Journal of Marine Systems 24 Ž2000. 233–248
Transport of radionuclides by sea-ice and dense-water formed in
western Kara Sea flaw leads
D. Dethleff a,) , H. Nies b, I.H. Harms c , M.J. Karcher d
GEOMAR Research Center for Marine Geosciences, Wischhofstraße 1-3, D-24148 Kiel, Germany
Federal Maritime and Hydrographic Agency, Hamburg, Germany
Institute of Marine Sciences, UniÕersity of Hamburg, Hamburg, Germany
Alfred-Wegener-Institute for Polar and Marine Research, BremerhaÕen, Germany
Received 9 June 1998; accepted 20 September 1999
A transport assessment of particle-bound and dissolved artificial radionuclides Ž137Cs and 239,240 Pu. by sea-ice and
dense-water formed in western Kara Sea flaw leads close to the Novaya Zemlya dumping sites is presented in this study. We
both performed a ‘‘best estimate’’ based on available data, and a ‘‘maximum assessment’’ relying on simulated constant
releases of 1 TBq 137Cs and 239,240 Pu from individual dumping bays. The estimates are based on a combination of Ži. the
content of particulate matter in sea-ice; Žii. analytical data and numerical simulations of radionuclide concentrations in shelf
surface deposits, suspended particulate matter ŽSPM., and the dissolved phase; and Žiii. estimates of lead-ice and
dense-water formation rates as well as modeling results of local ice drift pathways. In the ‘‘best estimate’’ case, 2.90 GBq
Cs and 0.51 GBq 239,240 Pu attached to sea-ice sediments can be exported from the lead areas toward the central Arctic
basin. The radionuclide burden of the annually formed dense lead water in the ‘‘best estimate’’ amounts to 4.68 TBq 137Cs
and 0.014 TBq 239,240 Pu. In the ‘‘maximum assessment’’, potential export-rates of ice-particle bound 137Cs and 239,240 Pu
toward the central Arctic would amount to 0.64 and 0.16 TBq, respectively. As much as f 900 TBq 137Cs and f 6.75 TBq
Pu could be annually taken up by 34.75 dense-water rejected in the lead area. Assuming the Žunlikely. instantaneous
release of the total 137Cs and 239,240 Pu inventories Žf 1 PBq and 10 TBq, respectively. from the Novaya Zemlya dumping
sites into the dissolved phase, the dense lead water locally formed during one winter season could take up f 90% of the Cs
and f 68% of the Pu released. q 2000 Elsevier Science B.V. All rights reserved.
Keywords: Kara Sea; flaw leads; ice sediments; dense-water; radionuclide transport; ablation area contamination
1. Introduction
From 1959 until the late 1980s the former Soviet
Union dumped large amounts of solid and liquid
Corresponding author. Tel.: q49-431-600-2805; fax: q49431-600-2941.
E-mail address: [email protected] ŽD. Dethleff..
radioactive waste in the Kara Sea. Besides the Novaya Zemlya trough ŽFig. 1., main dumping areas
were located in shallow bays Žmostly 10 to 50 m
water depth. along the eastern coast of Novaya
Zemlya ŽFig. 2. ŽYablokov et al., 1993.. Disastrous
leakages of radioactive material from the dumped
containers and reactors have not been reported yet
ŽJoint Norwegian–Russian Expert Group, 1996.,
0924-7963r00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 4 - 7 9 6 3 Ž 9 9 . 0 0 0 8 8 - 3
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
Fig. 1. Bathymetry in meters of the Kara Sea. The small inset shows the predominating ice drift systems ŽBG s Beaufort Gyre;
TPD s Transpolar Drift. of the Arctic Ocean. Circles indicate sampling sites of surface deposit along the east coast of Novaya Zemlya and
the investigation area off the coast of Cape Yugorskiy in April 1997. For more details see Fig. 2.
however, the total recent inventory of 1 PBq 137Cs
and 10 TBq 239,240 Pu may cause significant local and
regional ecological problems in the case of a catastrophic — i.e., instantaneous — release.
Different recent empirical and modeling studies
focus on the pathways of man-made radionuclides
with Arctic sea-ice and ocean water masses particularly originated from the Kara Sea Že.g., Pavlov and
Pfirman, 1995; Harms, 1997a; Meese et al., 1997;
Baxter et al., 1998; Krosshavn et al., 1998; Landa et
al., 1998; Mitchell et al., 1998.. Pavlov and Stanovoy
Žin press. pointed out that ice bergs calving from
northern Novaya Zemlya glaciers may damage
dumped containers in the bays along the eastern
coast of the island thereby causing dramatic releases
of radioactivity which could locally be entrained into
newly forming ice and exported from the Kara Sea.
Pfirman et al. Ž1997a. also pointed to sea-ice as a
rapid transport mechanism of dissolved and particlebound pollutants from the Kara Sea via the Transpolar Drift to Fram Strait and the northern North
Atlantic. Meese et al. Ž1997., Cooper et al. Ž1998.
and Landa et al. Ž1998. report enhanced 137Cs contaminations in central Arctic sea-ice sediments which
might have originated from the shallow Siberian
shelves and, particularly, from the Kara Sea.
2. Background and purpose
The occurrence of ice producing flaw leads Žzones
of open water between shore-fast ice and drifting
ice. along the eastern coast of Novaya Zemlya and
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
Fig. 2. Detailed map of the western Kara See showing the 1997 expedition area ŽA. and the bathymetry of the Novaya Zemlya dumping
bays ŽB, C and D.. The gray discolored regions - 50 m indicate the potential areas of sediment entrainment into newly forming sea-ice.
Vaygach Island ŽFig. 3. is attributed to predominating westerly winds Že.g., Martin and Cavalieri, 1989;
Dethleff and Reimnitz, 1996.. According to recent
Russian investigations the area east and south of
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
Fig. 3. NOAA-12 satellite image Ž4-10-1993. showing recurrent, coastal flaw leads in the western Kara Sea. Dashed lined boxes A–C
represent the east Novaya Zemlya lead sections Žadapted from Martin and Cavalieri, 1989., box D contains the AmdermarVaygach flaw
lead investigated during April 1997 field work. The main dumping bays along the east coast of Novaya Zemlya are indicated by
radioactivity-labels marking south to north Abrasimov, Stepovogo, Tsivolky and Technyia Bays. Dark gray and black lines display modeled
seasonal forward ice trajectories as reworked from Nies et al. Žin press.. Dots mark starting points.
