1000 km) and long time-scales

---

Chapter 7 ท Potential Accident Scenarios
85
(up to centuries or millennia). The model included terms
7.3.1.2. Komsomolets
that describe the dispersion of radionuclides into the
marine environment over time (Iosjpe and Strand, 1999;
The Komsomolets submarine sank in 1989 in the Nor-
Iosjpe et al., 1997, 2002).
wegian Sea, south of Bear Island (Bj๘rn๘ya). The ra-
Transport, transfer to fish, and collective doses to
dioactive inventory at the time of the accident is esti-
humans were modelled for a range of radionuclides
mated at 2.8 1015 Bq of 90Sr and 3.1 1015 Bq of 137Cs
present in the reactors. However, most attention was fo-
in the reactor, and 1.6 1013 Bq of plutonium in the
cused on 137Cs because this has a relatively long physical
warheads. Minor releases of radioactivity from the reac-
half-life (30 years), readily dissolves in water, and accu-
tor compartment have been detected but large-scale re-
mulates in edible parts of fish and shellfish. For 137Cs
leases are thought to be unlikely as the containment bar-
dispersion in oceanic surface water for the worst case
riers will prevent corrosion of reactor fuel for at least a
potential accidental release, with immediate release of
thousand years.
spent fuel and high burn-up (Scenario 1, Version 2), the
model predicted that 0.5 years after a hypothetical acci-
7.3.1.3. Other nuclear submarines
dental release of 100% of the inventory, the average ac-
tivity concentration in Barents Sea water would be 160
There are 70 decommissioned submarines moored around
to 210 Bq/m3 in the vicinity of the submarine. Activity
the Kola Peninsula at the bases from which they oper-
concentrations would decrease rapidly and after ten years
ated, some close to international borders. Fifty-two are
the average incremental water activity concentration in
waiting to be defuelled and are in various states of re-
the Barents Sea was estimated at 0.1 to 2.8 Bq/m3.
pair. Some have damaged cores due to accidents, which
For 137Cs activity concentrations in fish from the
has prevented the removal of the fuel from the reactor
Barents Sea region (also for Scenario 1, Version 2) the
compartment. Decommissioning submarines with dam-
calculations indicate that during the first few years of
aged cores is a major problem requiring large investment
potential dispersion, the activity concentrations would
and often significant radiation risk to workers.
vary widely depending on the habitat of the fish. During
In 1993, the International Arctic Seas Assessment Pro-
the early stages of dispersion, the Barents Sea would
ject (IAEA, 1997) began a study of the radiological and
contain areas with relatively high levels of contamina-
environmental hazard posed by the reactor compart-
tion and areas that were completely unaffected. The cal-
ments dumped in the Barents and Kara Seas in the 1960s
culated transfer to fish is subject to large uncertainties
and 1970s. Six were dumped with spent nuclear fuel on-
and other transfer pathways, such as particle ingestion,
board (two being complete submarines) and ten were
were not considered. The maximum 137Cs activity con-
dumped without fuel. An environmental survey of the
centration in fish was calculated as between 0 and 100
disposal sites found limited evidence of contamination