Novaya Zemlya is characterized by very high probabilities of flaw lead recurrence Žabout 50%; V.K.
Pavlov, 1996, AARI St. Petersburg, personal communication.. The dynamics, and, particularly, the
production of new ice and dense-water of Siberian
flaw leads contributing to the Arctic Ocean ice and
deep water budgets, are described in more detail by,
e.g., Zakharov Ž1966., Martin and Cavalieri Ž1989.,
Cavalieri and Martin Ž1994., and Dethleff et al.
Ž1998a.. Accordingly, Siberian flaw leads produce as
much as 8 times Žf 17 m. more new ice than the
remainder of the shelf areas through the mechanism
of permanent leeward advection. The total drift-ice
volume annually produced in the Kara Sea is roughly
900 km3 Žafter Pavlov et al., 1994.. The annual
volume of the net ice-export from the Kara Sea
toward the Arctic Basin varies between 170 km3
ŽPavlov et al., 1994. and 270 km3 ŽZakharov, 1976.,
while the exchange through Kara- and Vilkitskyi
Straits toward the adjacent shelves of the Barents
and Laptev Seas is less important.
The water exchange of the Kara Sea with the
central Arctic Basin shows a northward flow of as
much as 0.6–0.7 Sv, which is mainly compensated
by the inflow of Atlantic water through Kara Gate
ŽPavlov et al., 1994.. Mitchell et al. Ž1998. investigated the radionuclide content of water masses passing through St. Anna and Voronin Troughs and
found no evidence of enhanced activities arising
from the waste dumping in the Kara Sea. However,
extremely little is known about the uptake and transport of radioactive contaminations by dense brines
rejected subsequent to ice extraction in flaw leads.
According to Reimnitz et al. Ž1993b. the entrainment of Arctic shelf surface deposits into newly
forming ice mainly occurs in shallow, near coastal
regions - 50 m water depth. The process of suspension freezing Žscavenging of fine-grained suspended
particulate matter ŽSPM. from the water column
through buoyant rising frazil ice crystals, and the
entrainment of sediment by upward floating material-laden anchor ice; Fig. 4. is generally evaluated
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
Fig. 4. Hypothetical cross-section of the AmdermarVaygach flaw lead displaying possible hydrodynamic processes of turbid ice formation.
as the most effective entrainment mechanism on the
shallow Arctic shelves Že.g., Reimnitz and Bruder,
1972; Osterkamp and Gosink, 1984; Reimnitz et al.,
1992, 1993a; Dethleff et al., 1994; Nurnberg
et al.
1994; Dethleff et al., 1998b.. Extended coastal areas
of the Kara Sea with water depths less than 50 m are
mainly covered by fine-grained surface deposits ŽGeogruppen, 1994. and, thus, are predominated for
sediment entrainment into new ice. The entrainment
of sediment into newly forming flaw lead ice in the
southwestern Kara Sea was reported by Dethleff et
al. Ž1998b..
The purpose of this study is to provide a crude
assessment of sea-ice rafted vs. water-column transport Žexport. of artificial radionuclides Ž137Cs and
Pu. from western Kara Sea flaw leads. In this
context, ‘transport’ means the pure dislocation of
radionuclides from the lead areas after entrainment
into ice and dense-water without considering the
further fate of the transport media in more detail
Že.g., ice melt during summer or ice drift toward the
Arctic basin; dissolution or remixing of dense-water,
etc... On the other hand, the term ‘export’ exclusively considers the leave of lead-ice entrained, particle-bound radionuclides from the western Kara Sea
toward the Arctic basin.
In this paper we estimate Ži. the entrainment and
export of particle-bound radionuclides from the
western Kara Sea with newly formed lead-ice and
Žii. the uptake and dislocation Žtransport. of dissolved radionuclides by lead dense-water formed
subsequent to the ice extraction. We give a ‘‘best
estimate’’ based on available sedimentological and
radiochemical data from the western Kara Sea, and
we further provide a ‘‘maximum assessment’’, which
relies on simulated release scenarios from the No-
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
vaya Zemlya dumping bays ŽHarms, 1997a.. In order
to achieve our goal we use Ži. sedimentological data
of surface deposits and sea-ice incorporations, Žii.
radiochemical data of surface sediments, and Žiii.
empirical, simulated, and calculated particle-bound
and dissolved radionuclide concentrations. These data
sets are combined with 4-year mean-rates of ice and
dense-water formation ŽMartin and Cavalieri, 1989.
from the east Novaya Zemlya leads under consideration of own sea-ice drift modeling-results ŽNies et
al., in press..
3. Material and methods
3.1. Surface deposits and sea-ice sediments
Surface deposits and sea-ice sediments from the
AmdermarVaygach flaw lead area ŽFigs. 2 and 3.
were collected during the joint RussianrGerman
KaBaEx ’97 expedition ŽDethleff et al., 1998b.. The
samples were taken at water depths ranging from 6
to 41.5 m. A total of 15 surface Žinterval: 0–5 mm.
and 7 mixed surfacersubsurface Žinterval: 0–30 mm.
sediment samples were obtained at eight sites. The
Norwegian Radiation Protection Agency provided
surface sediments from the Novaya Zemlya dumping
Ice cores were taken at seven stations. After
melting, the core sections were filtered using preweighted, mixed-ester membrane filters with 0.45
mm pore diameter. The filtered material was freezedried for further sedimentological investigations. Additional sea-ice sediments were collected from pressure ridges.
Both shelf surface deposits and sea-ice sediments
were prepared for smear slide analyses in order to
determine the quantitative and qualitative sample
composition under the microscope.
3.2. SPM
SPM was collected at eight sites in the AmdermarVaygach lead area during the April 1997
fieldwork. A total of 19 SPM samples was taken at
seven sites by Niskin bottle in three different water
depths: Ža. close to the surface, Žb. in the middle of
the water column and Žc. near the bottom.
3.3. Radionuclides
Water volumes varying between 76 and 93 l were
collected at 3 stations during KabaEx ’97 expedition
in order to determine 137Cs. The sampled water was
acidificated with HCl to a pH of as low as 2 and run
over an exchanger resin of potassium-hexacyano-ferrate-ŽII.-cobaltateŽII. ŽKCFC., thereby absorbing Cs
ions from sea-water with a chemical yield of ) 95%.
Mixed surface sediment samples Žca. 400–800 g
wet weight. were freeze-dried and homogenized.
Both sediment and dissolved Žion-exchanged. samples were filled into beakers with a calibrated geometry. For g-spectrometric analysis we used high purity germanium detectors ŽHPGe.. The samples were
analyzed for artificial g-emitting radionuclides such
as 137,134 Cs, 60 Co, 241Am and different natural radionuclides. Simulated radionuclide data used in our
calculations were taken from the literature Žsee below..
3.4. Ice and dense-water
Ice and dense-water formation rates as well as
ice-drift modeling results were taken from the literature and our own work Žsee below..
4. Results
4.1. Sedimentological data
4.1.1. Bottom deposits
Smear slide analyses reveal that surface sediments
in the AmdermarVaygach area vary between silt
and sand, while the clay fraction is less abundant.
Highest silt and sand percentages amount to 85% or
even more, whereas the clay percentage generally is
- 20% ŽFig. 5.. The qualitative sample composition
reveals high percentages Ž65–80%. of mainly angular to subrounded quartz and feldspar, while rounded
clastic particles, rock fragments, mica, biogenic
components and opaque minerals are less abundant
or even absent.
Surface deposits of the Novaya Zemlya dumping
bays are extremely fine-grained ŽFig. 5. with up to
99% in the fraction - 63 mm. Sand contents vary
between 17% and 65%, while in ) 90% of the
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
Fig. 5. Sand–silt–clay content of surface deposits and sea-ice sediments of the AmdermarVaygach lead area, the Novaya Zemlya dumping
bays, and ice of the central Arctic Ocean.
samples silt Ž48–73%. and clay Ž25–49%. are the
dominant grain size fractions. The fine-grained material consists mainly of angular clasts.
4.1.2. SPM
SPM concentrations in the AmdermarVaygach
lead area ranged from 0.55 mgrl in the under-ice
surface layer to 13.77 mgrl in the nepheloid layer
close to the shelf bottom. Results from binocular
investigations reveal that the SPM consists mainly of
silt-sized clasts. Percentages of the coarse fraction
vary, and clay-sized material is less abundant. The
silt fraction is dominated by angular quartz particles.
The mean SPM load amounts to 3.17 mgrl.
4.1.3. Sea-ice sediments
Drift ice in the AmdermarVaygach flaw lead
area was generally more turbid than the shore-fast
ice, indicating that drift ice was formed under more
turbulent conditions. All ice samples Žfor locations
see Fig. 2A. consisted of first year ice and thicknesses varied between 70 and 170 cm, where the
latter was the result of recent deformation ŽDethleff
et al., 1998b.. Most of the particulate material was
contained in the uppermost 30–60 cm Žgranular ice.
of the cores. The visible, fine-grained sediments
occurred either evenly disseminated, were enriched
in layers or appeared in patchy aggregates. The
material content in the cores ranged from 2 to 35
mgrl, reaching an extreme of f 140 mgrl at station
5 Žcore section 12–22 cm, Fig. 6.. The mean sediment load of all core sections considered was 9.91
Smear slide analyses of particulate matter extracted from both fast- and drift-ice cores reveal
extremely high percentages in silt and clay fractions
Ž85–95%., while the sand fraction generally is less
abundant Žsee Fig. 5.. Thus, sea-ice sediments in the
SW Kara Sea are much finer grained than the underlying shelf deposits. According to our binocular estimates, as much as 80% of each individual sample is
composed of angular to subrounded clastic material
Žmainly quartz and feldspar. with highest abundances in the silt fraction, which represents on average 64.2% Žrange: 30–85%. of the material. The
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
suspension freezing and the sediment composition in
Beaufort Sea ice, thus, is finer than that of the
underlying shelf source deposits. According to what
we learned from the Reimnitz et al. studies, the
sedimentological character of the SW Kara Sea iceparticles described above implies that the material
was directly entrained from local bottom sources by
turbulent interaction with frazil ice crystals. To the
contrary, indications for anchor ice entrainment were
sparse in the SW Kara Sea. As deduced from the
results gained in the AmdermarVaygach flaw lead
area, the fine-grained bottom material in the Novaya
Zemlya dumping bays must also be regarded as
predestinated for resuspension and turbulent entrainment into newly forming ice.
4.2. Radionuclide data
Fig. 6. Vertical distribution of particulate matter in ice cores of the
AmdermarVaygach region. Note that the lower x-axis refers to
station 5 only due to enhanced particle concentration.
degree of roundness Žangular to subrounded. of fine
particles in sea-ice sediments resemble well the particle shape we found in bottom sediments and SPM.
Well-rounded particles generally do not occur in the
sea-ice sediments. Organic-clay–iron aggregates,
clay minerals and microorganisms partly appear in
slightly enhanced portions of 5–25%. Other clastic
material or biogenic components, such as heavy and
opaque minerals as well as plant debris, are generally
rare. The less abundant coarse fraction Ž5–15%. is
mainly composed of aggregates consisting of finegrained material, idiomorphic gypsum minerals and
varying biogenic material.
According to laboratory studies ŽReimnitz et al.,
1993a. and investigations from the Beaufort Sea
ŽReimnitz et al., 1998., fine-grained clastic material
is preferentially entrained into newly forming ice by
The water samples taken in the SW Kara Sea lead
area have 137Cs activities between 4.9 and 5.3
Bqrm3. The determined Cs activities were slightly
above the expected ‘‘background’’ concentrations of
about 2.5 Bqrm3 in the Northern hemisphere, which
still originates from the global radioactive atmospheric fallout subsequent to the nuclear weapon
tests performed mainly during the 1960s.
The surface sediments of the AmdermarVaygach
flaw lead area contained 137Cs activities between 0.3
and 20 Bqrkg dry weight of the bulk sample. The
slightly enhanced 137Cs contaminations must be regarded as remnants of stronger polluted Sellafield
discharges in former times. Thus, the transport of
radioactivity from the Irish Sea to the Arctic Ocean
is still seen in Kara Sea surface deposits. A pollution
of the AmdermarVaygach area due to leakages in
the Novaya Zemlya dumping sites or by discharge
from the adjacent Ob and Yenisej rivers is — according to our data — not evident.