Bq/kg during the first year after a hypothetical leak from
that could be attributed to the reactor compartments
the Kursk. By comparison, the intervention level for
(Strand et al., 1997). Transport and dispersion models
137Cs in basic foodstuffs, as recommended by the EC
using isotope release rates indicated that the maximum
and adopted by several countries, including Norway, is
annual dose would be received by local populations, al-
600 Bq/kg.
though this was < 0.1 Sv/yr. However, military person-
For Scenario 1, the collective dose to man is domi-
nel that patrol Novaya Zemlya were projected to receive
nated by the contribution from 137Cs. Calculations show
a potential annual dose of up to 700 Sv (comparable to
that a collective dose of 61 manSv would be attributable
natural background doses). The global collective dose
to the intake of 137Cs from the Barents Sea alone, while
over the next 1000 years from 14C in the inventory was
the total collective dose from all radionuclides from the
estimated at 8 manSv.
whole marine area would be 97 manSv. For the latter,
NATO has considered accident analysis in some de-
contributions from 137Cs and 239Pu correspond to col-
tail (NATO, 1998). For an environmental release to
lective doses of 69 and 5.5 manSv, respectively. For com-
occur, an event with sufficient energy to dislodge the ra-
parison, collective doses from other radionuclides for
dioactive material from its normal location and a failure
Scenario 1 are estimated at 6.5, 4.3, 2.2, 0.37, and 0.27
of the containment boundary are required. Fuel within
manSv for 90Sr, 134Cs, 241Am, 147Pm, and 106Ru, respec-
the reactor compartment is the most probable area for
tively. For Scenario 2, Version1, the total collective dose
such an event to take place due to existing defects, me-
was estimated at 8.4 manSv. Furthermore, approxi-
chanical damage, or overheating. Events can be internal
mately 80% of the collective dose from the Barents Sea
or external (specific to the mooring location) and the
was attributable to 137Cs exposure. There is no signifi-
main concern is core overheating or a LOCA. This is
cant contribution from 239Pu to the collective dose for
used as the reference event and indicates the upper limit
Scenario 1. For Scenario 2, however, the contribution of
for consequences arising from other events. The activity
239Pu is comparable to that of 137Cs. This mainly results
release from a core containing spent fuel is estimated at
from the comparatively short radioactive half-life for
100 TBq of 90Sr and 600 TBq of 137Cs immediately after
137Cs of 30 years.
the event. The NATO report does not attempt to esti-
No indications of leakage from the Kursk submarine
mate the quantity of radioactivity that could be released
were observed during the expeditions to the site in Au-
for each internal event analyzed.
gust and October 2000. Elevated levels of radioactivity
The number of potentially hazardous radionuclides
were not observed in any dose-rate readings or in any
likely to be dispersed following a criticality accident on a
environmental samples from close to or inside the sub-
decommissioned but non-defuelled submarine is rela-
marine, even after the submarine was taken ashore in
tively small, taking into account core activity, the release
Roslyakovo.
fraction, and exposure pathways for radiological effects.