The highest concentrations of 137Cs were found at
station 8 ŽFig. 2. where the surface deposits contained ) 90% coarse material, so that the fine fraction Ž- 10%., which is supposed to bond most of the
radionuclides, at this site must have been significantly enriched in Cs. Since fine particles are preferentially entrained from the shelf bottom into newly
forming ice, sea-ice sediments thus may be generally
stronger polluted than the shelf source sediments.
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
This was already hypothesized by Cooper et al.
Ž1998. and can be supported by findings from Landa
et al. Ž1998., who determined 137Cs concentrations of
) 70 Bqrkg in fine-grained central Arctic sea-ice
sediments, while the surface deposits in the western
Laptev Sea source area identified by backward trajectories has much lower radionuclide burdens ŽCooper et al., 1998; Pavlov et al., 1999.. Additionally,
a positive correlation between the percentage of fine
fraction and the 137Cs burden in sea-ice sediments
was found at varying sites in the central Arctic
Ocean ŽBaskaran et al., 1996; Føyn and Svaeren,
1997; Landa et al., 1998.. This relationship means
that the more polluted fine fraction sea-ice sediments
contain, the higher is their radionuclide burden, which
underlines again the importance of fine-grained Arctic sea-ice sediments for contaminant transport.
5. Assessment
5.1. Assumptions
Since ice formation and sediment entrainment
processes in the shallow dumping bays along the east
coast of Novaya Zemlya are still unknown and cannot be directly estimated yet, we apply a set of
simplifying rules and assumptions on which our
assessments of radionuclide-entrainment, transport
and export are based.
In the ‘‘best estimate’’, sediment-bound and dissolved radionuclides will be entrained into newly
forming lead-ice and dense-water as available in the
flaw lead areas according to recent data. In the
‘‘maximum assessment’’ we allow radioactively polluted water masses and fine-grained particles both
released from the Novaya Zemlya dumping bays to
reach the lead areas ŽFig. 3; sections A, B, C..
Harms Ž1997a. simulated that the surface waters off
the Abrasimov Bay showed enhanced activities of as
much as 1000 Bqrm3 Žvs. 4–5 kBqrm3 inside the
bay. subsequent to a constant release of 1 TBqryear
Cs. Under offshore Žsouthwesterly. wind conditions the adjacent flaw leads are maintained ŽMartin
and Cavalieri, 1989., and particle-bound and dissolved radionuclides released from the bays ŽHarms,
1997a. may enter the lead area and will be entrained
into newly forming ice and rejected dense-water.
According both to results of forward ice trajectory
simulations ŽFig. 3. and numerical model simulations
ŽRigor, 1997, personal communication. we assume a
probability of 80% for the Žradionuclide-laden. leadice formed in section A to leave the Kara Sea toward
the Arctic Basin. Ice from sections B and C was
assumed to have probabilities of 60% and 30%,
respectively, to leave the Kara Sea. Lead-ice formed
in section D has only a probability of about - 10%
ŽRigor, 1997, personal communication. of leaving
the Kara Sea toward the north within one winter.
Thereby, lead section D is excluded from the assessments and will only be used to show the recent
radionuclide situation in the western Kara Sea, and
to illustrate the sedimentological processes active in
local flaw leads. Conclusively, the calculations of
radionuclide export from the western Kara Sea toward the Arctic Basin are based merely on sections
A, B and C.
We adopted 4-year mean ice formation and
dense-water production rates of the east Novaya
Zemlya flaw lead sections A–C ŽFig. 3, Table 1.
from Martin and Cavalieri Ž1989.. We furthermore
considered the sediment to be entrained into newly
forming lead ice only by frazil crystals, since the
entrainment through anchor ice in average amounts
to - 10% ŽEd Kempema, 1998, personal communication.. Since by definition at maximum the upper
60–70 cm of an ice sheet was potentially formed
under turbulent conditions in a flaw lead Že.g., Zakharov, 1966., we considered only the material load
incorporated in the uppermost part of the ice cores
obtained from lead-section D for the sediment entrainment and transport budgets. The resulting mean
value of 11 mgrl Žf 11 grm3 f 11 = 10 3 trkm3 .
Table 1
Seasonal sediment entrainment and export rates of the western
Kara Sea flaw leads
Lead Ice
Sediment Sediment
section area
volume concentration entrained exported
Žkm2 . Žkm3 .
1200 17
1800 11
3450 35
4-year mean from Martin and Cavalieri, 1989.
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
was combined with the ice volumes formed in the
proper lead sections Žsee Table 1.. Since no sea-ice
sediment data from the east coast of Novaya Zemlya
are available, we assumed the sea-ice sediment content in section D also as representative for sections
Table 3
Worst case parameters for the east Novaya Zemlya flaw lead area
Kd factor
Water concentration
Specific activity sediment surface
1 PBq
3000 Žlrkg.
1000 Bqrm3
3000 Bqrkg
10 TBq
50,000 Žlrkg.
7.5 Bqrm3
750 Bqrkg
5.2. Lead-ice sediments
The results of our estimates ŽTable 1. show that
the lead area B entrains the highest amount of sediment into new ice, whereas sections A and C produce less sediment-laden sea-ice due to lower ice
production rates. A total of roughly 0.39 = 10 6 t of
sediment can be annually entrained into eastern Novaya Zemlya lead ice. The portion of lead-ice sediments which may leave the Kara Sea toward the
north and will be exported to the central Arctic basin
amounts roughly 210,000 tryear. The difference between entrainment and export rates of lead-ice sediments is due to the above-assumed reduced probabilities for the lead-ice produced to leave the Kara Sea.
Table 2
‘‘Best estimate’’ and ‘‘maximum assessment’’ of 137Cs and
Pu export rates in sea-ice sediments from the western Kara
Sea flaw leads
Lead section
Best estimate
Measured data ŽBqrkg. Export ŽGBqa .
Cs b
Lead section
Maximum assessment
Calculated data ŽBqrkg. Export ŽGBqa .