---

86
AMAP Assessment 2002: Radioactivity in the Arctic
Actinides and fission products provide the greatest po-
NATO Committee on the Challenges of the Modern So-
tential hazard. Short-lived radionuclides may dominate
ciety, Norway initiated a study on cross-border defense-
immediately after an accident and their presence is im-
related environmental problems. At an early stage, the
portant in the vicinity of the accident site, but they do
working group decided to focus on decommissioned but
not cause extensive spatial contamination. The majority
still fuelled submarines. Operational submarines and nu-
of the dose from an atmospheric release is contributed
clear weapons were beyond the scope of the working
by 137Cs, 134Cs, and 90Sr, with source terms estimated at
group. A direct comparison of the risks was not under-
350, 35, and 70 TBq, respectively.
taken.
A decommissioned, non-defuelled moored subma-
Three scenarios were used to examine releases to the
rine can sink and release radioactivity to the sea as a re-
sea: sinking of an undamaged submarine; sinking of a
sult of lapsed maintenance. The consequences are not
damaged submarine; and a criticality accident followed
necessarily severe since reactors and submarines are de-
by sinking. Using the release rate model established dur-
signed to withstand considerable pressure. There may be
ing the International Arctic Seas Assessment Project, the
some activity release from corrosion of the outer sur-
dose rate to an individual on a small craft in the harbor
faces of the nuclear reactor. However, if the reactor com-
of Ara Bay, chosen as the location of the sunken subma-
partment was breached, as for example in the event of a
rine, was 100 Sv/hr. At 2 km north of the site, average
collision, corrosion of the fuel could occur rapidly and
dose rates from the water surface to personnel in a small
release fission products to the sea. Estimates based on
craft decreased to about 10 Sv/hr. At the mouth of Ara
models using data from Ara Bay, near Murmansk, sug-
Bay, the level decreased to 1 Sv/hr. This work did not
gest that the release in the year of the accident would be
include uptake by edible fish species; however, Klopkhin
1.6 PBq, with actinides providing <1 TBq of the release
et al. (1997), based on model considerations of a ra-
and fission products dominating. Over time, the pre-
dionuclide plume in water, suggest that fish swimming in
dominant isotopes would change owing to differential
the plume do not accumulate enough activity to justify
decay and mobilization.
restricting their consumption.
It is also possible for an undamaged submarine to
Of the many scenarios discussed, only criticality acci-
sink and such a scenario was examined for Ara Bay.
dents, LOCAs, and hull damage due to sinking or ship
Sinking in such shallow water (a few tens of meters) is
collision were considered potential causes of cross-bor-
unlikely to damage a submarine's primary systems. Re-
der contamination. Weather conditions during an inci-
leases of the four major activation isotopes, 60Co, 59Ni,
dent may lead to contamination of foreign territory. For
63Ni, and 14C, were estimated at around 300 MBq one
example, using Ara Bay as the accident venue a Gauss-
year after sinking, decreasing to 180 MBq after 20 years
ian puff model was used to calculate the dispersion of
(IAEA, 1997). Except for 14C, these isotopes would ad-
radioactivity for stable weather conditions with winds
sorb onto coastal sediments and ultimately settle to the
toward Kirkenes and the county of Finnmark in Nor-
sea floor or remain on the hull. Models indicate little ac-
way. Kirkenes is an urban environment, whereas Finn-
tivity in the waters of Ara Bay and even less 1.5 km from
mark represents a critical group with a high consump-
the release site.
tion of locally produced foodstuffs. For dry deposition
Such studies indicate that recovery of sunken sub-
only, the 137Cs deposition at Kirkenes was about 10
marines or reactor compartments is not too difficult if
kBq/m2 and for 134Cs and 90Sr was typically a factor of
the reactor is undamaged but that the effects of a criti-
ten lower. The maximum annual effective dose for adult
cality accident are difficult to predict, making the conse-
members of the public for the two cross-border receptor
quences difficult to estimate. Nevertheless, the risks of
areas assuming a `worst-case' scenario (NATO, 1998) is
radionuclide release to the Arctic are considered to be
shown in Table 7ท4. With dry conditions during the pas-
negligible.
sage of the radioactive cloud, the average individual ef-
fective radiation dose received in the first year is <1 mSv.
Rainfall during cloud passage may lead to enhanced
7.3.2. Civilian icebreakers
deposition of radioactivity, which would cause signifi-
The Murmansk Shipping Company operates the Russian
cantly higher long-term radiation doses.
icebreaker fleet. According to ุlgaard (2001) the fleet
currently comprises six operational icebreakers (Ark-
7.3.4. Storage of spent fuel
tika, Rossiya, Sovetskiy Soyuz, Yamal, Taymyr, and
Vaigach), and one icebreaking container ship (Sevmor-
Russian marine reactors and spent fuel are of interna-
put). These are stationed at the Atomflot Repair Techni-
tional concern. One hundred and eighty-eight nuclear
cal Plant near Murmansk. Two icebreakers have been
submarines have been decommissioned in Russia. Of
decommissioned and defuelled (Lenin and Sibir). A new
these, 48 have been dismantled, 28 are being disman-
icebreaker, 50 let Pobyedy (50 Years of Victory), is cur-
tled, and 112 are waiting for dismantling to start. Most
rently under construction at the Baltiysky shipyard in St.
still contain loaded reactors. While the focus on military
Petersburg.
nuclear issues and spent fuel began around 1990, the in-
famous service ship Lepse, containing more than 600
spent fuel assemblies, is still harbored near Murmansk.
7.3.3. Decommissioned,
According to the CEG (2003), fuel arising from 130
currently-fuelled submarines
submarine nuclear reactor cores is currently being stored
The decommissioning of Russian nuclear submarines in
in northwestern Russia, while fuel from an additional 20
the Arctic has caused considerable concern since the end
cores is located in far eastern Russia. An average reactor
of the Cold War. In 1992 and under the auspices of the
core has approximately 455 fuel assemblies. In Decem-

---

Chapter 7 ท Potential Accident Scenarios
87
Table 7ท4. Maximum annual effective dose estimates for adult members of the public for two cross-border re-
ceptor areas assuming a `plausible worst-case' accident scenario (NATO, 1998). `Short-term' refers to the first
24 hr of the event (cloud passage) and `long-term' to the first year excluding the first 24 hr. `Wet' refers to the as-
sumption of moderate rainfall during passage of the radioactive cloud.
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
Kirkenes (urban)
Finnmark (rural)
--------------------------------------------------------------------
dry
wet
dry
wet
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
Ground deposition of 137Cs, kBq/m2
10.*
250.*
1
25
Integrated air concentration of 137Cs, MBq s/m3
10
10
1
1
Effective dose from exposure pathway, mSv
Short-term**
Inhalation*** 0.19
0.19
0.02
0.02
Cloud-shine ญ
ญ
ญ
ญ
Ground-shine ญ
0.02
ญ
ญ
Short-term subtotal
0.19
0.21
0.02
0.02
Long-term
Ground-shine****
0.08
1.9
0.02
0.5
Ingestion***
0.03* 0.9*
0.19
4.5
Long-term subtotal
0.11
2.8
0.21
5.0
First-year
Total annual dose
0.30
3.0
0.23
5.0
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
**** average contamination of the wider surroundings of Kirkenes is set equal to 30% of the Kirkenes value;
**** no protection assumed in the early phase of the incident;
**** effective dose commitment;
**** corrected for runoff (urban environment only) and shielding (rural area lower than urban environment).
ber 2000, there were reports of large quantities of fuel
doses to a maximally exposed individual that might be
ready to be taken ashore (Moltz, pers. comm., 2000).
caused by the loss of a flask at sea were estimated to
The scenarios for an accident or inappropriate use of
range from 5 10 ญ12 Sv/yr for the loss of a vitrified high
a Russian marine reactor or its fuel are numerous, as ev-
level waste flask to the deep ocean, to 2 10ญ6 Sv/yr for
idenced by various incidents throughout the 1990s. For
the loss of a high burn-up irradiated fuel flask to shallow
example, the sinking of Komsomolets and the Kursk,
coastal waters.
several thefts of fresh fuel from bases in northwest Rus-
It is difficult to predict the long-term trend in such
sia, and an attempt to blow up the Vepr, an Akula Class
traffic. However, if mixed oxide fuel is increasingly used
submarine, by a distressed Russian sailor after a serious
as a means of safeguarding surplus weapons-grade plu-
hostage situation at the Gadzhiyevo Naval Base on 11
tonium and if investment in nuclear power generation
September 1998. The scenarios include: releases to air,
increases as a means of reducing dependence on fossil
sea and/or the terrestrial environment; sabotage and other
fuels and emissions of carbon dioxide to the atmos-
radiological incidents initiated deliberately; and thefts or
phere, the quantities and frequency of such shipments
other illegal, organized acquisitions of radiological or
may increase substantially. A seminar on the transport
fissile material by terrorists.
of spent nuclear fuel in Norwegian coastal areas con-
Earlier impact assessments concentrated on releases
vened for Norwegian senior officials in March 2002
from sunken submarines to the marine environment
concluded that, even if the calculated risk is low, there is
(Eriksen, 1990; IAEA, 1997) or releases from decommis-
a need for consideration of possible release scenarios
sioned, non-defuelled submarines to sea and air (NATO
and for detailed impact assessments. The possible trans-
1998). There is need for additional understanding of
fer of spent nuclear fuel through Arctic areas has caused
criticality issues related to remediation and clean-up ac-
controversy, for example in Norway, and will continue
tivities; damaged cores; and the types of spent fuel con-
to do so if such concerns are not addressed properly.
figurations currently stored at naval bases such as those
In the case of transport of spent fuel within, for ex-
at Andrejeva Bay and Gremikha Bay.
ample, Russian territory, there are potential problems
associated with Russian transport ships not adhering
strictly to international transport regulations. Any for-
7.4. International transport
of spent nuclear fuel from commercial use
eign assistance, such as the provision of Norwegian
transport ships for assisting Russian authorities in the
Between 1992 and 1999 there were six shipments of
dismantling of nuclear submarines, would probably de-
plutonium and vitrified high level radioactive waste
mand and ensure adherence to international regulations
from France to Japan and one shipment of mixed oxide
and standards (IAEA, 2001).
reactor fuel from the United Kingdom to Japan. Such
shipments, if carried out in a manner consistent with in-
7.5. Reprocessing and production plants
ternational guidance and existing IAEA Conventions
paying specific attention to the prevention of criticality
7.5.1. Mayak
accidents, pose only minor risks to human health. The
Operations at the Mayak PA installation have resulted
risk of accidents for such transport has been reviewed
in serious nuclear environmental contamination. Two
extensively over recent years in a comprehensive cooper-
accidents have resulted in severe contamination outside
ation between the IMO and IAEA (IAEA, 2001). The
the Mayak site boundary. In 1957, an explosion in a