Cs d
GBqs1=10 exp 0.9 Bq.
Shelf surface deposits contamination, own analyses.
Surface deposits, mean data from Joint Russian–Norwegian
Expert Group Ž1996; Figs. 4.10, 4.22, 4.27, 4.28..
Calculated from worst case release scenarios.
In order to give an estimate of ice-particle bound
Cs and 239,240 Pu export from the east Novaya
Zemlya leads toward the Arctic Basin, we combined
our sea-ice sediment export-calculations ŽTable 1.
with radionuclide levels in shelf source sediments
close to the flaw lead sections A–C ŽTable 2.. The
‘‘best estimate’’ contamination levels of 137Cs and
Pu ŽBqrkg. in surface sediments Župper 2 cm.
are based on measurements. For 137Cs we used individual data from own analyses, while 239,240 Pu values were taken from Joint Norwegian–Russian Expert Group Ž1996.. In the ‘‘maximum assessment’’,
the concentrations of 3000 Bqrkg 137Cs and 750
Bqrkg 239,240 Pu in ice-entrained material Žformer
suspended particles from the dump bays. were calculated from the water concentrations Ž1000 Bqrm3
Cs and 7.5 Bqrm3 239,240 Pu. according to the
modeled constant release of each 1 TBqryear
ŽHarms, 1997a.. According to IAEA Ž1985. we considered Kd factors of 3000 Žlrkg. for 137Cs and
1 = 10 5 Žlrkg. for 239,240 Pu, respectively ŽTable 3..
The ‘‘best estimate’’ shows that 2.90 GBq 137Cs
and 0.51 GBq 239,240 Pu attached to sea-ice entrained
sediments can be exported from the lead areas toward the central Arctic Ocean ŽTable 2.. Strongest
contributor is again lead section B. In the ‘‘maximum assessment’’, the radionuclide export rates of
Table 4
Production of dense-water in the east Novaya Zemlya flaw leads
Lead section
34.75 salinity water
Volume Ž10 3 rkm3 .U
Flux ŽSv.U
Sv sSverdrups10 6 m3 rs.
4-year mean from Martin and Cavalieri, 1989.
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
sea-ice particle bound 137Cs and 239,240 Pu would
amount 0.64 and 0.16 TBq, respectively.
TBq 239,240 Pu may be rejected downward from the
lead area with 34.75 salinity dense-water.
5.3. Dense-water
6. Discussion
Based on the 4-year mean calculations of locally
expelled salt rates, annually about 900 km3 of 34.75
salinity dense-water ŽMartin and Cavalieri, 1989; see
their Table 6. is produced in the flaw lead sections
A–C ŽTable 4.. Since the dense-water is formed
from lead water, the descending brines are assumed
to contain the same radionuclide burden as the lead
source water. The concentrations of dissolved radionuclides in the dense lead water are displayed in
Table 5.
In the ‘‘best estimate’’ based on measured data,
the total radionuclide burden of the annually formed
dense lead water amounts to 4.68 TBq 137Cs and
0.014 TBq 239,240 Pu. To the contrary, the burden-results derived in the ‘‘maximum assessment’’ from
modeled release-data show significantly higher values of potential Cs and Pu contents in dense-water
since we considered extreme source water contaminations in our estimations ŽTable 5.. The estimates
reveal that as much as 900 TBq of 137Cs and 6.75
Table 5
‘‘Best estimate’’ and ‘‘maximum assessment’’ of radionuclides in
34.75 salinity dense-water released from the western Kara Sea
flaw leads
Lead section
Best estimate
Measured data ŽBqrm3 . Burden ŽTBq.
Cs a
Lead section
Maximum assessment
Modeled data ŽBqrm3 . Burden ŽTBq.
After Joint Norwegian–Russian Expert Group Ž1996..
6.1. Radionuclide transport and export
Parts of the Kara Sea have to be regarded as
potential source regions for the dispersal of radioactive contaminants by sea-ice and dense-water. These
areas are located close to the shallow radionuclide
dumping sites along the eastern coast of Novaya
Zemlya, where turbid new-ice is formed under turbulent conditions. In the present study, we tried to
deliver an estimate of potential transport and export
rates of particle-bound and dissolved radionuclides
with ice and dense-water formed in western Kara
Sea flaw leads. However, due to lack of data, observations and modeling results, the above assessments
can only give a crude estimate of the radionuclide
transport from the dumping sites.
The potential ‘‘maximum assessment’’ export
rates of particle-bound radionuclides of 0.64 TBq
Cs and 0.16 TBq 239,240 Pu with lead-ice from the
western Kara Sea represent 0.0006% and 1.6% of the
total Kara Sea 137Cs and 239,240 Pu inventories. However, the annual export of 0.64 TBq 137C s with lead
ice would balance 64% of the constant release of 1
TBq from Abrasimov Bay simulated by Harms
Ž1997a.. The amount of 900 TBq 137Cs potentially
taken up by 34.75 salinity dense lead water in the
‘‘maximum assessment’’ would balance 90% of the
total Novaya Zemlya 137Cs inventory of 1 PBq,
provided that the inventory will be released instantaneous into the dissolved phase. The concentration of
6.75 TBq 239,240 Pu in 34.75 salinity dense-water
balances roughly 68% of the total Kara Sea Pu
inventory of 10 TBq.
The final fate of cold, dense-water masses released from the Kara Sea leads is still unclear, yet. In
first approximation, the dense-water will descend
toward deeper layers which correspond to its temperature and salinity. According to model results by
Harms Ž1997b. and Harms and Karcher Ž1999., a
potentially strong perennial stratification of the local
water column may prohibit deep convection of
dense-water produced in the flaw leads along the
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
Fig. 7. Cross-section of the western Kara Sea shelf showing potential pathways of particle bound Žsea-ice. and dissolved Ždense-water.
radionuclides in the ‘‘best estimate’’ Župper. and the ‘‘maximum assessment’’ Žlower.. For profile line ŽF–G. see Abrasimov Fjord area in
Fig. 1.
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
east coast of Novaya Zemlya. Thus, the intrusion of
polluted, dense brines from the leads towards the
deeper Novaya Zemlya Trough seems to be unlikely
or at least questionable. Therefore, we speculate that
the annually formed Žand potentially polluted. 34.75
salinity water would be rather remixed to the surrounding shelf water ŽFig. 7. which will be flushed
constantly Že.g., Harms, 1997a..