---

88
AMAP Assessment 2002: Radioactivity in the Arctic
high level waste storage tank caused severe 90Sr contam-
2. A tornado in the Lake Karachay area. A tornado
ination of a 1000 km2 area within the Chelyabinsk,
passing over Lake Karachay lifts and disperses con-
Sverdlovsk and Tyumen regions. This is referred as the
taminated water and sediment over the surrounding
`Kyshtym accident'. In 1967, wind dispersal of contami-
area in a similar manner to the events of 1967.
nated sediment from the dried-out bed of Lake Kara-
3. Inflow of water from Reservoir 11 to the Techa River
chay (a storage reservoir for liquid radioactive waste) re-
due to:
sulted in 137Cs deposition over 1800 km2 surrounding
a. a dam break, which brings dissolved and particu-
the site. Between 1949 and 1956, authorized discharges
late radionuclides as well as washout from the Techa
of intermediate-level radioactive waste directly into the
riverbed and floodplain into the river system;
Techa River resulted in severe contamination downstream
from the release point. Although operational procedures
b. a controlled release that results in a discharge of
have been revised extensively since the late 1950s, as has
dissolved radionuclides from Reservoir 11 into the
also been the case at other nuclear installations, the pos-
Techa River.
sibility of accidents remains. The human population in
4. Release of radionuclides from the Asanov Swamp.
the vicinity of Mayak is at most risk from an accident
This was heavily contaminated by early operational
and has, together with the environment, suffered the ad-
discharges of radionuclides into the Techa River, due
verse effects of previous accidents. However, since the
to flooding.
Mayak installation is sited at the head of the Techa
5. An accident at the reprocessing plant. This is compa-
River, which is a tributary of the major Ob River, there is
rable to scenario 1, although on a smaller scale and
also the possibility of long-range transport of radionu-
with other radionuclides being involved.
clides to Arctic areas. AMAP has therefore recommended
studies on the transport of radionuclides from land-
6. Groundwater contamination from Lake Karachay
based sources through river catchments (AMAP, 1998).
reaches the river system.
The possible consequences of far-field transport of ra-
dionuclides released as a result of various hypothetical
The accidents vary in size, impact, and duration. Some
accidents at the Mayak installation have been assessed
allow time for the introduction of measures to reduce
by the Joint NorwegianญRussian Expert Group on Ra-
their severity; others represent serious, acute accidents
dioactive Contamination (JNREG, 2003). The study fo-
(e.g., a dam failure) that allow little possibility of mitiga-
cused on six accident scenarios.
tion. All incidents have the potential to release radionu-
clides that could result in impacts on biota and humans
1. An explosion in a storage tank for high level waste.
in the surrounding area, both in the near and far field.
This is a modern analogue of the Kyshtym accident. It
Because of the concern regarding long-range river
results in radioactive contamination of the environ-
transport, a major focus has been to model the transport
ment and subsequent washout of radionuclides into
of radionuclides through the Techa-Iset-Tobol-Irtysh-Ob
the river system.
River system to Ob Bay and the Kara Sea. In some cases,
Table 7ท5. Consequences for Arctic areas of six hypothetical contamination scenarios at Mayak PA (JNREG, 2003).
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
Maximum
Collective
dose
Release to
Discharge to
dose,
per person,
Total inventory
environment
the Techa River
manSv
mSv/yr*
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
Current runoff
1.2 TBq/yr 90Sr 0.6
TBq
90Sr for 50 yr
0.01
0.009
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Scenario 1. Waste tank explosion 370 PBq ญ single tank;
15.2 PBq 90Sr + 90Y
180 TBq 90Sr
0.39
1.9
20000 PBq ญ total
20.4 PBq 137Cs
59 TBq 137Cs
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Scenario 2. Tornado
4400 PBq
4.4 PBq 90Sr + 137Cs
5 TBq 90Sr, 0.005
0.006
0.5 TBq 137Cs
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Scenario 3a. Dam burst
In water: 650 TBq
300 TBq 90Sr 300
TBq
90Sr
1.0
4.8
3.7 TBq 137Cs
3.7 TBq 137Cs
In sediment: 1500 TBq
205 TBq 90Sr 205
TBq
90Sr
150 TBq 137Cs
150 TBq 137Cs
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Scenario 3b. Controlled release
650 TBq 90Sr + 137Cs
13 TBq 90Sr
13 TBq 90Sr
0.009
0.05
0.16 TBq 137Cs
0.16 TBq 137Cs
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Scenario 4. Asanov Swamp
19-22 TBq 90Sr
3.2 TBq 90Sr
3.2 TBq 90Sr
0.002
0.01
170-190 TBq 137Cs
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Scenario 5. Plant accident
1.1 TBq**
0.0007
0.004
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Scenario 6. Groundwater
4400 PBq
22 TBq/yr 90Sr 0.6
TBq
90Sr for 50 yr
0.00008
0.00007
contamination
ญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญญ
*for a diet containing 28 kg fish per year; **estimate for all radionuclides.

---

Chapter 7 ท Potential Accident Scenarios
89
other transport processes (e.g., atmospheric transport of
design features that can mitigate an accident before it
radioactive debris) have also been considered. The mod-
reaches a critical state. In the event of a design-based ac-
els, developed by scientists at Mayak PA and SPA Ty-
cident, the safety assessment principles state that `there
phoon, include radionuclide transport in river systems,
is no release of radioactivity except in the most severe of
and tornado, flood, and groundwater contamination.
cases, and even then, no person will receive an effective
Where possible, the models are based on existing scien-
dose of 100 mSv or more'.
tific knowledge concerning the transport and behavior
The storage of Highly Active Liquor (HAL) is an im-
of radionuclides in the area surrounding Mayak and on
portant source of concern for severe accidents at the Sel-
the outcome of previous accidents, both at Mayak and
lafield site. HAL is a waste product from the reprocess-
other installations. Information on the physico-chemical
ing of irradiated nuclear fuel and is currently stored on-
forms of radionuclides and the influence of speciation
site in water-cooled storage tanks. It is converted into
on transport processes and mechanisms was also in-
solid form via the process of vitrification (incorporation
cluded. Finally, major uncertainties, variability, and model
into borosilicate glass) at a rate limited by the capacity
sensitivity has been assessed.
of the vitrification plant. The Nuclear Installations In-
The models required particular variables (release in-
spectorate has instructed British Nuclear Fuels (BNFL)
ventory, radionuclide composition, meteorological con-
to vitrificate HAL from a current volume of about 1300 m3
ditions, etc.) for each scenario. Each scenario will vary
(1999) to a buffer volume of 200 m3 (to feed the vitrifi-
according to the course of events, particularly concern-
cation process) by 2015 in response to the potential haz-
ing the quantity of radionuclides released, which could
ard associated with these wastes (HSE, 2000). The main
be more or less than the hypothetical estimate. There-
part of the activity in a typical HAL tank is due to 137Cs
fore, the estimates derived using the models have large
and 90Sr. There are in total 21 tanks containing about
uncertainties. Worst-case conditions were generally con-
7 1018 Bq of 137Cs and 4.8 1018 Bq of 90Sr (Turvey
sidered for each scenario.
and Hone, 2000). Vitrified wastes are also stored on-site
The outcome of the modelling exercise is compared
at Sellafield and are generally thought to be safer than
with current run-off in Table 7ท5. For each scenario, the
HAL because the fission products are immobilized in a
table presents estimates for the total radionuclide inven-
solid matrix and cooled by the circulation of air. They
tory and estimates of radioactive releases to the environ-
are thus not dependent on an active cooling system.
ment outside the Mayak PA area. Radioactive discharges
A major BNFL safety case for HAL stores was com-
to the upper Techa-Iset-Tobol-Irtysh-Ob River system
pleted in 1994. This was followed by a Nuclear Installa-
(mainly into the Techa River) can be much lower than
tions Inspectorate assessment (HSE, 2000). The assess-
those to the environment, if for example, contamination
ment concluded that the BNFL approach to accident
is due to washout from the water catchment area. Doses
analysis was incomplete and not best practice. In 1999,
were estimated for Ob Bay and the Kara Sea for a 50-
BNFL completed the Continued Operation Safety Re-
year period after the hypothetical accidents.
port (COSR), which is the latest safety analysis associ-
Scenarios 1 and 3a result in a very high radioactive
ated with the HAL stores. Although this report is not
discharge to the Techa River. The other scenarios result
publicly available, the Radiation Protection Institute of
in much lower radioactive contamination of the Techa
Ireland (RPII) was given access to the BNFL safety docu-
River.
mentation and has published an evaluation report of the
The models indicate that 90Sr transport through the
COSR (Turvey and Hone, 2000). The objectives of the
river system will lead to a significant increase in contam-
RPII examination of the safety material were to deter-
ination in the lower reaches of the Ob River compared
mine whether the COSR includes all significant hazards;
to current levels. For example, the additional 90Sr activ-
to evaluate the conclusions of the COSR on the proba-
ity concentration for the first year after the dam break is
bility of occurrence of a number of accident scenarios; to
estimated to be five times higher than background. Con-
determine whether confidence can be placed in the data-
tamination of Ob Bay and the Kara Sea by other ra-
base used in the COSR and to assess the significance of
dionuclides is much lower. The longer-lived radionu-
any shortcomings; and to assess the need for further im-
clides released, 137Cs and Pu, are less mobile in river sys-
provements in safety. Turvey and Hone (2000) conclude
tems than 90Sr. For all six scenarios, the estimated activ-
that the risks of a severe accident associated with the
ity concentration is much lower than the norms regu-
HAL stores are low but identify some areas where the
lated by modern radiation safety standards. Overall, it
risks could be reduced further. The report also states
was concluded that the potential doses to Arctic biota
that the risk of damage from a severe earthquake has not
and human populations from hypothetical accidents at
been fully analyzed. According to Turvey and Hone
the Mayak PA installation are very low. However, for
(2000), all other major accident scenarios appear to
the local population, the consequences may be severe.
have been considered in the COSR. Despite the proba-
bility of an accident involving a significant release of ra-
dioactivity being considered low, Turvey and Hone
7.5.2. Sellafield
(2000) identify certain safety weaknesses, e.g., that the
The U.K. Health and Safety Executive have produced
water supplies for cooling the tanks are not fully inde-
safety assessment principles for nuclear plants (HSE,
pendent of each other, that there is no instrumentation
1992) that address safety issues, including accident sce-
for detecting possible hydrogen build-up in the storage
narios for the Sellafield site. There are a number of prin-
tanks, and that the consequences of very severe acci-
ciples to ensure that safety is maintained throughout op-
dents have not been adequately assessed.
erations and in the event of design or beyond-design
Low probability but high consequence events appear
based accidents. During operation there are a number of
to pose the greatest environmental risk at Sellafield, and