The export of potentially radioactively contaminated sea-ice sediments from the Kara Sea may
directly affect the Barents Sea and parts of the
European North Atlantic ŽPfirman et al., 1997a,b..
As shown by Nies et al. Žin press. through regional
scale modelings, potentially sediment- and
pollutant-laden sea-ice formed off the eastern coast
of Novaya Zemlya may be transported directly toward Fram Strait and Barents Sea within a period of
2–3 years ŽFig. 1.. Main portions of the ice melt
south of Svalbard where the particle — and possible
contaminant — load will be released.
Fig. 8. Behavior of radionuclides in Arctic Ocean water masses
ŽA. and sea-ice ŽB..
6.2. EffectiÕity of transport mechanisms
Due to dilution through intrusion and mixing of
less polluted or even ‘‘clean’’ water, the oceanic
radionuclide concentrations may strongly decrease
with increasing distance from the contamination
source ŽFig. 8.. On the contrary, sea-ice inclusions
— such as sediments and attached pollutants — may
be transported directly and without dilution from the
sources towards the sinks ŽCentral Arctic Ocean,
Barents Sea and European North Atlantic.. Sea-ice
entrained material experience minor changes in
quantitative and qualitative composition during
transport toward the sinks due to conservation within
ice. Thus, sea-ice sediments and attached pollutants
will be released in the ablation areas in approximately the same concentration they were rafted in
the shelf source regions.
Since the transport and dispersal of radionuclides
in the western Kara Sea may be closely connected to
new ice formation and dense-water rejection, the
season of a potential pollutant release is of particular
significance. During the short summer period, potentially released radionuclides would be mainly dispersed in shelf-mixed water, while during winter
contaminants would be rather dispersed by sea-ice
and dense-water. To investigate these processes in
more detail, winter expeditions should be carried out
to the Kara Sea.
6.3. Final remarks and outlook
After Nies et al. Žin press. and Kassens et al.
Ž1998., main ablation-areas of sea-ice from the Kara
Sea — and thus, main release-areas of potentially
polluted ice sediments — are located south of Spitsbergen in the western Barents Sea Žsee Fig. 1..
According to recent studies Že.g., Føyn 1998., accumulation-areas of biologically and sedimentary enriched 137Cs could be identified south of Spitsbergen.
Føyn and Svaeren Ž1997. found that 137Cs concentrations in Barents Sea surface sediments are correlated
with the portion of fine fraction. Wright Ž1974.
identified the area south of Spitsbergen as a region
where the clay mineral assemblage of bottom sediments differs significantly from the adjacent clay
provinces. The ‘unusual nature’ ŽWright 1974. of the
shelf surface sediments south of Spitsbergen as well
as the slightly enhanced local 137Cs activities may
imply the deposition of sea-ice rafted, contaminated
fine-grained material from the Siberian shelves Ži.e.,
Kara Sea..
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
If highly polluted dumping bay sediments Že.g.,
100,000 Bqrkg; Joint Norwegian–Russian Expert
Group, 1996. would be entrained into western Kara
Sea lead ice, the annual export of Cs attached to
sea-ice sediment would increase to as much as 20
TBq. This radionuclide concentration would be released in the ablation areas. Modern 137Cs concentrations in surface waters of the ablations areas amount
to 0.002–0.004 Bqrl ŽKellermann et al., 1998.. Cs
concentrations in fish are in the range of 0.2–0.4
Bqrkg, thereby exceeding the water concentration
by a factor of 100. However, even in the case of an
‘‘instantaneous’’ release of annually 20 TBq 137Cs
from the ice in the ablation areas, we can assume
that modern Cs concentrations in fish will not be
significantly exceeded.
7. Conclusions
Ž1. We have shown that the entrainment and
export of potentially radioactively contaminated particles from the Kara Sea by sea-ice is possible.
Sea-ice transport from the entrainment regions to the
ablation areas is very rapid Ž1–3 years. compared to
the dislocation of dissolved radionuclides by, e.g.,
dense-water or surface water.
Ž2. Two different exportrtransport estimates of
particle-bound and dissolved artificial radionuclides
by sea-ice and dense-water formed in western Kara
Sea flaw leads close to the Novaya Zemlya dumping
sites were presented. The ‘‘best estimate’’ was based
on present data from the western Kara Sea, while the
theoretic ‘‘maximum assessment’’ was derived from
modeled continuous radionuclide releases from the
dumping bays.
Ž3. According to our ‘‘best estimate’’, both the
particle-bound radionuclide export by lead ice from
the western Kara Sea and the radionuclide uptake by
dense-water in that region seems insignificant. In the
‘‘maximum assessment’’, the export of ice-particle
bound 137Cs and 238,239 Pu from the western Kara Sea
leads would amount to 0.64 and 0.16 TBq, respectively. As much as f 900 TBq 137Cs and f 6.75
TBq 239,240 Pu could potentially annually be taken up
by 34.75 salinity dense-water rejected in the lead
areas. This would represent f 90% of the Cs and
f 68% of the Pu inventories Žf 1 PBq and 10 TBq,
respectively. dumped in the eastern Novaya Zemlya
Ž4. The season of a potential release in the western Kara Sea is of substantial importance for the fate
of radionuclides in that area. While during summer
the radionuclides will be dislocated mainly by surface-mixed water, in winter the transport and dispersal of radionuclides in the western Kara Sea would
be closely connected to new-ice formation and
dense-water rejection.
Ž5. The total amount of radioactivity transported
by sea-ice represents merely a small fraction of the
activity potentially dislocated by dense lead water.
However, sea-ice sediments are expected to experience minor changes in concentration and composition during transport, and, thus, the particle-bound
radionuclide burden in sea-ice will not be significantly diluted on the path from the source areas
toward the regions of ablation. On the contrary,
dissolved radionuclides will be widely dispersed
throughout the entire Arctic Ocean during longterm
Ž6. We can state that a contamination of European
Arctic fishing grounds due to the melt-release of
potentially highly or extremely contaminated sea-ice
sediments from the western Kara Sea ŽNW coast of
Novaya Zemlya, Ob and Yenisei mouths. is principally possible. However, even in the case of enhanced radionuclide release from melting sea-ice, the
Cs-concentrations in North Atlantic water masses
and fish will probably not significantly exceed the
modern levels.