---

90
AMAP Assessment 2002: Radioactivity in the Arctic
could even impact upon the Arctic. These include: seis-
may be compared in order to help identify where to
mic events; fire or explosion due to hydrogen generation
focus efforts for risk reduction. The outcome of risk as-
as a result of radiolysis of HAL or red oil reactions (hot
sessments and actual accidents indicate that the conse-
organic liquid and aqueous nitrate solution); extreme
quences of releases to the atmosphere, and subsequent
weather conditions; aircraft crashes; other man-made
fallout to the terrestrial environment, are greater than
hazards (toxic gases); criticality; beyond-design basis ac-
for releases to the marine environment.
cidents; and accidents as a result of human factors.
This assessment has addressed the unintentional po-
These issues have been considered by BNFL and the
tential releases from reprocessing plants in central and
U.K. Health and Safety Executive (HSE, 1992, 2000)
southern Russia in detail. The first AMAP assessment
but estimates of the radiological consequences (i.e., radi-
concluded that possible consequences of accidents at
ation doses) of each accident scenario have not been as-
these plants should be assessed for the Arctic population
sessed.
and environments, owing to the possible transport of ra-
After 11 September 2001, the possibility of a terror-
dionuclides through the river systems. The present as-
ist attack on, or an airplane crash into, nuclear plants
sessment shows that the consequences of such accidents
has received much attention. A report prepared for the
for the Arctic are likely to be much less than previously
European Parliament by an external contractor, WISE-
expected.
Paris (WISE-Paris, 2001a), mentions this briefly and ad-
That many of the sources to be evaluated in risk as-
dresses the subject in greater detail in a later report
sessments are within the military domain, e.g., naval re-
(WISE-Paris, 2001b). The HAL stores are identified as
actors and nuclear weapons, is a problem. Necessary in-
the major risk for radioactivity releases. An assessment
formation is often restricted. Openness regarding mili-
is made based on an estimated release of half the total
tary sources should be promoted, such that risks to soci-
137Cs content in the HAL tank, and then compared with
ety as a whole can be compared and resources for risk
consequences from the Chernobyl accident. However,
reduction programs used optimally.
these estimations are controversial and have received
The increased awareness of terrorist activities since
some criticism.
2001 has also forced the nuclear industry to reassess the
probability and consequences of a terrorist event. Al-
though AMAP does not address security issues, and this
7.6. Conclusions
matter has therefore not been discussed further, it should
Risk assessments are important for establishing priori-
be noted that, with negative intentions, the results of an
ties. Even though the absolute results from these assess-
`accident' could be worse than those estimated in the
ments have large uncertainties, their relative magnitudes
present scenarios.

---