This study was funded by the Bundesminister fur
Umwelt, Naturschutz und Reaktorsicherheit ŽBMU,
project StSch 4101. and the Bundesminister fur
¨ Bildung, Forschung, Wissenschaft und Technologie
ŽBMBF, project 02-E-87054.. The scientific content
of this manuscript does not necessarily reflect the
opinion of the BMU and the BMBF. We are indebted to all MMBI colleagues, particularly to Prof.
Dr. Matishov, Dr. Denisov, Dr. Tarasov, and Dr.
Dmitri Matishov, who were decisively involved in
the preparation and conductance of the KaBaEx ’97
expedition. Many thanks are due to Per Strand and
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
Bjørn Lind from the Norwegian Radiation Protection
Authority ŽNRPA. for kindly providing surface sediment samples of the eastern Novaya Zemlya dumping bays. Warmest thank is given to Peter Lowe
¨ and
Dominik Weiel for their great support during preparation and conductance of the KaBaEx ’97 expedition. We appreciate the helpful comments from two
anonymous reviewers that substantially improved the
manuscript. We are also grateful to Ortrud Runze for
spell-checking the text.
Baskaran, M., Asbill, S., Santschi, P., Brooks, J., Champ, M.,
Adkinson, D., Colmer, M.R., Makeyev, V., 1996. Pu, 137Cs
and excess 210 Pb in Russian Arctic sediments. Earth Planet.
Sci. Lett. 140, 243–257.
Baxter, M.S., Harms, I., Osvath, I., Povinec, P.P., Scott, E.M.,
1998. Modelling the potential radiological consequences of
radioactive waste dumping in the Kara Sea. J. Environ. Radioact. 39 Ž2., 161–181.
Cavalieri, D., Martin, S., 1994. The contribution of Alaskan,
Siberian, and Canadian coastal polynyas to the cold halocline
layer of the Arctic Ocean. J. Geophys. Res. 99 ŽC9., 18343–
Cooper, L.W., Larsen, I.L., Beasley, T.M., Dolvin, S.S., Grebmeier, J.M., Kelley, J.M., Scott, M., Johnson-Pyrtle, A., 1998.
The distribution of radiocesium and plutonium in sea ice-entrained Arctic sediments in relation to potential sources and
sinks. J. Environ. Radioact. 39, 279–303.
Dethleff, D., Kleine, E., Loewe, P., 1994. Oceanic heat loss,
sea-ice formation and sediment dynamics in a turbulent
Siberian flaw lead. In: Proceedings of the Summer School of
Physics of Ice-Covered Seas, Savonlinna, Finland. Rep. Ser.
Geophys. 28pp. 35–40, Dept. of Geophysics, University of
Dethleff, D., Loewe, P., Kleine, E., 1998a. The Laptev Sea flaw
lead — detailed investigation on ice formation and export
during 1991r92 winter season. Cold Reg. Sci. Technol. 27
Ž3., 225–243.
Dethleff, D., Loewe, P., Weiel, D., Nies, H., Kuhlmann, G., Bahe,
C., Tarasov, G., 1998b. Winter expedition to the southwestern
Kara Sea — investigations on formation and transport of
turbid sea-ice. Rep. Pol. Res. 271, 1–40.
Dethleff, D., Reimnitz, E., 1996. Remote sensing of flaw leads in
the eastern Arctic — with a special emphasis on the Siberian
Laptev Sea. Fourth Circumpolar Symposium on Remote Sensing of the Polar Environment, 29 April–1 May 1996, Lyngby,
Føyn, L., 1998. The Barents Sea, distribution and fate of radioactive contaminants. In: International Symposium on Marine
Pollution, International Atomic Energy Agency ŽIAEA.,
Monaco, 5–9 October 1998. p. 491.
Føyn, L., Svaeren, I., 1997. Distribution and sedimentation of
radionuclides in the Barents Sea. ICES J. Mar. Sci. 54,
Geogruppen, A.S., 1994. Barents Sea and Kara Sea — Holocene
sediment thicknesses, surface sediment distribution. Report no.
93024.02, GeoGruppen as, Strandtorget 2B, 9008 Tromsø,
Harms, I.H., 1997a. Modelling the dispersion of 137Cs and 239 Pu
released from dumped waste in the Kara Sea. J. Mar. Syst. 13,
Harms, I.H., 1997b. Freshwater runoff and ice formation in Arctic
shelf seas: results from a high resolution Kara Sea model.
Proceedings of the WCRPrACSYS Conference on Polar Processes and Global Climate, Rosario Resort, Orcas Island,
Washington, USA, 3–6 November 1997. International ACSYS-Office, Oslo, Norway.
Harms, I.H., Karcher, M.J., 1999. Modelling the seasonal variability of circulation and hydrography in the Kara Sea. J. Geophys. Res. 104 ŽC6., 13431–13448.
IAEA, 1985. Sediment Kds and concentration factors for radionuclides in the marine environment. Tech. Rep. Ser. 247, 1–73.
Joint Norwegian–Russian Expert Group, 1996. Dumping of radioactive waste and investigation of radioactive contamination
in the Kara Sea. Norwegian Radiation Protection Authority,
˚ Norway, 1–55.
Kassens, H., Dmitrenko, I., Rachold, V., Thiede, J., Timokhov,
L., 1998. Russian and German scientists explore the Arctic’s
Laptev Sea and its climate system. EOS, Trans., Am. Geophys. Union 79 Ž27., 317, 322–323
Kellermann, H.J., Kanisch, G., Kruger,
A., Vobach, M., 1998.
Radioecological studies in the Barents Sea. In: International
Symposium on Marine Pollution, International Atomic Energy
Agency ŽIAEA., Monaco, 5–9 October 1998. pp. 235–236.
Krosshavn, M., Carroll, J., Engøy, T., Grennes, Ø., Johnson, A.,
Johnson, D., King, S.E., Ruud, A.B., 1998. Transport pathways of radionuclides and chemical contaminants in the Kara
Sea. Radiat. Prot. Dosim. 75 Ž1–4., 229–236.
Landa, E., Reimnitz, E., Beals, D., Pochkowski, J., Rigor, I.,
1998. Transport of 137Cs and 239,240 Pu by ice rafted debris in
the Arctic Ocean. Arctic 51 Ž1., 27–39.
Martin, S., Cavalieri, D.J., 1989. Contributions of the Siberian
Shelf polynyas to the Arctic Ocean intermediate and deep
water. J. Geophys. Res. 94 ŽC9., 12725–12738.
Meese, D.A., Reimnitz, E., Tucker, W.B. III, Gow, A.J., Bischof,
J., Darby, D., 1997. Evidence for radionuclide transport by
sea-ice. Sci. Total Environ. 202, 267–278.
Mitchell, P.I., McMahon, C.A., Leon
´ Vintro,
´ L., Ryan, R.W.,
1998. Plutonium in Arctic surface and sub-surface waters at
the St. Anna and Voronin Troughs. Radiat. Prot. Dosim. 75
Ž1–4., 247–252.
Nies, H., Harms, I.H., Karcher, M.J., Dethleff, D., Bahe, C.,
Kleine, E., Loewe, P., Oberhuber, J.M., Kuhlmann, G., Matishov, D., Stepanov, A., Vasiliev, O.F. in press. Anthropogenic radioactivity in the Arctic Ocean — results of a joint
project., Dtsch. Hydrogr. Ztschr.
D., Wollenburg, I., Dethleff, D., Eicken, H., Kassens,
H., Letzig, T., Reimnitz, E., Thiede, J., 1994. Sediments in
D. Dethleff et al.r Journal of Marine Systems 24 (2000) 233–248
Arctic sea ice — entrainment, transport and release. Mar.
Geol. 119, 185–214.
Osterkamp, T.E., Gosink, J.P., 1984. Observations and analysis of
sediment laden sea-ice. In: Barnes, P.W., Schell, D.M.,
Reimnitz, E. ŽEds.., The Alaska Beaufort Sea: Ecosystem and
Environment. Academic Press, San Francisco, pp. 73–94.
Pavlov, V.K., Pfirman, S.L., 1995. Hydrographic structure and
variability of the Kara Sea: implications for pollutant distribution. Deep-Sea Res. 42 Ž6., 1369–1390.
Pavlov, V.K., Stanovoy, V.V., in press. To the problem of transfer
of radionuclide pollution by sea-ice. Trudi AARI, St. Petersburg, Gidrometeoizdat, 1–9, in Russian.
Pavlov, V.K., Stanovoy, V.V., Nikitin, A.I., 1999. Possible causes
of radioactive contamination in the Laptev Sea. In: Kassens,
H., Bauch, H.A., Dmitrenko, I., Eicken, H., Hubberten, H.-W.,
Melles, M., Thiede, J., Timokhov, L. ŽEds.., Land Ocean
Systems in the Siberian Arctic: Dynamics and History.
Springer-Verlag, Berlin, pp. 65–72.
Pavlov, V.K., Timokhov, L.A., Baskakov, G.A., Kulakov, M.Yu.,
Kurazhov, V.K., Pavlov, P.V., Pivovarov, S.V., Stanovoy,
V.V., 1994. Hydrometeorological regime of the Kara, Laptev,
and East Siberian Seas: Part 1. Processes of transfer and
transformation of contaminants inflowing by continental discharge into the Kara, Laptev and East Siberian Seas. Federal
Service of Russia for hydrometeorology and monitoring of the
environment. The Arctic and Antarctic Research Institute,
Pfirman, S.L., Colony, R., Nurnberg,
D., Eicken, H., Rigor, I.,
1997a. Reconstructing the origin and trajectory of drifting
Arctic sea-ice. J. Geophys. Res. 102 ŽC6., 12575–12586.
Pfirman, S.L., Kogele,
J.W., Rigor, I., 1997b. Potential for rapid
transport of contaminants from the Kara Sea. Sci. Total Environ. 202, 111–122.
Reimnitz, E., Bruder, K.F., 1972. River discharge into an ice-
covered ocean and related sediment dispersal, Beaufort Sea,
coast of Alaska. Geol. Soc. Am. Bull. 83, 861–866.
Reimnitz, E., Clayton, J.R., Kempema, E.W., Payne, J.R., Weber,
W.S., 1993a. Interaction of rising frazil with suspended particles: tank experiments with applications to nature. Cold Reg.
Sci. Technol. 21, 117–135.
Reimnitz, E., Marincovich, L. Jr., McCormick, M., Briggs, W.M.,
1992. Suspension freezing of bottom sediment and biota in the
Northwest Passage and implications for Arctic Ocean sedimentation. Can. J. Earth Sci. 29, 693–703.
Reimnitz, E., McCormick, M., Bischoff, J., Darby, D.A., 1998.
Comparing sea-ice sediment load with Beaufort sea shelf
deposits: is entrainment selective?. J. Sediment. Petrol. 68 Ž5.,
Reimnitz, E., McCormick, M., McDougall, K., Brouwers, E.,
1993b. Sediment-export by ice rafting from a coastal polynya,
Arctic Alaska. Arct. Alp. Res. 25 Ž2., 83–98.
Wright, P.L., 1974. The chemistry and mineralogy of the clay
fraction of sediments from the southern Barents Sea. Chem.
Geol. 13, 197–216.
Yablokov, A.V., Krasev, V.K., Rumyantsev, V.M., Kokeyev,
M.Y., Petrov, O.I., Lystsov, V.N., Yemelyanenkov, A.F.,
Rubtsov, P.M., 1993. Facts and Problems Related to Radioactive Waste Disposal in Seas Adjacent to the Territory of the
Russian Federation. Office of the President of the Russian
Federation, English edition by Small Word Publishers, Albuquerque.
Zakharov, V.F., 1966. The role of flaw leads off the edge of fast
ice in the hydrological and ice regime of the Laptev Sea.
Oceanology 6, 815–821.
Zakharov, V.F., 1976. In: Cooling of the Arctic and the Ice Cover
of the Arctic Seas. Arctic and Antarctic Research Institute,
Leningrad, Russia, p. 96, ŽEnglish translation by the Norwegian Polar Institute..
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF