AAR-Ch03 Influence of Physical and Chemical Processes On Comtaminant Transport Into and Within The Arctic

Go back to opening screen 25
Chapter 3
The Influence of Physical and Chemical Processes
on Contaminant Transport
into and within the Arctic
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Contents 3.5.3.1. Ice and icebergs . . . . . . . . . . . . . . . . . . . . . . . . . . 62

River ice . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Sea ice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.2. Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Formation and melting . . . . . . . . . . . . . . . . . 63
3.2.1. Single and multi-hop pathways . . . . . . . . . . . . . . . . . . . . . 27 Fast ice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2.2. Atmospheric transport . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Polynyas . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2.3. Atmosphere-surface exchange . . . . . . . . . . . . . . . . . . . . . . 30 Glaciers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2.3.1. Particle deposition . . . . . . . . . . . . . . . . . . . . . . . . 30 Ice drift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.2.3.2. Particle resuspension . . . . . . . . . . . . . . . . . . . . . . 31 Residence time . . . . . . . . . . . . . . . . . . . . . . . 65
3.2.3.3. Wet deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Fluxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.2.3.4. Gas exchange between atmosphere and snowpack/ice 33 3.5.3.2. Contaminant incorporation in ice . . . . . . . . . . . . 66
3.2.3.5. Gas exchange between the atmosphere and soil, River ice . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
water, and vegetation . . . . . . . . . . . . . . . . . . . . . . 33 Sea ice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.2.3.6. Snow and ice melt . . . . . . . . . . . . . . . . . . . . . . . . 33 Ice scouring . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.5.3.3. Transport and release of contaminants by ice . . . 67
3.3. Terrestrial/freshwater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Marginal ice zone and seas . . . . . . . . . . . . . . 67
3.3.1. Introduction and scope . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.5.4. Transport pathways and hydrographic conditions in the
3.3.2. Sources of contaminants . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Arctic seas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.3.2.1. Atmospheric deposition . . . . . . . . . . . . . . . . . . . . 34
3.5.4.1. The Arctic Ocean . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.3.2.2. Direct wastewater discharges . . . . . . . . . . . . . . . . 34
The Arctic Surface Layer . . . . . . . . . . . . . . . 68
Russia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
The Atlantic Layer . . . . . . . . . . . . . . . . . . . . 70
Alaska . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
The Arctic Deep Water Layer . . . . . . . . . . . . 71
Canada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.5.4.2. The Arctic shelf seas. . . . . . . . . . . . . . . . . . . . . . . 72
3.3.2.3. Municipal wastewater sources . . . . . . . . . . . . . . . 36
The Barents Sea . . . . . . . . . . . . . . . . . . . . . . 72
Russia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
The Kara Sea, the Laptev Sea, and the
Canada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
East Siberian Sea . . . . . . . . . . . . . . . . . . . . . . 73
Greenland . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
The Chukchi Sea . . . . . . . . . . . . . . . . . . . . . . 73
Iceland . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
The Beaufort Sea . . . . . . . . . . . . . . . . . . . . . 73
Norway. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.5.4.3. The Nordic Seas . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.3.2.4. Agricultural sources . . . . . . . . . . . . . . . . . . . . . . . 37
The Norwegian Sea . . . . . . . . . . . . . . . . . . . . 74
3.3.3. Terrestrial processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Atlantic flow through the Nordic Seas . . . . . 75
3.3.3.1. Snowpack and snowmelt . . . . . . . . . . . . . . . . . . . 38
The Norwegian Coastal Current . . . . . . . . . . 75
3.3.3.2. Terrestrial drainage systems . . . . . . . . . . . . . . . . . 40
The Greenland Sea . . . . . . . . . . . . . . . . . . . . 76
3.3.3.3. Wetlands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Polar flow through the Nordic Seas . . . . . . . 76
3.3.4. Aquatic processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
The Iceland Sea . . . . . . . . . . . . . . . . . . . . . . . 76
3.3.4.1. Hydrology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Overflow waters . . . . . . . . . . . . . . . . . . . . . . 77
Headwater systems . . . . . . . . . . . . . . . . . . . . 42
3.5.4.4. The Labrador Sea, Baffin Bay, Canadian
Large river systems . . . . . . . . . . . . . . . . . . . . 43
Archipelago, and Canadian inland seas . . . . . . . . 77
3.3.4.2. Lakes and reservoirs . . . . . . . . . . . . . . . . . . . . . . 45
Labrador Sea . . . . . . . . . . . . . . . . . . . . . . . . 77
Small lakes . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Baffin Bay . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.3.4.3. Suspended sediments, sedimentation, and
Nares Strait . . . . . . . . . . . . . . . . . . . . . . . . . . 79
resuspension . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Canadian Archipelago . . . . . . . . . . . . . . . . . 79
Suspended sediments . . . . . . . . . . . . . . . . . . 47
Foxe Basin . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Sediment deposition and resuspension . . . . . 48
Hudson Bay, James Bay, and Hudson Strait . 80
Effects of floods . . . . . . . . . . . . . . . . . . . . . . 49
3.5.4.5. The Bering Sea . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.3.4.4. Humic matter in river systems . . . . . . . . . . . . . . . 49
3.5.4.6. Vertical mixing . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.3.4.5. Ice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Shelf convection in the Arctic . . . . . . . . . . . . 83
Sources of contaminants to the ice . . . . . . . . 51
Open ocean convection in the Arctic . . . . . . 84
Transport processes . . . . . . . . . . . . . . . . . . . 51
Upwelling . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Dissolved oxygen (DO) . . . . . . . . . . . . . . . . . 51
Vertical particle fluxes . . . . . . . . . . . . . . . . . . 85
3.3.5. Mass transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.5.4.7. Oceanic time scales . . . . . . . . . . . . . . . . . . . . . . . 86
3.3.5.1. Water and suspended matter . . . . . . . . . . . . . . . . 52
Ages and residence times . . . . . . . . . . . . . . . 86
Norway. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Russia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 3.6. Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Canada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 3.6.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.3.5.2. River/lake ice . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 3.6.2. Atmospheric modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.6.2.1. Three-dimensional atmospheric models . . . . . . . . 87
3.4. Estuaries, deltas, and fjords . . . . . . . . . . . . . . . . . . . . . . . . . 54
Models for meteorological fields . . . . . . . . . . 87
3.4.1. Estuaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Dispersion models . . . . . . . . . . . . . . . . . . . . . 88
3.4.2. Deltas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Off-line dispersion models . . . . . . . . . . . . . . 89
3.4.3. Fjords . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
On-line dispersion models . . . . . . . . . . . . . . 89
3.4.4. Particle and nutrient transport . . . . . . . . . . . . . . . . . . . . . 57
3.6.2.2. Multi-compartmental models . . . . . . . . . . . . . . . . 90
3.4.5. Sedimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
The Bergen and Toronto models . . . . . . . . . . 90
3.4.6. Ice cover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.6.3. Freshwater systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.4.7. Contaminant fate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.6.3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.5. Ocean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 3.6.3.2. Norway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.5.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Physical, chemical, and biological processes
3.5.2. Sources of contamination . . . . . . . . . . . . . . . . . . . . . . . . . 60 in rivers and lakes . . . . . . . . . . . . . . . . . . . . . 92
3.5.2.1. Rivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Consequences for humans and ecosystems . . 92
3.5.2.2. Glaciers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Examples of the use of River System Simulator 93
3.5.2.3. Atmospheric deposition . . . . . . . . . . . . . . . . . . . . 60 Gjengedal River system . . . . . . . . . . . . . . . . 93
Volatilization . . . . . . . . . . . . . . . . . . . . . . . . 61 Stjørdal River system . . . . . . . . . . . . . . . . . . 93
3.5.2.4. Dumping, direct discharge, and accidents . . . . . . 61 3.6.3.3. Canada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.5.2.5. Geology: resuspension and mass flows . . . . . . . . . 62 3.6.3.4. United States . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.5.3. Ice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 3.6.3.5. Russia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
26 AMAP Assessment Report
A t m o s p h e r e
A Atmosphere AI AO
AT AS
TF Terrestrial-Freshwater A-SO
A-TF
SO Surface ocean
DO Deep ocean
BW Bottom water
S Sediments Snowpack-River ice- River ice - Shore ice Sea ice
Lake ice-Glacial ice
E Estuary-Delta-Fjord
F Fluvial TF-A SO-A
TF T TF S E SO T SO S SO I
Terrestrial - Freshwater Estuary-Delta-Fjord Surface ocean SO O
Subscripts
FI FO
Shelf Central
I : Input
TF-S E-S SO-S SO-DO
O : Output
T : Transformation
S-TF S-E S-SO
S : Internal sources
Sediments Sediments Sediments
DO-SO
DO T DO S
3.6.4. Marine system modeling . . . . . . . . . . . . . . . . . . . . . . . . . . 94 DO I
Deep ocean DO O
3.6.4.1. Objectives of modeling. . . . . A. . . . .Atmosphere . . . . . . . . . . . 94 Subscripts
DO-BW
3.6.4.2. Types of models . . . . . . . . . . TF . . . . .Terrestrial-Freshwater
SO . . . . .ocean
Surface . . . . . . 94
Ice models . . . . . . . . . . DO . . . . .Deep
. . . ocean
. . . . . . . . 94 I : Input
Thermodynamics . . . . . BW . . . . .Bottom
. . . . water
. . . . . . . 95 O : Output
S
Requirements and constraints E .Sediments
. . . . . . . . . . . 95
Estuary-Delta-Fjord
T : Transformation BW-DO
Bottom
Quality and availability of F data. . . . . . . . . . . 95
Fluvial Figure 3·1.S A : Internal sources
multi-compartment
schematic dia- water BW I
Availability of models . . . . . . . . . . . . . . . . . . 95 gram of the major pathways of contaminants into BW-S BW O
3.6.4.3. Ocean models . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 and within the Arctic environment.
Specific ocean models . . . . . . . . . . . . . . . . . . 96
Coupled ice-ocean, Arctic modeling studies. . 96 S-BW
Partially diagnostic models . . . . . . . . . . . . . . 96 Sediments
Prognostic models . . . . . . . . . . . . . . . . . . . . . 96
Evaluation and verification of model results . 97 The contaminants found in the Arctic as a result of trans-
3.6.4.4. Modeling of oceanic contaminant transport in port over long distances are those that have been produced
Arctic seas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Examples of dispersion modeling . . . . . . . . . 97 in large quantities and are environmentally stable. Properties
3.7. Summary and information gaps . . . . . . . . . . . . . . . . . . . . . . 100 which promote the long-range transport of pollutants into
3.7.1. Atmospheric pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 the Arctic via the atmosphere are: large-scale and geographi-
3.7.2. Terrestrial/freshwater pathways . . . . . . . . . . . . . . . . . . . . 101 cally widespread emissions; low water solubility; the pres-
3.7.3. Ocean pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
3.7.4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 ence of the chemical in the gas phase, or on small particles
3.7.5. Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 with atmospheric residence times of several days or longer;
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 and, chemical stability. These properties are also important
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 with respect to long-range riverine and marine transport,
with the additional requirement that these compounds must
be biologically stable.
The phenomenon of Arctic haze was the first indication
3.1. Introduction of long-range transport to the Arctic, occurring from De-
The Arctic region is a seemingly pristine, remote environ- cember through April each year. Arctic haze was reported as
ment, yet there is increasing evidence that it is greatly im- early as 1956 (Mitchell 1956), and was later shown to be a
pacted by global pollution. It is a highly sensitive ecosystem. manifestation of the atmospheric long-range transport of
Compounds found in the Arctic are of concern not just be- acid sulfur pollution (Rahn and Shaw 1982, Barrie 1986,
cause they are transported long distances, but also because Shaw 1995). This was followed by convincing evidence of
they are bioaccumulated and biomagnified in the food chain. the atmospheric transport of persistent organic pollutants
Consequently, an understanding of pollution pathways to (POPs) to the Arctic (Oehme 1991, Barrie et al. 1992).
and within the Arctic is of special concern. The Arctic region, as defined in chapter 2, encompasses a
The sources of most contaminants of interest to AMAP total of 13.4 106 km2 of land and 20 106 km2 of ocean
generally lie outside of the Arctic region; however, there are (Gloersen et al. 1992). Clearly, there is a large degree of var-
some internal regional sources, especially from mineral ex- iability across this huge area. To provide a framework for
traction and related activities. Pathways within the three ma- this discussion, therefore, a simple box model has been used
jor environmental compartments, specifically the atmosphere, to illustrate the dominant processes and pathways of conta-
terrestrial/freshwater, and marine compartments, all contri- minants for the Arctic region. The principal physical conta-
bute to the delivery of contaminants to the Arctic (Barrie minant pathways into and within the region are described
1986, Ottar 1989, Barrie et al. 1992). Both the physical pro- schematically in Figure 3·1. This figure represents a simplifi-
cesses and the general nature of the main pathways within cation of the complex transport pathways and reservoirs op-
these three compartments are reviewed in this chapter. This erating in the Arctic system. The main compartments are the
complements the next chapter on biological systems and pro- atmosphere, terrestrial/freshwater, and ocean. The ocean
cesses and subsequent chapters on specific contaminant groups. compartment is separated into three components, specifi-
The objective of this chapter is to provide a general under- cally, the surface ocean (0-200 m), the deep ocean (> 200 m
standing of the physical processes and pathways, both within depth), and the bottom water. The surface ocean can be
and between the main environmental compartments, which further subdivided into a stable upper layer of 50-60 m
determine the fate of contaminants in the Arctic environment. depth and characterized by low salinity, on top of a pycno-
General consideration is given to sources and emissions, trans- cline layer which extends down to approximately 150 m
port mechanisms, and the processes affecting the inter- and depth and is located above the shelf and central ocean wa-
intra-compartmental interactions of contaminants of concern. ters. Both the shelf and bottom waters are connected to the
Chapter 3 · The Influence of Physical and Chemical Processes on Contaminant Transport 27
marine sediments. The terrestrial/freshwater compartment Arctic, removal processes, and the exchange of compounds be-
also contains a sediment subcompartment associated with tween the different compartments. Generally, the atmosphere
rivers, lakes, and reservoirs. The main interfaces between contains a relatively small amount of a contaminant compared
the compartments are indicated by arrows. to the total amount in other environmental compartments
Both the terrestrial/freshwater and ocean compartments (e.g., for HCH, see Strand and Hov 1996). The importance of
have seasonally variable subcompartments, i.e., snowpack, the atmosphere is that it provides a significant mode of rapid
river/lake ice, and glacial ice for terrestrial/freshwater sys- transport of contaminants from source areas to the Arctic re-
tems, and sea ice in oceans. Deltas, estuaries, and fjords are gion. Transport times of contaminants via air currents are
much smaller spatially than the three main compartments, rapid compared to those in the oceans. The time for an air
but have been identified as a key interface between the ter- parcel to completely mix in the troposphere of the northern
restrial/freshwater and ocean compartments. Large amounts hemisphere is of the order of six months. Mixing in the region
of sediment transported by rivers are deposited in deltas north of 30°N is of the order of three months (Plumb and
and estuaries, and thus, sediments are also identified here. Mahlman 1987). In contrast, transport times of water parcels
In all cases, water transfers include, by definition, organic in northern marine systems are measured in years and decades.
and inorganic particulate matter. Atmospheric contaminants that are transported to the Arc-
The pathways shown in Figure 3·1 illustrate transport tic can be introduced from point, area, and regionally-distri-
by air (‘A’) or water (‘F’ for freshwater, including suspended buted sources. Point sources include power generating stations,
sediments; ‘SO’ for surface ocean; and, ‘DO’ for deep smelters, incinerators, and dump sites. This is the predominant
ocean, with the ‘I’ or ‘O’ subscript indicating input to and source type in the northern regions of Eurasia and North Ame-
output from the system). Chemical and biological transfor- rica. Dump sites are of greatest concern for semi-volatile conta-
mations resulting in removal of material from the compart- minants which volatilize once they are exposed to the air. Area
ment are indicated by a ‘T’ subscript, while source terms sources consist of urban-industrial complexes. In heavily pop-
(‘S’ subscript) represent anthropogenic or natural sources ulated regions, a number of urban-industrial complexes mixed
within the compartment and exchange processes between with point sources constitute a regionally-distributed source.
the compartments. The exchange processes involve trans- In the following sections, the sources, processes and fac-
port, as well as chemical and physical reactions, as conta- tors affecting atmospheric transport, as well as the exchange
minants move across the interface. Note that the relative processes between the atmosphere and other media (ice, land,
size of the boxes is not of significance in this figure. In most ocean) are described.
cases of inter-compartmental transfer, double arrows are
used to indicate two-way movement. This model is a useful
3.2.1. Single and multi-hop pathways
tool in the quantitative assessment of pathways. For in-
stance, it has been applied to a group of pesticides that are Atmospheric transport pathways can be subdivided into two
commonly found in air in the Arctic (hexachlorocyclohexa- types: one-hop pathways and multi-hop pathways. One-hop
nes or HCHs) in order to understand the sources of this pathways describe the movement of compounds that are emit-
POP to the surface ocean (see chapter 6). ted to the atmosphere, transported, and then deposited to the
Following this brief introduction, section 3.2 explores surface, never to return to the atmosphere. In such cases, the
the atmospheric compartment and its direct exchange with source region of an Arctic contaminant is simply defined by its
snowpack, lake ice, sea ice, and the surface ocean. Also con- source distribution, its lifetime in the atmosphere (governed
sidered are the indirect exchanges between the atmosphere by removal processes), and atmospheric circulation. This ap-
and the terrestrial/freshwater compartments through the de- plies to acids, heavy metals with the exception of mercury, and
livery of contaminants via precipitation (rain, fog, and snow) nonvolatile, particle-bound organics, such as benzo(a)pyrene
and dry deposition, as well as direct atmospheric exchange [B(a)P], black carbon particles, and radionuclides. The path-
with rivers, lakes, soil, and vegetation. Section 3.3 is con- ways of these constituents follow that of Arctic haze from mid-
cerned more specifically with the transport and fate of con- latitudinal sources into the Arctic (Barrie 1986, Barrie 1995,
taminants in the terrestrial/freshwater compartment, with Barrie 1996b). Single-hop compounds are marked by longer
emphasis on the role of freshwater in transporting these atmospheric residence times in winter (≈ 20-30 days) compared
contaminants. Discussion includes the delivery of pollutants to summer (≈ 3-7 days), as well as stronger south to north
from the land to the aquatic system, and sedimentation and transport into the Arctic from Eurasia in winter than in sum-
sediment recycling in fluvial and limnic systems. Section 3.4 mer. With multi-hop pathways, a compound re-enters the at-
reviews processes in deltas, estuaries, and fjords, as these mosphere after initial deposition to the Earth’s surface, and
environments serve as an interface between rivers and lakes continues over time to move through the environment in mul-
in the case of freshwater deltas, and between the terrestrial/ tiple hops. Processes by which this can occur include volatili-
freshwater and ocean compartments for marine deltas, estu- zation from the Earth’s surface under temperatures warmer
aries, and fjords. Section 3.5 explores various components than during initial deposition; sudden exposure to the atmos-
of the ocean system, including the surface and deep ocean phere of ocean water saturated with a volatile contaminant
compartments, sea ice, and the major intra- and inter-com- after being covered by ice; and, resuspension by wind, dust or
partmental exchange processes. Lastly, the role of modeling snow. For multi-hop compounds, the source region affecting
as a pivotal tool in developing our understanding of the the Arctic is not only defined by atmospheric transport, re-
complex pathways into and within the Arctic is briefly removal and circulation, but also by surface processes that con-
viewed in section 3.6. trol its re-entry into the atmosphere. Mercury, most organo-
chlorines (OCs), and many PAHs fall into the multi-hop group.
3.2. Atmosphere 3.2.2. Atmospheric transport

The occurrence of contaminants in the Arctic is influenced Barrie (1992) and Iversen (1989a, 1989b, 1996) have dis-
by the nature and rate of emissions from sources, the pro- cussed the role of atmospheric circulation in the transport of
cesses by which these compounds are transported to the contaminants from mid-latitudes to the Arctic region. At-
January July
Figure 3·2. The mean circulation of the lower atmosphere during January and July as depicted by mean streamlines of the resultant winds (Source:
Wendland and Bryson 1981).
mospheric winds deliver contaminants to the north. This is W/m 2 (a)

illustrated by streamlines of average winds in the northern 150
Arctic 70°N
hemisphere in January and July (Figure 3·2). In winter, the
lower tropospheric circulation of the northern polar region
is dominated by high pressures over the continents and low 100
pressures over the northern Pacific and Atlantic Oceans. In TE+SE+MMC
MMC
particular, the intense Siberian high pressure cell tends to
force air on its western side northward into the Arctic. The TE+SE SE
high pressure ridge over North America generally drives air 50
out of the Arctic southward. The mean flow in winter is out TE
TE
of Eurasia into the Arctic, and out of the Arctic into North
America. Some air is also exchanged with the south when 0
low pressure vortices along the Arctic front mix warm south- J F M A M J J A S O N D J
ern air with cold northern air in a large-scale turbulent eddy.
W/m 2
In summer (Figure 3·2), the continental high pressure 200
cells disappear, and the oceanic low pressure cells weaken, Antarctic (b)
particularly in the north Pacific. Northward transport from
mid-latitudes decreases accordingly. Mean circulation in the 70°S
lower atmosphere gives way to a more circular clockwise 150
flow around the pole at a higher altitude prompted by polar
low pressures prominent in both summer and winter. The
winter Siberian high pressure cell at the lower elevations Model
100 TE+SE+MMC
disappears at an altitude of 3 km.
The mean flow maps in Figure 3·2 present a simplified
picture of the circulation. Three types of flow regimes are TE+SE MMC
50
involved in the exchange of mass or heat between polar re-
TE SE
gions and extra-polar regions: 1) mean meridional circula-
tion (MMC), 2) standing eddies (SE), and 3) transient eddies TE
(TE). To illustrate that these can all contribute substantially 0
to north-south mass exchange, consider the seasonal distrib- J A S O N D J F M A M J J
ution of energy exchange for each flow regime calculated by Figure 3·3. Modeled annual variation in poleward energy flux through lat-
Nakamura and Oort (1988) for the Arctic and Antarctic (Fig- itude 70°, as effected by transient eddies (TE), standing eddies (SE), and
ure 3·3). Heat exchange is least in summer and greatest in mean meridional circulation (MMC) for the atmosphere from the surface
to 25 hPa, for a) the Arctic and b) the Antarctic (adapted from Nakamura
winter. Quantitative parallels between heat and mass exchange and Oort 1988).
are erroneous, since the exchange depends on the lifetime and
spatial distribution of the sources of a contaminant. Neverthe- of year. The difference between the Arctic and Antarctic is
less, this analogy is valid qualitatively, and hence instructive. due to the very different distribution of continents and oceans,
Standing eddies, such as the Siberian high pressure cell as well as the major mountain ranges between the two hemi-
(marked by a source of surface winds in Figure 3·3), tend spheres. These cause an almost rotational symmetric gyre
to play a more prominent role in the northern hemispheric around Antarctica, while long planetary waves forced by the
winter than in summer, or than in the Antarctic at any time ground surface influence the circulation pattern around the
Frequency (%)
80 Low-cloud
(a)
70
5 - 15 % 60
50
Middle-cloud
40
Arctic Precipitation
Ocean
10 - 40 %
5 - 25 % 30
20 High-cloud
North North
Pacific Atlantic 10 Clear skies
Ocean Ocean
0
J F M A M J J A S O N D
Precipitation, water equivalent (mm)

35
(b)
Arctic Front Winter 30
Alert Mould Bay Resolute
Major south to north air transport
Arctic Front Summer routes into the Arctic 25
Figure 3·4. The mean position of the Arctic air mass in winter (January) 20
and summer (July) from Li et al. (1993), superimposed on the (summer
value % (orange) – winter value % (blue)) frequency of major south-to- 15
north air transport routes into the Arctic (Source: Iversen 1996).
10
Arctic. The stationary planetary wave pattern in the north-
5
ern hemisphere is more pronounced during winter, when the
temperature contrast between the tropics and the polar re- 0
J F M A M J J A S O N D
gions, and the strength of the jet streams are at their maxi-
mum. Furthermore, the total eddy exchange (SE+TE) is Figure 3·5. Seasonal variation in a) cloud cover and precipitation occur-
rence and b) precipitation amount in the Canadian High Arctic, 1951-
stronger than MMC at all times in the northern hemisphere, 1980 norms (Source: Barrie 1986).
while in the southern hemispheric winter, the exchange is
dominated by MMC. The thermal contrast between oceans during their atmospheric transport to Arctic areas. This is a
and continents during the northern hemispheric winter is consequence of the temperature distribution in the atmos-
believed to be the major cause of stationary eddy features in phere and the energy conversions driving the air motions.
the lower atmosphere close to the ground, where the main With few exceptions, atmospheric flow systems convert po-
part of the transport of anthropogenic contaminants takes tential energy to kinetic energy through rising flows of rela-
place. These features include the Aleutian and Icelandic lows, tively warm and light air and sinking flows of relatively cold
as well as large northern continental anticyclones. In con- and heavy air. The reservoir of potential energy remains un-
nection with transport to the Arctic, the occurrence of the changed as the tropics are warmed by solar radiation, while
Siberian winter high pressure over large portions of Eurasia the polar areas are cooled by terrestrial outgoing radiation.
is of particular importance, as it causes a deep extension of As a consequence, contamination at higher altitudes (2-8 km)
the very stable Arctic boundary layer to parts of Eurasia in the Arctic generally originates from lower latitudes than
which contain a large number of anthropogenic sources. contamination in the lower Arctic atmosphere (0-2 km), the
Winter flow patterns are markedly different from those latter originating in colder and more northerly regions (Carl-
in the summer. In winter in the lower atmosphere, air move- son 1981, Iversen 1984). Since many contaminants are wa-
ment is driven by both the strong quasi-stationary cyclones ter-soluble, and rising motions frequently are connected with
over the Aleutians and near Iceland, and anticyclones re- precipitation release, upper level atmospheric Arctic conta-
siding over the North American and Eurasian continents, mination is generally less concentrated than in the lower at-
resulting in the transport of air over Siberia into the High mosphere. Nevertheless, it is characteristic of the Arctic that
Arctic, and south over North America. Consequently, cont- the vertical extent of anthropogenic aerosol particles is great-
aminants which are introduced into the atmosphere in Sibe- er than at mid-latitudes (Hansen and Rosen 1984, Radke et
ria or Eastern Europe (in either vapor form or bound to al. 1984, Schnell and Raatz 1984, Barrie 1986, Pacyna and
small particles which have a low deposition velocity) are Ottar 1988). In summer and autumn, when temperature
frequently transported into the Arctic region. Anticyclones contrasts between major pollution source areas and the Arc-
are also characterized by relatively low wind speed, and thus tic are smaller, dispersion of contaminants in the vertical is
stagnant conditions. Near the surface, the relative lack of less pronounced than at other times of the year.
cloud cover and low incident solar radiation during the Arc- Figure 3·4 illustrates the mean position of the Arctic air
tic winter can produce extended periods of surface radiation mass and the frequency of south-to-north transport in both
inversions. These factors reduce the effectiveness of vertical summer and winter at various points on the Arctic Circle. In
mixing and removal at the surface, resulting in the accumu- contrast to the winter period, in the summer, south-to-north
lation of contaminants in the lower Arctic atmosphere. transport from Eurasia is much weaker as the Siberian high
Contaminants having ground-level sources at mid- and dissipates. Marked variation in cloud cover and precipita-
low latitudes will, on average, experience rising air motions tion accompanies this seasonal variation (Figure 3·5). In
(a) summer, temperatures in the High Arctic are near 0°C, al-
lowing much more drizzling marine stratus than in winter,
when temperatures between –25 and – 45°C prevent build-
up of moisture in the atmosphere. Photochemical activity
during summer months is also higher, and provides an addi-
tional loss mechanism for some contaminants. Thus, during
summer, transport from mid-latitudes to the Arctic is least
for soluble, cloud-reactive or particle-bound compounds
that are more easily removed by precipitation scavenging
than insoluble or cloud-unreactive gases.
To illustrate the interplay between atmospheric circula-
tion and the spatial distribution of contaminant emissions
to the atmosphere, consider the case of anthropogenic sul-
fur transport to the Arctic. In Figure 3·6, three panels are
SO2 shown. The top one (a) is the spatial distribution of annual
emission
(Gg S/y) anthropogenic sulfur dioxide emissions to the atmosphere,
100 the center one (b) is the frequency of south-to-north flow of
75 air as a function of longitude in summer and winter (Iversen
50
25 1989a), and the bottom panel (c) is the flux of anthropo-
10
7
genic sulfur into the Arctic for a one-year period (July 1978
5 to June 1979), as a function of longitude, modeled by Bar-
2
1 rie et al. (1989b). The greatest inputs of these one-hop com-
pounds occur in the Eurasian longitude sector (0-70°E) in
winter. North American sources are not well connected to
Frequency (%) the Arctic by winds. Air tends to move eastward from North
(b)
50 America over the stormy north Atlantic Ocean, where cont-
aminants tend to be lost to the North Atlantic Ocean rather
Winter
than northward into the Arctic air mass (Figure 3·4).
40
Summer
3.2.3. Atmosphere – surface exchange
30
Figure 3·7 summarizes schematically the major contaminant
pathways between the atmosphere and the surface, as well
20
as our current ability to calculate exchanges using flux equa-
tions. Transfers from the atmosphere to the Earth’s surface
10 in the polar regions are complicated by the presence of a
seasonally varying intermediate media, namely, snowpack
and ice. In the Arctic Ocean, the marginal seas undergo the
0
180° 120°W 60°W 0 60°E 120°E 180°E
greatest variations in percentage of ice cover.
The snowpack plays an important role in atmosphere –
kt-S/y surface exchange. During the winter, as snow accumulates
600 on the surface as snowpack, it stores contaminants gained
(c)
during its formation in addition to those delivered by particle
dry deposition and gaseous exchange. The latter component
500
may be negative (surface loss), or positive (surface gain). At
present, there is no way of measuring it directly. However,
400
there are indirect means which can be used to infer gas ex-
change (see below). The net exchange between the atmos-
300 phere and the Earth is the result of all these processes.
200
3.2.3.1. Particle deposition
(processes 2 and 7 in Figure 3·7)
100
Turbulent transport and/or sedimentation bring airborne
particulate contaminants to the near-surface layer from the
0
free atmosphere. Then processes of convection, diffusion, or
180° 120°W 60°W 0 60°E 120°E 180°E
inertial impaction carry them across a viscous sublayer where
Figure 3·6. An illustration of the interplay between a) contaminant emis-
sions distribution for sulfur dioxide and b) frequency of atmospheric south-
chemical and physical processes trap them on or in the sur-
to-north transport, yielding c) a net annual input of sulfur to the Arctic face (Voldner et al. 1986). The rate of particle transfer de-
as a function of longitude that favors Eurasian sources in the winter half pends on the type of surface, the physical/chemical charac-
of the year. This contaminant has a one-hop pathway in contrast to more teristics of the contaminant, and the state of the atmosphere.
volatile persistent organochlorines, PAHs, and mercury. a) Gridded global
emissions inventory of sulfur dioxide for 1985 as compiled on a 1° 1° The flux of contaminants from the atmosphere to the
grid by the Global Emissions Inventory Activity (GEIA) of the Internatio- Earth’s surface via particle dry deposition (F) is calculated as
nal Global Atmospheric Chemistry Program (IGACP); b) the frequency of the product of a particle dry deposition velocity (Vd) and the
south-to-north transport of air in summer and in winter as a function of
contaminant concentration on particles in air (Cp) at some
longitude (from Iversen 1989a); and, c) the annual input of anthropo-
genic sulfur to the Arctic at the Arctic Circle as a function of longitude reference height above the surface (usually 10 m). Vd depends
as calculated for July 1978 to June 1979 (Source: Barrie et al. 1989b). on particle size and physical properties such as water vapor
Atmosphere of soil aluminum measured at Alert during a blizzard in
April 1986 were 5-7 times higher than during low wind con-
ditions (Barrie et al. 1989a). Thirteen years of routine aero-
2 4 5 6 7
Particle Gas Gas Rain Particle sol aluminum observations at Alert show much higher con-
deposition exchange exchange deposition
1 3 8
centrations of suspended soil during windy periods than dur-
Snow Particle Particle ing low-wind weeks (Barrie 1995).
resuspension resuspension
The erosion of surface snow occurs when the shear force
at the air/snow interface exceeds inter-particle cohesive forces
and particle inertia. The two primary modes of snow trans-
Snowpack Lake / port are saltation and suspension. Saltation is a near-surface
Sea Ice phenomenon which occurs at low wind speeds, while sus-
pension can disperse snow to heights of tens of meters and
9 10
dominates as wind speed increases. Both are important phe-
Snowmelt Snow and nomena in the Arctic, especially in the tundra and barren
runoff ice melt
lands. Blowing snow grains become fragmented during colli-
sion with other snow particles or surface snow, resulting in a
change to a more amorphous particle shape (Pomeroy et al.
1991, Pomeroy and Gray 1995) and an associated decrease
Earth’s Surface : Land / Ocean
in surface area.
The effect of surface area changes during blowing snow
F L U X E Q U AT I O N S Flux equations events on contaminant concentrations is not well under-
1. Css
1. C x PPss 6. Cr Pr
5. K(Cg - Cw /H) water 8. not defined stood. Pomeroy and Gray (1995) have shown that wind ve-
2. Vd Cp 7. Vd Cp
2. V
3. d xdefined
not Cp not defined land
8. not defined 9. Crunoff x Prunoff locity, temperature, and relative humidity are important fac-
3. not defined
4. 6. Cr x Pr 9. Crunoff Prunoff 10. Crunoff x Prunoff tors in snow sublimation, which can cause increased concen-
5. K(Cdefined
4. not g – C w/H) water, not defined land
7. V x C 10. Crunoff Prunoff
d p trations of major ions and the loss of volatile organochlo-
Cs = concentration of contaminant in snow Cw = concentration of contaminant in water rines (VOCs). More research is needed on the resuspension
Ps = snowfall rate H = Henry’s Law constant
Vd = particle deposition velocity Cr = concentration of contaminant in rain of contaminants during high winds.
Cp = concentration of contaminant on particles Pr = rainfall rate
K = gas exchange coefficient Crunoff = concentration of contaminant in runoff
Cg = concentration of contaminant in gas Prunoff = rate of runoff
3.2.3.3. Wet deposition
Figure 3·7. Schematic diagram of processes of atmosphere-surface ex- (processes 1 and 6 of Figure 3·7)
change (Source: Barrie et al. 1997).
Wet deposition of contaminants occurs when contaminants
sorption (hygroscopicity). The flux of contaminants is being are incorporated into rain or snow which subsequently falls
studied intensively in many research programs, including to the ground. The process, which is highly complex (Barrie
European and North American acid rain research programs. 1991), includes in-cloud removal and below-cloud removal.
A rigorous treatment requires a knowledge of Cp and Vd for In clouds, nucleation scavenging of particles is followed by
each particle size. However, in practice, a knowledge of the uptake of soluble and reactive gases, as well as Brownian dif-
size distribution of particulate matter measured in the north fusion of particles to form liquid cloud droplets. Then cloud
can be used to obtain an average Vd to estimate dry deposi- water is converted to precipitation (rain or snow) in complex
tion. For the Arctic, most anthropogenic aerosols that reach ways that may be either efficient or inefficient in removing
the region are smaller than 1 m in diameter (Barrie 1986, contaminants, depending on the process of precipitation for-
Barrie 1996b). In this case for snow, a Vd for submicrome- mation. During unstable atmospheric conditions, below-cloud
ter particles, typical of those carrying most contaminants, scavenging is generally thought to be of less importance than
of 0.05-0.1 cm/s has been determined in field studies (Ibra- in-cloud scavenging (Murakami et al. 1983, Schumann et al.
him et al. 1983, Davidson 1989) and applied to estimate 1988). However, when stable atmospheric stratification causes
contaminant deposition (Giorgi 1986, McVeety and Hites contaminants to accumulate within the surface boundary
1988, Jantunen and Bidleman 1995). layer, below-cloud scavenging may contribute significantly
to wet deposition, especially of the larger particles (> 2 m)
(Zinder et al. 1988). The processes of diffusion and intercep-
3.2.3.2. Particle resuspension
tion may be of greater significance for precipitation in the
form of snow, since snowflakes have a larger surface area
It is well established that desert soils can be suspended to than rain droplets.
great heights in the atmosphere and then transported long The concentration of condensation nuclei collected in pre-
distances. For instance, Gobi desert dust is observed in Ha- cipitation may be significantly reduced if conversion from
waii in the central Pacific, and occasionally in the Canadian liquid to solid phase dominates the aggregation of cloud drop-
Arctic (Welch et al. 1991, Barrie 1995, Barrie 1996). Con- lets during the growth of snowflakes. Relatively large sedi-
siderable quantities of snow can be suspended in the atmos- menting hydrometeors (e.g., raindrops, snowflakes) can in-
pheric boundary layer together with suspended soil. How- tercept cloud droplets as they sweep through the cloud, an
ever, little is known quantitatively about the resuspension accretion process known as ‘riming’ when associated with
of snowbound contaminants during blowing snow condi- snowflakes. Since cloud droplets contain most of the aerosol
tions (Pomeroy and Jones 1996) when surface wind speeds mass within the cloud, riming can yield concentrations in
exceed 3 m/s in the Arctic. Surface soil is not covered by precipitation that are similar to cloud droplet concentrations
snow in many regions of the High Arctic which receives little (Barrie 1985, Parungo et al. 1987, Borys et al. 1988, Collett
precipitation during the winter (see Figure 3·5). These ex- et al. 1993). The accretional sweep out of cloud droplets
posed soils commonly exist throughout the year allowing (riming) is considered to be the dominant process controlling
soil dust to be mobilized and mixed with surface snow dur- atmospheric removal by precipitation (Scott 1981, Borys et
ing wind events (see also section 3.3.3.1). Air concentrations al. 1988, Barrie 1991). In Greenland, rimed snow has been
estimated to contribute about 5% of the annual snow mass, It may be that both adsorption and dissolution mecha-
while accounting for approximately 30% of the annual de- nisms operate in rimed snow. Interfacial adsorption to the
position of atmospheric contaminants (Borys et al. 1993). surface water film of snow crystals occurs as does Henry’s
In northern regions, in-cloud scavenging is thought to be Law dissolution into cloud droplets scavenged by rimed
more important than below-cloud scavenging (Scott 1981, snowflakes. Because droplets rimed to snowflakes, freeze on
Murakami et al. 1983, Rehkopf et al. 1984, Schumann et contact with the snowflake, Henry’s Law partitioning behav-
al. 1988). Generally, in-cloud scavenging incorporates sub- ior may be important only if there is negligible loss of con-
micron particles into cloud droplets and precipitation ele- taminants during freezing of the droplet. Some fraction of
ments, while below-cloud scavenging captures supermicron both soluble and particulate contaminants may be excluded
particles (> 2 m) and submicron particles (diameters < 0.01 from the crystalline lattice upon freezing (Uhlmann et al.
m) (Slinn et al. 1978). However, field observations show 1964, Hoekstra and Miller 1967, Gross et al. 1975, Iribarne
that below-cloud rain scavenging coefficients for submicron et al. 1983, Lamb and Blumenstein 1987, Iribarne and Pysh-
particles are about an order of magnitude larger than pre- nov 1990, Iribarne and Barrie 1995).
dicted by theory (Volken and Schumann 1993). Additional Another atmosphere-surface exchange pathway is fog wa-
factors encountered in the ambient atmosphere, such as ter deposition (Barrie and Schemenauer 1986, 1989). It is a
phoretic and electrostatic forces and turbulent diffusion, hybrid of precipitation scavenging and dry deposition proces-
may enhance particle scavenging by snow over theoretical ses. Fogs are essentially clouds at the ground. In some cases,
predictions which neglect these mechanisms (Martin et al. this is literally true, for example when a layer of cloud pas-
1980, Murakami et al. 1985a, Murakami et al. 1985b). ses over a mountain. In other cases, the fog is formed by
The efficiency of the particle scavenging process depends processes (different than those of cloud formation) that take
on a number of meteorological and physical factors, such as place near the Earth’s surface. If air is cooled sufficiently, the
the hygroscopic nature of the particles, the size distribution water vapor in it will condense into liquid drops. The cool-
of both the hydrometeors and atmospheric particles, precip- ing can be produced by several different meteorological situ-
itation intensity and cloud base height (Doskey and Andren ations leading to different types of fogs. At night, the Earth’s
1981, Pruppacher 1981, Slinn 1983, Sauter and Wang 1989, surface loses heat through the transmission of terrestrial in-
Mitra et al. 1990, Sparmacher et al. 1993). The ratio of frared radiation to the atmosphere above. Sufficient surface
pollutant concentration in precipitation to that in air feed- cooling leads to saturation of the near-surface layer and ‘ra-
ing the cloud is the scavenging ratio (Wp). Use of this is diation fog’. Light winds (< 10 km/hr) promote fog develop-
fraught with difficulties (Barrie 1992). However, as a rough ment by mixing cool, moist surface air in the vertical. At
measure of relative scavenging efficiencies, it is useful. It has higher wind speeds, mixing with drier air aloft is so great
been estimated that the coalescence of cloud droplets to rain that fog formation is inhibited.
drops produces particle scavenging ratios (Wp ) of about ‘Advection fog’ is produced when warm, moist air is trans-
106 (on an equivalent volume basis), while Wp values of ported over a colder surface. Most coastal fogs are of this
≤ 105 are indicative of unrimed snow, or scavenging of ei- type. ‘Upslope fog’ is produced when air is forced to rise
ther insoluble particles or particles with diameters of 0.1- over a topographical barrier decreasing atmospheric pressure
1.0 m (Scott 1981). These latter ratios reflect the ineffi- that produces expansion and cooling of the air. Other types
ciency of unrimed snow to scavenged particles within cloud of fogs are less important vehicles of deposition. ‘Steam fog’
droplets, or the inefficiency of accumulation-mode particles or ‘Arctic sea smoke’ is produced by the condensation of wa-
to be removed below the cloud base. By comparison, Wp ter vapor from a relatively warm sea surface in much colder
values for rain, based on field measurements, range from air aloft. When Arctic Ocean ice cracks to form open leads
≈ 103 to 106 (Bidleman 1988). in winter and the atmospheric temperature is much below
Snow is likely more efficient than rain at below-cloud 0°C, this type of fog is observed. ‘Ice fog’ is produced at tem-
scavenging of particles because of the larger size and sur- peratures below about –30°C when water vapor sublimes di-
face area of the snowflakes (Redkin 1973, Graedel and Fra- rectly onto atmospheric ice nuclei (particles) thus forming
ney 1975, Murakami et al. 1983, Raynor and Hayes 1983, tiny ice crystals. This illustrates a characteristic of fogs (and
Parungo et al. 1987, Leuenberger et al. 1988, Schumann et clouds) that is important in the scavenging of contaminants
al. 1988, Nicholson et al. 1991). Particle scavenging at low from the atmosphere, namely that fogs remain liquid (con-
temperatures, typical of the early fall and late spring Arctic taining supercooled water drops) down to temperatures as
seasons (0 to –30°C), may be more effective than dry snow- low as –20°C, and occasionally to much lower temperatures.
fall during the winter, as water film on the crystal surface The means by which particles and gases are incorporated
enhances the probability that particles contacting the sur- into fog droplets of 1-20 micrometer diameter are the same
face will stick (Mitra et al. 1990). In addition, the particle as for cloud droplets. They are then deposited to the Earth
scavenging efficiency of snow depends on the crystalline by processes similar to that of dry deposition of particles:
shape, with needles and columns being less effective than sedimentation for droplets larger than 5 micrometers and
stellar plates, dendrites and snowflakes (Takahashi 1963, impaction on surface roughness elements for smaller par-
Miller and Wang 1991). Both snowflakes, which are aggre- ticles. The microstructure of fogs is strongly influenced by
gates of individual crystals, and dendritic crystals tend to the characteristics of atmospheric cloud condensation nuclei
have a ‘filtering effect’ on atmospheric particles during sedi- (i.e., hygroscopic aerosol particles dominated by sulfates and
mentation as a result of the passage of air through the crys- sea salt). Polluted air masses with high aerosol concentra-
tal pores. This ventilation increases the scavenging efficien- tions generally produce fogs with high fog droplet number
cies, especially for small particles (0.2-2 m) which would concentrations and smaller drop sizes. Sea fogs generally
otherwise follow the streamlines around a nonporous hy- have lower total droplet concentrations and larger droplets.
drometeor, thereby reducing the probability of being scav- Knowledge of the chemical composition of fogwater as a
enged (Redkin 1973, Mitra et al. 1990). Field studies have function of droplet diameter is important in calculating the
demonstrated that snow may be five times more efficient deposition of contaminants in fogs. Unfortunately, reliable
than rain at below-cloud scavenging (Murakami et al. 1983, size-dependent chemical composition data are not available
Sparmacher et al. 1993). for fogs in the polar regions.
nificant potential source for the Arctic atmosphere. The air–
3.2.3.4. Gas exchange between atmosphere
water partition coefficient decreases with decreasing tempe-
and snowpack/ice (process 4 in Figure 3·7)
ratures, and thus, the highest equilibrium concentrations of
Material can be added to the snowpack by snowfall and by OCs in surface waters are expected to occur at cold water
particle dry deposition, while gas exchange can both add temperatures (Wania and Mackay 1993).
and remove it. The partial vapor pressure gradient of an Gas exchange to soils depends on the sorption of conta-
organochlorine (OC) between ambient air and interstitial minant vapors on soil surfaces. It is poorly understood and
air in snowpack or ice determines the potential for exchange. is being studied in other programs focused on the movement
This is strongly dependent on temperature. In general, com- of contaminants in agricultural areas, though usually not at
pounds with relatively high vapor pressures, such as hexa- Arctic temperatures (Eitzer and Hites 1989, Hart et al. 1993,
chlorobenzene (HCB) and - and -HCH, can volatilize Staehelin et al. 1993, Leister and Baker 1995). Models for
substantially from the snowpack, while polyaromatic hy- such exchange are available, but are very uncertain. In forested
drocarbons (PAHs) and many of the polychlorinated biphe- areas (leaf:ground area ratios of 3-19), leaves or needles may
nyl (PCB) congeners with lower vapor pressures may not. represent the primary terrestrial contact for airborne chemi-
Supporting evidence for volatilization from snow is drawn cals (Schulze, 1982) and thus for gas-phase transfers. Uptake
from past and present studies in the Arctic. For instance, at and loss of OCs by Arctic plants, such as lichens, and by
the Canadian Ice Island, HCB and chlordane were detected frozen soils are poorly understood, necessitating further re-
in snow collected from May-June, 1986 (33 and 91 pg/L, search if the air-plant-terrestrial food chain pathway is to be
respectively), though not found in snow (< 1 pg/L) from the better quantified.
same area several months later in August-September (Har-
grave et al. 1988).
3.2.3.6. Snow and ice melt
Several observational studies in the Canadian Northern
Contaminants Program (Barrie et al. 1997) indicate that
the process of post-depositional gas exchange between the Runoff from snowpack melt and lake- and sea-ice melt de-
snowpack and atmosphere is a significant pathway for livers water, particulate matter, and dissolved and particle-
many of the relatively volatile OCs. These include: 1) Agas- associated contaminants from their winter storage reservoirs
siz Icecap studies (Gregor 1990, Gregor 1991, Franz et al. to terrestrial, freshwater, and ocean environments. While at-
1997), 2) Amituk Lake basin studies (Semkin 1996), and mospheric processes are still important (e.g., temperature,
3) snowpack–snowfall chemistry surveys. precipitation), terrestrial and fluvial processes (e.g., runoff,
hydrology, infiltration, permafrost) tend to predominate
with respect to the delivery of contaminants between com-
3.2.3.5. Gas exchange between the atmosphere and soil,
partments. Consequently, these processes will be considered
water, and vegetation (process 5 in Figure 3·7)
in section 3.3.
Gas exchange between the atmosphere and liquid water
(fresh or salt water), vegetation, or soil is commonly a re-
versible process. Most is known about the exchange to wa- 3.3. Terrestrial /Freshwater
ter, that is usually described by the two-film model (Liss
3.3.1. Introduction and scope
and Slater 1974), in which the rate of transfer is assumed to
be limited by molecular diffusion across thin air and water The Arctic land mass, which includes freshwater lakes and
films at the interface. The net gas flux can be estimated by: river systems as well as the islands of the Arctic Ocean and
adjacent seas, measures a full 13.4 106 km2 or about 40%
F = K [ C w – CaRT / H ] (1)•
of the total AMAP area. This compartment, referred to as
1/ K = 1/ k w + RT / H k a (2)•
the ‘terrestrial/freshwater’ compartment in Figure 3·1, is im-
In these equations, C w and C a are the concentrations of portant in its own right in that it supports terrestrial and
dissolved and gaseous chemicals in water and air, kw and ka freshwater ecosystems, but also because its large surface area
are the mass transfer coefficients for the individual water serves as a receptor for atmospherically transported conta-
and air films, H is the Henry’s Law constant at the water minants. The emphasis of this section is on the general fea-
temperature, R is the ideal gas constant, and T is the air tures of the sources of contaminants to this compartment
temperature (°K). K is the overall mass transfer coefficient and the processes that determine their fate, including the
that takes into account the resistance to transfer offered by connective function that the terrestrial/freshwater compart-
both the air and water films. ment provides between the atmosphere and the oceans. For
These relationships have been applied to the exchange of example, while surficial geology generally determines infil-
OCs by Bidleman and McConnell (1995) and Cotham and tration rates and groundwater conductivity, permafrost is an
Bidleman (1991) as well as to the calculation of Arctic Ocean overriding factor with respect to groundwater in cold re-
HCH budgets (Barrie et al. 1992, 1997). It is important to gions. Similarly, ice and snow play a dominant role in the
recognize that gas exchange takes place even under equilib- physical and chemical processes influencing contaminant
rium conditions when invasion and evasion are balanced transport and fate.
and the net flux is zero. Thus, the mass of material trans- Different types of development associated with urbaniza-
ferred by gas exchange is much larger than is apparent from tion and industrialization, including construction of roads
the net flux. For this reason, Murphy (1995) suggested that and pipelines, directly contribute to erosion (Claridge and
gas exchange be calculated as two separate terms, one for Mirza 1981). Where the ground cover has been disturbed
volatilization and the other for deposition. or removed, subsidence and trapping of water may follow,
The oceans represent a large reservoir of some contami- resulting in an accelerated degradation of permafrost. If
nants, with total burdens thousands of times greater than water begins to flow, hydraulic erosion may follow and re-
those found in air. Given this, and the fact that the Arctic sult in the formation of erosion gullies (Claridge and Mirza
Ocean covers the majority of the area considered to be with- 1981) and export of particulate matter from the area. Defor-
in the Arctic region, the air/water exchange represents a sig- estation is often the cause of increased erosion and in tem-
n n
p o s it io ti li z a ti o
e ri c d e R e v o la
d ry a tm o s p h
d
Wet an R e v o la
ti li z a ti o
n
n
ti li z a ti o
R e v o la
W eo gt la nd )
(b /s w a m p
ta l
n C o ausn o ff
hed rm a ti o r
W a te r sr T ra n s fo
Lake te n
n rm a ti o
n wa
und nage Ocea
e n s io T ra n s
fo r o i
esusp G dra
R u n o ff
R
nd o u tf lo w
e n ta ti o
n W e tl a
S e d im +
tation in
W ate imenrmation zone
S e dn
pa rti cu lat e r + n fo e
ma sp e n s io s
tra /estuar in
tra ns po rt tte r tr ia l Resu delta
us
ip a l In d
M u n ic ti o n
e n ta
at io n S e d im
g sform
Riv M in in Tran
erin
e tr
a nsp
ort
A g ri c u
lt u re
Figure 3·8. Conceptual model of the terrestrial/freshwater compartment

showing the main subcompartments and contaminant transfers, and
exchanges with the atmosphere and oceans.
perate areas can result in an increase in sedimentation of

3.3.2. Sources of contaminants
between 20 and 50 times.
3.3.2.1. Atmospheric deposition
Rather than considering each and every process and its
role with respect to contaminant transport, this discussion The most important, large-scale contaminant delivery pro-
will emphasize the processes that are unique or specific to cess to the terrestrial/freshwater compartment is deposition
the Arctic. In addition, the mass transport of water, sedi- from the atmosphere through wet and dry particle scaveng-
ment, and organic matter, which are important determi- ing as discussed in section 3.2. The influence of atmospheric
nants for the transfer of contaminants in fluvial systems are contaminants on Arctic freshwater systems starts with their
quantified where possible. This will set the stage for discus- deposition onto surface waters (lakes, rivers, wetlands) or
sions on the delivery and fate of these contaminants to the land surfaces in their catchment area. While the total surface
marine system including deltas, estuaries, and fjords which area of some lakes, reservoirs, and rivers is large, it remains
provide an interface between the terrestrial/freshwater com- small relative to the total land surface area (< 1%), and thus,
partment and the ocean compartment. deposition directly to the water/ice surface will not be con-
The terrestrial/freshwater compartment is multi-faceted sidered separately from the land mass. Reports describing
and Figure 3·8 provides an overview of the components of the input of contaminants to remote freshwater systems by
this compartment, the transfers and processes that will be long-range atmospheric transport include Gregor and Gum-
considered, and the main sources of contaminants. Note mer (1989), Barrie et al. (1992), Mackay and Wania (1995),
that precipitation, including snowfall, is part of the atmos- Oehme et al. (1995), and Barrie et al. (1997).
pheric compartment discussed in section 3.2, whereas snow Land and freshwater can also be contaminated via the at-
on the ground (snowpack and ice) and snowmelt are included mosphere by sources within the Arctic. These contaminants
here. The other main components of the terrestrial/freshwa- are often of the single-hop variety as defined in section 3.2,
ter system are the land surface and subsurface, wetlands, although the heavier, multi-hop contaminants could also ini-
rivers, and lakes/reservoirs. Both water and suspended and tially be deposited close to the emission source (e.g., higher
bottom sediments are considered. These components are re- chlorinated PCB congeners and PAHs), especially during the
viewed below, presenting their unique character and magni- colder winter months, and then selected, more volatile com-
tude in the context of the Arctic system. pounds could be remobilized and transferred over greater
As illustrated in Figure 3·8, contaminants are delivered distances during warmer periods. This is most important in
to the terrestrial/freshwater environment by the atmosphere the vicinity of large urban and industrial complexes and
and from direct discharges of wastes to land and water. consequently, this pathway will be dealt with as part of the
Large river systems that drain to the Arctic, but that have a point source discussion below.
major portion of their drainage basin outside of the Arctic,
including many of the rivers of Russia (e.g., Yenisey, Lena,
3.3.2.2. Direct wastewater discharges
Ob) as well as several Canadian Rivers (e.g., Mackenzie,
Churchill and Nelson), are important conduits to the north- Development has a direct influence on water quality. In ad-
ern marine environment. While details of specific contami- dition to the effects on the environment from land clearing,
nant loadings from point and diffuse sources will be consid- construction, and decreased permeability, the quality of wa-
ered in subsequent chapters, general features of sources rel- ter returned to the basin is usually diminished even after
evant to their interaction with physical processes in a cold treatment. The disposal of industrial and municipal waste
environment will be provided in this section. waters is always a concern. Storm and meltwater runoff,
which is usually routed directly to receiving water bodies els of heavy metals in wetland ecosystems are generally low,
without treatment, may be highly contaminated due to with the exception of areas with intensive heavy industrial
spills and localized atmospheric fallout. These factors can development (Zhulidov et al. 1977). (See chapter 7 for addi-
exacerbate the problems of providing clean water to basin tional details regarding heavy metals).
residents. This and the subsequent section provide some ex-
amples of the industrial and municipal sources of contami- Alaska
nation within the Arctic. Large volumes of used drilling fluids have been released into
Arctic wetlands during exploration. Increased concentra-
Russia tions of common and trace metals and organic hydrocar-
The large Arctic drainage basin within Russia has been de- bons have been found in ponds both near and at a distance
veloped, especially along the major rivers which have pro- from drill site sumps (Garland et al. 1988). In addition, there
vided a transportation corridor. Information on discharges are mining activities which are potential sources to the envi-
from many of the industrial sources was not available. ronment of acid mine drainage and trace elements.
Oil and gas extraction activities have led to significant
contamination of vast areas of adjacent land containing Canada
wetlands. For instance, in northwest Siberia, a wide range Over 1200 oil and gas exploratory wells have been drilled
of contaminants have been discharged, both from natural on land in the Yukon and NWT (COGLA 1989), mostly in
seeps which are typical for oil and gas deposits, and from the western portion of the NWT and adjacent portion of the
anthropogenic sources. The anthropogenic contaminants Yukon, from the 60th parallel to the Arctic Islands. Drilling
are typically used in the processes of extraction, pre-treat- fluids, containing a range of contaminants from common
ment and transportation of raw products (e.g., petroleum metal salts, surfactants, and petroleum hydrocarbons, have
hydrocarbons (PHC), phenols, nitrogen compounds, heavy, commonly been disposed of in sumps adjacent to the rigs.
alkaline and alkaline-earth metals, diethylenglycol (DEG), French (1978) observed that approximately one-third of the
and methanol). These waste waters are usually discharged sumps had problems, including:
in an untreated condition into landscape depressions. Under
• non-containment of fluids during drilling;
northern temperature and light conditions, the degradation
• melt-out problems during summer operations;
rates for many organic contaminants are likely lower than
• sump subsidence and collapse;
farther south. Horizontal and vertical drainage rates are
• non-containment of fluids during in-filling of sumps; and,
also low in these areas due to the low relief and presence of
• subsurface leakage of fluids.
permafrost. Hence, it is expected that in the areas surround-
ing oil and gas extraction activities, polluted wetlands can Nonetheless, detailed studies of abandoned sumps in the
represent long-term sources of secondary contamination of NWT by Smith and James (1979) indicated that contami-
surface waters, particularly during seasonal and rain floods, nants in soils and plants were localized to within 100 m of
even after cessation of the extraction activities (NEFCO the sump.
1995). These releases and their effect on the environment As of 1986, 14 mines were producing lead, zinc, silver,
will be considered further in chapter 10. gold, copper, tungsten, cadmium, or arsenic. Prior to this
Forests destroyed by air pollution primarily from large time, 68 mines had been developed and abandoned (Thomas
metal smelters in Pechenga and Monchegorsk, have become et al. 1991). The concern with these abandoned mine sites is
typical of the Kola Peninsula (Igamberdiev et al. 1995, primarily the refuse that has been left behind, including fuels
NEFCO 1995). This impact on forests extends into the and waste oils. The main active mines in the Northwest Ter-
northern areas of Lapland and Finnmark, Norway. The ritories (NWT) are the base metal mines at Polaris on Little
deforestation results in more runoff and reduced infiltra- Cornwallis Island and at Nanisivik on Baffin Island, and
tion. In turn, this increases the amount of erosion in the gold mines at Contwoyto Lake and Yellowknife. Perhaps the
basin, increasing the sediment load and any associated con- largest single source of anthropogenic contaminants are the
taminants to the fluvial systems. Within the river channel, gold mines in the vicinity of Yellowknife, where the presence
higher flows result in greater erosion and transport of sedi- of sulfides requires roasting prior to the cyanide process.
ments. Investigations of the effects on water quality and The fumes produced from the roasting process can contain
contaminant transport in these areas are scarce. significant amounts of arsenic trioxide, and sulfur dioxide,
The problem of poor drinking water supply in Monche- some of which are released to the atmosphere (Thomas et al.
gorsk City on the Kola Peninsula is one example of signifi- 1991). Processing of the ores often results in the pollution of
cant local industrial impact on water bodies. The tap water both land and water with heavy metals, but this has been
supply of this city is taken from Lake Moncha, which is lo- relatively small in magnitude (see also chapter 7) (Allan 1979,
cated a few kilometers from the city, where a large nickel Mudroch et al. 1989, Mudroch et al. 1992).
smelter, Severonickel, is located. Channeled waste waters are Base metal mines commonly involve the processing of sul-
not discharged into the lake. However, contaminants are trans- fide rocks. The resultant acid mine drainage from the tail-
ported into the lake either by direct atmospheric deposition ings can often solubilize and consequently mobilize metals.
or with meltwaters and surface runoff. As a result, nickel To date, more than 1011 kg of tailings have been produced
concentrations in this lake reach 0.05 mg/L or five times the (Thomas et al. 1991). Mine tailings can become a long-term
maximum allowable concentration (NEFCO 1995). source of metals to surface waters, while the acid drainage
Wetlands are known as efficient scavengers of many types in and of itself can be a concern in poorly buffered aquatic
of wastes, and, indeed, have been used under various cir- systems. For example, acid mine drainage is not a major
cumstances as surface water quality cleansers. In areas with concern in the vicinity of Nanisivik on Baffin Island due to
strong anthropogenic impact, wetlands can serve as accu- presence of dolomite bedrock, whereas acid mine drainage
mulators of contaminants, but subsequently may serve as and the mobilization of metals (especially Zn) are a definite
sources of significant secondary contamination of river wa- and ongoing concern at the Faro mine in the Yukon, which
ters. Based on a study of more than 250 wetlands in the began production in the 1970s (R. Nicholson, pers. comm.,
Russian sector of the Arctic, it has been shown that the lev- Department of Earth Sciences, University of Waterloo, 1996).
pean part and in western and central Siberia. It is only re-

3.3.2.3. Municipal wastewater sources
cently that attempts have been made to quantify the most
Rivers draining to the Arctic Ocean receive contaminants hazardous sources in the Arctic region of Russia (Igamber-
from different sources within and outside their catchment diev et al. 1995).
areas. Direct discharges of industrial, mining, and urban The impact of anthropogenic contamination on Arctic wa-
wastes into water bodies are the main sources of surface terways can be illustrated by the condition of waste water dis-
water contamination, unless stringent discharge regulations charged to waterbodies in Murmansk province (Kola Penin-
are in effect and enforced. The Arctic rivers with vast basins sula), which is the most industrialized area of the whole Arc-
may receive contaminants discharged in heavily industrialized tic region. In 1994, only 4.7% of the waste water in the pro-
zones and large cities located far to the south of the Arctic vince was treated to specified standards. Approximately 6.1%
region. For example, in the territory of Russia, southern of the waste water (103.5 106 m3) was discharged into wa-
reaches of the large Arctic rivers include large industrial terbodies without any treatment and 13.6% (229.1 106 m3)
complexes and cities with populations often exceeding a was considered insufficiently treated. The remaining 75.6%
million inhabitants. (1269.8 106 m3) was classified as ‘conditionally clean’ and
The Barents region of the Russian Arctic is the most high- allowed to be discharged without treatment. Most of this
ly populated and industrially developed part of the circum- latter group is generated by cooling waters of the Kola nu-
polar Arctic. The province of Murmansk has over one mil- clear power plant (NEFCO 1995). The distribution of
lion inhabitants with many in major industrial cities, includ- wastewater discharge rates in Russia are summarized in
ing Murmansk, Apatity, Kirovsk, Kandalaksha, and Mon- Table 3·1. Amounts of contaminated waters and releases
chegorsk. To the east of Murmansk is the province of Arch- of selected contaminants in the Murmansk province are
angel, which, including the Nenets autonomous area, in given in Table 3·2. Murmansk city discharges its waste wa-
1993 had a total population of 1561 000 inhabitants with ters partly into the Kola Fjord and partly into freshwater
over 70% in major cities (including the cities of Archangel bodies. The Kola River, which serves as a source of drink-
and Severodvinsk). In highly urbanized areas, surface runoff ing water for Murmansk, receives discharges of municipal
waters, including organized industrial and communal waste- and agricultural waste, significantly affecting the quality of
water discharges, can contribute up to 40-50% of total pol- this supply.
lution entering the water bodies (WHO/UNEP in press). There are also the so-called fugitive discharges of un-
In contrast, the large drainage basins of North America known content and size. These include dump sites and spills
have relatively small populations with limited industrial de- and require further investigation to determine the scale of
velopment and relatively stringent environmental controls. potential and real emissions.
The largest cities in the Canadian Arctic are Yellowknife
and Iqaluit in the NWT, with populations of approximately Canada
12 000 and 3000, respectively, and Whitehorse in the Yukon, There are only five major cities in Canada that discharge ef-
with about 15 000 inhabitants. fluents into rivers which drain to the AMAP region. All of
these, identified in Table 3·3, are within the Nelson River sys-
Russia tem, with the closest, Winnipeg, being more than 1000 km
Arctic rivers in Russia are affected by untreated or partially from the outlet to Hudson Bay. It should be noted that all in-
treated waste water (Table 3·1) due to the high level of industries in these cities are required to discharge through the
dustrial development in their basins, especially in the Euro- sewage treatment plants, except for uncontaminated wastes,
Table 3·1. Range of annual discharges of waste waters to the river basins such as cooling waters. As mentioned above, within the Arc-
of the Russian Arctic between 1988 and 1993 (106 m3/y) (Source: CIP 1995). tic, the cities of Whitehorse, Yukon and Yellowknife, NWT
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– are relatively small in population and have little industry.
Territory Range
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
POPs and radioactive material are thought to reach the
Northwest Region Arctic areas of North America mainly through atmospheric
Republic of Karelia a 206.1-288.1 long-range transport. However, a few local sources of OCs
Komi Republic 168.7-193.7
Archangel Province 642.5-696.7 have been disclosed. Military ‘DEW Line’ (Distant Early
Vologda Province 240.5-325.6 Warning Line radar) installations deployed between 1950
Murmansk Province 260.3-382.1 and 1970, used DDT mixtures, as well PCB-containing de-
Urals vices (electrical transformers and capacitors) widely. Follow-
Chalyabinsk Province a 608.9-707.9
Sverdlovsk Provincea 728.3-864.1
ing disposal, vandalism and natural processes resulted in the
Kurgan Province 13.1-25.5 release of some of the PCBs to the environment (see chapter
Western Siberia 6 for additional details).
Republic of Altai b 0.1-0.3
Altai Territory 61.2-117.8 Greenland
Kemerovo Province 567.4-768.3
Novosibirsk Province 326.0-360.6 Apart from mining, there are no other major industrial en-
Omsk Province 136.7-291.1 terprises in Greenland. A few mines are known to discharge
Tomsk Province 47.3-68.3 heavy metals directly into the marine environment. How-
Tyumen Province 69.1-183.8
ever, according to available sources, there is only an insigni-
Eastern Siberia
Republic of Tuva 6.1-13.4 ficant level of discharge into rivers (M. Holm, pers. comm.,
Republic of Khakassiab 71.5-78.1 Greenland Hjemmestyre, 1996).
Krasnoyarsk Territory 785.4-1405.5
Irkutsk Province 1353.7-2056.8
Iceland
Far East
Republic of Sakha (Yakutia) 44.8-95.4 Only a very limited amount of heavy metals, POPs, and
Chukchi Autonomous Area a,b 10.2-11.0 PAHs are suspected to be directly discharged into lakes and
Magadan Province a 31.4-53.4 rivers of Iceland. There are no major mining activities in-
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
a. Only a portion of the Territory belongs to the Arctic basin. volving metals, and hence the risk of heavy metal discharge
b. Data available only from 1988-1990. into freshwater is low.
Table 3·2. Annual wastewater discharges in the cities and counties in Murmansk province in 1994 (Source: ECMP 1995).
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Amount of contaminants discharged (tonnes/y)
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Annual Petrol Sus-
wastewater hydro- pended
City, district volume, 106 m3 BOD carbon matter N-NH4 N-NO3 Al Fe, total SO4 P-PO4 F CI –
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Murmansk 65.2 7429 56.4 5347 871 45.9 27.5 1531 116 4960
Kirovsk and Apatity area 1356 872 41.7 1637 416 4051 15.0 8191 87 297 2357
(Kola nuclear power plant) (1172) (8.6) (1.3) (9.3) (4.6) (110) (0.3) (22)
Kandalaksha 41.5 438 5.5 436 30.9 28.6 0.03 3.0 0.8 12.1 313
Monchegorsk 24.6 139 3.0 721 53.0 11.3 0.02 34679 14.4 7344
Olenegorsk 14.7 80 3.4 97.2 47.5 77.1 1.6 122 6.2 235
Polyarny 16.1 430 8.2 445 167 18.8 0.04 377 22.6 1088
Severomorsk 17.6 903 6.1 512 146 15.7 0.1 236 18.5 373
Kovdor County 39.5 251 2.4 265 26.7 107 0.27 14.0 4543 32.4 876
Kola County 44.8 246 1.7 202 72.5 46.8 0.48 1.4 80.5 16.6 1.3 222
Lovozero County 17.7 103 2.3 317 25.4 7.3 0.2 103 4.6 115.6 131
Pechenga Country 37.7 279 31.3 249 80.7 101 0 0.8 2252 0.1 0 377
Murmansk Province (total a) 1680 11487 162 10387 1955 841 0.78 63.9 52170 332 415 18343
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
a. This total includes additional discharge areas to those listed in the table.
Table 3·3. Wastewater treatment plant effluent discharge data for major Canadian cities discharging into rivers draining to the AMAP region. (Data
sources: Gold Bar Waste Water Treatment Plant (Edmonton); City of Calgary Engineering and Environmental Services Department (Calgary); Pollution
Control Plant (Saskatoon); Treatment Plant and Public Works Department (Regina); City of Winnipeg and Wastewater Department (Winnipeg).
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Treated Amount of contaminants discharged (tonnes/y)
Population wastewater ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
City served 6 3
10 m /yr N-NH4 P-PO4 N-NO3 Al Cl Fe, total SO4 Mn2+ Ni2+ BOD
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Edmonton 740000 940 n.r. n.r. n.r. n.r. n.r. n.r. n.r. n.r. n.r. 8460
Calgary a 767059 146 461.0 104.0 1055.4 51.6 n.r. 12.9 17245.4 n.r. 2.7 699.3
Regina 183000 26 641.3 71.8 2.6 20.5 5900 10.3 7952.1 4 < 0.5 359.1
Saskatoon 200000 33 n.r. 2404 n.r. n.r. n.r. n.r. n.r. n.r. n.r. 2800
Winnipeg a 607000 122 2708 n.r. 219.8 n.r. n.r. n.r. n.r. n.r. n.r. 2629.6
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
n.r..: not reported
a. Since Calgary and Winnipeg both have more than one wastewater treatment facility (two and three plants, respectively), the contaminant level totals
from the different plants are summed.
aquatic media. There are no other Norwegian industrial ac-

Norway tivities in the area, except for a thermal electric plant using
The few sources of direct discharges of contaminants by in- local coal and a waste incineration plant. All of these dis-
dustrial and mining wastes in northern Norway are listed in charges are directly to the ocean. Russia operates mines and
Table 3·4. The production and use of POPs are negligible, related facilities on Spitsbergen, but no information has been
and most of the emissions are not discharged into freshwater provided with respect to discharges.
systems, but directly into the marine environment or the at-
mosphere. Mines represent the main local sources of heavy
3.3.2.4. Agricultural sources
metal pollution. Although most mines have been closed down,
many of the abandoned sites are still a source of contamina- The potential for agricultural activities to act as sources for
tion. The Pasvik River on the Russian-Norwegian border is pesticides to the Arctic environment is real. The global culti-
contaminated with heavy metals originating in the nickel plants vation intensity, as a percent of the total land area within
of the Kola Peninsula. However, the actual amount of heavy each zone, is shown in Figure 3·9. Soil can serve as both a
metals discharged directly to the river system is unknown. long-term and short-term storage reservoir of, for example,
Contaminated effluent from coal mines and fuel tanks on pesticides, if they are not degraded. There are two pathways
Spitsbergen represent the major hazardous discharges into for agricultural pesticides to be transported long distances.
Table 3·4. Large industrial plants and mines in northern Norway with emissions of hazardous substances. Estimated total yearly releases (1985 and
1993) to freshwater courses from mining areas (Source: Arnesen and Iversen 1995) and amounts discharged yearly by industries (1989 and 1993)
(Source: Industrial Database, Norwegian State Pollution Control Authority).
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Discharges, tonnes/y a
Company Place Plant / mine Operating status 1985 b/1989 c 1993 Receiving water body
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Løkken Gruber Trondheim Mines Closed down 1987 48 Cu 60 Zn 18 Cu 64 Zn Orkla watercourse
Skorovas Gruber Namsos Mines Closed down 1984 12 Cu 35 Zn 5 Cu 26 Zn Local freshwater (to Namsen River)
Elkem Mosjøen Aluminum plant 1 PAH 0.2 PAH Vefsnafjord (marine)
6.1 PAH Air
Elkem Rana Smelting works 0.1 Cr < 0.003 Cr6+, Ranafjord (marine)
0.2 kg PAH < 0.1 Cr-T,
Pb, Cu, Zn 0.05 PAH,
Pb, Cu, Zn
Koks-verket Rana Chemical industry Closed down 18 PAH
Sulitjelma Gruber Sulitjelma (Fauske) Mines Closed down 1991 40 Cu 54 Zn 28 Cu 25 Zn Sulitjelma watercourse
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
a. Unless otherwise indicated.
b. Data for mines are from 1985.
c. Data for industries are from 1989.
north from agricultural use areas upstream in the drainage

basin (McCarthy et al. 1997).
Another indirect agricultural effect is the impact of rein-
deer farming on vegetation and erosion. Overgrazing by
reindeer is an increasing problem in Norway, Sweden, Fin-
land, and Russia. Locally, overgrazing has resulted in a total
disappearance of the lichen cover (Johansen and Tømmervik
1992). This can lead to soil erosion in those areas where re-
establishment of the vegetation is slow, causing mobilization
of particle-bound contaminants. Quantitative data on the
impacts of agricultural activity on the chemistry of the Rus-
sian Arctic rivers were not available. There are also indica-
tions that areas in North America might be overgrazed by
Global cultivation intensity caribou (Bergerud 1990, CAFF 1994).
20% cultivated 75% cultivated
50% cultivated 100% cultivated
3.3.3. Terrestrial processes
3.3.3.1. Snowpack and snowmelt
Figure 3·9. Global distribution of cultivation intensity based on a 1°1° lat-
itude/longitude grid (Source: Canadian Global Emissions Interpretation Although snow is just one form of precipitation in the Arc-
Center (CGEIC) 1996). tic, it merits separate discussion for a number of reasons.
First, snow is an important medium for the wet deposition
The first is via the atmosphere. Transport can occur during of contaminants in the Arctic, not only because winters are
application and for extended periods thereafter, especially long and snow is the main form of precipitation, but also
following precipitation and cultivation. The second path- because snowflakes are effective scavengers (Gregor 1996).
way is riverine. This pathway tends to be slower, often with Nonetheless, concentrations of contaminants in the atmos-
transport only in the spring. phere are, on average, higher in summer than in winter (Bar-
While there is little agricultural activity within the bound- rie 1995 and 1996b, Pacyna 1995, Fellin et al. 1996, Oehme
ary of the Arctic as defined for this report, the agricultural 1996). Second, the snowpack is a major storage variable in
lands of the entire northern hemisphere are potential sources the overall water budget of Arctic systems, and contaminants
for atmospheric delivery of contaminants. As well, the large accumulated in the snowpack during the winter are released
rivers in Canada and Russia which drain to the Arctic, often to the underlying land and water during a short period of
have drainage basins extending far outside of the Arctic that snowmelt in spring. Third, snowfall is to a large extent de-
contain extensive agricultural activities, which represent a coupled from the entire terrestrial compartment throughout
potential source of contaminants. the winter season because of frozen soil.
Especially noticeable in Figure 3·9 is the intense agricul- The measurement of snowfall quantity is difficult. Con-
ture in the southern midwest of Canada which lies within sequently, snowfall data should be used with caution as a
the drainage basins of the Nelson River system. This river result of problems related to the wind-induced under-catch
system is routinely monitored for agricultural contaminants of various instruments, the sparseness and location of the
by Environment Canada, but there has been no intensive sampling network, the redistribution of snow by wind, and
study of the Nelson River with respect to the delivery of the effect of sublimation (Kane 1994, and see also, for ex-
agricultural chemicals. Although the agricultural activity of ample, Larson and Peck 1974, Woo and Marsh 1977, Ben-
the Mackenzie River basin is not as intensive as in the Nel- son 1982, Woo et al. 1983, Sturgess 1986, Clagett 1988,
son, there has been some investigation of the transfer of Thomsen 1994). Difficulties in snowfall measurement are il-
contaminants from the southern part of the watershed into lustrated for Alert, Canada in Figure 3·10. Using standard
the Arctic as part of the Slave River study. While this is con- Canadian snow gauges, a large number of the days when
sidered further in chapter 6, there is little evidence of signi- snow did fall without wind would have actually been re-
ficant transport of selected current-use pesticides to the ported as trace quantities and not quantified, because the
Precipitation
mm/d
2.5
Alert
1.5
Standard Gauge
1 Detection Limit
0.5
No
Wind Wind snow Wind Wind No snow Wind
0 26 27 28 29 30 31 32 33 36 37 38 40 41 42 43 44 45 47 48 49 50 51 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 102 104 106 108 110
Day of year
Figure 3·10. Precipitation rate measurements at Alert from January through April 1992 showing periods of no snow, periods of no snow measured due to wind
events, and the relatively large quantities of precipitation on days during which precipitation rates were less than the detection limit of the standard gauge
and which would normally be unreported. During periods marked ‘wind’, no measurements were possible due to the blowing snow (Source: Barrie 1996a).
accumulated snowfall was less than the standard gauge forces and particle inertia. The two primary modes of snow
detection limit. transport are saltation and suspension. Saltation is a near-
Snowpack measurements, pioneered extensively in Scan- surface phenomenon that occurs at low wind speeds, while
dinavia and in Finland in particular, where records of snow suspension can disperse snow to heights of tens of meters
accumulation extend back to the early and mid-18th cen- and dominates as wind speed increases. Blowing snow grains
tury (Kuusisto 1994), tend to be more reliable for determin- become fragmented during collision with other particles or
ing snow accumulation. Extensive research and develop- surface snow and may experience sublimation during sus-
ment work into measurement techniques have been under- pension (Pomeroy et al. 1991). Franz et al. (1996) noted
taken in Norway and Sweden (Bengtsson 1994, Killingtveit that to their knowledge the effect of wind transport on re-
1994), including the development and calibration of snow tention or, conversely, volatilization of contaminants, has
pillows (an automated system for measuring snow water not been investigated. As noted by Pomeroy et al. (1991),
equivalents), snow measurements by means of gamma radi- sublimation may increase the chemical concentration of the
ation, and airborne surveys using surface and airborne radar blowing snow if the contaminants do not volatilize while
and remote sensing (Kuusisto 1994). being blown about as has been observed for conservative
In general, annual snowpack depths range from 20 to 40 substances including acids. Volatilization of semivolatile
cm in the Canadian eastern Arctic and Arctic islands and in OCs may be attributable to the reduction in crystalline sur-
central Siberia to several meters in Iceland, the coastal re- face area that begins to occur immediately upon deposition
gions of Norway, and the west coast of North America (in- of contaminants (Gregor 1996). While additional scavenging
cluding parts of Alaska and the Yukon). This regional var- may occur during suspension as the snow grains encounter
iability is mainly determined by large-scale meteorological atmospheric vapors and particles, it is not likely to offset the
patterns, proximity to relatively warm oceans, and topo- loss resulting from the reduction of surface area.
graphical extremes (Killingtveit and Sand 1991). During spring melt, percolation of meltwater through the
Snowpack accumulation is also susceptible to mesoscale snowpack preferentially elutes soluble compounds from the
variability. Wind drift of snow is the most important pro- snowpack. Volatilization to the atmosphere may occur at
cess at this scale, causing scouring at exposed locations and this time due to both the continued metamorphosis of the
deposition of densely packed snow at sheltered sites. Den- snowpack reducing surface area and the warming tempera-
sity of snow redistributed by wind ranges between 0.25 and tures. However, it has been suggested that deep snowpacks
0.50 g/cm3, prior to the onset of melting (Grigoriev and So- are able to retain volatile contaminants, such as HCHs, more
kolov 1994). The resultant snow distribution is largely re- effectively than shallow snowpacks (Semkin 1996). In addi-
lated to local variation in topography (Killingtveit and Sand tion, degradation of compounds, such as PAHs, may occur
1991), with the bulk of the snow accumulating in ravines in the snowpack as a result of photolysis or through liquid
and river valleys, frequently to depths of many meters, phase reactions with hydroxyl radicals (Jaffrezo et al. 1994).
while ridges are snow free, and intermediate areas have rel- Chromatographic fractionation of contaminants in the
atively shallow snowpacks. This redistribution of the snow- snowpack occurs by meltwater percolation that preferenti-
pack is important, because the net result is that much of the ally elutes soluble compounds, leaving residues that are less
land surface area has little snowcover, which frequently sub- soluble or bound to solid surfaces within the snowpack. For
limates, generating little or no runoff. On the other hand, inorganic ions, the initial 20-30% of meltwater may remove
deep snowpacks within the depressions and watercourses 40-80% of the snowpack chemical burden (Johanessen and
serve to retain surface meltwater, thus increasing the water Henriksen 1978, Davies et al. 1982, Cadle et al. 1984a, Cadle
content of this snowpack during the ripening stage. At et al. 1984b, Brimblecombe et al. 1985, Tranter et al. 1986,
spring melt, the Arctic snowpack disappears, releasing large Brimblecombe et al. 1987, Williams and Melack 1991). The
quantities of water in a period of only a few weeks. Melt scavenging processes responsible for incorporating atmos-
rates and total runoff are enhanced if spring melt coincides pheric contaminants into snow influences their elution from
with rainfall (Woo and Sauriol 1980). In addition to the di- snowpack (Tsiouris et al. 1995). Soluble contaminants scav-
rect impacts on the hydrograph of Arctic river systems, this enged during snowfall by attachment to the surface of the
phenomenon influences the quantity of specific contami- crystal are more available for removal during the initial melt-
nants released to surface waters (see chapter 6). ing period than soluble contaminants within the crystal.
The most extreme manifestation of snowpack accumula- Fractionation of organic compounds (such as PAHs, OCs,
tion is the development of polar and alpine icecaps and gla- etc.) depends upon their solubility and dissolved particle par-
ciers (see chapter 2). Icecaps can provide a long-term sink titioning behavior. Relatively soluble compounds (e.g., HCHs)
for snow-scavenged contaminants, while glaciers influence may be enriched in initial meltwaters, whereas those that are
the hydrologic regime of rivers and can deliver snowmelt adsorbed to particles (e.g., PCBs, PAHs) may remain in the
contaminants to surface waters, including oceans. snowpack (Schöndorf and Herrmann 1987). The Amituk
In addition to the physical weathering of snow crystals by Lake study in the Canadian Arctic Archipelago (Semkin 1996,
wind transport, fresh snow crystals metamorphose as they Barrie et al. 1997) has shown that most OCs left in the snow-
mature. The intricate crystalline structure changes into a more pack immediately prior to melt enter the lake in streams ex-
compact structure because of the instability caused by large periencing little snowmelt-soil interaction, and leave in the
surface-to-volume ratios (LaChapelle 1969). The surface free river outflow. The exception was PCB which showed 40%
energy of the snow crystal is reduced by transferring water retention in Amituk Lake. Mercury cycling was similar to
molecules through the vapor phase to achieve a structure that of a volatile OC. Post-depositional and snowmelt pro-
that is rounded and more compact. This process, referred to cesses of semivolatile contaminants in snowpacks require ex-
as destructive metamorphism, occurs in the absence of any tensive and careful experimental design and interpretation of
temperature gradient within the snowpack and is most rapid results to adequately quantify contaminant loads to surface
near the freezing point. The rate diminishes with tempera- waters (Franz et al. 1996, Barrie et al. 1997).
ture, and below – 40°C, it does not occur (LaChapelle 1969). The transport and migration of compounds through the
The erosion of surface snow occurs when the shear force snowpack is highly dependent on the ‘premelt’ condition of
at the air/snow interface exceeds the inter-particle cohesive the snow. Shallow snowpacks exposed to above freezing
temperatures may show extensive fractionation of contami- and surficial geology of the entire Arctic region are provided
nants by meltwater elution. Deep snowcovers, however, may in chapter 2, it is not possible to generalize the impact of
retain most of the meltwater either within the snowpack or surficial and bedrock geology on runoff and infiltration, due
within ice lenses that form at depth, thus delaying the melt- to the heterogeneous nature and complex interaction of vari-
water wave front from exiting the snowpack until extensive ables that determine these processes in large Arctic water-
melting has occurred (Franz et al. 1996). Arctic snowmelt sheds. The integrated effect of physical controls on runoff in
chemistry is further complicated by the presence of perma- large watersheds is illustrated by a comparison of hydro-
frost which can refreeze meltwaters, creating a basal ice graphs for northern rivers (see section 3.3.4.1).
layer. Refreezing of early meltwater can delay the release of Just as unconsolidated material has different infiltration
the more water-soluble compounds to surface flow until and transport coefficients, so does bedrock. Generally, bed-
late in the melting cycle. Particulate matter may also accu- rock aquifers are found in sedimentary sandstone forma-
mulate in the lower portions of the snowpack by filtration tions and in limestone and dolomite. Limestone and dolo-
through less permeable bottom snow. Permafrost also chan- mite may be especially important due to the dissolution of
nels more snowpack contaminants into runoff, preventing the rock by water, which opens up joints and fractures, thus
interaction with soil. increasing the permeability and storage capacity. Sandstone,
The snowpack also has a direct influence on the biologi- which is either interbedded with shale or highly cemented,
cal cycle of Arctic freshwater systems. The combination of has low hydraulic conductivity, and is thus less important as
accumulated snowpack and the reduced sunlight during the a groundwater aquifer. Similarly, most igneous and meta-
winter season combine to make light the limiting factor for morphic rocks have low hydraulic conductivities and limited
biological productivity. Even with increased light in the storage potential except along fracture zones. Bedrock aqui-
spring, the snowpack on the river and lake ice surface effec- fers have generally not been well documented in the Arctic
tively reduces light penetration (Prowse and Stephenson region, since surface aquifers have been adequate to meet
1986). Light remains limiting until the surface snow melts water demands. The exception is Iceland where aquifers form
after which biological productivity increases and frequently in highly permeable lava (Elíasson 1994).
reaches a maximum while the ice surface is still intact. The amount of water available for groundwater move-
In summary, snowpacks reflect the net influence of depo- ment is restricted by low precipitation over large areas of the
sition (wet and dry) and post-deposition processes; conse- Arctic. Active layer (i.e., the upper soil layer which seasonally
quently, for semivolatile substances, snowpack measure- melts) thawing and water drainage may only occur during as
ments cannot be employed to monitor atmospheric deposi- little as two months of the year. The base flow in streams
tion. Nevertheless, snowpacks can be useful to assess net without lake storage in their drainage basins consists pri-
deposition. Snowpack chemistry surveys can also be used marily of the discharge of groundwater. In the zone of dis-
to examine regional variations in deposition, provided continuous permafrost, the contribution of groundwater to
that diagenetic processes (i.e., post-depositional chemical base flow ranges between 2 and 5 L/s/km2 (van Everdingen
changes) are assumed to be similar in magnitude among 1990). In the zone of continuous permafrost the winter base
regions. Such surveys may yield information on atmos- flow approaches zero. Only 10-19% of the snowpack vol-
pheric transport and help determine the relative source ume adds to the moisture of the underlying soil (Kane et al.
strength of atmospheric contaminants to which an area is 1991), and may, under the right geomorphological circum-
exposed (Franz et al. 1996). stances, be transported as drainage water. Horizontal and
vertical drainage intensities are low in areas with low relief
and permafrost layers close to the surface.
3.3.3.2. Terrestrial drainage systems
The role of lakes in influencing groundwater is important
Snowmelt and rainfall, and any associated contaminants, due to the suppression of permafrost beneath the lake (i.e.,
are delivered to the surface waters directly or become part talik). Based on a study of a small watershed in the north-
of the terrestrial drainage system. Terrestrial drainage intro- eastern Canadian Arctic (62°41'N, 97°03'W), which is un-
duces a significant lag time into the transfer of the water derlain by continuous permafrost, it has been estimated that
and depending upon the nature of the subsurface material, snowmelt runoff accounts for 45% of the total annual run-
can greatly influence the water chemistry. As well, transport off, which is largely the proportion of annual precipitation
through the porous subsurface media can result in losses represented by snow. This is a result of the restricted infiltra-
and/or gains of contaminants. This discussion will consider tion of snowmelt into the frozen active layer (Bursey et al.
the influence of bedrock geology, surficial geology, soils, 1991). Depending on the saturation state at freezing the pre-
and permafrost on groundwater and drainage water. Drain- vious fall, some snowmelt will penetrate the frozen ground,
age water is defined here as the part of the groundwater and remain in temporary storage. Groundwater flow be-
which drains to a river. comes progressively more important as the frost table de-
Geology influences natural water chemistry and runoff clines, releasing stored water and increasing the capacity to
yield principally through the magnitude of storage potential accept recharge from summer rain. Peak groundwater flow
and related water transmission properties. In the Arctic, is probably attained by late August and slowly diminishes
subsurface storage and transmission properties are further through September. Bursey et al. (1991) estimated that ground-
complicated by the presence of permafrost. Permafrost can water comprises 55% or more of the average annual water
influence infiltration, surface and subsurface storage, and balance for this site. These subsurface waters are a mixture
transmission of water. of ground-ice melt derived from late season rain of the pre-
With respect to surficial geology, which includes the soils vious year, rainfall from the current year, and local recharge
that have developed on the parent material, the infiltration from streams and ponds. Farther north, the role of ground-
rate into unfrozen, unconsolidated surface layers is complex water in lake water budgets decreases, while larger lakes
and dependent on a number of variables, including vegeta- may have a greater contribution from groundwater due to
tion cover, particle size distribution, porosity, layering, soil the larger talik beneath the lake.
packing properties, and organic matter content (Gray et al. The physical and chemical characteristics of groundwater
1973). Thus, while generalized maps showing topography in the permafrost region of the Yukon and NWT (van Ever-
dingen 1990) are similar to those in non-permafrost areas. The flows of melt and rain water from palsas and poly-
Limited data from Russia (Runnells et al. 1992) support gons are mainly filtered through the active layer. Swamp
this. The chemical composition of suprapermafrost water water flows down inter-polygonal cracks and discharges to
(drainage water in the active layer, undergoing seasonal rivers and lakes. Wetland runoff is characterized by perma-
freezing) reflects the influence of rainfall, snowmelt, and frost, a prolonged seasonal frost, a shallow active layer, rapid
surface runoff. This water is often characterized by a high and considerable release of snowmelt water, limited ground-
organic content (humic acid) derived from the peat sub- water capacity, and the prevalence of surface flow on wet-
strate. Where there is a strong influence of intra- and sub- lands. The frozen condition of wetlands during the spring
permafrost water, the concentration of dissolved solids can impedes infiltration of snowmelt water, resulting in a rapid
be extremely high (van Everdingen 1990). In areas rich in increase to maximum surface flow across the wetlands.
mineral deposits, natural concentrations of metals in the Later, when the active layer thaws and water recedes, the
characteristic acidic stream waters and shallow groundwa- water-retention capacity increases and summer flows are
ter may be naturally high (Runnells et al. 1992). leveled out (Woo 1992). There is no water discharge from
The ion chemistry of Arctic rivers is strongly influenced swamps during the winter period.
seasonally through the proportional increase in the contri- Ombrotrophic bogs receive water only from rainfall and
bution of groundwater. Groundwaters have a different hy- snowfall, there being no exchange of material with ground-
drochemical composition to that of surface waters. For ex- water. The water is nutrient poor, contains low levels of dis-
ample, during the greater part of the year, bicarbonates and solved solids and becomes acidic in contact with peat (Lock-
calcium ions predominate in the lower reaches of the Lena hart et al. 1992). Hence, cores or samples from ombrotrophic
River. However, during the winter period, when the concen- bogs can be very appropriate for the measurement of the de-
tration of dissolved solids exceeds 250 mg/L, Lena River position trends of organic contaminants from the atmos-
waters are transformed to a chloride class, with sodium and phere, though little work has been completed using this nat-
potassium ions dominating calcium. This is due to the in- ural archive.
creasing influence of groundwater input. The chemical com- Minerotrophic bogs, receiving water both from rain and
position of the groundwater is controlled mainly by the wide- from surface runoff, contain mineral-rich waters, and the
spread limestone and dolomite deposits in the upper and water is less prone to acidification in contact with peat (Gos-
middle reaches of the river (Gordeev and Sidorov 1993). selink and Turner 1978). In general, however, Arctic wetlands
Contaminants originating from rainfall, snowmelt, and are characterized by a very low pH value (3-5.5). There does
surface runoff might be present in the active layer (suprap- not seem to be any information available on the overall ef-
ermafrost water). The transport of contaminants by drain- fect of these natural water quality attributes on the degrada-
age water is probably very limited, owing both to the strong tion and transformation of contaminants.
association of many of these compounds with organic mat- The volatilization of hydrophobic contaminants is proba-
ter or particles, as well as to the seasonally short and re- bly limited from wetlands, compared with clear water lakes
stricted water flow through an often thin active layer. Har- under similar climatic conditions, because wetlands are rich
disty et al. (1991) reported no known cases of groundwater in organic matter, which binds hydrophobic contaminants.
contamination in the Canadian Arctic. On the other hand, Bursey et al. (1991) have shown that
In summary, processes and levels of contamination of wetlands can have high evaporation losses, of the order of
drainage waters and their discharge to Arctic rivers have 40% of the total runoff from the area, which may result in
rarely been studied. The variables controlling the transport revolatilization of more water-soluble compounds.
and fate of contaminants in surface waters (such as pH, or- In addition to the input of contaminants from the atmos-
ganic matter, ions, etc.) also influence drainage waters, but phere, wetlands are also subject to local activities with asso-
these processes and the overall extent or potential for con- ciated contamination risks, such as mining, pipeline opera-
tamination of Arctic groundwater have apparently not been tions, and facilities for petroleum exploration. Some exam-
investigated. This is not surprising, at least for the Cana- ples are outlined below.
dian Arctic, due to the sparse population and limited indus- Oil and gas extraction activities have led to significant
trial development in the region, although this may not be contamination of vast areas of adjacent land containing wet-
the case in some parts of Scandinavia and Russia. lands, as is the case in northwest Siberia. In this area, a wide
range of contaminants are discharged, both through natural
seeps (typical for oil and gas deposits) and through extrac-
3.3.3.3. Wetlands
tion, processing, and transportation activities. These wastes
The term ‘wetland’ is commonly used for a large variety of are usually discharged in an untreated condition into natural
biotopes, including freshwater meadows, salt marshes, reed depressions in the landscape. Under northern temperature
swamps, bogs, ponds, shallow lakes (depth below 2 m), and conditions, the decay rates for many organic contaminants
tidal and some other coastal zones (Cowardin et al. 1979). are likely low. Horizontal and vertical drainage rates are
Their nature and distribution in the Arctic have been dis- also low in these areas due to the relatively flat land and the
cussed in chapter 2. closeness of permafrost to the surface. This is probably a
Deposition of contaminants from the atmosphere, along positive situation, as the contaminants would tend to remain
with direct discharges from mining, oil and gas extraction, concentrated in a localized area rather than being spread
and other industry and urban wastes in the Arctic may con- horizontally and vertically. However, it is expected that in
tribute to the accumulation of significant amounts of conta- the areas surrounding oil and gas extraction activities, pol-
minants in wetlands. Wetlands may serve as sinks for conta- luted wetlands can represent long-term sources of secondary
minants, and in turn, become sources of secondary contami- contamination of surface waters, even after cessation of the
nation through indirect discharges to river systems. Contami- extraction activities, particularly during seasonal and rain-
nants can be carried into wetlands by flooding waters. The induced floods.
frequency and velocity of flooding is important in determin- Large areas in northeast Canada and Alaska are domi-
ing the total amount of contaminants in wetlands and their nated by lakes and rivers surrounded by land saturated with
availability for flora and fauna (Gosselink and Turner 1978). water, forming different types of wetlands. The small amount
of industrial development in the Canadian Arctic is likely in- Little is known about the rate of snow sublimation which
dicative of minor anthropogenic input to wetlands except in can cause a moderate amount of snow loss relative to the
the local vicinity of DEW Line military sites (see chapter 6). total snowpack (Woo 1992). Evapotranspiration can be sig-
In Alaska, more than 40 placer gold operations exist, nificant in the summer, ranging from about 150 mm/y in the
mainly on previously mined tailings in and along the wet- northeastern Canadian Arctic to > 200 mm/y in the Mac-
lands and riparian habitat areas. Aside from the obvious kenzie Valley and the Yukon. In a single watershed in the
physical disturbance of the surficial material, there is no in- eastern Canadian Arctic, it has been demonstrated that there
formation on significant emissions. can be a wide variation in the degree of evaporation from
various levels of the drainage system. For example, evapora-
tion can account for 40% or more of the discharge from
3.3.4. Aquatic processes
shallow tundra ponds, whereas the average annual vapor
3.3.4.1. Hydrology
discharge from the entire watershed is likely about 10%.
River systems are a major pathway for contaminants from Evaporative losses are minimized in part by the relative im-
the Arctic terrestrial compartment to the oceans, as a result portance of subsurface storage and flow within the active
of snowmelt runoff in addition to contaminants from local layer (Bursey et al. 1991). Evaporation is greatest in the
sources. In general, the quality of river waters is highly de- early summer, immediately following the spring snowmelt,
pendent on the water flow regime. Many dissolved compo- during the period with maximum radiation, but not neces-
nents tend to be diluted during high flows, whereas particu- sarily the maximum air or soil temperatures (Kane et al.
late matter and associated contaminants tend to increase 1990). Analysis of stable hydrogen and oxygen isotope com-
with water discharge. Discharge is a major factor in assess- position in northern Alaska revealed that evaporation from
ing environmental quality, as the load of contaminants to soil, vegetation, stream flow, permafrost, and snowcover is
a receiving water body will increase with discharge even also a critical factor in the hydrological cycle (Cooper et al.
for chemical constituents that are diluted (Meybeck et al. 1993). Studies carried out over a five year period indicated
1989). Consequently, it is necessary to consider the hydro- that of the original snowpack water volume in the catch-
logy of Arctic river systems in some detail. Throughout ment, snowmelt runoff amounted to 50-66%, 20-34% evap-
most of this chapter, the emphasis is on large rivers. How- orated directly, and only 10-19% remained in the soil (Kane
ever, as it is the large land area and the strong influence of et al. 1991).
snowmelt that feeds these rivers and dominates the hydro- The measurement of precipitation, especially during the
logic cycle, it is also necessary to consider the hydrology of winter season, is believed to be the most significant source of
smaller rivers and headwater systems as part of the total error in applying the water balance equation to Arctic sites,
hydrologic regime. requiring positive adjustment by up to 100% (Haas 1991)
Rivers in the AMAP region can be classified into those (see also section 3.3.3.1).
that originate in the Arctic region (referred to here as Arctic
rivers) and those that have headwaters in the subarctic, bo- Headwater systems
real, or temperate zones (e.g., Mackenzie River in Canada Streamflow regime is defined by the average seasonal pattern
and the large Russian rivers including the Ob, Yenisey, Le- of runoff, as determined by the timing and quantity of water
na, etc.). These large rivers, which originate outside of the supplied to, or lost from the basin, as well as by the storages
Arctic and which are referred to here as Arctic drainage which modify water delivery from the basin to its channels.
basins, have different flow regimes than Arctic rivers and Streamflow is also affected by in-channel attributes, includ-
have often been significantly modified by impoundments ing the formation and dissipation of river ice, flow blockages
and diversions. Arctic rivers, on the other hand, tend to by deep snow accumulation in the channels, or transit across
have a short flow season, concentrated within the brief wetlands and lakes (Woo 1992).
spring and summer periods and with little or no flow dur- Most rivers in areas of continuous permafrost exhibit an
ing the long, cold winters. These rivers are very responsive Arctic nival regime (Church 1974) with pronounced spring
to snow and ice melt events and rainfall due to the presence high flow generated by snowmelt. During this period, stream-
of relatively impervious permafrost and a shallow active flow in headwater streams follows prominent diurnal cycles,
zone (Woo 1992), resulting in highly episodic flow regimes. which lag the daily pattern of incoming solar radiation. This
The discharge of a river system depends upon the water pattern is illustrated in Figure 3·11. Spring floods are accen-
gains and losses within the basin and can be summarized tuated by snow dams along the channels formed by the redi-
by the water balance relationship: stribution of the snowpack by wind (Woo and Sauriol 1980).
These dams temporarily block the flow of water until the
Q = P – E + S
dams are breached, resulting in a high flow pulse. After a ma-
which partitions precipitation (P) into runoff (Q), evapora- jor flood, generally of several weeks duration, runoff recedes
tion (E), and a change in basin storage (S) be it in the form rapidly, being sustained by melting of the deepest snow accu-
of snow and ice, soil moisture and groundwater storage, mulations and ground ice in the active layer. Rainfall during
or retention in lakes and wetlands. Annual snow storage this time can result in significant peak flows due to the lim-
changes are particularly significant in the Arctic, with snow ited infiltration capacity of the active layer. Flow ceases in
accumulations in winter causing an increase in basin stor- most small rivers by early fall as a result of the lack of avail-
age followed by storage depletion in spring and summer able groundwater to maintain base flow as the active layer
(Woo 1992). Arctic rivers are likely to have a higher unit begins to freeze and precipitation changes to snow. Small
area yield of runoff, relative to the amount of precipitation, and many medium-sized rivers freeze to the bottom during
than do the Arctic drainage basins, due to the limited infil- the winter season, resulting in no flow for several months.
tration capacity of frozen soil. In general, the surface runoff Basins draining glaciated regions display a proglacial flow
in Arctic river basins is relatively low compared to the mean regime (Church 1974). As illustrated in Figure 3·11, spring
global surface runoff. For example, the mean runoff in the snowmelt in the basin is superseded by high flows from gla-
Russian Arctic is 7 L/s/km2, versus 10 L/s/km2 on a world- cier melt, with the discharge often related again to the energy
wide basis (WHO/UNEP in press). input to the glacier. Melt runoff can be considerably reduced
m3/s (a) Examples of annual hydrographs are provided in Figure
15 3·13 (next page) for a number of Arctic rivers. Rivers hav-
ing a characteristic Arctic hydrograph, that is, very low
13.5 m3/s
flows during the winter and a spring snowmelt dominated
Snow melt runoff Mac Master river : by a peak discharge include: the North Dvina, Mezen, Pe-
Cornwallis Island, NWT chora, Ob, Yenisey, and Lena Rivers in Russia; and, the
Basin area 33 km2
10 Back and Caniapiscau Rivers in Canada. The Khatanga,
Outflow of Small Lake
Cornwallis Island Olenyok, and Indigirka Rivers in Russia form a special sub-
basin area 2 km2 set of this group, as they freeze entirely during the winter,
resulting in negligible flow. During the flood period, which
Peaks due to rainstorms usually happens in June-July, these rivers discharge more
5 than a half of their annual water flow. For example, average
June discharge of the Olenyok River (Laptev Sea basin) is
60% of the annual discharge (see Figure 3·13).
Many of the Canadian rivers retain relatively higher flows
Summer low flow
during the winter season than do the Russian Arctic rivers,
0 and thus, the spring peak is less dramatic. This is due to the
July 1978 August 1978
large storage capacity of lakes within these systems. These
m3/s (b) rivers include the Mackenzie, Coppermine, Kasan, and
220
Sverdrup River : Hayes Rivers. The remaining group, which includes the
Ellesmere Island, NWT Churchill, Nelson (see Figure 3·13), Rupert, and Grande
200
Basin area 630 km2 , 77% glacierized Baleine Rivers are indicative of systems which: 1) have
180 largely temperate and/or subarctic watersheds; 2) have a
Snow and large natural storage capacity; and, 3) demonstrate the
160 Snowmelt runoff glacier melt
runoff additional influence of flow regulation by controlled re-
140
leases from reservoirs.
120
Glacier melt
100 runoff Water discharge (thousand m3/s)
40 Daily mean
80 a
discharge
35
60 1973-1990 average
40 Cold spell
30
after
20 snowmelt 25
0 20
June 1974 July 1974 August 1974
Figure 3·11. Hydrographs showing two types of streamflow regimes 15
for Arctic rivers: a) Arctic nival regime and b) proglacial regime (Source:
Woo 1992). 10
5
by overcast conditions or periods of cold weather. From
time to time, small glacial lakes may be breached, causing 0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
periodic high flows.
Streams flowing through wetlands acquire a wetland Sediment concentration (mg/L)
or muskeg regime, which is typified by little or no winter 4 000 Daily mean 5 values exceeded
flow, followed by a spring snowmelt freshet when surface concentration 3 000 in 1988
runoff sweeps across the frozen ground with little infiltra- 3 500 1972-1991 average
tion (Woo 1992). Once the wetland starts to thaw, the b
uneven ground contains a high degree of storage, and 3 000
thus, even the response to rainfall is attenuated by poor
drainage and by water retention in ponds and peaty soils. 2 500
Normally, the shallow thawed layer does not yield much
drainage water, thus restricting the amount of summer 2 000
flow (Woo 1992).
1 500
Large river systems
The hydrologic regimes of large river systems reflect the cli- 1 000
mate and physiography of the total watershed, as well as
the nature of the river (i.e., reservoirs and lakes). As a result, 500
large river systems tend to have a less well-defined hydro-
graph pattern. A unique feature of some of these large rivers 0
is that because they flow north, their headwaters generally Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
melt first. Thus, the peak flow is transferred downstream Figure 3·12. a) The annual (1973-1990) and mean annual hydrographs
often resulting in a hydrograph significantly dominated by for the Mackenzie River above the Arctic Red River, NWT, Canada, indi-
snowmelt, which will have taken several weeks to progress cating the dominance of the spring freshet in May. b) Daily suspended
sediment concentrations (mg/L) and mean annual suspended sediment
from the upstream areas, during which time downstream
concentrations (1972-1991) for periods of high flow and high sediment
snowmelt will have supplemented the flood crest. This is concentration in the Mackenzie River above the Arctic Red River (Source:
well illustrated for the Mackenzie River in Figure 3·12. HYDAT 1994).
Mackenzie
Indigirka
Coppermine Lena
Back Khatanga
Churchill Yenisey
Grand-Baleine
Caniapiscau Pechora
Hydrographs of Russian Arctic rivers
N.Dvina
% of total % of total % of total
annual flow annual flow annual flow
35 35 40
30 30 35
25 25 30
20 25
20
20
15 15
15
10 10 10
5 5 5
0 0 0
J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D
North Dvina Yenisey Khatanga
% of total % of total % of total

annual flow annual flow annual flow
35 40 35
30 35 30
25 30 25
20 25
20
20
15 15
15
10 10 10
5 5 5
0 0 0
J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D
Pechora Lena Indigirka
Hydrographs of Canadian Arctic rivers

Hydrographs
% of total of Canadian Arctic rivers % of total % of total
annual flow
% of total annual flow
% of total
annual flow
16 annual flow
18 annual flow
12
16
14 18
16 12
10
14
12 16
14 10
12 14
12 8
10
12
10 8
10
8 6
10
8
68 6
68 4
46 46 4
2
24 24
2
02 02 0
0 J F M A M J J A S O N D 0 J F M A M J J A S O N D 0 J F M A M J J A S O N D
J F M AMackenzie
M J J A S O N D J F M A Coppermine
M J J A S O N D J F M A MChurchill
J J A S O N D
% of total Mackenzie % of total Coppermine % of total Churchill
annual flow
% of total
annual flow
40 16
annual flow annual flow
30
40
35 16
14 30
25
35
30 14
12 25
30 12
10 20
25
10 20
25
20 8 15
20
15 68 15
10
15
10 46 10
10
5 24 5
5
05 02 0
0 J F M A M J J A S O N D 0 J F M A M J J A S O N D 0 J F M A M J J A S O N D
J F M A MBack
J J A S O N D J F M AGrand-Baleine
M J J A S O N D J F M A M J J A S O N D
Caniapiscau
Back Grand-Baleine Caniapiscau
Figure 3·13. Monthly mean discharge expressed as a percent of annual discharge from available data for selected Russian and Canadian rivers draining
to the Arctic (Source for Canadian river data: HYDAT 1994).
Table 3·5. Summary of selected physical and chemical characteristics of major lakes and reservoirs.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Maxi- Drainage Residence Trans- Suspended
Lake (L.) Surface Volume, mum Mean area, time, parency, DO, TDS, solids,
/Reservoir (Res.) area, km 2 km 3 depth, m depth, m 3
10 km 2 y m mg/L pH mg/L mg/L
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Russia
L. Baikala 31500 23000 1741 740 560 –.0 ~ 5-23 9.6-13 0 6.8-8.5 – –
Krasnoyarskoye Res.a 2000 73.3 105 37 28700 0.81 1.2-6.0 7.7-14.2 7.2-8.4 – –
Uste-Ilimskoye Res.a 1920 58.9 97 30.7 785080 0.58 2.3-4.1 2.7-16.4 7.8-7.5 – –
Bratskoye Res.a 5478 169.3 150 31.1 757200 1.81 2.6-11.0 9.2-14.2 7.1-8.9 – –
Canada (within Arctic)
Great Bear L.b 31153 2236 446 71.7 114717 124.0.0.0 20-30 Saturation 7.8-7.9 78-81 –
Great Slave L.c 28600 –0 614 41 –00 –.0 < 1.0-1700 Saturation 6.6-6.9 22-150 –
Hazen L.e 541.8 –0 280 –0 –00 –.0 – 10.41-15.11 – – –
Dubawnt L. 3830 –0 –0 –0 –00 –.0 – – – – –
Laberge L.d 201 10.8 146 54 3500 1.0 4-11 9.7-11.6 7.7 – –
Atlin L.f 588 –0 283 85.6 –00 –.0 10 9.7-9.91 7.5 – –
Marsh L.f 64.5 53 –0 –0 –00 –.0 5 – – – –
Tagish L.f 340.8 214 –0 –0 –00 –.0 1.0-6.1 – – – –
Canada (within drainage basins draining into Arctic)
Winnipeg L.e 23750 284 36 12 953250 2.9-4.3. 0.35-2.00 – 8.1-8.4 – –
Reindeer L. 6650 –0 –0 –0 –00 –.0 – – – – –
L. Winnipegosis 5370 –0 –0 –0 –00 –.0 – – – – –
L. Athabasca 7940 –0 –0 –0 –00 –.0 – – – – –
g
S Indian Res. (Churchill) 2391 23.4 –0 9.8 242 0.72 – – – – 1.2-11.0
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Data sources: a. ILEC 1993; b. Johnson 1975a; c. Rawson 1947, 1950; d. Gray et al. 1992; e. Allan et al. 1994; f. Lindsey et al. 1981;
g. Newbury et al. 1984.
and lakes (Albright et al. 1980, Schreier et al. 1980, Prowse

3.3.4.2. Lakes and reservoirs
and Stephenson 1986, Whitfield and McNaughton 1986).
Arctic lakes and reservoirs, like those in temperate climates, The phenomenon of biological productivity in Arctic lakes
reflect loadings of contaminants supplied by the rivers and occurring in the spring and being controlled by the light
streams flowing into them, as well as exchanges of contami- penetrating a stable ice cover after snowmelt is likely typical
nants directly with the atmosphere at the lake surface, the of many Arctic lakes. Recently, the biomagnification and
bottom sediments, biota, and subsurface drainage. Most fate of POPs in Yukon lakes have been investigated for Lake
Arctic lakes were formed recently following the retreat of Laberge and other Yukon River system lakes (Kidd et al.
glacial ice, and, in contrast to temperate lakes, permafrost 1996, Lockhart and Muir 1996) (see chapter 6), but whether
plays an important role, in that it controls the depth of the or not the timing and physical processes associated with the
active layer and thus, the groundwater in the vicinity of lakes. period of maximum productivity plays an important role in
Few large Arctic lakes have been studied in detail. John- the fate of contaminants delivered to these lakes has not
son (1975a, 1975b) described the physical and biochemical been considered.
limnology of Great Bear Lake in the Mackenzie River basin Hydroelectric power production on Arctic rivers and with-
in Canada (see Table 3·5) between 1963 and 1965. Great in Arctic drainage basins has produced large reservoirs in
Slave Lake (also in the Mackenzie River system) was stud- both Canada and Russia (see Table 3·5). Reservoirs signifi-
ied extensively by Rawson during the summers of 1944 and cantly change the hydrologic regime of rivers, intercept the
1945 (Rawson 1947, Rawson 1950), but has not been stud- sediment load, and can affect the river ecology through in-
ied since. The characteristics of Great Slave Lake are more terrupting species migration and increasing contaminant
complex than Great Bear Lake, as the former receives turbid concentration (e.g., mercury). Also, the loss of waterfalls
water from the Slave River, which mixes into the western and rapids can greatly reduce the natural aeration of the wa-
part of the lake to a distance of about 40 km from the delta ters and increase ice coverage. In turn, this could result in in-
(Mudroch et al. 1992). In 1994/95, an investigation of the creased contaminant concentrations under the ice due to re-
biomagnification of POPs in the Great Slave Lake system duced exchange with the air (Dynesius and Nilsson 1994).
was initiated by Evans (1996) (see chapter 6). Lastly, Lake Reservoirs can have a significant effect on downstream
Hazen on Ellesmere Island in the Canadian Arctic, the largest temperatures. The storage and release of large quantities
and most northern Arctic lake, was studied in the 1960s of water at about 4°C throughout the winter can keep the
and 1970s (Hunter 1960, McLaren 1964, Johnson 1994b). reaches immediately downstream of the reservoir free of ice
Several studies have investigated the limnology of Yukon and retard the formation of ice cover farther downstream.
Territory lakes (Lindsey et al. 1981). Four of these lakes, Even ice cover type and the nature of its hydraulic effects
Atlin, Laberge, Marsh, and Tagish, were investigated in can be modified. Further, the dam virtually eliminates the
greater detail in 1982 and 1983, in an effort to understand supply of ice to the downstream river reaches from the reser-
better the processes controlling the limnology of these lakes, voir and upstream. Fluctuating flows caused by releases from
and to assess the effect of potential developments within the the reservoir can result in increased ice thicknesses as a re-
basin (Kirkland and Gray 1986, Gray 1994). These authors sult of surface ice overflows or they can induce premature
concluded that the biological productivity in these lakes is ice break-up during the winter. The magnitude of the effect
nutrient limited, the optimal time for algal uptake of nitrate of regulation on spring break-up varies according to the ratio
being between snowmelt and ice-out, when there is ade- of the regulated flow to the total river discharge. Depending
quate light and an absence of wind-induced mixing. As an on this ratio, break-ups and associated ice-jam flooding could
example of this, Shortreed and Stockner (1986) observed be more or less severe compared to unregulated conditions
that chlorophyll levels under ice were higher than average (Prowse and Conly 1996).
summer chlorophyll in 5 of 10 Yukon lakes sampled. Pro- The generation of hydroelectricity from the Churchill-
duction under ice has also been indirectly measured by dis- Nelson River systems (in Manitoba, Canada) began in the
solved oxygen (DO) concentrations in both Arctic rivers 1950s. The first reservoir on the Nelson River was com-
m3/s the vicinity of Resolute, NWT, Canada. Char Lake was stud-
3 000 ied intensively between 1969 and 1973, with the aim of de-
Pre-1973 Post-1976 scribing for the first time the complete annual cycle of physi-
2 500 cal and chemical events in a High Arctic lake (Schindler et
al. 1974, Welch 1994). In most years, Char Lake is a typical
polar lake never rising above 4°C, and having a brief sum-
2 000
mer overturn period of up to six weeks duration. The lake is
extremely oligotrophic, with soluble reactive phosphorus
1 500 measurements (SRP) consistently lower than the detection
level (< 0.7 mg/L); nevertheless, winter oxygen depletion at
1 000
the sediment/water interface was considerable, probably re-
flecting respiration by the benthic community. Dissolved oxy-
gen concentrations were greatest just under the ice when it
500 was thickest in late May (Schindler et al. 1974). This prob-
ably reflects increased photosynthesis in the surface waters
0 in association with increased light penetration (24 hours of
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec sunlight and melting snow) at this time of year (Albright et
Figure 3·14. Monthly mean discharges for the Nelson River, Canada for
al. 1980, Prowse and Stephenson 1986, Whitfield and Mc-
the period 1960 to regulation in 1973, compared to post-regulation dis- Naughton 1986). Schindler et al. (1974) also noted that mix-
charges from 1976 to 1990 (Source: HYDAT 1994). ing continues after the lake is frozen over, as evidenced by
the general uniformity of specific conductance with depth.
pleted in 1961, and in 1976 most of the flow of the Church- This mixing delays reduced oxygen levels at depth.
ill River was diverted into the Nelson River basin through There was evidence at Char Lake of a freeze out of sol-
the enlargement of South Indian Lake (1977 km2). The com- ute-rich water which concentrates gases and dissolved solids
bined annual mean flow of the Churchill and Nelson Rivers in the deepest part of the lake over the course of the winter.
amounts to 3500 m3/s (Rosenberg et al. 1987) and the ef- This cryoconcentration can be quite substantial in Arctic
fect of this regulation on hydrology is illustrated in Figure lakes. For example, a lake 5.2 m deep at 64°N was 77% fro-
3·14. The effect has been a change to peak flows in Novem- zen with 2 m of ice cover, and even a lake 34.2 m deep had
ber, December, and January (corresponding to peak hydro- 23% of its volume frozen (Welch et al. 1987). Cryoconcen-
electric demand) as opposed to the natural peak flows in tration likely affects the fate of contaminants in the water
June, July, and August. These developments have been stud- column of these lakes. However, the cryoconcentration of
ied for their environmental impacts, and have been consid- contaminants and the effect of anoxia at the sediment/water
ered elsewhere (e.g., Hecky et al. 1984, Rosenberg et al. interface on contaminant recycling has not been considered
1985) and will be discussed further in chapter 7 in the con- for this environment.
text of the fate of mercury in reservoirs. Limnological research into the effect of anthropogenic ac-
The James Bay project is Canada’s largest hydroelectric tivities on water quality in northern Canada has also been
development, and it affects rivers flowing both into James undertaken as part of a multidisciplinary research project at
Bay and southeastern Hudson Bay. The first phase of the Sagvagjuac, near Chesterfield Inlet on the west coast of Hud-
La Grande River development included three main diver- son Bay. This work began in 1976 to parallel a project at the
sions, specifically, of the Caniapiscau, Eastmain, and Opi- Experimental Lakes Area in southern Manitoba (Welch 1985).
naca Rivers into the La Grande River. The limnology of La Sagvagjuac is an area of numerous glacial ice-scoured lakes,
Grande 2 reservoir, filled in 1978 and 1979, is reviewed in most of which are small (< 50 ha). The lakes are cold and mo-
Kislalioglu and Berkes (1994). nomictic, circulating once annually during the summer with
In Russia, the Kransnoyarskoye Reservoir on the Yeni- temporary thermal stratification in midsummer. The water is
sey River is one of the largest man-made lakes in Siberia generally soft, with total dissolved solid concentrations between
(Table 3·5). Construction began in 1967 and the reservoir 15 and 30 mg/L, and oligotrophic and clear. Water circula-
was filled in 1970 (Data Book of World Lake Environments tion and winter respiration of these lakes has been discussed
1994). Also within the Yenisey River basin, are the Ust-Ilim- by Welch and Bergmann (1985a, 1985b), while the chemical
skoye and Bratskoye Reservoirs which were filled in 1974 limnology has been reviewed in Welch and Legault (1986).
and 1967 respectively. The Lena River has dams in only one This was the first detailed investigation of element input/out-
large and one very small tributary and as such retains a nat- put on continuous permafrost watersheds. The year-to-year
ural hydrologic regime (Dynesius and Nilsson 1994). variability in solute output from these permafrost watersheds
was high, and shown to be determined by the timing and mag-
Small lakes nitude of the snowmelt runoff. Welch and Legault (1986) con-
There are innumerable small lakes in the Arctic, as described cluded that the solute output was low in years with little mid-
in chapter 2. Unfortunately, to date there has been little at- to late-summer rainfall, causing the runoff into the lake to
tempt to characterize these large numbers of lakes region- consist mostly of snowmelt, which occurred before the active
ally, in order to provide a rational basis for a systematic in- layer was well developed. The importance of mid- to late-
vestigation of the operation of Arctic lakes and the fate of summer rainfall was demonstrated by its coincidence with
contaminants within them. The exceptions to this are in high solute outputs from the basin into the lake. This was
Alaska, where regional characterization has commenced, also an important factor with respect to lake element loading
and in Norway, where this is being considered (D. Landers, and retention. If most of the runoff occurred prior to ice-off,
pers. comm., US Environmental Protection Agency, Envi- the inflowing water and solutes flowed out from the surface
ronmental Research Laboratory, Corvallis, Oregon, 1996). of the lake prior to mixing with the bulk of the lake water
Some of the major studies are described below. where biological removal could take place (Welch 1974).
Early investigations into northern limnology were under- Rivers draining into a lake tend to flow before the ice of
taken in Alaska (e.g., Hobbie 1973) and at Char Lake, in the lake has begun to melt and before there is any outflow
Melt M el t 3.3.4.3. Suspended sediments, sedimentation,

inflow o u tf l o w
and resuspension
Ice
Many contaminants are hydrophobic and consequently tend
to be adsorbed to and transported with sediments in rivers.
Thermocline Sedimentation processes in lakes, deltas, and other quiescent
zones can remove contaminants from the water column and
Lake store them temporarily or permanently. The resuspension of
wa t e r these sediments can result in the reentrainment of the associ-
ated contaminants and ongoing transport. These processes
will be considered in the context of Arctic river systems.
Sediment Suspended sediments

The particulate matter transported by a river is commonly
divided into wash and suspended load and bed load. The
Figure 3·15. Illustration of the movement of less dense spring freshet latter is the coarse material (sand size > 63 mm particle di-
water moving through a small Arctic lake underneath the surface ice ameter) moving on or near the stream bed, whereas the for-
cover, but not mixing with the water column.
mer is continuously suspended in the water and is carried
from the lake. As a result, the lake temporarily stores this with approximately the same velocity as the water (Meybeck
water, causing the water level of the lake to rise, and even- et al. 1989).
tually float the ice clear of the lakeshore, creating a moat Particulate matter, or suspended solids, originate from
around the lake. The lake level will continue to rise until soil and bank erosion in the watershed and by the resuspen-
the level of the outflow is exceeded. Because the incoming sion of particles deposited in the river bed. A general rela-
meltwater is relatively warm, compared to the lake water tionship can often be seen between water discharge and the
which is stratified under the ice, this first flush of meltwater concentration of suspended solids (Meybeck et al. 1992).
tends to pass through the lake with little or no mixing with Resuspension of fine-grained bottom sediment with increas-
the lake water, as illustrated in Figure 3·15. Thus, contami- ing discharge is often a major cause of increased suspended
nants transported by early snowmelt probably move through solids (Thomas and Meybeck 1992). Particle size and or-
a lake without mixing with the water that has been stored ganic carbon content are acknowledged as playing key roles
there from the previous fall (Welch 1974, Barrie et al. 1997, in the adsorption of chemical compounds to particulate mat-
Freitas et al. 1997). This suggests that: 1) rivers will tend to ter (Knezovich et al. 1987). It is the surface area of the parti-
transport the main load of soluble contaminants from snow- cles and the amount of organic carbon present that controls
melt beyond headwater lakes; and, 2) lake water chemistry adsorption capacity. This capacity is inversely proportional
will generally reflect that of summer rain and ground-ice to grain size. Thus, the finest clays are generally those with
melt, as modified by movement through the active layer and the highest concentrations of contaminants (Thomas and
mixing with the perennial lake waters. Meybeck 1992). Also important is the fact that these finer-
Recently, a contaminant mass balance study of Amituk grained particles settle less quickly, and thus, there is poten-
Lake (75°02'57'' N, 93°45'51''W) in the Canadian Arctic tial for riverine transport over long distances.
Archipelago was initiated by Environment Canada. This is Contaminants also bind to suspended organic particulate
the first integrated study of contaminants in freshwater sys- matter as well as aggregates of various origins and chemical
tems in the Arctic (Semkin 1996). While work is still pro- characteristics. These carrier particles may consist of the fol-
gressing on this study (the contaminant details are reported lowing fractions/phases: organic, carbonate, Mn and Fe, de-
in chapter 6), it is appropriate to note here that even in such trital/non-detrital, adsorption, and cation exchange phases
a small study area (drainage area ≈ 26 km2 ), considerable (Håkanson and Jansson 1983). Humic matter is known to
variability occurs with respect to stream delivery of water be an important carrier for contaminants in freshwater sys-
and sediments due to the basin geography and surficial ge- tems (Håkanson and Jansson 1983).
ology. In this case, the importance of a particular stream to Pulp mill effluents have been shown to affect the physical
the delivery of atmospheric-derived contaminants seems to transport characteristics of sediment. The effects are most
be primarily determined by the presence of depressions pronounced during low-flow periods (i.e., winter) when the
where wind-blown snow is trapped, which in turn delivers ratio of the effluent discharge to the flow discharge is great-
contaminants to the system at snowmelt (Semkin 1996). est. The effluent contains organic fibers and bacteria that
A number of other lakes in the Canadian Arctic have have affinity for ambient inorganic sediment resulting in sed-
been investigated in less detail than Amituk Lake to deter- imenting particles larger in size than the ambient material
mine depositional trends of contaminants by means of dated (Krishnappan et al. 1995). The nature of this material may
sediment cores. In 1994/95, a number of Russian Lakes result in its subsequent remobilization during periods of
were added to this spatial/temporal study (Lockhart 1996). high flow. This interaction can have important consequences
A comparable study has been reported on two lakes in for contaminant transport.
Alaska (Gubala et al. 1995). Sedimentation in small Arctic The most important stream characteristics which deter-
lakes is considered further in section 3.3.4.3, while conta- mine the input of contaminants to the river, their transport
minant studies are synthesized in chapters 6 and 7. down the river system, and their concentration in the water
Many Arctic water bodies are very shallow (frequently include the rate of water flow from slopes and in channels,
less than 2-3 m maximum depth) and are characterized by as well as the capacity of the flow to transport suspended
being unstratified during the summer. These lakes tend to particles. The concentration of suspended matter in the flow
be warmer in summer due to their shallow depth and early depends on the nature of the soils and sub-soils of the water-
loss of ice, but are also likely to be totally frozen during the shed, vegetative cover, slope, infiltration capacity, depth of
winter. Lakes of this type have not been investigated to any active layer, etc. Land runoff reflects the geology and phys-
great extent in the Arctic. iography of the drainage basin. In Canada, for example,
Table 3·6. Contributions of suspended sediments from rivers draining phase determined by concentrations on the suspended mat-
the three physiographic regions of the Northwest Territories, Canada. ter. Nevertheless, if the concentration of particulate matter
(Source: Brunskill 1986).
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– in rivers, for example headwater systems, is low, contami-
Average suspended nants will be transported in the dissolved phase. Field stud-
sediment load, ies in two Canadian rivers have shown that, under the ma-
Drainage region 103 kg/km2/y
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– jority of flow conditions, the main transport medium for
Precambrian Shield <3 most hydrophobic contaminants is in the dissolved phase
Interior Plains 56 (Carey 1995).
Cordillera 101
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Most metals have low water solubilities, but high natural
concentrations of metals can occur as a result of suspension
rivers draining the Precambrian Shield, the Interior Plains, and of the transport of fine-grained sediments as part of the
and the Cordillera can be distinguished by their suspended natural weathering process. In addition, factors such as the
sediment burdens (Table 3·6) (Brunskill 1986). pH, redox, alkalinity, salinity, and trophic level are also im-
The highest concentrations of suspended matter gener- portant in the spread and fate of metals (Means et al. 1979).
ally occur near the bottom, and are related to the presence Regardless of the hydrophobicity of many of the contami-
of coarser grains in suspension (Eisma 1993). The condi- nants, hydrological measurements of snowmelt runoff indi-
tions for generating high sediment loads are usually maxi- cate a poor mixing of meltwater with underlying ice-rich
mized in the spring. For example, the amount of sediment soils, and only minor amounts of this water remain in the
transported during the 10 days/year with the highest water watershed (Cooper et al. 1991, Cooper et al. 1993). Simul-
discharges was on average 47% of the total annual sedi- taneously, substances scavenged from the atmosphere and
ment load transported in a mountainous river course in deposited in the snowpack are often strongly retained by
northern Norway (Calles 1977). In the Mackenzie River, sorption or chemical interaction with vegetation and soil
approximately 65% of the total annual sediment load is during snowmelt (Cooper et al. 1991). It has, for instance,
delivered in two months (July and August), representing been shown that over 90% of the atmospheric load of the
about 36% of the annual flow (Figure 3·16). isotope 7 Be and about 70% of the isotope 35S, were adsorbed
Most POPs are hydrophobic (and lipophilic) and adsorb onto the surficial 2-3 cm of frozen organic soils and tundra
and partition (dissolve in a polar phase) to particles. The vegetation in Alaska. In contrast, in the Canadian Arctic,
amount of contaminants adsorbed on particles depends on little loss to soil was observed in the Amituk Lake basin for
the sorption potential (number of surface sites) of the parti- OCs and mercury (Semkin 1996), probably because these
cles, the binding energy and the hydrophobicity (expressed contaminants are being sourced from deep snowpacks with-
as KOW – see chapter 6 for details) of the compound, plus in stream valleys, allowing little opportunity for loss to soils
its chain length and its molecular configuration (Means et following snowmelt.
al. 1979). The hydrophobicity of most organic compounds
is such that the mass transport in the dissolved phase (water Sediment deposition and resuspension
concentration) tends to be less than that of the particulate In general, the processes of transport and deposition of sedi-
ment are governed by the water flow rate and the size of the
50 000 particles. In a turbulent water stream, upward forces are re-
Suspended sediment sponsible for the transport of the fine fractions in suspension.
load (thousand tonnes) Quartz-density solids with a diameter less than approxi-
45 000
Discharge (m3/s) mately 0.015 cm are mainly transported as a suspended load
(Allen 1970). Erosion and deposition are results of the vary-
40 000 ing ability of stream water to carry sedimentary particles.
The flow force and the sediment transport rate are propor-
tional to the mean water flow velocity (Allen 1970). Heavier
35 000
particles do not become suspended, but they can be trans-
ported as a bed load (bounce and roll) by the water flow.
30 000
Although the upper reaches of river systems experience
erosion, the lower parts are commonly marked by a state of
net deposition, expressed as a thick, laterally extensive spread
25 000 of alluvial sediment in which the active river courses are em-
bedded. The mean size of deposited particles decreases down-
stream, as the transporting capacity and turbulence decline
20 000
(Moss and Walker 1978). The finest particles or wash load,
which usually contain the highest contamination levels, will
15 000 therefore travel the longest distance. These will only accu-
mulate in large lakes or reservoirs. Consequently, the sus-
pended matter transported by rivers is not all destined for
10 000 the sea bed.
Sedimentation can result in the accumulation in the river
bed of particles and surficially active substances, such as
5 000
heavy metals and POPs, including heavy fractions of oil hy-
drocarbons, polyaromatic hydrocarbons such as B(a)P, as
0 well as phenols. Resuspension of contaminants occurs when
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec the bedload is disturbed, for example, during the spring
snowmelt flood. When bottom sediment particles become
Figure 3·16. Comparison of monthly water and suspended sediment loads
to the Mackenzie Delta plain for the period 1974-1983 (Source: Lewis suspended, the sorbed contaminants may be transported
1988). over long distances by the river.
Lakes and reservoirs can have a significant effect on river foreset area) and on the depth conditions of the river mouth
hydrology and on sediment accumulation in downstream area. When bed-load transportation ceases, the deposition of
deltas and estuaries. Eight reservoirs with a total combined the suspended load (particles less than 0.18 mm) will sys-
volume of 474 km3, and which together control about 23% tematically reduce the height of the foreset slopes. The bot-
of the total river flow have been constructed in the Yenisey tomset slopes are generally much more stable than the fore-
River basin. Together they have caused a reduction by a set slopes, where mass movement is comparatively frequent.
factor of three in sediment transport to the river mouth. In The delta advance can be very rapid in lakes receiving large
1961, the load of suspended matter near the river mouth at loads of material (Håkanson and Jansson 1983).
the community of Igarka was 13 106 tonnes, but this had Sediments are resuspended when the bottom shear ex-
been reduced to 4116 tonnes per year by the end of the ceeds a critical shear stress for the sediment bed. The critical
1970s (Alabyan et al. 1991). In the Peace River, a tributary shear stress is a function of the properties of the bottom ma-
of the Mackenzie River, the gates of the W.A.C. Bennett terial, such as water content and grain size (Håkanson and
Dam were closed in 1967, creating the Williston Reservoir. Jansson 1983). For example, recently deposited, unconsoli-
Flow routing revealed that while there would be no signifi- dated material is resuspended more easily than compacted
cant change in the total annual inflow to the Mackenzie material (Bengtsson et al. 1990).
Delta, the dam and reservoir would result in an average
flow increase of 20-30% to the Delta during the low flow Effects of floods
months, especially February to May. Conversely, summer During snowmelt, the water surface in the river rises, and at
and fall flows would be diminished by as much as 5-10% the same time, the flood waters can mobilize and disperse
(Wiens 1991). Subsequent measurements have shown that deposited bottom and bank material. The settled material
the monthly discharge of the Peace River has declined 16% which previously lay immobile at the bottom becomes di-
for the ice-off period, but increased 40% during ice-on per- lated, and some may go into turbulent suspension (Friedman
iods. Further, it is estimated that the average annual sedi- and Sanders 1978).
ment load downstream has decreased by 33% (373 103 Under normal flow conditions, most of the suspended
tonnes/y) (English et al. 1996, Prowse and Conly 1996). material moves downstream with the water, without being
In most cases, sedimentation in large lakes is more or deposited. However, as a result of ice jams and increased
less a permanent sink. Nevertheless, in many shallow lakes, water levels during spring melting, the water may overflow
wind induced resuspension of the sediments can occur with the river banks. The effect of flooding on the transport of
subsequent transport downstream. This is likely of signifi- suspended matter is variable. When the plain is flooded, ma-
cance in lakes with large surface areas and which are ori- terial may be picked up (Scrimgeour et al. 1994), but both
ented with their long axis parallel to prevailing winds. The when flow is reduced and in the case of falling water, sus-
presence of surface ice throughout much of the year will, pended matter is deposited (Eisma 1993).
however, reduce the extent of wind turbulence. Many Arc- The development of vegetation on the tundra can pro-
tic lakes retain extensive ice cover throughout most of the duce hummocky surfaces, which reduce the rate of overland
open water season, and some lakes are only free of ice peri- flow by enhancing surface storage, thereby reducing soil ero-
odically. To our knowledge, there has been no investigation sion and enhancing sedimentation. Sedimentation in flood
of sediment resuspension in Arctic lakes within the context plains and deltas is also affected by vegetation. Sedimenta-
of contaminant transport. tion rates in the Mackenzie Delta can be predicted from the
The rate of sedimentation in, for example, Scandinavian type of vegetation on the delta (Pearce 1993).
lakes in wooded areas is usually between 0.7 and 1.7 mm/y
(Rognerud and Fjeld 1993). Age determinations in lake se-
3.3.4.4. Humic matter in river systems
diments in subarctic parts of Norway show that the sedi-
mentation rate is usually less than 1 mm/y (Norton 1986, While considerable attention in this chapter has been given
Rognerud et al. 1993). Similar rates have been reported for to the role of inorganic sediment in transporting contami-
Canadian lakes by Lockhart (1996). nants in aquatic systems, particulate organic matter (POM),
The most important variables regulating gross sedimen- measured as particulate organic carbon (POC), can also play
tation to glacial lakes in Sweden were found to be: the depth an important role in this regard. Especially when inorganic
relative to wind fetch which together influence resuspen- sediment concentrations are low, autochthonous POM can
sion; the form and size of lakes; the proportion of the lake provide an appropriate substrate for the adsorption of con-
catchment containing forest and open land; the distribution taminants. Nevertheless, there is an absence of information
of mires and lakes in the catchment; the relief of the drain- available on the role of POM in transporting contaminants
age area; and, the theoretical lake water retention time (Hå- in temperate rivers and lakes, let alone in Arctic river sys-
kanson 1994). tems, due in part to the low concentrations of POM relative
The hydraulic conditions change rapidly at the mouth of to that of total (organic and inorganic) suspended particu-
rivers. The flow pattern in freshwater delta areas is very com- late matter.
plex, but bears similarities to the expansion of submerged Large amounts of dissolved organic carbon (DOC) are ex-
jets with complementary zones of reverse flow (Jopling 1960). ported from wetlands and taigas, where organic matter accu-
The form of the delta front depends both on fluvial and mulates due to slow decomposition rates (Hope et al. 1994).
shore processes, on bed load and suspended load, and on The DOC released from wetlands appears to influence the
the original topography of the basin. Deltas are generally transport of, for example, total and methyl mercury (Dris-
divided into a topset area, a foreset area constituting the coll et al. 1994). The extent to which contaminants bind to
delta front, and a bottomset area beyond the foreset slopes. dissolved humic materials depends on the concentration and
Bed load transportation of particles larger than 0.18 mm is type of humic materials present, pH, calcium concentration,
a necessity for the formation of foreset slopes, and such ma- and ionic strength of the water (Carter and Suffet 1982).
terials, often highly sorted, constitute the bulk of the foreset Humic matter is known to be an important carrier for
deposits. The size of the foreset area depends on the grain contaminants in freshwater systems (Håkanson and Jansson
size of the material (the coarser the material the larger the 1983). OC contaminants were present in the water of five
major rivers of the Hudson Bay lowland, but were found to Stable ice cover is formed at river velocities below 0.6 m/s,
be almost absent in bottom sediments (McCrea and Wick- while the water surface generally remains free of ice at velo-
ware 1986). The water also contained high concentrations cities greater than 1.5 m/s (Elíasson 1994). In late winter, ice
of organic carbon, iron, and aluminum. It is conceivable cover on lakes can be of the order of 2 m thick in the Cana-
that the hydrophobic contaminants were maintained in the dian Arctic Archipelago. River ice can be substantially thicker
water column by being bound to humic substances. The than this due to the incorporation of slush layers and overflow,
seepage of organic-rich water originating in the surround- and has been reported at up to 2.5 to 3.0 m thick in some
ing wetlands is a major source of aquatic humic matter to Russian Arctic river reaches (Grigoriev and Sokolov 1994).
the rivers. Dissolved organic carbon has been found to be a Also important is the nature of the ice, with so-called
major constituent of these northern waters, and concentra- white ice forming if snow or slush is incorporated, and black
tions in the area of 10-14.5 ng/L have been reported (Mc- ice forming when freezing occurs rapidly and under calm
Crea and Wickware 1986). conditions (or at least without incorporating snow or slush).
Due to their complexing ability, DOC and POC can act Thus, black ice is often more typical of open water in a lake
as carriers of heavy metals and organic compounds and in- proper, away from the effects of shore ice (Bengtsson 1994).
fluence their partitioning. This can change the solubility, The nature of the ice determines the ability of light to pene-
mobility, and bioavailability of adsorbed compounds (Mc- trate it – white ice severely limits transmission, although
Crea and Fischer 1986, Hope et al. 1994). Hydrophobic snowcover attenuates light more than either black or white
compounds can remain in the water adsorbed to humic ice (Prowse and Stephenson 1986). Some lakes retain sub-
substances. Thus, in the presence of high concentrations of stantial ice cover throughout most years (e.g., Lake Hazen).
organic carbon (DOC and POC), sorption of organic con- This can impact on light transfer during the summer and can
taminants to sediments may not readily occur, although result in earlier freeze-up in the fall.
this is an area requiring further investigation. Humic sub- Anchor ice forms when the entire water column is super-
stances and humic and fulvic acids can also reduce the rate cooled and ice nucleates on particles on the bed of the river
of volatilization of compounds, such as PCBs (McCrea and (Reimnitz et al. 1992). A change in river temperature or tur-
Fischer 1986). bulence, or accumulated buoyancy resulting from the build-
In the Lena River, 10-11 mg DOC/L has been recorded up of ice can eventually dislodge the combined ice and sedi-
during summer floods (Martin et al. 1993). This is usually ment mass, rafting it to the surface. Here it is included in the
accompanied by increased levels of iron, complexed to the overlying ice cover. Rocks as heavy as 30 kg have been in-
humic fraction (McCrea and Fischer 1986). Available data corporated in river ice by this mechanism (Martin 1981).
on separate determinations of DOC and POC concentra- Frazil ice also forms when the water column is supercooled.
tions and fluxes show that the northern rivers discharge While actively growing, frazil crystals tend to collect parti-
mostly DOC (Table 3·7). The organic matter content is cles which can subsequently be incorporated into the ice
lowest during the summer and autumn periods. The organic cover. Ice formed in this way appears turbid, with patchy
matter content is highest in the swampy river basins of the discolorations due to entrained material (Reimnitz et al.
western Siberian lowland. In the weakly mineralized wa- 1992). In addition, wind-blown dust and dirty snow accu-
ters, rich in organic substances, which are typical for these mulate locally on the surface of frozen rivers. As a result of
Arctic regions, substances such as dissolved ions, copper, these processes, river ice often contains large amounts of se-
lead, and manganese are mostly found in the form of or- diment (Zubov 1943).
ganic complexes. The surface ice can also freeze into the bed and/or bank,
and the rising river stage in the spring will dislodge both the
ice and the attached frozen material (Beltaos et al. 1993).
3.3.4.5. Ice
In Arctic regions, in-channel ice plays a more critical role
Ice plays a unique and important role in the Arctic. River in the control of hydrological processes than landscape run-
and lake ice restrict the exchange between the atmosphere off (Prowse 1994). The presence of floating ice increases the
and the water. River ice incorporates particles when it freezes wetted surface of the channel perimeter and the under-ice
to the bottom, as well as during anchor and frazil ice for- surface roughness increases the drag on the water. Assuming
mation. River and lake ice also receive contaminants from that the bed and ice-cover underside have identical rough-
atmospheric deposition throughout the ice-on season. Dur- ness, then the uniform flow depth will increase by something
ing the spring melt, ice transport can gouge sediments and of the order of 30% (Beltaos et al. 1993). For the same dis-
create ice jams which can result in accelerated stream flows charge, the presence of an ice cover generally reduces the
and associated sediment erosion. For these reasons, the role sediment ‘driving’ variables of shear stress, velocity, and dif-
of ice in transporting contaminants must be considered. fusivity, and thus, the sediment transport capacity ought to
be reduced, though there has been only limited investigation
Table 3·7. Dissolved and particulate organic carbon concentrations in of this effect. This is especially true for fine particles (< 62
selected Arctic rivers (Source: WHO/UNEP in press).
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
µm) (Beltaos et al. 1993).
POC, DOC/ Seasonal low flows and floods are primarily the result of
DOC a, POC b, % in TOC d, TOC, river ice freeze-up and ice jams, respectively. Most of the an-
River mg/L mg/L TSS c mg/L %
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
nual discharge of water and particles in Arctic rivers takes
North Dvina 20.1 3.2 23.4 23.3 86 place during the short spring freshet. The large hydrodyna-
Pechora 12.7 0.3 16.0 13.0 98 mic forces combined with the flow of broken ice result in
Ob 09.1 0.9 02.0 10.0 91
Lena 06.6 1.1 03.8 07.7 86 erosion of the embankments, sediment re-distribution, and
Mackenzie 05.3 7.3 01.7 12.6 42 other major changes of river geomorphology. A dynamic
Average for world rivers 05.3 4.6 01.0 09.9 55 break-up with associated ice jamming and release will pro-
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– duce higher suspended sediment flux than a thermal break-
a. Dissolved organic carbon. up (Milburn and Prowse 1996).
b. Particulate organic carbon.
c. Total suspended sediment. Under open water conditions, it is the atmospheric inputs
d. Total organic carbon. that dominate the heat budget of rivers. Once ice forms, the
Table 3·8. Average dates for the beginning of river ice drift, and the length al. 1993). During a freezing period, supercooled water in
of ice drifting inthe river mouth regions of selected Russian Arctic rivers turbulent reaches can form extensive covers of anchor ice on
(Data source: V. Vuglinsky, pers. comm., State Hydrological Institute, St.
Petersburg, Russia, 1996).
the river bed. When anchor ice releases from the bottom, it
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– lifts the underlying bed material. The significance of anchor
Average date of Length of ice on the erosion of the riverbed has been only poorly as-
the first day spring ice
River of drifting drifting sessed to date (Prowse 1994).
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
North Dvina 2 May 8 days Transport processes
Pechora 24 May 4 days
Ob 27 May 4 days The contaminants in the water, suspended matter, and bot-
Yenisey 30 May 10 days tom sediments are subjected to processes of ice trapping,
Lena 3 June 9 days
Yana 27 May 6 days transfer, and release. These processes are controlled by the
Indigirka 3 June 4 days varying patterns of freezing, ice growth, and ice cover decay,
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– as well as by the water regime in the autumn-winter period.
convective exchanges are decoupled. Also, ice quickly ab- During freeze-up or ice jams, flow is reduced downstream of
sorbs incoming long wave radiation resulting in strong the ice accumulation. Ice-induced flood waters upstream can
cooling of the water at the ice surface, thereby creating the replenish water in the flood plain (Prowse 1994). The reduced
potential for high rates of ice growth (Beltaos et al. 1993). water velocity on the flood plain can result in sedimentation
Ice cover reduces the flux of short wave radiation through of fine-grained material and associated contaminants.
greater surface reflection and attenuation. Snowcover on There are two major patterns of ice cover break-up in
the ice further reduces this flux, as noted by Prowse and rivers, mechanical or dynamic break-up and thermal break-
Stephenson (1986). up (Prowse 1994). In some rivers, the peak of the spring
Break-up of the river ice affects water temperature and flood wave is accompanied by rapid ice movement and for-
other important processes, such as nutrient processing and mation of a large number of strong jams, resulting in a criti-
the availability of carbon to the biota (Scrimgeour et al. cal rise in water level. A largely irregular river system with
1994). Average dates for when river ice first begins to drift many branches contributes to the formation of numerous ice
in the river mouths of selected Russian Arctic rivers are jams, which significantly extend the flood period. Dynamic
given in Table 3·8. In general, ice break-up occurs in late break-ups are observed in rivers such as the Yenisey, Kotuy,
May or early June and continues for one to two weeks, as Anabar, Olenek, and Yana. Moving ice is subjected to mech-
considered above in section 3.3.4.1. anical decay and melting, gradually releasing particles and
contaminants into the water phase, or depositing them on
Sources of contaminants to the ice shores and sand banks. The thermal type of ice break-up is
River ice stores and transports contaminants, and may there- less energetic and does not create strong jams in the rivers
fore be considered an indirect source of contaminants. Fac- due to the general decrease in the flexural strength of the ice
tual information about the rates of uptake of contaminants through the melting and increased porosity of the ice along
to the ice or the extent of their subsequent accumulation crystal intersections (Prowse 1994).
and transport is scarce. The magnitude of these processes The breakup season within the Mackenzie Delta area
can only be estimated through combining existing informa- lasts 4-7 weeks, beginning in late April or early May and
tion on contaminant accumulation and physical behavior in ending in mid to late June. In the Mackenzie River, the flood
the ice with the more traditional and ample knowledge on water from the southern part of the basin arrives at the delta
transport processes of sediments, particles, and solutes. before the melting of the ice cover in the delta lakes takes
Contaminants are incorporated into the ice cover from place. Thus, in early May, there is a rapid increase in dis-
the atmosphere, water masses, and sediments. The surface charge and suspended-sediment concentrations which ac-
of the ice accumulates snow which has scavenged contami- company the initial period of flow and ice movement along
nants from the atmosphere during deposition, and retains shallow distributary channels. Delta front areas less than 1 m
particles and associated contaminants deposited as dry de- in height become submerged as large volumes of water be-
position. Contaminants which are sorbed to fine suspended come confined at the interface between river mouth and sea-
particulate matter in the river can be scavenged during the ward bottom-fast ice. The melting process continues until
growth of frazil ice (Prowse 1994). The uptake of contami- much of the ice has been cleared along major channels. Max-
nants takes place along the entire period of ice cover (8-10 imum discharge occurs in early June as a result of the melt-
months in most Arctic rivers), and their release occurs during snow within the basin. Ultimately, the high volumes of
ing a short melting period in summer, when biological pro- meltwater overflow bottom-fast ice out to the 2 m isobath,
duction is at its highest. There does not seem to be any evi- accelerating the melting process. Delta-front flooding per-
dence to suggest whether or not water-soluble contaminants sists into late June at which time the landfast ice is virtually
are frozen into the ice during freezing, or preferentially re- melted and the river discharge begins to decline.
leased to the underlying water. Nonetheless, Lesack et al.
(1991) report an increase in solute concentrations in the Dissolved Oxygen (DO)
residual water by a factor of 4.3 from the time of initial The long periods of ice cover in northern rivers may result in
freeze-up to the development of maximum ice thickness. severe DO depressions during the winter season. Investigations
In rivers, the effects of freezing on the water composition within the Yukon River basin (Schreier et al. 1980), revealed
could be less apparent than in lakes, because of constant that severe DO depressions occurred throughout the basin
mixing and through-flow (Cheng et al. 1993). with lowest levels in late winter. Open water reaches during
Where the growing ice pack comes into contact with the winter season did not necessarily aerate the river, as these
shores and sand banks, the bank and bottom sediments were frequently linked to groundwater inflows, which often
may freeze into the ice. This process happens both in au- exhibited low oxygen concentrations as well. Nevertheless,
tumn, during the period of stable ice cover formation, and recovery from low DO was rapid with ice break-up.
in winter, if the flow of bottom water decreases or ceases The Takhini River, a groundwater-dominated system, and
and the ice cover lies on the bottom of the river (Beltaos et the Nordenskiold River, a river rich in organic matter and
with many bogs and marshes, (both in the Yukon) were sam- chapters (e.g., chapter 6 for POPs and chapter 7 for metals).
pled to further investigate DO under ice (Whitfield and Mc- Discussion here will be limited to water, sediments, and total
Naughton 1986). DO decreased very rapidly in November organic carbon (TOC) which are summarized in Table 3·9.
and continued to decline in both rivers until February or Most dissolved contaminants are not conservative; their
March with both rivers showing DO recoveries during April concentrations in the water flow are changed as a result of
under full ice cover. The fact that DO increased while ice was discharges from various sources to the river and due to vari-
present, concurred with the conclusion of Albright et al. ous transformation processes in the river. In order to describe
(1980) that the Ogilvie and Swift Rivers had greatest micro- and understand the fate of contaminants in river water, con-
algal and bacterial biomasses in spring and summer, which centrations of contaminants should be measured at different
were likely controlled by light and dissolved organic carbon. points within the entire river basin. However, this is not al-
Lakes receiving this deoxygenated water may become ways possible for remote sites. The difficulties in accessing
anoxic themselves. Anaerobic bacteria have been identified many Arctic rivers frequently limits even the collection of
as being capable of mobilizing contaminants (e.g., Hg) from suspended sediment data in sufficient detail to permit the
lake sediments (Bloom and Effler 1990, Mason and Fitzger- sediment transport estimates for these rivers.
ald 1991, Regnell and Tunlid 1991, Gillmow et al. 1992, Choi The unit area discharges (L/s/km2) and unit area loads
and Bartha 1994, Regnell 1994, Zhag and Plonas 1994). (tonnes/km2/y) of suspended sediments and total organic car-
Char Lake, which is a headwater lake and does not re- bon (TOC) are compared for major Arctic rivers in Table 3·9.
ceive oxygen-deficient waters, is in most years a typical po- While these are large, heterogeneous river basins, there are at
lar lake, never rising above 4°C and with a brief summer least some general patterns evident. Relatively high unit area
overturn period of up to six weeks duration (Schindler et al. discharges with generally low sediment and TOC aerial loads
1974, Welch 1994). The lake is extremely oligotrophic, are evident for the rivers of Norway and Finland. By compar-
with soluble reactive phosphorus (SRP) measurements con- ison, discharges tend to decrease to the east in Russia. The
sistently less than the detection level (< 0.7 mg/L); neverthe- highest unit area sediment loads in Russia are for the Pechora
less, winter oxygen depletion at the sediment/water inter- and Indigirka Rivers, essentially on opposite sides of the coun-
face was considerable, probably reflecting respiration by the try, while the highest TOC aerial loadings are in the western
benthic community. DO concentrations in Char Lake appear drainage basins. The large drainage basins of Canada tend to
to recover in early spring and were greatest just under the have relatively low aerial discharges. Unit area sediment loads
ice when it was thickest in late May (Schindler et al. 1974). are also low except for the Mackenzie and Nottaway Rivers.
This probably reflects increased photosynthesis in the sur- A full explanation of these riverine characteristics and the
face waters, in association with increased light penetration impact on the receiving marine waters would require a much
(24 hours of sunlight and melting snow) at this time of year more intensive assessment than is possible here.
(Albright et al. 1980, Prowse and Stephenson 1986, Whit- Large Arctic rivers tend to have a high water discharge and
field and McNaughton 1986). Schindler et al. (1974) also a low sediment load, compared to other large rivers of the
noted that mixing continues after the lake is frozen over, as world. For example, Russian Arctic rivers represent 67% of
evidenced by the general uniformity of specific conductance the total river drainage from the territory of Russia, but only
with depth, which may mitigate the DO depression. 32% of the suspended matter flux from the same area (WHO/
The impact of ice on contaminant transfer has not been UNEP in press). Unit area discharges and sediment loads are
studied to any great extent, but it is clear that ice will form highly variable, depending upon the nature of the watershed.
a reasonably impermeable barrier between the atmosphere The Mackenzie River has a higher total suspended matter dis-
and the water throughout large parts of the year. This obvi- charge than do any of the reported Eurasian rivers (Table 3·9),
ously limits the exchange of atmospheric contaminants with although the unit area load of suspended matter from the Mac-
the water, but also retains any contaminants introduced be- kenzie River is well within the range of the Eurasian rivers.
low the ice within the limited water volume. This may also Most of the other large North American rivers actually have
be exacerbated by low DO resulting in release of these con- very low sediment loads, especially on a unit area basis, due
taminants to the water column or a change to a more bio- to the minimal surficial material within the watershed and the
logically reactive form (e.g., methylmercury). Sources of con- presence of lakes and reservoirs within the system. In Nor-
taminants that may be of concern include contaminated way, the rivers have a much higher rate of water delivery per
groundwater drainage, industrial and municipal discharges unit area than any other part of the Arctic, due to the high
below the ice, and natural seeps of hydrocarbons such as precipitation and relatively short rivers with minimal basin
occur in the Mackenzie River. The potential impact of con- storage. Sediment loads per unit area in Norway are quite low
taminant sources on ice-covered ecosystems has not been and comparable to many of the rivers in the Canadian Arctic.
adequately addressed. While particulate matter is responsible to a great degree
for the transport of hydrophobic contaminants in river sys-
tems, it does not necessarily follow that higher concentra-
3.3.5. Mass transport
tions of total suspended solids means that there is a greater
3.3.5.1. Water and suspended matter
load of anthropogenic chemicals transported by the river, as
The mass of water transported in Arctic river systems is this is determined by the actual inventory of contaminants
directly relevant to the transport of contaminants, either available to the river system from atmospheric and direct
those which are water soluble or those which are hydropho- sources. Even without a high sediment load, rivers have the
bic and adsorbed to suspended particulate matter. Increased potential to transport contaminants either in the dissolved
water flow is often directly correlated with suspended mat- or particulate phase if they are delivered to the system. Rivers
ter load and adsorbed contaminant fluxes. Anthropogenic contribute a large proportion of the sedimentary and by ex-
compounds may be present in different speciations and par- tension, the contaminant budgets of the Arctic Ocean and
titioned between liquid and solid phases, which significantly marginal seas. For example, the estimated annual discharges
influences their transport by river flow. The subject of ac- of suspended sediments by the Lena and Mackenzie Rivers
tual transport of contaminants by rivers and loadings to are 17.6 106 tonnes (Martin et al. 1993) and 125 106
oceans and marginal seas will be dealt with in the relevant tonnes (Lewis 1991), respectively.
Table 3·9. Drainage area, discharge, surface runoff, concentrations and estimated loads of suspended solids, and TOC for selected Arctic and subarctic
rivers (Sources: Gordeev et al. 1995, Meybeck and Ragu 1995, Holtan et al. 1994, and ENVIRODAT 1996).
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Suspended sediment transport Total organic carbon transport
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Drainage Mean annual Surface Concen- Annual Unit Concentra- Annual Unit
area, discharge, runoff tration, load, area load, tion, load, area load,
River 103 km2 m3/sec L/sec/km2 mg/L 103 t/y t/km2/y mg/L 103 t/y t/km2/y
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Norwegian Sea (from 62°N 5°E to 70°N 20°E)
Gaula 3.7.0. 96.6 26.4 3.90 13.9 3.5* 3.2 11.9 3.0
Orkla 3.1.0. 66.3 21.7 1.50 3.9 1.4* 2.8 6.2 2.0
Rana 3.8.0. 172.70 44.9 7.10 45.9 11.7* n.a. n.a. n.a.
Vefsa 4.1.0. 181.30 44.0 4.00 23.9 5.6* 3.4 20.9 5.2
Whole Basin .94.7040 7353.500 77.6 –.9 224.9 3.7* –.9 61.9 1.02
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Barents Sea (Norway, from 70°N 20°E to Russian Border)
Alta 7.4.0. 86.9 11.8 1.80 5.9 0.7* 3.5 11.9 1.49
Tana 16.4.00. 188.50 11.5 2.10 23.9 0.8* 3.0 n.a. n.a.
Pasvik 18.4.00. 171.20 9.3 1.40 7.9 0.4* 3.0 20.9 1.09
Whole Basin 73.1410 1316.000 17.9 –9– 36.9 0.7* –.9 33.9 0.62
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Barents and White Sea (Russia)
Onega 57.1.00. 501.60 8.8 18.90 300.9 4.9* n.a. n.a. 5.8
North Dvina 357.100.0. 3462.9 9.7 35.90 3800.9 10.6* n.a. n.a. 7.2
Mezen 78.10.0. 868 11.9 32.90 900.9 11.1* n.a. n.a. 2.4
Pechora 322.100.0. 4089 12.7 80.10 13500.1 32.4* n.a. n.a. 5.2
Whole Basin 1236.1.0000. 13596 11.1 50*0 22000.1 17.5* n.a. n.a. 4.1
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Kara Sea
Ob 2990.1000.0. 12857.1 4.3 38.10 16500.1 6.4* n.a. n.a. 1.2
Yenisey 2580.1000.0. 19866 7.7 10.10 5900.1 2.3* n.a. n.a. 1.8
Pyasina n.a..00. 2727 n.a. 40.10 3400.1 18.8* n.a. n.a. n.a.
Whole Basin 6248.1.0000. 42486.4 6.8 22*0 33200.1 5.0* n.a. n.a. 1.5
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Laptev Sea
Khatanga 364.1.000. 2693.6 7.4 20.10 1700.1 4.6* n.a. n.a. 1.5
Anabar 100.1.000. 550.00 5.5 n.a.00 n.a.** n.a.* n.a. n.a. 0.9
Olenjok 219.1.000. 1138.800 5.2 31.10 1100.1 5.1* n.a. n.a. 1.2
Lena 2486 16656 6.7 34.10 17600.1 7.1* n.a. n.a. 2.1
Yana 238.1.000. 1094.800 4.6 103.10 3500.1 14.8* n.a. n.a. 1.0
Whole Basin 3643.1000.0. 23679.5000 6.5 34*0 25100.1 6.9* n.a. n.a. 1.8
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
East Siberian Sea
Indigirka 362.100.0. 1918.600 5.3 210.10 12900.1 35.6* n.a. n.a. 1.3
Kolyma 660.1.000. 4158.100 6.3 120.01 16100.1 24.3* n.a. n.a. 1.6
Whole Basin 1342.0001.0. 7917.800 5.9 134*0 33600.1 25.0* n.a. n.a. 1.3
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Chukchi Sea (excluding Alaska)
Whole Basin 94.2.00. 640.60 6.9 34*0 700.1 7.4* n.a. n.a. 1.0
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Bering Sea
Anadyr n.a..0. 1902.6 0 n.a. 59.10 1800.1 17 .1* n.a. n.a. n.a.
Yukon n.a..0. 6659.001 n.a. 286.07* 60000.1 70.4* n.a. n.a. n.a.
Whole Basin n.a..0. n.a. n.a. n.a.00 n.a.** n.a.* n.a. n.a. n.a.
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Beaufort Sea (Mackenzie is totally dominant)
Mackenzie 1787..0000.1 10543.3000 5.9 127.07* 125000a.1 24.1* 7.8 2597.1 1.5
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Arctic Archipelago
Coppermine b 19.3.00. 105.01 5.4 01 3.3 0.2* 2.3 7.6 0.39
Back b 93.9.00. 4970.1 5.3 04 62.7 0.6* 3.3 51.7 0.55
Hayes b 18.1.00. 195.01 10.0 11.07 68.0 3.8* 3.1 18.4 1.1
Whole Basin n.a..0. n.a. n.a. –.1 n.a.0 n.a.* –.1 n.a. n.a.
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
*Hudson Bay and Hudson Strait
Churchill b 287.1.000. 882.01 3.1 06.50 180.2 0.6* 9.8 272.6 0.9
Nelson b 1010.0001.0. 2170.100 2.1 10.8000 736.3 0.7* 1.1 68.4 0.06
Moose c 60.1.00. 780.01 12.9 09.20 400.1 3.7* 21.1 516.5 8.6
Nottaway b 57.5.00. 1040.001 18.1 26.70 1000.1 15.2* n.a. n.a. n.a.
La Grande b 96.6.00. 170000.1 17.6 n.a.00 n.a.0 n.a.* n.a. n.a. n.a.
Koksoak n.a..0. 80.4 n.a. n.a.00 n.a.0 n.a.* n.a. n.a. n.a.
Whole Basin 3607.0001.0. 16952.90000000 4.7 –.1 n.a.0 n.a.* –.1 n.a. n.a.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
* Mean Value
a. Lewis 1991.
b. Data for this station were calculated using information provided from HYDAT (1994) and ENVIRODAT (1996).
c. Data for this station were calculated using information provided by HYDAT (1994) and ENVIRODAT (1996). In addition, the Total Organic Car-
bon data, and the sum of Particulate and Dissolved Carbon (1977-1978) were provided from Water Quality Investigation in Ontario’s Arctic Water-
shed, Inland Water Directorate, Environment Canada, Ontario Region (unpublished).
Norway northern Norway to the sea is in excess of 150 103 ton-

Water courses north of the Arctic Circle cover an area of nes/y, or 1.6 tonnes/km2/y (Holtan et al. 1994). A model
approximately 93 000 km2, and their discharge to the Nor- has been developed to calculate area runoff of total phos-
wegian and Barents Seas is approximately 70 109 m3/y, phorus, total nitrogen, phosphates, nitrates, and ammonia,
corresponding to a total runoff of 23.9 L/s/km2. The total using area-specific runoff coefficients (see Holtan and Åste-
amount of suspended particulate matter transported from bøl 1991).
e p o s it io n
h e ri c d change
tm o s p G a s e xn o c e a n
d d ry a b e tw e e o s p h e re
Wet an a n d a tm
g
M in in
l
C o a snta
t
en
y ru o ff
rr
str
cu
u
Ind
an
ce
O
ti o n
s ta l e n ta
C o an S e d im
ru o ff c ti o n
O il p ro d u
Agric ulture ic e
Sea
Sh
ip t
raff
i c
+
ti o n
e n taa ti o n ino n e
n s io n S e dnim fo r m ez
spe tr a as/e s tu a r in
Resu d e lt
re
ti o n cultu
e n ta Aqua
r in e
R iv e s p o r t S e d im
tr a n C it y
Figure 3·17. Conceptual model of the coastal zone and marine environ-
ments showing the main subcompartments and contaminant transfers,
and exchanges with the atmosphere.
The rivers in northern Norway are small, relative to the and stereols. The PAH and pentacyclic triterpanes showed,
major rivers in northern North America and Russia, and respectively, alkyl homologue and hopane distributions char-
the quantity of particulate matter transported is relatively acteristic of natural seeps of fossil fuels (Yunker et al. 1991a).
small. The total amount of suspended particulate matter The particulate hydrocarbon flux from the Mackenzie River
transported to the Norwegian Sea and the Barents Sea is is by far the most important terrestrial source of hydrocar-
calculated to be 166 106 tonnes per year in a total water bons to the Beaufort Sea. The flux of total alkanes is esti-
runoff of 70 109 m3/y (Holtan et al. 1994). The river with mated at 440 94 103 kg/y and for PAH 49 8 103 kg/y
the highest annual sediment load is the Rana, transporting (Yunker et al. 1991b). There is a strong seasonality in the hy-
45 103 tonnes/y with an average suspended particulate drocarbon flux, with an estimated winter contribution of less
matter concentration of 7.1 mg/L (Holtan et al. 1994). than 0.6% of the total annual flux (Yunker et al. 1991a).
Other rivers in Canada have not been studied in the same
Russia detail, due to their much lower suspended sediment loads
The mean annual runoff of particulate matter in most Arc- and more difficult access. As noted in Table 3·9, the flows,
tic rivers of Russia does not exceed 20 tonnes/km2/y. The sediment, and TOC loads in the other rivers are relatively
lowest values (5 tonnes/km2/y) are characteristic of rivers in small compared to the Mackenzie River. Unit area loads are
the forests of the western (European) region, north Cauca- also small by comparison, except for the sediment load from
sus, west Siberia, Kazakhstan, and the mid-Siberian plateau. the Nottaway River and TOC from the Moose River. The
A moderate runoff of particulate matter (from 5 to 20 ton- high TOC load of the Moose River may be an artifact result-
nes/km2/y) is characteristic in the eastern part of the Euro- ing from the use of a different analytical method and having
pean territory, the Ural Mountains, the mountains of south had more detailed data available for this river system from
Siberia, the northeast and Far East, and also in the desert the intensive study of McCrea and Wickware (1986).
regions of Central Asia.
With the exception of Mezen River, rivers of the basins
3.3.5.2. River/lake ice
of the Barents Sea and White Sea flow through forests and
bogs and the TOC runoff to the rivers is relatively high Although ice on Arctic rivers and lakes is considered to be
(mean values of 5.2-7.2 tonnes/km2/y). The TOC runoff to an important phenomenon with respect to the fate and path-
Siberian rivers is not as high (0.9-2.1 tonnes/km2/y) despite ways of contaminants, there seems to be no information avail-
the occurrence of fairly high TOC concentrations in the wa- able with respect to the mass transport of ice in the freshwa-
ter (5-10 mg/L). The TOC runoff from the tundra zones of ter environment. Some preliminary investigations of this
the basins of the Laptev, East Siberian, and Chukchi Seas is transport pathway should be undertaken to determine if
low (0.5-0.8 tonnes/km2/y), compared to the average global quantification of ice transport and associated contaminants
value of 3.7 tonnes/km2/y (Meybeck 1993). is required relative to ocean loadings. This is considered fur-
ther in the Ocean section (section 3.5.3.1).
Canada
The Mackenzie River has been extensively studied with respect
to the sediment load to the delta and the Beaufort Sea. The
3.4. Estuaries, deltas, and fjords
delta is maintained by the Mackenzie and Peel Rivers, which
together have a mean annual discharge of almost 10 000 m3/s Oceanographic conditions in coastal waters will differ from
and deliver in the order of 125 million tonnes of clayey silt those in the open ocean, and some of the factors causing these
and sand to the delta annually (Lewis 1991). The suspended differences are river runoff, tidal currents, and the effect of
particles in the Mackenzie River are of mixed biogenic and shore boundaries. The interfaces between the terrestrial/fresh-
petrogenic origin. The Mackenzie River receives only minor water processes and the oceanic processes are highly complex
anthropogenic hydrocarbon inputs (Yunker et al. 1991b). and unique to both compartments. Deltas and estuaries play
Higher plants and peat were sources for the alkanes, alcohols, an important role in sedimentation in freshwater systems.
Nevertheless, deltas and estuaries have a close connection their extreme seasonality in water discharge, winter flows
with the marine environment and have important influences being only about 5-10% of the annual average flow (Anto-
upon oceanic transport and contaminant fate (Figure 3·17). nov 1970), and because they are covered with ice from Oc-
tober through June (e.g., Pfirman et al. 1995a). The Ob and
the Yenisey Rivers have the largest estuaries in the Arctic.
3.4.1. Estuaries
Most estuaries were formed following the last deglacia-
Estuaries are semi-enclosed, coastal bodies of water, having tion, when the rise in sea level inundated coastlines and
a free connection to the open ocean, and within which sea- drowned the mouths of river valleys. However, when sedi-
water is measurably diluted by freshwater derived from ment discharge is high, the estuary fills, and a delta forms in-
land drainage (Lauff 1967, Pritchard 1967). From a geo- stead. For existing estuaries, either the sediment discharge is
chemical point of view, any semi-enclosed body of water, low, the inlet is deep, or the tidal action is large enough that
where marine waters are mixed with freshwaters, should be the inlet does not fill in.
considered as estuarine zones. Consequently, river deltas, All estuaries contain a transitional zone where the salt
fjords, bays, and tidal swamps may be considered as estuar- water of the ocean and the freshwater of the river runoff
ies. Because of their high productivity, estuaries often serve meet. Because freshwater is less dense, it tends to flow out
as breeding and nursery grounds for many species. The fate over the saline seawater. Entrainment of some seawater in
of contaminants released from local as well as up-watershed the freshwater outflow results in a return, replacement flow
activities is influenced by the characteristics of the estuaries. of seawater up the estuary at depth. In the lower, marine
Arctic estuaries differ from those in temperate regions in part of the estuary, there is free connection with the open
Salinity
STN 1 STN 2 STN 3 STN 4
0 10 20 30
River
0
5 10 15
STN 1
Seawater
STN 2
33
salt wedge
STN 3
STN 4
0 10 20 30
River
0
10
15
20 25
30
STN 1
STN 3
partially-mixed
STN 2
STN 4
0 10 20 30
River
0
5 10
15
20
STN 1
25
30
STN 2
STN 3
STN 4
well-mixed
0 10 20 30
River
0 5 10 STN 2 3 4
15 1
33
Sill
33
fjord-type
STN 2, 3
Figure 3·18. Basic circulation and salinity distribution in salt wedge, partially- mixed, well-mixed and fjord-type estuaries as defined by Wollast and Duinker
(1982). Numbers and shading show salinity values.
ocean. In the middle part of the estuary, salt water and

freshwater mix. Dominated by freshwater river runoff, the
upper part of the estuary is also influenced by tides. In some
Arctic rivers, tidal influence penetrates as much as hundreds
of kilometers upstream (Antonov 1970).
Wollast and Duinker (1982) defined three types of estu-
aries based on the way that seawater mixes with river run-
off. These are salt wedge, partly mixed, and well mixed es-
tuaries. The basic circulation in each of these types of estu-
aries is illustrated in Figure 3·18. Subsequently, fjord-type
estuaries were added by Duxbury and Duxbury (1994). Salt
wedge estuaries occur where there is a large amount of river
runoff. The less dense freshwater tends to run over the more
dense seawater that would otherwise flow upstream at depth,
but is held back by the river discharge. A strong front forms
between the freshwater and salt water, resulting in the salt
wedge. The salt wedge moves upstream when the tide rises,
or when the river discharge is low. Conversely, it moves
downstream when the tide falls, or the discharge is high.
Partially mixed estuaries have both strong seaward flow
of freshwater and strong inward flow of seawater. In well-
mixed estuaries, the combined effects of strong tidal mixing
and low river discharge result in a slow net seaward flow of
water at all depths. The isohalines are nearly vertical. Figure 3·19. Advanced Very High Resolution Radiometer (AVHRR) satellite
Fjord-type estuaries (see also Figure 3·20) are deep with image (June 4, 1986) showing a color composite of the visible data band
a sill at the fjord entrance, and have little tidal mixing. The for the Mackenzie Delta. River water overflowing the landfast ice can be
seen as purple regions extending offshore from the mouths of channels
river water flows out in a thin surface layer over the seawa- (Source: Dean et al. 1994, used with permission of K. Dean and K. Ahlnäs).
ter, with only minor entrainment of the underlying seawa-
ter. Because circulation is confined to the surface waters, shelves. The best examples are the Lena River in Russia and
and the deep waters are isolated by the sill, the deep waters the Mackenzie River in Canada (Figure 3·19), where shal-
may stagnate and become anoxic. low inshore areas of the shelf provide a wide platform upon
If the estuary is broad relative to its length, the circula- which large deltas have been built. Seasonal ice potentially
tion pattern may have three distinct regimes. Due to the Co- has important effects on the morphology of some deltas in
riolis force, the out-flowing river discharge within an estu- the Arctic. For example, Reimnitz et al. (1988) note a flat
ary in the Northern Hemisphere will be forced to its right (< 2 m vertical variation in morphology), extensive (16 km)
as it moves out of the estuary. Similarly, the incoming flow platform in the subaqueous ice-stressed deltas on the Alas-
is forced to the right looking upstream in the estuary from kan Arctic shelf. This feature, however, is not seen in the
the perspective of the ocean. As a result, seawater will tend Mackenzie Delta. Because the freshet precedes ice break-up
to flow in on one side of the estuary and the river water will in Arctic deltas, the water level rises rapidly in spring (re-
flow out on the opposite side. In between, a zone of lateral portedly more than 10 m for the Lena) and floods the
mixing between the freshwater and seawater is established. grounded ice in the delta. At this time, water and sediments
are transported on top of the ice and in channels under the
ice (Figure 3·19 and see Dean et al. 1994). Over-flooding
3.4.2. Deltas
may also be associated with ‘strudel-scour’ which is caused
The term ‘delta’ refers to the depositional plain formed by a by rapid drainage of ponded waters on the ice surface
river at its mouth (Encyclopedia Britannica 1990). Inherent through melt-holes and fractures in the ice (Reimnitz and
in this definition is the concept that sediments delivered by Kempema 1987).
the river are deposited at or near the coast, and thereby ac- Arctic deltas exhibit the characteristics of drowning coast-
tively build out the land area into the sea. In addition to the lines. These include features like coastal cliffs showing mass
subaerial portion of the delta (i.e., above the mean water level wastage, debris slides, and surface erosion; cliffs showing
and consequently flooded only during the freshet), a subaque- ground-ice slumps and mud flows; sediment deposits on
ous delta (i.e., below the mean water level of the receiving coastal tundra; inundated tundra; and, breached lakes (e.g.,
body of water and consequently subjected to coastal pro- see Harper 1990). This is due to a protracted rise in rela-
cesses rather than terrestrial processes) also builds outward tive sea level during the Holocene (approximately the past
and is usually evident as a bulge in the adjacent bathymetry. 15 000) which is estimated to total 70 m for the Beaufort
Delta morphology exhibits a lot of variation resulting shelf (Hequette et al. 1995) and possibly as much as 120 m
from: drainage basin characteristics (climate, lithology, basin for the Laptev shelf (Nurnberg et al. 1995) and other Arctic
size), sediment loading and factors affecting it (Milliman and regions (Pirazzoli 1991, Rasch and Nielsen 1995). Sea-level
Syvitski 1992), energy along the shoreline (tides, current, rise continues globally at a present rate of about 3 mm/y
and waves), bathymetry of the shelf onto which the river dis- (Douglas 1991). As a consequence of the rising water level,
charges, and human impacts on river systems (GESAMP parts of the outer Mackenzie Delta are estimated to be erod-
1993). The largest deltas are created by major rivers draining by as much as 20 m/y, although 2 m/y is more typical
ing large areas and carrying abundant sediment loads. Large (Harper 1990). Similarly, the Lena delta is generally retreat-
subaqueous delta plains are best developed where the conti- ing (Zenkovitch 1985). Coastal retreat can be expected to be
nental shelf is shallow and the sediment loads are great. most rapid in regions of low relief, containing poorly con-
Within the Arctic, deltas are most obviously formed solidated sediments bonded by permafrost (i.e., Arctic del-
where rivers discharge directly onto broad continental taic environments). Even though portions of Arctic deltas
are presently being lost, there may still be a net accumula- the intermediate layer (between the upper layer and sill
tion of sediment in deltas. For example, thaw subsidence depth) in narrow stratified fjords takes place as a pressure-
could be occurring with subsequent replacement by new driven, mainly two-layer circulation (Svendsen 1977 and
sediments. Alabyan et al. (1996) found that most of the 1981, Cannon and Holbrook 1981, Klinck et al. 1981,
21 106 tonnes/y of sediments supplied by the Lena River Stigebrandt 1990).
is presently settling within the delta. Similarly, Macdonald Due to the rotational dynamics, the exchange in broad
et al. (1996) estimate that about half of the 127 106 ton- fjords takes place somewhat differently than in narrow
nes/y of sediments from the Mackenzie River are presently fjords. The transient behavior due to the distortion of the
stopping in the Mackenzie Delta. density field at the coast will propagate into the fjord as an
Deltas are important transition zones between the rivers internal Kelvin wave, with the shore to the right when look-
and the coastal Arctic Ocean and seas. Since deltas capture ing in the direction of its propagation (Proehl and Rattray
sediment, they potentially provide a significant sink for par- 1984). Preliminary simulations indicate that the wave en-
ticle-reactive contaminants transported by the river and, un- ergy, which is dispersed in the fjord, is directed down-fjord
der conditions of erosion or resuspension, a source. The im- on the opposite side (Asplin 1995, Svendsen 1995).
portance of deltas in the delivery of riverborne contaminants The density of the basin water, below the intermediate
to the Arctic Ocean has been recognized in a number of re- water, gradually decreases with time if renewal is not taking
cent studies (e.g., Martin et al. 1993, Peulvé et al. 1993, place. The speed of this process varies substantially between
Yunker et al. 1993, Coquery et al. 1995). Despite their im- fjords as a function of topography and the effect of driving
portance, modern deltaic sediment budgets are only poorly forces. In some fjords there is an annual renewal, while in
known, and establishing such budgets would be a crucial others it may take several years before the density of the
first step to the construction of regional and local budgets basin water is sufficiently reduced to condition a renewal.
for particle-reactive contaminants discharged to Arctic However, special persistent wind conditions causing strong
rivers. upwelling on the coast can cause unexpected basin water
renewal.
Sometimes the outflow from fjords can be very pronounced,
3.4.3. Fjords
and fjord water can be traced far from the coast. The trans-
Figure 3·20 schematically demonstrates the general circu- port of anthropogenic nutrients and contaminants from the
lation in connection with a fjord. With a sill at the mouth fjords to coastal and open waters strongly depends on their
and a river commonly at the head, the water masses may be residual time in the fjords.
split in three depth zones. The freshwater runoff results in
the formation of an often shallow, brackish surface layer.
3.4.4. Particle and nutrient transport
The seaward flux of brackish water gradually mixes with
the underlying seawater such that its volume may increase Much of the sediment discharged by rivers is deposited
5-10 times before it reaches the fjord mouth. As a conse- within the confines of the estuary. This is due to two factors:
quence, an inflowing compensation current is established to 1) the circulation itself, and 2) flocculation at the freshwa-
make up for the loss of seawater from the fjord. The inter- ter – saltwater front.
mediate water occupies the zone between the brackish wa- Estuarine circulation is characterized by river water and
ter and the sill. The water exchange in this layer is mainly entrained seawater flowing out at the surface, with a re-
governed by seasonal and short-term variations in the placement flow of seawater at depth (Figure 3·21). This
coastal water. means that particles suspended in river discharge will also
River flow out at the surface. But, as they settle, they will be
transported into the in-flowing seawater at depth, and will
Ocean Coastal Fjord water be advected back upstream some distance before they fi-
water water
nally deposit. In addition, suspended marine sediments will
Brackish water
be transported into the estuary by the upstream flow of
Intermediate layer seawater. Both nutrients and plankton may also remain in
the estuary by this circulation pattern (Laws 1993). As a
result of this recycling of nutrients, estuaries tend to sup-
Basin water port high levels of productivity. Fjords represent a special
case because of the presence of a sill. For the most part,
materials that sink below the level of the sill become trap-
ped in the fjord.
Figure 3·20. Sketch of main circulation pattern in a fjord. Flocculation is the result of molecular attractive van der
Waals forces. In fresh river water, clay minerals carry a net
Recirculation is a characteristic feature of fjord circula- negative charge and repel one another. However, in salt wa-
tion driven by the local forces of runoff and wind. Some re- ter, free cations neutralize the clay minerals, allowing mole-
newing of water masses takes place due to exchange with cular attractive forces to dominate when clay minerals are
the water masses adjacent to the ‘open’ end of the semien- brought close enough. As the clay minerals attach to one an-
closed circulation in the frontal area at the mouth of the other, their settling velocity increases, leading to increased
fjord. However, the main part of the exchange is associated deposition.
with nonlocal forcing. Several mechanisms governing the Similarly, coagulation and biological aggregation result
exchange have been studied through data analysis, numeri- in deposition of particles. At these low salinities, riverborne
cal models, and theory. Common to all of these studies are colloidal iron and humic acids are known to coagulate and
the assumptions of displacement of the vertical density field settle out of suspension (Sholkovitz 1976, Boyle et al. 1977,
at the coast caused by the coastal wind; and, that the ex- Sholkovitz et al. 1978). Organisms feeding in the estuary ex-
change mainly affects the intermediate and deep-water crete particle aggregates as fecal pellets (up to 5 mm long),
masses in the fjords. It has been shown that the exchange in which have rapid settling velocities.
River runoff Freshwater
Null Entrainment
point
Saltwater
Estuarine circulation
River discharge
Outward transport of
sediments and plankton
Turbidity Sinking of
maximum
Inward and upward transport of detritus
sediments and nutrients
Sedimentation
Estuarine cycling
Figure 3·21. Illustration of water, sediment and nutrient cycling in estuaries. Freshwater flows outward, transporting sediments, plankton, and also con-
taminants. Material that sinks into the inflowing seawater may be trapped in the estuary. Estuaries such as those of the Ob and Yenisey are covered by
ice during much of the year (Source: adapted from Laws 1993).
As a result of estuarine circulation, a turbidity maximum geographical conditions of their basins (see section 3.3.4.3).
with accumulation of fine-grained sediment often forms at Sediment can be deposited in deltas upstream in lakes and
the front between river water and seawater (Figure 3·21). reservoirs, such as the Slave River delta upstream in the Mac-
Here, particle concentrations are high because clays brought kenzie River system and the numerous reservoirs in the up-
downstream by the river flocculate, and particles transported per and middle parts of the largest Siberian rivers. In the
upstream in the marine flow deposit at the null point in the Yenisey basin, for instance, eight reservoirs have been con-
saltwater flow. Because of the strong seasonality in Arctic structed, which together control about 23% of the total
river discharge (see section 3.3.4.1), the location of the tur- river flow. Their impact on suspended matter flow is parti-
bidity maximum will shift, potentially resulting in remobi- cularly pronounced, with a reduction by a factor of three in
lization of particle accumulations on a seasonal basis. Dur- sedimentation subsequent to the reservoirs being put in op-
ing the spring flood, the rise in waters of the lower regions eration (Alabyan et al. 1991).
of the Yenisey River may exceed 20 m, while in the Lena it Recent field studies of particulate matter carried out in
attains as much as 30 m (Antonov 1970). the Ob and Yenisey estuaries (Lisitsin et al. 1994) showed
significant sedimentation of riverine suspended matter in the
mixing zone of river and sea waters as well as in the shelf
3.4.5. Sedimentation
zone of the Kara Sea. Estimates show that a maximum of
Much of the sediment load transported by Arctic rivers dur- 10-20% of the particulate matter discharged by the rivers
ing the high flow, high sediment load season is deposited in Ob and Yenisey passes beyond the borders of estuaries and
river mouth deltas. Overbank sedimentation rates in the the Kara Sea shelf. This is a typical situation for rivers which
Mackenzie Delta can be of the order of several centimeters transport sediment material into wide shelves, such as that
per year (Pearce 1993). This is likely typical of other deltas of the Kara Sea (Milliman 1991).
and is much greater than the 100-1000 mm/1000 years in
estuaries and 1-3 mm/1000 years for ocean sedimentation
3.4.6. Ice cover
rates (Lisitsin 1988). Huge amounts of suspended matter
are deposited in estuaries and excluded from further trans- Arctic estuaries are influenced by ice cover during most of
port to the open ocean. Sharp moderation of river flow the year. Usually in October, fast ice develops in the estuary,
rate, followed by decrease of a vertical component of turbu- along the coast and in the river itself and the ice remains in
lent flow and deposition of large particles is the main rea- place until break-up in June (Ingram 1981, Ingram and La-
son for this effect. Sedimentation of finer particles, which rouche 1987, Macdonald et al. 1995, Pfirman et al. 1995a).
have colloidal properties, is caused by coagulation due to Thus, for most of the year in the Arctic, river runoff enters
mixing of freshwater with seawater electrolytes. the ocean under an ice cover. In shallow estuaries (< 2 m) ice
Suspended matter concentrations in large Arctic rivers growth may interfere with inflow in late winter. Under these
are rather low, as a result of the geological and physical- conditions, the water ‘leaks’ into the offshore through ero-
sional channels (strudel scour), or in pulses modulated by of suspended and dissolved matter (55-60% of total annual).
tidal lifting of the ice (Reimnitz and Kempema 1987). De- The major portion of the suspended matter is trapped and
spite substantially reduced inflow over winter (Antonov deposited during this season. These sediments may be remo-
1970, Ingram 1981, Pavlov and Pfirman 1995), estuarine bilized during autumn storms, especially in the nearshore
river plumes may be thicker, extend farther offshore and (Hill and Nadeau 1989). The winter regime is characterized
maintain integrity over a larger area because of the ice cover by increased ice cover of river and sea surfaces, this reducing
(Ingram 1981). wind-induced mixing. In addition, a sharp reduction of river
The lack of wind mixing when ice has formed leads to discharge and fluxes of suspended and dissolved substances
very sharp vertical stratification with an almost completely (10-15% of total annual) occurs. The wintertime reduction
fresh layer floating above the salt water. For example, Mac- of biological activity results in less bio-filtration. Although
donald et al. (1995) were able to account for virtually all of material fluxes are reduced in winter, concentrations of con-
the winter inflow from the Mackenzie River as a large, 5 m taminants may be higher than in summer.
thick, 80 km3 lake under the ice on the inner shelf at the end Conservative contaminants dissolved in river water mix
of winter. The salinity gradient in the estuary becomes ex- in the same way as salinity, and are discharged from the es-
tremely strong in the vertical (30 psu over several centime- tuary. The flushing time of the estuary, that is the length of
ters) and extremely weak in the horizontal (10 psu over 60 time required for the estuary to exchange its water, deter-
km) (Macdonald and Carmack 1991). The strong vertical mines how rapidly such wastes are moved out to sea and
stratification can lead to the production of frazil ice at the diluted.
saltwater interface. This occurs because the freezing temper- Fast ice often melts in place, releasing its sediment and
ature of salt water is lower, by as much as 2°C, causing heat associated contaminant load (e.g., Reimnitz and Bruder
to diffuse out of the freshwater layer which is already at its 1972). Fast ice that does drift away from the Kara Sea coast
freezing point. Because particle supply is reduced at this during the summer is likely to melt within the confines of
time of year, it is expected that contaminant scavenging will the Kara Sea, perhaps redistributing some incorporated con-
be weak and the river water will maintain much of its contaminants. Some river-influenced ice may also be exported
taminant load. A significant portion (15% or more) of the from the sea, resulting in a wider distribution of incorpo-
winter inflow is incorporated into the growing ice cover in rated contaminants.
the estuary during winter (Macdonald et al. 1995).
During spring break-up, the passage of the powerful flood
wave is often accompanied by the deposition of large quan- 3.5. Ocean
tities of ice on the banks and floodplains of rivers. However,
3.5.1. Introduction
the fast ice cover remains in place in the Ob and Yenisey es-
tuaries through the month of June (e.g., Pfirman et al. 1995a). The ocean has been for a long time the final resting place for
Due to the massive discharge of water, the turbidity maxi- many anthropogenic waste materials. Because of its huge
mum will move seaward resulting in sediment and contami- volume, the ocean has been thought a safe place for disposal
nant accumulation farther out in the estuary. With the arri- which would not pose any danger to humans. This view has
val of break-up, the nearshore opens, ice melts, and the ac- to be reconsidered. Today, the ocean water column carries
cumulated winter inflow is then free to mix as a pulse into too many signatures of our society: radioactive isotopes,
the outer shelf during the spring biological bloom. such as strontium and cesium, produced during nuclear-bomb
detonations and in nuclear fuel reprocessing plants; pesti-
cides, including DDT and its degradation products which
3.4.7. Contaminant fate have been found in all ocean organisms analyzed; chloroflu-
Contaminants strongly sorbed to particles will tend to de- oromethanes (freons) which have been measured in the deep
posit with the particles within the estuary, and are often waters of the ocean; petroleum products which soil the sur-
found to accumulate at the turbidity maximum. Contami- face of the ocean; and so on (e.g., Goldberg 1975, Dahlgaard
nants with strong particle affinities include PCBs, DDT, et al. 1986, Krysell and Wallace 1988).
lead, and many other metals. Coagulation of colloids also Once a contaminant has been introduced into the marine
may be implicated in the removal of some dissolved metals environment it becomes a matter of interest to determine
(e.g., for Pu, Shen et al. 1983, Sholkovitz 1983). For exam- where this contaminant goes and how its concentration var-
ple, although Baskaran and Naidu (1995) observed elevated ies with time and location. The processes by which contami-
levels of Pu and 137Cs in surface sediments of the Yenisey nants are dispersed in the ocean fall into two basic groups:
estuary, they noted that it was not possible to determine advection and dilution. Advection transports contaminants
whether the Pu came from coagulation of dissolved Pu, or from one place to another, while dilution reduces its concen-
from simple deposition of particulate material. Such non- tration by mixing (Williams 1979). To be able to understand
conservative contaminants (Dai and Martin 1995) have environmental changes in the Arctic Ocean, it is important
complicated behavior in estuaries, where the conditions to have a good assessment of the present situation, as well as
change so radically (i.e., from freshwater to salt water, from data on which chemicals are being added, subtracted, and
high- to low-particle concentration, from well-mixed to transported from one place to another. The following sec-
stratified, from a shallow regime with constant resuspen- tions describe transport routes and mixing processes that
sion in the river to a deeper regime, often dominated by play a role in the redistribution of contaminants.
settling). This complexity is compounded in the Arctic by Contaminants can be delivered, distributed, or removed
the strong seasonality, resulting in shifts in physical and from the Arctic by different pathways. Atmospheric circula-
chemical conditions. tion brings contaminants into and out of the area on a very
There is a significant difference between summer and short time scale (section 3.2). Rivers, streams, and ground-
winter regimes of estuarine zones which affect contaminant water from the surrounding land (section 3.3) carry their
fate. The summer regime is characterized by the most inten- contaminant burdens to the margins of the ocean, where
sive chemical and biological processes and sedimentation they will be further distributed by ocean currents or by ice.
rates, as well as maximal discharge of river waters and fluxes Drifting ice may transport the contaminants into or out of
the Arctic Ocean (section 3.5.3). The North Atlantic Drift Ten percent of all riverine discharge to the world oceans
(the extension of the Gulf Stream) brings contaminated occurs in the Arctic. The total discharge is about 3300 km3/y
water from the Atlantic Ocean into the Arctic; the East (Aagaard and Carmack 1989). The Arctic rivers Yenisey,
Greenland Current and currents of the Canadian Arctic Ob, and Lena are ranked as 5, 6, and 7 in the world, respec-
Archipelago bring the water back to the Atlantic Ocean tively, in order of annual discharge, and the Canadian rivers
once having circulated in the Arctic; a small inflow of water Mackenzie and Yukon are ranked as 11 and 18, respectively
of North Pacific origin enters the Arctic Ocean through the (Goldberg 1976).
Bering Strait, passes through, and leaves primarily through Siberian rivers discharging into the Kara, Laptev, and
the Canadian Archipelago (section 3.5.4). Surface-inserted East Siberian Seas have a huge combined drainage area of
contaminants enter the deeper water column at locations of 9 000 000 km2 extending far to the south (Shiklomanov and
deep-water formation (section 3.5.4.6). In addition, living Skakalsky 1994), encompassing many industrial and agricul-
organisms within the marine environment may transport tural regions. The Ob River and its tributaries originate as
contaminants from one place to another. far south as 45°N (Futsaeter et al. 1991). While most smal-
From scattered observations and deductions based on in- ler rivers are frozen in the winter, some of the major rivers
terpretation of Arctic processes, the following can be postu- discharge year round as discussed extensively in section 3.3.
lated (Roots 1982): A unique characteristic of the Arctic Ocean is that the river-
ine waters can be traced throughout the Arctic Basin due to
1. A considerable proportion of the dissolved and colloi-
the extensive ice cover which minimizes mixing.
dally suspended material, organic or inorganic, which is
delivered to the North Atlantic Drift eventually reaches
the Arctic Ocean. 3.5.2.2. Glaciers
2. Material reaching the Arctic Ocean in solution or sus-
Atmospheric tests of nuclear weapons and nuclear acci-
pension normally spends several years (and probably
dents contributed to generally elevated levels of radionu-
much longer if entrained in the Beaufort Gyre) circulat-
clides in snowfall during the 1950s and 1960s. Because the
ing under very stable oceanographic conditions, mainly
former Soviet Union used Novaya Zemlya as a nuclear test
under an ice cover where photosynthetic reactions and
site, nearby regions, especially glaciers on Novaya Zemlya,
gas exchange are much reduced. The extreme stability of
Franz Josef Land, Svalbard, and Severnaya Zemlya may be
the oceanic stratification probably retards the mixing of
expected to contain elevated levels of radionuclides. Jawo-
contaminants with deeper layers of water, but ensures
rowski (1989) provides some data on levels of radionuclides
that those which do reach deeper levels stay there.
in glacier ice from Spitsbergen showing that there are ele-
3. Airborne contaminants which are deposited on the sur-
vated levels of 137Cs in ice layers dating from 1955, 1958,
face of the Arctic Ocean (typically originating from north-
and 1965.
ern mid-latitude industrialized areas) become trapped on
Corresponding with these peaks in radioactivity are ele-
the surface of the sea ice, get flushed into the layer of
vated levels of uranium, vanadium, lead, and cadmium due
comparatively fresh water that floods the ocean surface,
to sorption onto particles. Heavy metals deposited from
undergo repeated freezing and melting, and, in general,
Arctic haze similarly are known to accumulate in glacier ice
have less interaction with other oceanic constituents than
in the circumpolar Arctic (Jaworowski 1989). These conta-
would similar contaminants in most other oceans.
minants are discharged from glaciers when icebergs calve,
4. Contaminants and introduced material which remain in
and are released to the marine environment upon melting.
or on the surface layers eventually get delivered to the
However, this period of contaminated ice accumulation (i.e.,
North Atlantic Ocean, less reduced or altered by biologi-
the last 50 years) represents a minor fraction of the total ice
cal action or chemical interaction than would be the case
in these glaciers (perhaps < 1%) so the contribution from
after the same length of time in oceans at lower latitudes.
this source is likely minor.
5. Material that reaches abyssal depth in the Arctic Ocean
appears likely to stay there for very long periods of time
due to the semiclosed nature of the basins. 3.5.2.3. Atmospheric deposition
Emissions to the atmosphere are returned to the land or the
sea as wet and dry fallout as discussed above. Atmospheric
3.5.2. Sources of contamination transport is recognized as a major route for the transfer of
Contaminants are delivered to the Arctic marine environ- contaminants to the ocean (GESAMP 1985). The time it
ment by rivers, ocean currents, atmospheric deposition, takes for a contaminant released to the atmosphere on any
domestic and industrial out falls and direct storm runoff, continent to reach northern waters ranges from days to a
dumped material, and contamination from ships. The ma- few weeks, as compared to years by means of oceanic circu-
rine environment also receives contaminants from second- lation (Gaul 1989).
ary sources, such as glaciers and snow and ice melt. In this Radioactive cesium, strontium isotopes, tritium, and
section, the most important sources for contamination of krypton 85 have been deposited from the atmosphere to
the ocean are briefly described. A more detailed description the ocean as a result of atmospheric weapons testing and
of atmospheric and riverine sources is provided in sections nuclear power plant operations in the 1950s and 1960s
3.2 and 3.3. (Rozanski 1979, Broecker et al. 1980, Smethie and Swift
1989). Atmospheric contributions of OC compounds, such
as PCBs and DDT, remain significant today. They are re-
3.5.2.1. Rivers
leased into the atmosphere in the vapor phase and are
Being nearly landlocked, the Arctic Ocean receives a variety transported as gases, aerosols, and adsorbed to particles
of materials from the surrounding continents, including (Duce et al. 1983). Eventually, the more persistent conta-
both anthropogenic and naturally-occurring constituents. minants are permanently deposited in the sea floor sedi-
Contaminants discharged to rivers will most noticeably in- ments, but only after cycling through the water, biota, se-
fluence the coastline and nearby seas. diments and ice, perhaps many times. Once introduced
Atmospheric transport
G A S E X C H A N G E
Sea birds
Deposition Volatilization Deposition Volatilization
Depth (m) Polar
bear
0
Rivers Turbid plume Flocculation Ice melt
10 Plankton Seals
Particles from
Bottom Biogenic particules
20 melting ice
resuspension Salt wedge
by waves Arctic cod
30
Walrus
40
Upwelling Narwhal
50 and Beluga
60
70
80
90
Spring / summer
100
Particle
deposition
G A S E X C H A N G E
Depth (m) Rivers

0 Landfast ice
Anchor ice Shear zone
10
Suspension B r i n e d r a i n a g e Polar
20 Strong stratification mixed
freezing
layer
30 Ice scour zone
40
50
60
70
80
90
Winter
100
Figure 3·22. Schematic representation of shelf processes involving transport of contaminants and sea ice.
into aquatic environments, synthetic organics partition to acts like a lid, keeping volatile contaminants introduced
all components of the ecosystem. The insecticide DDT, below it from entering the atmosphere as discussed for river
found in the fat of polar bears, is a good example (Ball- and lake ice in section 3.3.4.6. In summer, stratification of
schmiter and Zell 1980). the surface ocean from sea-ice melt or river runoff limits at-
In winter and spring, the Arctic atmosphere contains high mospheric exchange to the top 5-10 m. Contaminants dis-
levels of contaminants from Eurasia, and to a lesser extent charged in river water and frozen into the sea ice may not
from North America, known as Arctic haze. As contami- be released until the ice breaks up and melts. For example,
nated particles settle out of the atmosphere, sea ice acts as a oil released under the ice is degraded slowly until the ice
lid on the surface of the Arctic Ocean (Figure 3·22). Each breaks up (see chapter 10). Polynyas and leads – places
year that ice floes drift, contaminants and other materials where there are breaks in the ice cover – may represent re-
are deposited on its surface from the atmosphere in the gions for exchange of volatile compounds including OCs
form of snow, rain, fog, and dry deposition (Pfirman et al. (Barrie et al. 1992).
1995a, Chernyak et al. 1996). Heavy metals accumulating Because volatilization increases with temperature, some
in the snowcover of central Arctic sea ice can reach values chemicals deposited from the atmosphere on the ice surface
that are characteristic of snow deposits on sea ice near Si- in winter are released back to the atmosphere in the summer
berian industrial areas (Melnikov 1991). Contaminants de- when the surface warms and the snow melts (Barrie et al.
posited on sea ice by atmospheric transport could percolate 1992). If deposition occurs on drifting sea ice, release may
into the ice surface when meltwater refreezes, and could occur far from the original incorporation location.
also be added to the ice underside when meltwater runs off
and refreezes. During drift, contaminants concentrated in
3.5.2.4. Dumping, direct discharge, and accidents
the oceanic surface microlayer may also be incorporated in
the ice (Gaul 1989). Contaminants are dumped at sea in designated dumping
grounds. The 1972 London Dumping Convention regulates
Volatilization all dumping at sea. Sea dumping of radioactive solid wastes
The presence of sea ice in the Arctic Ocean generally in- has been practiced since 1946. The Kara and Barents Seas
hibits ocean-atmosphere exchange. This is important behave been the dumping grounds for radioactive waste for the
cause many contaminants, such as OCs, are semi-volatile. former Soviet Union. Both shipping and oil production con-
In winter, when surface water is being convected, the ice tribute to the input of petroleum hydrocarbons to the sea.
The introduction of radioisotopes to the seas of north- et al. 1994). Also, particle-laden river ice has a lower albedo,
western Europe from industrial installations began in 1952 and therefore melts more quickly than cleaner ice.
with discharges from Windscale Works (now Sellafield Another factor contributing to retention of river ice in the
Works) into the Irish Sea and from Cap La Hague into the nearshore zone is extensive fast ice cover. This ice is anchored
English Channel/North Sea. Municipal and industrial wastes, both to the coast and to shallow offshore banks, forming a
often untreated, are discharged directly to the marine envi- barrier to offshore transport of the ice that is released be-
ronment. Records of the chemical nature of these discharges hind it. Particle-laden river ice, as well as ice influenced by
are frequently not available. accumulation of coastal sediments, is largely retained near
Contaminants may also be introduced to the marine en- shore in this way.
vironment through accidents. Industrial ship traffic and oil Some river and shore fast ice does make its way past
exploitation, especially in icecovered areas, should be con- these barriers, survives transport across the shelf and is in-
sidered as potential sources. corporated in the large-scale drift of the Arctic ice pack. At
present, there is not enough information on this transport
pathway to actually quantify the amount of river ice that is
3.5.2.5. Geology: resuspension and mass flows
incorporated in the Arctic ice pack.
Many contaminants of concern are particle reactive, mean-
ing that they are commonly attached to mineral or organic Sea ice
material. The sedimentation of these particles transports The area covered by sea ice in the Arctic varies seasonally
contaminants to the sea floor. If accumulations of contami- (Figure 3·23), from 16 106 km2 in March to 9 106 km2 in
nated sediments are disturbed, contaminants may be resus- September (Gloersen et al. 1992). The maximum limit of the
pended, transported, and deposited elsewhere. marginal ice zone extends as far south as 50°N in the Sea of
Some contaminants have a low solubility in water and
tend to be adsorbed on particulate matter. Fine-grained
sediments, with their larger surface area, tend to have
greater concentrations of such contaminants than do Pacific
coarse-grained deposits. Ocean
Sedimentation rates in the open ocean may be a thou-
sand times less than coastal deposits. Particles transported
5
to the central parts of basins are generally associated with 6
4
atmospheric transport, deep ocean currents, ice transport
(sea ice and/or icebergs), and biological productivity. In ad- A
dition to these sources, coastal waters receive matter directly Beaufort
from terrestrial runoff and therefore usually have higher Gyre 3
2
sedimentation rates (Goldberg 1976).
Transpolar
Potentially contaminated particles may be transported Drift
off Arctic shelves via mass flows (e.g., submarine landslides, 0 1
such as slumps, slides, and turbidity currents), as well as in B Fram
bottom nepheloid layers caused by resuspension by near- Strait
bottom currents.
3.5.3. Ice Atlantic

Ocean
3.5.3.1. Ice and icebergs
River ice
Sea ice extent
The potential for contaminant delivery by river and lake Number of years for sea
Minimum Maximum 1
ice transported downstream by rivers has been discussed in ice to exit Fram Strait
section 3.3.4. Spring break-up of river ice can be a violent
event in addition to being the largest annual hydrologic
A T M O S P H E R I C D E P O S I T I O N
event. The release of water stored behind ice dams can re-
sult in massive discharges of water and ice causing gouging East Greenland Current Transpolar Drift
and erosion of river bed and overbank sediments, especi- B A
Surface sediment accumulation
ally in deltas. Sediment -laden sea ice
River ice discharged to the marine environment most Clean
sea ice
likely melts and deposits incorporated material in the near-
shore zone (Reimnitz and Bruder 1972). This happens be-
cause the main period of discharge is during June, when the
Arctic summer starts and ice in the marginal seas is just be-
ginning to melt. Where river break-up occurs before there is
much melting of the shorefast ice, river water may flow out Particle release
over the ice, depositing its sediment load on the surface

Fram Siberian
(Reimnitz and Bruder 1972). Because the presence of this Strait Eurasian Basin shelves
surficial sediment effectively lowers the albedo of the ice, it
melts rapidly in early summer. In part because of the dis- Figure 3·23. Average annual maximum and minimum sea ice extent. The
charge of river water and ice, river estuaries are usually numbered lines show the expected time in years for the ice at that loca-
centers of initial ice melting (Zubov 1943). The river water tion to exit the Arctic Ocean through the Fram Strait, based on drifting
buoy data during 1979-1990. The lower panel shows a schematic repre-
is warmer than the < 0°C shelf water, and rapidly melts sea sentation of ice growth, surface melting, and sediment accumulation dur-
ice in a region near the river mouth (Antonov 1970, Dean ing drift of a hypothetical ice floe (Source: Rigor 1992).
Okhotsk and the Labrador Sea, but also as far north as tirely. Particles, entrained within the ice during formation on
75°N in the Norwegian and Barents Seas. the shelves, eventually melt out and often accumulate on the
There are decadal and inter-decadal fluctuations in the ice surface (Figure 3·23). Zubov (1943) stated that every par-
areal sea-ice extent in the Arctic. Shpaiker and Yankina ticle frozen into the ice from below will appear on the sur-
(1971), Mysak and Manak (1989), Zakharov (1994), and face in two to three years. Because of their darker color, par-
Gloersen (1995) attribute these changes to atmospheric ticles near the ice surface absorb more solar energy and melt
pressure anomalies, variations in continental discharge, and the ice around them, forming accumulations in pits, called
El Niño/Southern Oscillation events which influence the in- cryoconite holes.
flux of Atlantic and Pacific waters. Zakharov (1994) notes Winds and ocean currents keep most of the ice in con-
that there appears to be a four-year lag between variations stant motion. As a result, the ice is broken up into floes,
in total continental discharge and ice extent in the northern with open water spaces between the floes. In the central Arc-
Eurasian Basin. tic drift ice, such leads – roughly linear regions of open wa-
ter between floes – reduce the ice cover to 80-90% in sum-
Formation and melting mer from 99% in winter (Gow and Tucker 1990).
Late summer storms are critical to the freezing process,
which depends on the temperature contrast between air and Fast ice
ocean water, as well as on the magnitude and distribution Fast ice grows seaward from a coast and remains in place
of heat in the surface layer. Storms extract heat from the throughout the winter (Figure 3·24). Typically, it is stabilized
ocean and mix the remaining heat more uniformly and to a by grounded pressure ridges at its outer edge, and therefore ex-
greater depth in the water column. New ice formation is tends to the draft limit of such ridges, usually about 20-30 m
first evident in shallow bays, where heat is quickly extracted water depth (Wadhams 1986a). Fast ice is environmentally
as the rate of heat loss to the atmosphere increases (Barrie important for several reasons: 1) it hinders and increases the
et al. 1992). risk to vessel traffic along the coast and through straits re-
During sea-ice formation, individual ice crystals develop sulting in, for example, spills; 2) contaminants in sea floor
first, forming spicules called frazil ice. The crystals coagu- sediments may be protected from redistribution under the
late to form a solid sheet, or, where agitated by waves or shelter of the ice cover (Pfirman et al. 1995); 3) freshwater
currents, a soupy surface layer called grease ice. With con- discharged under the fast ice by some rivers in winter is im-
tinued freezing, the grease ice forms nilas – thin elastic sheets pounded by the fast ice to a near shore plume (Macdonald
of ice, or, where there are waves, pancake ice. Pancake ice et al. 1995); 4) it provides a platform for migration of spe-
consists of circular pieces about 0.3-3 meters in diameter cies; and, 5) grounded fast ice disturbs the sea floor.
with raised rims on the edge produced when the ‘pancakes’
strike against one another (Parkinson et al. 1987). Polynyas
Continued cold temperatures lead to consolidation of the Polynyas are open water regions ranging in area up to thou-
ice cover. Further thickening of the cover occurs by down- sands of square kilometers. Flaw leads (also called coastal
ward growth of ice crystals, forming columnar ice under- polynyas) occur at the fast ice border where offshore winds
neath the surface layer of frazil ice. When the ice section is separate the drift ice from the pack ice (Figure 3·24). Polynyas
about 0.3 m thick, it is called first-year ice. Because sea salt and flaw leads are environmentally important for several rea-
is largely excluded during the freezing process, the salinity sons: 1) the open water provides opportunities for air/sea
of young sea ice far from river outlets is about 5 in compar- exchange of semi-volatile contaminants, such as organo-
ison with surface ocean waters which typically have salini- chlorines (Barrie et al. 1992); 2) they are a source of mois-
ties greater than 34. As sea ice ages, it continues to lose salt ture to the surrounding areas (Müller et al. 1976); 3) they
because brine that was trapped in the rapidly growing fra- typically form the locus of spring ice break-up in late-April
zil ice drains from it (Parkinson et al. 1987). to mid-May as an expansion of the flaw leads and polynyas
By the end of the winter, first-year sea ice, generally less (Barrie et al. 1992); 4) they are often locations of intense bi-
than 2 m thick, covers much of the shelf seas (Figure 3·23). ological activity where contaminants may be taken up into
In summer, this ice melts, thins, and often moves offshore. the food chain; 5) sea ice often forms there in great quanti-
If the ice survives the next summer’s melt season, it is called ties, perhaps incorporating contaminants (Pfirman et al.
second-year ice. Multi-year ice refers to ice that has sur- 1995a); and, 6) dense brines developing from sea-ice forma-
vived two or more melt seasons. Multi-year ice covers much tion may result in redistribution of contaminants to the bot-
of the central Arctic and east Greenland areas, and is often tom waters in basins (Pfirman et al. 1995a).
thicker than 3 m. Important polynyas include (Stirling and Cleator 1981,
During several years of melting and freezing, the original Parkinson et al. 1987): the North Water polynya of Smith
ice floe is modified substantially (Pfirman et al. 1990). Each Sound (northern Baffin Bay); a polynya off the Bering and
summer, all of the snow and between 32 and 70 g/cm2 of Chukchi Sea coasts of Alaska and adjacent islands; the flaw
ice melts off the surface (average is 40 g/cm2, Hanson 1965, lead off east Greenland; a polynya north of the New Siberian
Romanov 1992). Water produced by the melting snow per- Islands; the flaw leads of the Kara, Laptev, and East Siberian
colates into the floe surface, accumulates in melt ponds, Seas; a polynya around Franz Josef Land; and, a polynya in
runs off the floe, and may refreeze on the ice underside, the Beaufort Sea (Figure 3·24).
redistributing some contaminants originally located on the
ice surface. Extensive surface melting results in formation Glaciers
of meltwater ponds which may cover 25% of the ice sur- The largest glacier ice mass in the Arctic is on Greenland,
face by mid-July. where ice covers an area of 1800 000 km2, which is nearly
Some dissolved and particle-associated contaminants are 85% of the land area (Encyclopedia Britannica 1990). Along
also lost to the water column with the meltwater. Each win- the coasts of Greenland, Svalbard, Franz Josef Land, Sever-
ter, more ice is added to the underside of the floe. As a re- naya Zemlya, and the north island of Novaya Zemlya, out-
sult, while the floe thickens with time, the original first-year let glaciers terminate in the sea (Solheim and Pfirman 1985,
ice section, perhaps 1.5 m thick, eventually will melt en- Encyclopedia Britannica 1990, Romanov 1992). These glaci-
Major shore
lead polynya
Concentrations
Annual net sea of polynyas
ice flux (km2)
Shore fast ice
Net direction
of the ice flux
Maximum sea
1 000 000 278 000 150 000 78 000 28 000 ice extent
Figure 3·24. Approximate net sea ice exchange (see legend for key to arrow sizes), extent of shore fast ice and winter sea ice, and coastal polynyas and
main polynya concentrations. (Source of ice flux data: Zakharov 1976, Vinje 1987, Aagaard and Carmack 1989, Kvambekk and Vinje 1993, and
Kotchetov et al. 1994; shore-fast ice extent on Siberian shelf after Buzov 1991).
ers periodically break off, or calve, forming icebergs, many the shelf, the ice islands tend to circulate with the Beaufort
of which drift into the North Atlantic Ocean (Figure 3·25). Gyre, often completing several circuits before leaving the
Some 12 000 bergs are calved annually from tidewater gla- Arctic Ocean in the Transpolar Drift (Jeffries and Shaw
ciers on west Greenland. The most productive of these glac- 1993). Some of this tabular ice can enter the Canadian Ar-
iers is at Illulissat and it contributes an estimated 20 106 chipelago where it would eventually disintegrate and melt,
tonnes of freshwater per day. This is equivalent to a medium but in only one known example has an ice island drifted
size river with a mean annual discharge of 230 m3/s. eastward after calving (Jeffries and Shaw 1993). Icebergs
In the Canadian Arctic, tabular and glacier ice is formed ground close to shore in fjords or are melted on the shelves.
in several locations, the most important being the Ward They work their way out of fjords from July to September,
Hunt Ice Shelf off Ellesmere Island and fjords on the north when coastal regions are often free of sea ice. The role of
coast of Axel Heiberg Island (Jeffries 1987a, Jeffries et al. icebergs and ice islands in disturbing or moving sediments is
1989). The Ward Hunt Ice Shelf has been the source of a minor compared to sea-ice pressure ridges. Nevertheless, ice-
number of large tabular ice islands of deep draft (40-50 m) bergs and ice islands do pose environmental concerns, for
like T3 and, more recently, Hobson’s Choice. Between 1983 example: 1) they are a hazard to vessel traffic; 2) they can
and 1986, it is estimated that about 25 such ice islands disturb the sea floor when they ground, potentially stirring
were sighted along the coast of the Queen Elizabeth Islands up contaminants as well as disrupting dumpsites and per-
(Jeffries 1987b). Upon breaking off from the main body of haps rupturing containers (see section 3.5.3.2); 3) they can
similar to ice drift patterns (Colony et al. in press): a) a large
Beaufort Gyre (Figure 3·23), forcing the Transpolar Drift
Pacific
Ocean southward with a modest Laptev Gyre, resulting in increased
water and ice export to the Greenland Sea and export of ice
from the Siberian seas, or b) large gyres in the Chukchi Sea
and north of the Laptev Sea, with a reduced Beaufort Gyre,
resulting in recirculation of ice and expanded distribution of
multi-year floes.
Icebergs released in the Siberian seas may be advected off
Beaufort
the shelf to be transported in the Transpolar Drift across the
Gyre Arctic to the East Greenland Current in the Fram Strait. Here,
they join icebergs calved from east Greenland.
Sea ice as well as icebergs released along the west coast of
Transpolar Greenland are advected with the West Greenland Current
Drift
northward in Baffin Bay. Some ice crosses over to the west-
ern side of the bay, and returns southward along the coast of
Baffin Island. Continuing to the south, ice is advected in the
Labrador Current to the south of Newfoundland (Figure
3·25). Although many icebergs ground or melt during transit,
as many as 1400 icebergs have been documented to reach
48°N in one year (Groen 1969). The average number is 386,
and such bergs typically have lost about 90% of their mass
Atlantic
since calving along the Greenland coast.
Ocean
Residence time
Transport from the Laptev Sea to the Fram Strait typically
Sea ice extent Iceberg sources takes about three years in the Transpolar Drift (Figure 3·23).
Minimum Major Common iceberg tracks Kara Sea ice incorporated in the Transpolar Drift may exit
Maximum Minor through the Fram Strait in two years (Rigor 1992), while
transport to the Barents Sea takes less than a year. Ice in the
Figure 3·25. Main sources of icebergs and common iceberg drift trajecto- western Arctic, contributed from the Beaufort, Chukchi, and
ries. The main ice circulation pattern follows the Transpolar Drift in the East Siberian Seas (Figure 3·23), is often incorporated in the
eastern Arctic and the Beaufort Gyre in the western Arctic (white arrows)
(Source: Sugden 1982). Beaufort Gyre where it may circulate for more than five years
(Thorndike 1986).
pose a threat to bottom structures and other equipment Some ice from the Transpolar Drift may also be entrained
used for oil/gas exploration and exploitation (Sackinger et in the Beaufort Gyre and recirculate for long periods of time,
al. 1985); and, 4) they contain contaminants deposited however, the amount of exchange between the two circula-
from the atmosphere which will be released to the sea sur- tion systems is not well known.
face when they melt although we consider this to be minor
due to the fact that the bulk of the ice pre-existed the per- Fluxes
iod of major release of contaminants of concern to AMAP. The main exit for ice in the central Arctic is through the Fram
Strait (Figures 3·23 and 3·24). Each year, about 2600 km3 of
Ice drift sea ice (representing about 1 million km2) is exported through
Much of the central Arctic drift ice forms in winter in the this region in the East Greenland Current (Kvambekk and
marginal seas, especially in the wide Siberian Shelf seas, Vinje 1993).
such as the Laptev and the Kara. Ice usually forms over Within the Arctic region, the combined actions of winds
the outer shelf region, because along the coast there is a and currents cause large seasonal variations in the transport
band of fast ice anchored to the sea floor and/or the shore. of ice, both onto and off of the shelves. While Cavalieri and
A coastal polynya or a flaw lead often exists between the Martin (1994) have made synoptic assessments of sea-ice
fast ice zone and the drift ice (Figure 3·24). production in polynyas, estimates of shelf-basin ice fluxes
The two main ice circulation systems are the clockwise are much cruder, based on different years and use of varying
Beaufort Gyre in the Amerasian Arctic and the Transpolar methodology. The Laptev Sea discharges the most ice each
Drift in the Eurasian Arctic (Figure 3·25). The Transpolar year to the Arctic Ocean, from 256 000 km2/y (Rigor and
Drift transports ice primarily from the Kara, Laptev, and Colony in press) to 350 000 km2/y (Zakharov 1976). This
East Siberian Seas, toward the Fram Strait, where it exits is followed by the Kara Sea at 180 000 km2/y. Export from
the central Arctic Basin (Colony and Thorndike 1985). The the Barents Sea to the Arctic Ocean is estimated to range
Beaufort Gyre recirculates ice formed in the Beaufort, Chuk- from 17 500 to 100 000 km2/y (Zakharov 1976, Vinje 1985).
chi, and East Siberian Seas, as well as some ice from the The flux of ice (measured in km3/y) depends on the thickness
Transpolar Drift. of the ice.
Changes in ice circulation patterns may be related to There are seasonal variations in ice exchange. Both the
fluctuations in continental discharge and ice extent, as well Barents and Kara Seas export ice to the Arctic Basin in win-
as to variations in atmospheric conditions. Over approxi- ter and import ice in summer (Zakharov 1976). According
mately decadal periods, there is an out-of-phase relation- to Vinje (1987), approximately 629 km3/y of ice is exported
ship of atmospheric pressure between the Beaufort/Chukchi from the Kara to the Barents Sea through the strait between
Sea region and the Greenland Sea (Walsh and Johnson 1979, Franz Josef Land and Novaya Zemlya in winter. During
Mysak and Manak 1989). Kotchetov et al. (1994) have doc- June, July, August, and September, 72 km3 of ice are im-
umented two different ocean circulation patterns, roughly ported from the Barents Sea (Vinje 1987), for a net flux of
557 km3 (278 500 km2, assuming an ice thickness of 2 m). shore. River discharge can influence the ice pack on the shelf
Other sources indicate that the net annual flux from the in several ways. Firstly, sediments deposited from the river
Kara to the Barents through this strait could be as low as on the shelf are resuspended during fall and winter storms,
198 km3 (Pavlov et al. 1994). and are incorporated in the developing ice cover, as described
Ice exchange also occurs to the south in the Karskiye Vo- below. Secondly, the large river water discharge is primarily
rota Strait, through which there is a 98 000 km2 flow from in June and July, at which time there is still extensive ice
the Barents Sea to the Kara Sea each year from December to cover in parts of the marginal seas, and the river water (in-
April, and 21 000 km2 flow back to the Barents Sea during cluding sediment and contaminants) may flow under or over
the same months (Kuznetsov 1983). Zubakin (1987) esti- the fast ice, perhaps contaminating it. Thirdly, although
mated the annual net flux from the Barents Sea to the Kara river discharge declines in winter, it does not cease in all
Sea in this region to be about 16.8 km3/y assuming an ice rivers, so some water may be contributed to ice forming
thickness of 0.6 m (equivalent to 28 000 km2 of ice). Net along the flaw polynya during the rest of the year as well.
flux from the Kara Sea to the Laptev Sea through the Vilkit-
sky Strait is about 50 km3/y (ca. 28 000 km2 based on an Sea ice
ice thickness of 1.8 m). Contaminants deposited from the atmosphere are added to
Although much of the sea ice in the Barents Sea forms lo- the surface of sea ice during drift. Ice forming near major at-
cally, it also receives ice from the Kara Sea and the Arctic mospheric sources (such as in the Kara Sea north of Norilsk)
Ocean. More than 40% of the ice may be multi-year (Loeng or under main atmospheric transport pathways will have el-
and Vinje 1979). According to Vinje (1985), who assumed evated contaminant loads.
an average ice thickness of 2 m, the Barents Sea imports 37 Ice crystals growing in the sea exclude salt, resulting in
km3 from the Arctic Ocean, and exports 72 km3. Most of ice with a lower bulk salinity than the water from which it
the import from the Arctic Ocean is between the months of forms. In this way, soluble salts (Weeks and Ackley 1986)
April and June. Ice coring in the western Barents Sea in and some contaminants that are dissolved in the water col-
May 1989, confirmed that much of the sea ice sampled was umn are likely to be excluded from the ice (Weeks 1994).
imported from elsewhere. Pfirman et al. (1995a) suggest However, Arctic ice forming over shallow Siberian seas often
that some of the ice could have formed on the Siberian shelf includes sediments and organic material. Because many con-
in waters influenced by river discharge. Abelmann (1992), taminants of concern in the Arctic tend to sorb onto particu-
based on analysis of sea-ice diatom assemblages, also con- late material (Stumm and Morgan 1981), particle-laden ice
cluded that ice sampled east and north of Svalbard in 1987 may also be contaminant-laden (Pfirman et al. 1995a). On
probably originated in parts of the Kara or Barents Seas the other hand, sea ice without incorporated sediments or
that had some river influence. organic material, probably has less of a dissolved contami-
Import of ice into the Arctic through the Bering Strait is nant load than the water from which it grew due to the ex-
estimated at about 30 km3/y (Aagaard and Carmack 1989). clusion of salts and other impurities (Weeks 1994).
Up to 155 km3/y may advect southward through the Cana- Sea ice formed over the Siberian shelves incorporates par-
dian Archipelago (Aagaard and Carmack 1989). However, ticles predominantly during suspension freezing and frazil
because much ice in the Archipelago is formed locally, the ice formation, but also as a result of anchor ice rafting (Reim-
amount of Arctic ice exported through these passages is not nitz et al. 1992). Most particle-laden ice appears to form in
well known. water depths less than 50 m (Reimnitz et al. 1993). This is
Ice production in the Beaufort Sea has been estimated di- because the energy needed to resuspend sea floor sediments
rectly from records of ice thickness (Melling and Riedel 1995) through the water column increases as the depth increases.
and indirectly from salt budgets in the water column (Mel- Also, anchor ice growth requires that the entire water col-
ling and Moore 1995, Macdonald et al. 1995). The two meth- umn is supercooled. Another mode of sediment entrainment
ods yield similar rates of ice production from 3 to 4.5 m per is when sea-ice ridges moving over shallow areas act as sedi-
year. From ice thickness and velocity measurements, Melling ment traps, with resuspended sea floor sediments plating on
and Riedel (1996) calculated that the maximum net export the underside of the ice.
of ice from the flaw lead on the Beaufort Shelf during winter Processes associated with frazil ice formation tend to
was about 19 km3, which is only about 60% of the estimate cause elevated levels of suspended particulate matter in the
made by Cavalieri and Martin (1994). Prorating this num- water column (Kempema et al. 1989). Combined with wave
ber for the whole Beaufort Shelf, including the area west of activity and scavenging by ice crystals, the initial ice cover
Banks Island, gives a regional annual export of ice during may become enriched in particulate matter relative to normal
winter of about 60 km3. However, this estimate does not in- concentrations observed in the underlying ocean water (Ack-
clude ice advected to the interior ocean when the shelves ley 1982, Garrison et al. 1983, Reimnitz et al. 1990, Shen
clear at the end of winter. If, for example, two thirds of the and Ackermann 1990, Ackermann et al. 1994). Particles en-
ice beyond the landfast zone is advected off the shelf in spring, trained during frazil ice formation are silt-sized or smaller.
as much as 450 km3 of additional ice could be exported from The affinity of hydrophobic contaminants for particles tends
the Beaufort marginal seas to the interior ocean. to increase with decreasing grain size because of the greater
surface area of small particles. As a result, contaminant
loads of particle-laden sea ice may be elevated over that of
3.5.3.2. Contaminant incorporation in ice
the seabed, which typically contains a mixture of coarse as
River and sea ice can become contaminated from above by well as fine grain-sized material (Pfirman et al. 1995b).
deposition of atmospheric contaminants, as well as from be- Suspension freezing could also contribute to the elevated
low, for example, when contaminated sediments are entrained. levels of organic material observed in Arctic sea ice. Concen-
trations of suspended organic carbon in sea ice may be two
River ice orders of magnitude higher than in seawater (175-560 g/L
River ice may be an important localized source of contami- compared with 25-45 g/L in June to August and approxi-
nants to the nearshore zone where the ice is retained frozen mately 5 g/L during the remainder of the year) (Melnikov
or impounded into landfast ice so that it does not move off- and Pavlov 1978). According to these authors, the elevated
levels appear to result from in-freezing of organic material accumulated throughout the fall and winter are released
during ice formation on marginal Arctic seas which have during spring snowmelt and summer sea-ice decay. The melt
higher biological productivity than the central Arctic Basin. period coincides with the spring bloom of biological activity,
When ice melts in summer, some suspended organic mater- increasing the potential for biological uptake of contami-
ial is contributed to the surface seawater (Melnikov and nants (Melnikov 1991). Organisms feeding on the spring
Pavlov 1978). Dissolved organic carbon may also be en- blooms may be subjected to elevated levels of contaminants
riched within the ice cover due to adsorption. Because of released from the ice (Figure 3·22). Sea-ice transport of con-
association of many contaminants with fine-grained sedi- taminants also has the potential to influence other regions, if
ment (clay) and/or organic material, incorporation of such the ice exits the shelf and drifts within the central Arctic
material on the shelves provides a process for contaminant pack ice. In order to assess potential contaminant transport,
enrichment in sea ice formed there. detailed data are required on both import and export of ice
The formation of cryoconites during the melt season is that is formed in regions where it may entrain contaminants.
important because it concentrates the particles as well as Because of formation of cryoconites on the ice floe sur-
retains much, but not all, of the particle load on the ice surface, much of its particle load will be released when the en-
face, even when the ice meltwater runs off the floe, or the tire floe disintegrates during melting. Sediment traps on
floe is tipped or submerged during a rafting event. Therefore, moorings deployed across the Fram Strait show that the
drifting ice that originally contained dispersed contaminant- traps located in the marginal ice zone accumulated much
laden particles tends to form concentrated accumulations at more ice-rafted debris than the traps located underneath the
the surface as time progresses. Repeated melting at the sur- ice stream to the west, where there was a persistent ice cover
face and freezing at the bottom will cause a transport to the (Hebbeln and Wefer 1991). Therefore, release of contami-
surface of embedded contaminants at the same time that at- nant-laden particles is expected to be greatest along the mar-
mospheric deposition continues to add contaminants. ginal ice zone.
A mitigating factor that may be important for multi-year
Ice scouring ice is that during transport, freeze/thaw cycling tends to ag-
Another process that affects materials on the sea floor, is gregate particles into pellets on the ice surface (Barnes and
gouging by sea-ice pressure ridges and icebergs (Weeks Reimnitz 1974, Barnes et al. 1990, Goldschmidt et al. 1992).
1994). Sea-ice pressure ridges on the Siberian shelves often These pellets were also observed in sediment traps located
have drafts of 25 m (Zubov 1943), and have been docu- under the ice (Berner and Wefer 1990). Pelletization results
mented in the Eurasian Arctic extending down more than in increased sedimentation rates of the particles and their
40 m (Wadhams 1986b). While recent iceberg gouging in contaminant load, moving them out of the surface layer much
the Barents Sea by bergs calved from glaciers on Nordaust- more rapidly than if they were released as single particles.
landet (Svalbard) and Franz Josef Land have affected the Very few data are available on particle release from sea ice
sea floor to water depths of 120-130 m (Elverhøi et al. to determine the relative importance of particle aggregation.
1989), most icebergs observed today are generally less than
100 m thick (Vinje 1985). Marginal ice zone and seas
The icebergs documented by Pavlov (1993) to the east An important point is that the surface accumulation of par-
of the northern island of Novaya Zemlya, may come from ticles/contaminants is released to the sea surface when the
here, although according to Zubov (1943), most of the No- ice melts. This concept is emphasized by Pavlov and Volkov
vaya Zemlya icebergs are trapped in shallow fjords. Sever- (1993) and Pavlov (1993), who conclude that contaminated
naya Zemlya is also a source of icebergs to the Siberian drifting ice will ‘partially clean’ the area of contaminant in-
seas. Icebergs in the Laptev Sea reportedly ground in water corporation, but will also lead to contamination of surface
depths up to 183 m (Kovacs 1972). seawater in the region where ice melt occurs. Most melting,
While ice gouges may penetrate more than 5 m into the and therefore particle/contaminant release, occurs in the
sea floor, in the Barents Sea typical plough mark relief is marginal ice zone, where there is a great amount of biolo-
2-5 m deep and 10-50 m wide (Elverhøi et al. 1989). Plough gical activity in the surface waters. Here, fauna associated
marks are also common on the Greenland shelf and in the with the ice form an important link in the food web be-
Denmark Strait (S. Malmberg, pers. comm., MRI, Reykja- tween primary producers and fish, sea birds, and mammals
vik, Iceland, 1996). Presumably, the most important effects (Futsaeter et al. 1991). In particular, ice fauna are funda-
of ice gouging are: 1) the damage that it could do to waste mental to the Arctic marine food chain. If these fauna are
containers resting on the sea floor (Weeks 1994), and 2) the contaminated by pollutants in the ice, the contaminants
release of contaminants to the water column when sedi- may be passed on and accumulate in the higher trophic lev-
ments are physically reworked by the ice. Materials dumped els of the food web.
in the shallow fjords of Novaya Zemlya (Yablokov et al. The large-scale circulation of ice in the Arctic generally
1993a, 1993b) could be affected by these processes. Sedi- results in export of ice from the shelf seas, transport over the
ment transport via adhering and adfreezing to the ice mass central basin, and discharge through the Fram Strait, and to
is probably not as important for contaminant transport. a lesser degree, the Barents Sea and Canadian Archipelago.
The main exit for Arctic sea ice is through the Fram Strait.
Between 50-85% of the ice discharge consists of multi-year
3.5.3.3. Transport and release of contaminants by ice
and second-year ice (Vinje and Finnekåsa 1986), which po-
Melnikov (1991) considers processes related to sea-ice for- tentially contains accumulated contaminants. The marginal
mation, drift, and ice and snow melting to be among the ice zone extends southward from the Fram Strait, along the
main factors governing surface ocean metal concentrations eastern slope off Greenland. In winter, ice also continues
in the Arctic. Similarly, Pavlov and Volkov (1993) and Pav- around the southern tip of Greenland and extends up into
lov (1993) note that sea ice formed in the Kara Sea could Baffin Bay.
incorporate contaminants from the sea as well as those de- Ice also melts within the Arctic Basin. Modeling of ice
posited from the atmosphere and release these contami- motion (Colony and Thorndike 1985) indicates that much
nants when the ice melts. This means that contaminants of the ice which melts in shelf regions comes from the cen-
Beaufort Sea Chukchi Sea Oceanic circulation is driven by tidal forces, gravitational
90
70 forces resulting from horizontal density gradients, wind stress
50 directly on the water surface or through an intervening ice
90
70 10 sheet, the Coriolis force related to the Earth’s rotation, and
50 30 turbulent shearing. A particular force may dominate in a
10 particular geographical area. For example, wind stress is
30
most important for surface currents in the Canadian Basin
10 (Barrie et al. 1992).
Mesoscale eddies are potentially an important transport
and redistribution mechanism for water properties in the
Canadian Basin (D’Asaro 1988, Aagaard 1989). These fea-
tures, which are characteristically 20 km in diameter and
East Siberian Laptev Sea
Sea have maximum speeds of about 25 cm/s, are long-lived and
may be one of the most important ways for water properties
70
10
(including contaminants) introduced at the basin boundaries
50
90 to penetrate and mix into the interior ocean.
10 Knowledge of the amount of water transported in a cur-
30 rent is important for estimation and modeling of the total
30 70
30 transport of contaminants from one area to another. In Fig-
50 ure 3·27, the mean volume fluxes (given in Sverdrup (Svd)
= 106 m3/s) in and out of the Arctic Ocean are shown. These
fluxes provide the Arctic Ocean with a variety of anthro-
pogenic tracers, although the amount of contaminant car-
Beaufort Sea 50 Chukchi Sea
ried by the currents is still rather uncertain. Oceanic cur-
30 rents, hydrographic conditions, and physical processes in the
ocean are closely linked and are treated jointly in this chap-
10 10 ter in order to provide a complete description of the Arctic
70 Ocean and surrounding seas.
30
50
3.5.4.1. The Arctic Ocean
A simplified view of the vertical structure in the Arctic
Ocean is shown in Figure 3·28. A detailed classification of
water masses can be found in Anderson and Jones (1986),
Macdonald et al. (1989), and Carmack (1990). The Arctic
50 90 East Siberian Laptev Sea
Ocean is usually described as three layers, characterized by
70 Sea
30 different water masses and circulation patterns: the Arctic
10 10 Surface Water, Atlantic Water, and Deep Water.
10 30
50 The Arctic Surface Water
Surface water (0-200 m) is subdivided into the Polar Mixed
Layer (upper 30-50 m) and the halocline (50-200 m). The
temperature of the Polar Mixed Layer is kept close to the
freezing point because of the ice cover, whereas the salinity
exhibits seasonal and geographical fluctuations caused by
freezing and melting of sea ice and input of freshwater from
Figure 3·26. Above: contours of the asymptotic probability of ice formed the shelf seas by river runoff. The Halocline Layer consists
in the hatched region moving into different parts of the Arctic Basin. of water advected from the shelves (see section 3.5.4.6), and
Below: contours of the asymptotic probability of ice moving from differ-
ent areas into the hatched region and melting. Note that the Kara Sea is one can distinguish between Pacific Halocline Water (with
not included in this analysis (Source: Colony and Thorndike 1985). salinity 33.1, also called 33.1 water after its characteristic
salinity) originating in the Bering and Chukchi Seas, and At-
tral Arctic. For example, sea ice melting in the Beaufort Sea lantic Halocline Water (salinity 34.2) produced on the Eu-
is likely to contain some ice from the north (Figure 3·26), rasian shelves (Jones and Anderson 1986, Aagaard and Car-
which may have accumulated deposits of atmospheric con- mack 1989). Even though the mechanisms responsible for
taminants over several years during drift in the Beaufort the transfer of shelf water to the Arctic Ocean are not yet un-
Gyre. The northern Barents Sea is a particularly vulnerable derstood in detail, the signature of this transfer can be seen
area with a large amount of ice import and melting in rela- throughout the Arctic Ocean in the distribution of nutrients
tively close proximity to contaminant sources in the eastern and other tracers carried by the water masses (Aagaard et al.
Barents and Kara Seas (Pfirman et al. 1995b). 1981, Jones and Anderson 1986, Wallace et al. 1987).
The halocline is markedly stratified in salinity and den-
3.5.4. Transport pathways and hydrographic sity, and is maintained by lateral transport of water from the
shelves (Aagaard et al. 1981, Melling and Lewis 1982, Moore
conditions in the Arctic seas and Smith 1986, Schlosser et al. 1994a), where it may have
In the ocean, currents are the main transport mechanism of acquired its contaminant burden. The stratification prevents
contaminants. Once entrained in a current, contaminants winter convection and deepening of the Polar Mixed Layer
will be carried by it, rather than dispersing uniformly in all to more than about 50 m, and it effectively insulates the At-
directions. lantic Layer below from surface processes (Aagaard and
enters the Transpolar Drift and exits the central Arctic Basin
through the Fram Strait. The Transpolar Drift runs length-
wise across the Eurasian Basin from off the Siberian coast
Pacific water out through the western Fram Strait. Surface waters in the
Eurasian Basin tend to move from east to west, toward the
North Pole, following the Transpolar Drift. Mean speeds in
Beaufort Gyre 0.8 Transpolar Drift the central ocean are slow at about 2 cm/s, but increase as
the water exits the basin and becomes part of the East Green-
Precipitation land Current (Carmack 1986).
0.05 The processes described above are likely to cause varia-
tions in the storage of water in the Beaufort Gyre and shifts
1.9
in the amount of surface water exiting the Arctic either
0.04 1.7 through the Canadian Archipelago or through the Fram
2.0
Strait. The large freshwater flux from rivers into the Arctic
3.0 Ocean and subsequent transport of this freshwater into the
0.16 1.2 deep-water formation regions of the Nordic and Labrador
4.9 3.1 Seas might at least in part control the formation rate of
Atlantic water
North Atlantic Deep Water (Aagaard and Carmack 1989).
8.0 For example, when more surface water is diverted through
the Archipelago, surface waters in the Greenland Sea may
be less stable, enhancing deep-water formation. On the
other hand, Labrador Sea Water may form more readily
when there is a minimal flux of low-salinity surface water
Figures are estimated in- or outflows in Sverdrups (million m 3 per second).
through the Archipelago or advection of ice from the Nor-
Atlantic water + Intermediate dic Seas around the southern coast of Greenland (Mysak
Surface water circulation
layer, 200-1700 m
and Manak 1989).
Pacific water, 50-200 m River inflow
Tracer studies show that water discharged from Siberian
Figure 3·27. The predominant currents in the Arctic Ocean and their rivers comprises about 10% of the surface water in the main
major routes around the basin edges of the Arctic (Source: Macdonald
and Bewers 1996).
axis of the Transpolar Drift (Östlund 1993, Schlosser et al.
1995a). This means that if contaminants are conservative
Carmack 1994), since it inhibits the vertical transport of (that is, not modified by biological or chemical processes dur-
properties, including heat, from the deep water to the sur- ing transport) and introduced by rivers, they will only be di-
face (Carmack 1990, Jones and Anderson 1986), and pro- luted by about a factor of 10 by the time the river water exits
vides an effective barrier between the surface water and the Arctic through the Fram Strait. Outflow pathways for
deeper water, thus reducing contaminant transport. surface water from the Arctic Ocean (Figure 3·27) are mainly
Two main features characterize the surface water and sea- through the Fram Strait (1.0 Svd) and the Canadian Archi-
ice circulation in the Arctic Ocean: the Beaufort Gyre and pelago (1.7 Svd), but some surface water (0.4 Svd) also exits
the Transpolar Drift (Figure 3·29). The Beaufort Gyre is a through the Barents Sea (Rudels 1986a, Foldvik et al. 1988,
large clockwise gyre extending over the entire Canadian Loeng et al. 1997). Inflows of water to the Arctic Ocean Sur-
Basin. The circulation is rather slow between the pole and face Layer are through the shelf seas (Rudels 1987a, 1987b,
the Canadian Archipelago, but accelerates in the southern Blindheim 1989, Quadfasel et al. 1992) and through the Be-
and western regions. Coachman and Barnes (1961) report ring Strait (Coachman and Aagaard 1988, Coachman 1993).
velocities of 1-5 cm/s around most of the gyre and rather The shallowness of the Bering Strait (≈ 50 m) permits only
greater velocities (10 cm/s) immediately north of Alaska. surface water of Pacific origin to enter the Arctic Ocean.
From the Beaufort Sea, water is exported to Baffin Bay Mean annual flow through the Bering Strait is into the
through the Canadian Archipelago, but some water also Arctic Ocean and the mean transport is about 0.8 Svd
Atlantic Halocline
Ice Polar Mixed Layer Residence time ~10 years
A 75° 80° 85° 90° 85° 80° B
Pacific 0
Water
Pacific Halocline Atlantic
200 Residence time ~10 years Water
400
Residence time Residence time
600 ~30 years Atlantic Water ~25 years
800
1 000 Canada Basin Arctic Deep Water Eurasian Basin

Bottom Water Deep Water
2 000 Eurasian Basin Norwegian
Residence time Residence time Sea and
A Bottom Water ~75 years
~300 years Greenland
3 000 Residence time Sea Deep
~290 years Water
4 000
B
Depth Canada Basin Makarov Amundsen Nansen
(m) Basin Basin Basin
Bering Strait Alpha Ridge Lomonosov Ridge Nansen Gakkel Ridge Fram Strait
Figure 3·28. A schematic representation of the three-layer structure of the Arctic Ocean, with the Arctic Surface Layer above the Atlantic Water and
Arctic Deep Water (Aagaard and Carmack 1989). The residence time for the different water masses are also shown (Bönisch and Schlosser 1995).
West
Alaska C. Navarin C. Olyutorskiy C.
Yukon C.
Anadyr C.
Alaskan
Coastal C.
Lena C.
Beaufort New
Gyre Siberian C.
East
Taimyr C.
Transpolar West
Taimyr C.
Drift
Ob-
Yenisei C.
Saint
Anna C.
Persey C.
Yamal C.
Baffin C.
East
East East Spitz- Novozemel C.
Greenland C. bergen C.
West
Greenland C. West
Spitz-
bergen C. Bear Murman C.
Island C. Murman Pechora C.
Jan North Cape C. coastal C.
Labrador C.
Mayen C.
East
Icelandic C. Norwegian
Norwegian coastal C.
Atlantic C.
Warm current
Cold current Irminger C.
Figure 3·29. Surface currents in the Arctic region. Square boxes indicate that the denser inflowing (Atlantic and Pacific) waters are submerging under the
Polar Surface Water. The continuation of these flows can be seen in Figure 3·27.
(Coachman and Aagaard 1988). Interannual variability in the Arctic via the passages of the Canadian Archipelago (Co-
the transport is about 0.2 Svd, the seasonal variations in dispoti and Lowman 1973). The water mass called Bering
mean monthly transport values are 1-1.5 Svd in summer Sea Water in the Arctic Ocean is a mixture of Siberian Shelf
and 0.3-0.5 Svd in winter. Short-term variations in transport water (80%) and water from the Bering Sea (20%) (Coach-
have been observed to range between about 3.0 Svd north man and Barnes 1961).
and 5.0 Svd south (Coachman 1993).
The Bering Sea Water enters the Chukchi Sea as a surface The Atlantic Layer
current (Figure 3·29), but due to its higher density it intrudes In general, below the surface layer, circulation in the Arctic
under the Polar Mixed Layer and separates the Mixed Layer Ocean is counterclockwise or opposite to that of the sea ice
water from the Atlantic Halocline water. At Point Hope, the and surface waters (Figure 3·27). The flows into the Arctic
Bering Sea Water is split into two branches (Coachman et Ocean from the Atlantic Ocean are large and difficult to es-
al. 1975) with one turning northwestward to Harold Can- timate. This is partly because there are at least two main
yon (65%) and following the flow of the surface water branches into the Arctic Ocean (West Spitsbergen Current
above, the Beaufort Gyre, and thereby spreading out over and through the Barents Sea) and partly because the flows
much of the Canadian Basin. In contrast, the other branch through the Fram Strait undergo considerable recirculation
of the water (35%) enters the Barrow Canyon and follows obscuring the calculation of net import of water. When en-
the coast of Alaska as the Alaskan Coastal Current (Over- tering the Arctic Ocean through the Fram Strait, the Atlantic
land and Roach 1987). This branch exits the Arctic Ocean water of the West Spitsbergen Current submerges below the
through the Canadian Archipelago. Most of the Bering Sea Halocline Layer due to its higher density, and follows the
Water that enters the Beaufort Gyre will also eventually leave continental slope eastward. Along this path, some water re-
circulates toward the north and follows the direction of the Depth
(m)
Transpolar Drift (Coachman and Barnes 1963) while the A Potential temperature (°C) B
rest passes over the Lomonosov Ridge and enters the Cana- 0 -1 >2 -1.2 3 7
0 > 0.5
dian Basin. 0
0
Water of Atlantic origin is therefore found at intermedi- 1 000
ate depths between 200 and 900 m and is defined as saline -0.75 -1 <-1.2 -1
2 000
water with temperatures above 0°C. When Atlantic Water
enters the Arctic Ocean north of Spitsbergen and submerges
3 000
under the Surface Layer, it has a temperature well above -0.53
-0.96
-1.29
3°C and a salinity of about 35.0%. On its path around the 4 000
basins, both temperature and salinity decrease as a result of
diffusion and mixing with water masses above, below, and
alongside the current (Timofeyev 1961, Coachman and
Barnes 1963, Newton and Coachman 1974). The coldest
core temperature (0.38°C) of the Atlantic Water has been Depth
measured north of Ellesmere Island just before it exits the (m)
A Salinity (S) B
Arctic through the Fram Strait. 0 32.0 35.0
34.0
Atlantic Water from the Barents Sea can enter the Arctic 34.8
34.92
34.88
34.9
34.9 34.89
Ocean at the depth of the Atlantic Halocline and, when den- 1 000
34.92 > 34.91
sity is higher, mix with Atlantic Water or Deep Water (Midt-
34.95
tun and Loeng 1987, Anderson et al. 1994). Northeast of 2 000
34.89
the St. Anna Trough, the two branches of Atlantic Water
3 000
(that came through the Fram Strait and the Barents and Ka-
ra Seas) meet and merge before continuing eastward along
4 000
the slope toward the Laptev Sea. Here some of the Atlantic
Water turns north and flows as a broad band back toward
the Fram Strait, while a smaller fraction crosses the Lomo-
nosov Ridge. The water entering into the Canadian Basin Alaska
follows the continental slope counterclockwise around the
basin, branching off at various topographic features until it Depth
(m) Density (δ0)
reaches the Atlantic Water and flows back over the Lomono- 0
A B
26.00 27.90
sov Ridge and meets the other branch again north of Green- 27.90 28.00
28.00
land. Atlantic Water then exits the Arctic Ocean through 1 000 28.05
the Fram Strait, flowing below the cold Polar Water at the 28.08
28.08
surface as an undercurrent of the East Greenland Current 2 000
28.09
(Coachman and Barnes 1963, Aagaard 1989). Bathymetry
plays a major role in shaping the circulation patterns of At- 3 000
lantic Water and deeper waters. A sharp front in water
masses centered over the Lomonosov Ridge separates wa- 4 000
Greenland Norwegian
ters in the Eurasian Basin from the Canadian Basin (Ander- Canada Basin Eurasian Basin Basin Basin
son et al. 1994).

0 1 000 km 2 000 km 3 000 km 4 000 km 5 000 km
Alaska Lomonosov Ridge Norway
The Arctic Deep Water Layer Fram Strait Jan Mayen
Arctic deep water is divided into Canadian Basin Deep Wa-

ter and Eurasian Basin Deep Water, because of significant A
Figure 3·30. Distribution of potential tempera-
differences in the temperature and salinity found in the two
ture, salinity, and density across the Arctic Ocean
basins. Canadian Basin Deep Water is relatively warm and and the Greenland and Norwegian Seas (Source:
saline (– 0.5°C, 34.955) while Eurasian Basin Deep Water Aagaard et al. 1985).
B
is colder and fresher (– 0.95°C, 34.945). Eurasian Basin
Deep Water is subdivided into deep water and bottom wa- basins. Only the shallower, warmer, deep water of the Eura-
ter (Aagaard et al. 1981, Smethie et al. 1988). Only water sian Basin can enter the Canadian Basin (Coachman and Aa-
above the Lomonosov Ridge (sill depth about 1500 m) can gaard 1974). Higher salinities in the Canadian Basin are
be exchanged between the Canadian Basin and the Eurasian thought to be caused by the addition of brine-rich water
Basin, while between the Eurasian Basin and the Greenland from the shelves (Rudels 1987a, 1987b), or could represent
Sea, water down to a sill depth of 2600 m can be exchanged relict water derived from an earlier, more saline Arctic Ocean
through the Fram Strait. Both inflow and outflow of deep (Macdonald and Carmack 1991). The freshness of the Eura-
waters can only occur through the Fram Strait because of sian Basin is probably linked to the exchange of colder and
the shallow sill depths of the Bering Strait, the Canadian less saline deep water of the Greenland and Norwegian Seas.
Archipelago, and the Barents Sea. In the Fram Strait, deep- The nutrient content of the deep water is also higher on the
water inflow consists of Norwegian Sea Deep Water and Canadian side of the Lomonosov Ridge, due to the older age
Greenland Sea Deep Water (Swift et al. 1983, Aagaard et al. of the Canadian Basin Deep Water (Östlund et al. 1987).
1985) while Eurasian Basin Deep and Bottom Water flows The mean flow field in the interior of the Arctic Ocean ap-
out (Smethie et al. 1988, Swift and Koltermann 1988). pears to be very weak (Melling et al. 1984, Aagaard 1989).
The temperature difference between deep waters in the Current measurements near the Lomonosov Ridge on the
Canadian Basin and the Eurasian Basin, as seen in Figure Eurasian side show that speeds in the abyss are extremely
3·30, is caused by the existence of the Lomonosov Ridge low (mean 0.4 cm/s, maximum 3.8 cm/s), while currents
which prevents exchange of bottom water between the two over the crest are considerably higher, reaching 12 cm/s
(Aagaard 1981). The flow over the crest is characterized by During winter, with the formation of dense bottom water,
a series of pulses which probably represent overflow across entrained contaminants will be transferred off the shelves and
the ridge of deep water from the Eurasian Basin destined to into the interior of the Arctic Ocean. Shelf-derived contami-
sink adiabatically, and fill the Canadian Basin abyss. nants, injected mainly into the halocline (Aagaard 1994),
The principal large-scale advection in the Arctic Ocean is may spread throughout the entire Arctic Ocean (Macdonald
the subsurface boundary currents, trapped over steep topo- and Carmack 1991). Ten years after their injection, tritium
graphy which borders each of the major basins, including and helium had spread with shelf-derived waters throughout
the flanks of the Lomonosov Ridge (Aagaard 1981, Aagaard the upper Arctic Ocean (Östlund 1982, Schlosser et al. 1990).
1989). Similar to the flow of Atlantic Water, these deep and
apparently narrow currents flow around the basins in a cy- The Barents Sea
clonic (counterclockwise) sense and exit through the Fram The main circulation pattern in the Barents Sea consists of
Strait. Speed increases with depth through the Atlantic layer relatively warm Atlantic Water flowing eastward with the
and into the deep water. The boundary current is not unique- North Cape and Murman Currents in the south and cold
ly associated with a particular water mass. Brine-rich wa- Arctic waters flowing southwestward with the Bear Island,
ters flow off the shelves and mix with these boundary cur- Persey, and East Spitsbergen Currents in the north (Loeng
rents at different depths depending on density. Contami- 1991). The southern part of the Barents Sea is therefore char-
nants enter the deep Arctic Ocean mainly by the dense wa- acterized by relatively warm waters of the Coastal Water
ter flow off the shelves and can be transported over long (t > 3°C, S < 34.7) and Atlantic Water (t > 2°C, S > 35.0). The
distances relatively rapidly by these currents. Contaminant northern part of the Barents Sea is characterized by Arctic
concentrations (chlorofluoromethanes, radionuclides, etc.) Water (t < 0°C, 34.3 < S < 34.7). During summer, the Arctic
are higher in the boundary currents than in the center of the Water is covered with a thin layer, 5-20 m, of meltwater
Arctic Basin, and it seems that these currents provide a (t > 0°C, S < 34.2). The Arctic and Atlantic Waters meet and
rapid route out of the basins for newly formed deep waters mix at the oceanic Polar Front which is an important area
and associated contaminants (Schlosser et al. 1995b). for biological production (Loeng 1989, Pavlov 1993). The
One example of rapid transport with the subsurface bottom topography of this shelf sea strongly influences its
boundary current is the anomalously high 137Cs/90Sr ratio current distribution.
observed in 1979 near 1500 m over the Eurasian Basin Bottom Water (t < –1.7°C, S > 35.0) is formed over shal-
flank of the Lomonosov Ridge (Livingston et al. 1984). It low banks as a result of cooling and ice formation (see sec-
represents a signal from the Windscale nuclear fuel repro- tion 3.5.4.6), and flows down slopes to local depressions
cessing plant (now Sellafield) on the Irish Sea. This signal (Midttun 1985). This cold and saline Bottom Water will
was carried with the boundary current either from Spitsber- eventually leave the Barents Sea and enter the Arctic Ocean
gen or from the continental slope off the Barents Sea (Liv- through the strait between Svalbard and Franz Josef Land
ingston 1988) to the North Pole around the perimeter of or through Saint Anna Trough, or enter the Norwegian Sea
the Eurasian Basin (about 4000 km). With a speed of about through Bear Island Trough.
4 cm/s, which is well within the observed range, the Wind- Atlantic and Coastal Waters enter the Barents Sea from
scale-tagged water would have taken only three years to the west (Blindheim 1989), mainly between northern Nor-
travel that pathway. way and Bear Island, while Arctic Water enters between
The deep Arctic Ocean is mainly renewed from the adja- Svalbard and Franz Josef Land, as well as via the Kara Sea
cent shelves (Aagaard et al. 1985) which results in a net (Pfirman et al. 1994, Loeng et al. 1995). The main exit for
conversion of surface to deeper water (see section 3.5.4.6). water leaving the Barents Sea is between Novaya Zemlya
Measurements of the chlorofluoromethanes F-11 and F-12 and Franz Josef Land. Mean monthly volume transports
show a slight maximum near the bottom of the Eurasian through this strait, calculated from current measurements,
Basin, suggesting that high-salinity shelf water is transported show an increased outflow in winter with maximum in De-
to the ocean bottom. The F-11, F-12, temperature, and cember (3.3 Svd) and a weaker flow in summer (1.0 Svd). To
salinity measurements also indicate that Eurasian Basin the south of Novaya Zemlya, there is also an exchange of
Deep Water flows southward into the Greenland Sea along water with the Kara Sea. Between Svalbard and Franz Josef
the sea floor on the west side of the Fram Strait with a vol- Land there is only a small net outflow of water to the Arctic
ume transport of about 1.0 Svd (Smethie et al. 1988, Bö- Ocean. Along the western margin of the Barents Sea, the
nisch and Schlosser 1995), and from there into the Iceland main outflow occurs south of Bear Island. Water exchange
Sea all the way to the Denmark Strait and the Atlantic through the section Bear Island-Sørkapp (Svalbard) is negli-
Ocean (see section 3.5.4.3). gible - the Atlantic Water that enters the Storfjordrenna re-
circulates back into the Greenland Sea (Loeng et al. 1995).
The Murman Coastal Current is the northeastern exten-
3.5.4.2. The Arctic shelf seas
sion of the Norwegian Coastal Current and follows the coast
The marginal seas cover the continental shelves surrounding of Russia. Part of it enters the White Sea where the general
the Arctic Ocean. The shelf waters are relatively well mixed circulation is counterclockwise. The water exchange between
during the winter and stratified in summer during peak river the Barents Sea and the White Sea is very small. With an in-
discharge and ice melts. Shelf seas receive their contaminant flow of 0.167 Svd from the Barents Sea and an outflow of
load mainly via river discharge and atmospheric pathways 0.174 Svd from the White Sea, the net transport is 0.007 Svd
from the continents to the south. The shelf seas also provide to the Barents Sea. This value is balanced by river runoff
large open areas during summer where interactions with the into the White Sea (Uralov 1960, Altshuler et al. 1970, Po-
atmosphere can take place. Further, they are the sites of oil tanin and Korotov 1988). The freshwater inflow from rivers
development both in the Canadian and Russian sector, and to the Barents Sea is very small, 5000-6000 m3/s, of which
they support significant biological production in the spring the Pechora River runoff accounts for 90% of the total.
and summer. As a result of biological production, deposition The Barents Sea comprises a mixture of locally formed
occurs of organic material and its decomposition products, sea ice, and ice advected in from the Arctic Basin. While the
including incorporated contaminants (Wallace et al. 1987). southern part of the sea is kept free of ice by the warm North
Cape Current, locally formed and multi-year ice advected winds. Very little is known about the deep and near bottom
from the Arctic Basin cover much of the Barents Sea north currents (Timokhov 1994).
of about 74°N in winter. Beginning in December, ice growth Through these Siberian seas, vast amounts of river runoff
proceeds westward from Novaya Zemlya, so that by March enter the Arctic Ocean. The seasonal runoff cycle influences
most of the sea east of 45°E may be icecovered. Ice decay the current circulation which becomes stronger and flows
begins in May, with the ice edge retreating back to about more toward the sea when runoff is high, and becomes weaker
79-80°N by September. and flows closer to the coast when runoff is low (Pavlov et
al. 1993, Pavlov et al. 1994, Kotchetov et al. 1994). The
The Kara Sea, the Laptev Sea, and the East Siberian Sea strong continental runoff is considered to be one of the most
Inflow from the Barents Sea and the Arctic Ocean together significant features of the Kara Sea, the freshwater input to-
with the river runoff governs the water structure of the Kara tals 0.03 Svd, mainly from the Ob River (0.01 Svd) and the
Sea (Pavlov et al. 1993). In the southwestern Kara Sea, wa- Yenisey River (0.02 Svd). The average annual volume of
ter is exchanged in both directions through Karskiye Vorota river runoff to the Laptev Sea is about 0.02 Svd, most of it
(Pavlov and Pfirman 1995). The water entering from the from the Lena River, while freshwater input to the East Si-
Barents Sea submerges because of higher density under the berian Sea is smaller, less than 0.01 Svd (Pavlov et al. 1994).
surface water (t = –1.4°C, 22.0 < S < 25.0). The northern
Kara Sea is influenced by Atlantic Water entering from the The Chukchi Sea
Arctic Ocean through the deep troughs of Saint Anna and Relatively warm water enters the Chukchi Sea through the
Voronin. The southeastern Kara Sea is strongly influenced Bering Strait. One branch flows northward and then west-
by continental runoff from the rivers Ob and Yenisey. northwest in a broad stream, and enters the Arctic Ocean just
Ice formation begins in the northern Kara Sea in Septem- east of Herald Island; while a northeasterly branch narrows
ber and in the southern regions in October, persisting until into a high-speed jet-like stream closely following the Ala-
May (Pavlov et al. 1994). Fast ice developed along the coast skan coast into the Beaufort Sea (Paquette and Bourke 1981).
may remain in place until July. Flaw leads along the fast ice Water from the East Siberian Sea enters the Chukchi Sea
margin are the main areas of new ice production (Pavlov and mixes with the branch of water from the Bering Sea that
and Pfirman 1995). Typical ice thicknesses are 1.5-2 m. Ice flows northward, and this product is what is called Bering
is exported northward, into the Arctic Basin as well as into Sea Water in the Arctic Ocean (see section 3.5.4.1). The
the Barents and Laptev Seas. western and central parts of the Chukchi Sea are dominated
The surface water of the Laptev Sea and the East Siberian by this Bering Sea Water (31.0 < S < 33.5, 0°C < t < 8°C in
Sea is of low salinity (22.0 < S < 25.0) and cold (t = –1.4°C) summer), whereas the eastern part is dominated by Alaskan
and occupies the central part of these seas. To the north is Coastal Water (S < 31.0, 10°C < t < 15°C in summer), repre-
the Polar Water and to the south is the river water (only senting ‘unmixed’ water from the Bering Sea (Coachman et
present in summer). The local surface water is a mixture of al. 1975). In winter, the water column is near the freezing
these. In the deep northern part of the two seas, a tempera- point (t = –1.8°C) over the entire Chukchi Sea, and interme-
ture maximum is found between 100 and 400 m, from in- diate and deep water is formed by brine release. The deep
truding Atlantic Water. Inflowing Polar Water and Deep At- water settles to the sea floor and flows down the Barrow
lantic Water from the Arctic Ocean influence much of the Canyon. During summer, freshwater inflow (mostly from the
Laptev Sea (Kotchetov et al. 1994). In the East Siberian Sea, Yukon River) and ice melt provide a low-density surface
the eastern part is covered with Bering Sea Water in the win- layer which absorbs solar energy and only gradually mixes
ter (Pavlov et al. 1994). with the water below (Ahlnäs and Garrison 1984).
Ice begins to develop in the Laptev Sea in late September, The ice cover of the Chukchi Sea retreats in summer to
with persistent ice cover ranging from October to May (Pav- the shelf break north of the Bering Strait. It has characteris-
lov et al. 1994). Fast ice, extending over 30% of the sea, tics similar to the Arctic drift ice, but with a greater propor-
covers the regions that are less than 20-25 m deep. New ice tion of first-year ice.
forms along the persistent flaw lead at the northern margin
of the fast ice. Large amounts of ice formed in the Laptev The Beaufort Sea
Sea are exported to the Arctic Basin. As a result, the Laptev The Canadian Beaufort Shelf is the largest shelf on the North
is often called the ice factory of the Arctic. American side of the Arctic Ocean, but because it is so nar-
The East Siberian Sea has the most persistent ice cover of row it represents less than 2% of the total Arctic shelf area.
all the Siberian seas. It is completely covered with ice from Due to oil exploration in the 1970s and 1980s and the ease
October or November to June or July (Pavlov et al. 1994). of ship access, the Canadian Beaufort shelf has been the focus
Fast ice develops out to about the 25 m isobath, and reaches of many oceanographic and contaminant studies (Thomas et
thicknesses of 2 m. Although flaw leads and new ice develop al. 1990, Macdonald and Thomas 1991). The dominant ba-
in the western and eastern regions of the sea, large amounts thymetric feature is the Mackenzie Trough which separates
of ice are also imported from the Arctic Basin. In the sum- the Alaskan and Canadian shelves. Freshwater input comes
mer, multi-year ice often covers the northeastern portion of from the Mackenzie River, the fourth largest river draining
the sea. into the Arctic, discharging about 0.001 Svd, most of it (70%)
The Kara, Laptev, and East Siberian Seas all exhibit simi- between May and September (Carmack et al. 1989, Mac-
lar patterns of circulation (Figure 3·29). There is a counter- donald et al. 1995), providing a dominating influence on the
clockwise circulation with currents directed eastward along regional oceanography, sedimentology, and biogeochemistry.
the coast and westerly setting currents along the northern Water masses in the surface layer (< 200 m) are freshwater
limits where the seas adjoin the Arctic Ocean. The circula- from the Mackenzie River, icemelt water, and water of the
tion is greatly dependent upon local winds. Persistent east- Polar Mixed Layer and Pacific Halocline Layer (Macdonald
erly winds may reverse the usual easterly currents along the et al. 1989, Macdonald and Carmack 1991, Melling and
Siberian coast and cause westerly sets at the surface in any Moore 1995). In certain winters, the Mackenzie shelf pro-
of the marginal seas. The effect of winds on the subsurface duces sufficient dense water to ventilate the offshore halo-
currents depends upon the strength and duration of the cline (Melling and Lewis 1982).
Atlantic Water (t > 3°C, S > 35.0) and Polar Water (t < 0°C,
S < 34.4) (Figure 3·32). The Atlantic Water occupies the east-
ern side, whereas Polar Water is found on the western side
Floating Stamukhi Arctic of the Nordic Seas. A distinct hydrographic regime lies be-
land-fast ice zone pack ice
tween these two regions, and the general term Arctic Water
or Arctic Intermediate Water (t > 0°C, S < 34.9) is used to
Polynya distinguish the upper layer waters of this transition region
from those of more direct Polar or Atlantic origin. All three
River inflow and domains (Polar, Arctic, and Atlantic) overlie a vast layer of
impoundment cold (< 0°C), nearly isohaline (34.88-34.94) deep water.
Brine release and convection
Overviews of the Nordic Seas oceanography are given by
Swift (1986) and Hopkins (1991).
The Nordic Seas are thus a transition zone for warm and
Depth saline water on its way from the Atlantic Ocean to the Arc-
(m)
0 tic Ocean, and for cold and less saline water from the Arctic
5 10 15 30 Ocean to the Atlantic Ocean (Figures 3·32 and 3·33). The
5
25
20 Greenland-Scotland Ridge forms a barrier between the deep
30
10
31 31
waters of the Nordic Seas and the North Atlantic which con-
15 Tuktoyaktuk Cape strains the exchange of deep water.
Bathurst
20
The Norwegian Sea
25 The surface layer of the Norwegian Sea is dominated by
30 Mackenzie
River
31.6 warm, saline Atlantic Water, except for the coastal area of
Salinity
Norway where Coastal Water with low salinity and variable
35
1 2 3 4 4A 5 6 7 9 temperature is transported northward by the Norwegian
Stations Coastal Current (Figure 3·33). Norwegian Coastal Water is
Figure 3·31. A schematic diagram showing the inflow from the Macken- colder than the Atlantic Water in winter and early spring,
zie River trapped in the nearshore zone beneath the landfast ice. Beyond and warmer the rest of the year. Where these two water
the stamukhi zone, intermittent opening and refreezing in the flaw lead masses meet, a sharp surface front is formed. The lateral
produces brine which encourages mixing and convection. Numbers in
lower panel are salinity values. (Source: Macdonald and Carmack 1991). extent of Coastal Water varies during the year.
A slight salinity minimum (34.85 < S < 34.91) can be traced
Along the coast of the Beaufort Sea there is generally a all over the Norwegian Sea as a result of Arctic Intermediate
westward flow driven by the prevailing easterly winds, but Water (lower AIW in Figure 3·32) intruding between the At-
the circulation responds rapidly to changing wind condi- lantic Water and the deep water. Arctic Intermediate water
tions, such that under westerly winds the motion is east- enters the Norwegian Sea from both the Greenland and Ice-
ward. Significant flow on the inner shelf is primarily a sum- land Seas, and prohibits direct mixing of the Atlantic Water
mer phenomenon, when the shelf water is ice free. Strong and Norwegian Sea Deep Water (Blindheim 1990).
easterly winds can produce coastal upwelling which dis- The Norwegian Sea Deep Water has lately been catego-
places the Mackenzie plume offshore and replaces the brack- rized as two water masses: young and old Norwegian Sea
ish water with saline water (Giovando and Herlinveaux Deep Water. The young Norwegian Sea Deep Water forms in
1981). Seaward of about the 50 m isobath, the flow is sub- the periphery of the Greenland Gyre as a 50-50 mixture be-
stantially different: strong flow occurs throughout the year, tween Greenland Sea Deep Water and Eurasian Basin Deep
locally aligned with the isobaths. The dominant feature is
East Greenland
the Beaufort Undercurrent, a bathymetrically-steered east- Current
Spitsberg
A B
ward flow extending seaward from the 40-50 m isobaths to
West
at least the base of the continental slope (Paquette and Greenland Greenland Spitsbergen
Current
Bourke 1974, Aagaard 1984). Gyre
Norwegian
Jan Mayen
A zone of landfast ice progresses from the shore to the Current Atlantic Current
20 m isobath each winter. Between the landfast ice and the 20°W 15°W 10°W 5°W 0° 5°E 10°E 15°E
westward-moving pack ice, there is a shear zone where a
persistent lead forms. In this region, the ice cover opens and A B
refreezes several times during the winter, producing dense, Polar Arctic Surface Water Polar
Polar Atlantic Water
brine-enriched water. Ice formed in the Beaufort Sea is typi- Water
Intermediate Upper Water
Arctic
cally advected along shore to the west and may be incorpo- Water
Intermediate Lower
Arctic
rated in the westward circulating Beaufort Gyre. At the sea- Water
Intermediate
Water
ward edge of the landfast ice, a deeply ridged rubble zone
forms an effective barrier to Mackenzie River water (which Greenland Norwegian
Depth
Sea Deep Sea Deep

flows throughout the year), keeping the freshwater dammed Water Water
up near the shore under the ice as shown in Figure 3·31.
Minimum ice concentrations occur in September, when the
area near the Mackenzie Delta forms a region of open water.
3.5.4.3. The Nordic Seas

20°W 15°W 10°W 5°W 0° 5°E 10°E 15°E
The Nordic Seas include the Norwegian, Greenland, and
Figure 3·32. Schematic diagrams of the Greenland Sea showing the surface
Iceland Seas. All of the upper waters of the Nordic Seas are circulation (above) and water mass structure at a section across the cen-
supplied, either through mixture or local modifications, by tral gyre (below) (Source: Carmack 1986).
carried by the West Spitsbergen Current will either enter the
Arctic Ocean through the Fram Strait, or be carried across
the Fram Strait and enter deeper waters of the East Green-
land Current.
The Norwegian Coastal Current

Inshore of the Norwegian Atlantic Current flows the Nor-
wegian Coastal Current, which is a continuation of the Bal-
tic Current. Freshwater from the Baltic Sea mixes in the Sk-
agerrak with local water, North Sea Water, and Atlantic Wa-
ter to form the Norwegian Coastal Current (Blindheim and
Loeng 1981).
The current follows the coast of Norway into the Barents
Sea, and transports contaminants from both the Baltic and
the North Sea toward the north. For instance, radionuclides
discharged with waste water from Sellafield Works into the
sea are transported in part over the Central North Sea, through
90
Sr (Bq/m3)
25
20
15
Arctic Water Mixture of Atlantic 10
and Arctic Water
5
Atlantic Water Coastal Water o
50 E
0
N
o
o o o o o
0 10 20 30 40
80
Figure 3·33. Major surface currents in the North Atlantic Ocean. 90
Sr (Bq/m3)
0 1 2 3 4 5 6 7
0
Water, and enters the Norwegian Sea through gaps in the 50
100
mid-ocean ridge. The old Norwegian Sea Deep Water is 150
200
found in the deep basins of the Norwegian Sea and the Ice- 250
300
land Sea, and is an aged version of the mixture (Aagaard et 500
al. 1985, Swift and Koltermann 1988). 750
1 000
1 250
1 500
Atlantic flow through the Nordic Seas 2 000
The main inflow of Atlantic water into the Norwegian Sea 2 500 75
o
occurs along the continental slope west of the Shetland Is- Depth
(m)
4.1
5.2
lands and across the Iceland-Faeroe ridge north of the Fae- 5.9 4.8
roe Islands (Figure 3·33). Transport via these inflows is un- 6.3 7.4
certain, estimates have varied between 2 and 10 Svd (Tait 5.6 8.5
8.1
1957, Timofeyev 1963, summarized in Hopkins 1991).
7
From direct measurements, Gould et al. (1985) derived a
flux of 7.8 2.3 Svd, while Blindheim (1993), on a section 3.7 8.9
10 11
15
a little north of Shetland, found volume transports ranging 70
o
from 2.4 to 7.9 Svd. This section was situated north of the 6.3 11
offshoots from the Norwegian Atlantic Current into the 7
North Sea, and therefore covers the flow which continues 14
northward into the Nordic Seas. Transport was highest in 90
Sr (Bq/m3) 7
winter. Atlantic water also enters the Iceland Sea through 0
0 1 2 3 4 5 6 7 14
50
the eastern side of the Denmark Strait, as the Irminger Cur- 100 4.4
150 11
rent. This inflow is small (Kristmannsson et al. 1989), and 200 6.7 16
250
after mixing with the local waters north of Iceland its last 300 65
o
Depth 16
traces may be found in the shelf area east of Iceland (Ste- (m) 8.9
fánsson 1962). 24 16
The Norwegian Atlantic Current follows the Norwegian 3.7 26 6.3
continental slope north to about 70°, where the current bi- 3.7
15 16
furcates into the North Cape Current and the West Spits- 15 4.1
bergen Current (Figure 3·33). The West Spitsbergen Cur- 21
rent sends several branches westward on its way through 16 4.1
the Greenland Sea. Since this water is denser than the sur- 5.9
4.14.8 3.7
o
60 N
face water of the region, the Atlantic Water submerges to
intermediate depths. Much of the water of the West Spits-
bergen Current, after passing 79°N, recirculates in the Fram
Strait instead of entering the Arctic Ocean. This so-called
‘Return Atlantic Water’ eventually enters the southward
flowing water of the East Greenland Current, and could
Figure 3·34. Transport of 90Sr, originating from Sellafield, along the coast
represent as much as 70-80% of this flow (Bourke et al. of Norway. The highest values are found in the Norwegian Coastal Cur-
1988, Jónsson and Foldvik 1992). The contaminant load rent and decrease northward (Source: Kautsky 1987).
Skagerrak, and alongside the Norwegian Deep Water to- Ice extent reaches its maximum in February-March and its
ward the north (Dahlgaard et al. 1986, Kautsky 1987). minimum in August-September. The eastern portion of the
Close to the Norwegian Coast, one can follow a narrow Greenland Sea is kept free of ice by the warm waters of the
strip of water with relatively high transport of 90Sr (Figure West Spitsbergen Current.
3·34), with decreasing values westward and northward.
The ratio 137Cs : 90Sr is much higher in this area than the Polar flow through the Nordic Seas
ratio expected from atomic bomb tests, which indicates The East Greenland Current brings water from the Arctic
that the additional content of 137Cs in the water originates Ocean southward along the East Greenland coast to the At-
mainly from the Sellafield Works. Radiocesium from the lantic Ocean. This current consists of three water masses:
Sellafield Works is found mainly in the surface layer of the Arctic Surface Water (Polar Water), Atlantic Water that has
ocean (0-200 m). completed a cyclonic gyre in the Arctic Ocean (Intermediate
Atlantic Water) or has recirculated in the Fram Strait (Re-
The Greenland Sea turn Atlantic Water), and deep waters, mainly Eurasian Ba-
The most prominent feature of the Greenland Sea is the cy- sin Deep Water. Year-long moored measurements from the
clonic Greenland Sea Gyre, centered on the prime meridian East Greenland Current near 79°N show a mean southward
at about 74-75°N (Figure 3·32). The Greenland Sea Gyre transport of about 3.0 Svd in the upper 700 m, of which about
exhibits only a small variation in seasonal temperature, 1.0 Svd is Polar Water (Foldvik et al. 1988). The southward
varying between – 0.8°C and – 1.5°C, and both horizontal flux of Eurasian Basin Deep Water is about 1 Svd (Bönisch
and vertical temperature gradients are much smaller than in and Schlosser 1995).
the Atlantic Water (Quadfasel and Meincke 1987). The The East Greenland Current follows the edge of the
well-defined East Greenland Polar Front separates the Polar Greenland continental shelf as a southward boundary cur-
Water from warmer and more saline Arctic Water to the rent. Highest velocities are centered over the continental
east. There are marked changes in both temperature and slope. The main body of the current forms a surface layer
salinity across the front, which are usually identified with between 100 and 200 m thick. Seasonal variations are well
the 0°C isotherm (Paquette et al. 1985, Manley et al. 1987). pronounced, with the current being deflected offshore dur-
In addition, we have the Arctic Front between the Arctic ing summer because of runoff along the east coast of Green-
Water and the Atlantic Water. land and the surface broadening during winter.
The upper layer of the Greenland Sea is rather thin in Deep water entering the Greenland Sea through the Fram
the Greenland Sea Gyre compared to the other areas of the Strait can be separated into three different water masses
Greenland Sea. Greenland Sea Deep Water fills the basin (Aagaard et al. 1991, Rudels and Quadfasel 1991). The
from about 800 m to the sea floor in the peripheral areas densest of these water masses is found at a depth of approxi-
while in the center of the gyre it fills the water column from mately 2000 m in the Greenland Sea. Since the depth of the
50-200 m to the sea floor. Greenland Sea Deep Water is ridge between Greenland and Jan Mayen is about 1600 m,
nearly homogeneous, although deep water (–1°C < t < 0°C) the flow of the most dense water turns eastward north of the
has been distinguished from bottom water (t < –1.0°C). The ridge. Along the path, it mixes with Greenland Sea Deep
Greenland Sea Deep Water constitutes at least 85% of the Water and forms Norwegian Sea Deep Water. Part of the
total volume of water in the Greenland Sea, and is produced newly formed Norwegian Sea Deep Water will recirculate in
locally by convection during winter. Vertical mixing occurs the Greenland Sea or enter the Arctic Ocean below the West
in winter when the surface waters cools and sea ice forms, Spitsbergen Current (Smethie et al. 1988), while the rest
rejecting salt, resulting in increased surface density in the (about 0.9 0.1 Svd, Swift and Koltermann 1988) enters the
Greenland Sea Gyre (see section 3.5.4.6 and Figure 3·39). Norwegian Sea mainly through the Jan Mayen Fracture
Freshwater influx to the upper Greenland Gyre from the Zone (Sælen 1983). Some of this water eventually enters the
East Greenland Current may inhibit the formation of the North Atlantic through the Faeroe-Shetland Channel.
dense water during any given winter (Clarke et al. 1986, The two other deep-water masses pass through the
Muench 1990). This happened, for example, during the Greenland Sea as narrow cores along the East Greenland
winter of 1988-89 when convection reached just to an in- continental slope at depths of around 1000 m and 1500 m,
termediate depth of 1600 m. Hydrographic observations and enter the Western Iceland Sea. Signals of the densest
and measurements of the concentrations of chlorofluorocar- Eurasian Basin Deep Water, however, have also been seen in
bons (CFCs) have suggested that the formation of Green- the Western Iceland Sea (Malmberg et al. 1990, Aagaard et
land Sea Deep Water slowed considerably during the 1980s, al. 1991, Rudels and Quadfasel 1991). Transport between
related to a weakened convection in the Greenland Sea the Greenland Sea and the Iceland Sea through the passage
(Schlosser et al. 1991, Malmberg and Jónsson 1997). The between Greenland and Jan Mayen is about 2.5 Svd (Buch
formation of Greenland Sea Deep Water is estimated at 0.1- et al. 1996).
0.47 Svd (Schlosser et al. 1991, Bönisch and Schlosser 1995), The Jan Mayen and the East Icelandic Currents are
with the value of 0.47 Svd being the deep-water formation formed by branches of the East Greenland Current. The
rate from 1965 to 1980, and 0.1 Svd since 1980. A decrease Jan Mayen Current completes the southern rim of the coun-
in the salinity in the upper water column in the Greenland terclockwise gyre north of Jan Mayen, while the East Ice-
Sea could be the cause of this reduction, which seems to landic Current augments the counterclockwise gyre north
have started at the time of the appearance of the salinity of Iceland.
anomaly in the East Greenland Current between 1981 and
1982 (Dickson et al. 1988). The Iceland Sea
The western portion of the Greenland Sea has perennial Atlantic Water is found over the north Icelandic continental
ice cover, due to ice formed locally as well as ice advected shelf, where it interacts with shelf water (Stefánsson 1962,
from the Arctic Basin through the Fram Strait. The Jan Swift and Aagaard 1981). Polar Water is found over the
Mayen Gyre entrains some ice from the East Greenland Greenland continental shelf, but is also occasionally carried
Current into the center of the sea. This feature, known as into the Iceland Sea by the East Icelandic Current (Malm-
the ‘Odden’, is located at about 74°N (Parkinson et al. 1987). berg 1984). The Polar Front separates the Polar Water from
the Arctic Water. The location of this front depends on waters in the deeper layers. The currents in the Denmark
water transport in the Irminger Current, as well as in the Strait are strongest in the near-bottom layer over the Green-
East Icelandic Current. land slope. Current measurements indicate velocities be-
Hydrographic conditions in the Iceland Sea are highly tween 40-95 cm/s (Aagaard and Malmberg 1978, Meincke
variable because of its location near the boundary between 1983). The maximum observed velocity was 167 cm/s at
warm and cold currents. Changes in the inflow of Atlantic 100 m above the sea floor (Ross 1978). Over the Iceland
Water and variability in the extent of Polar Water lead to slope, the northward current flows at about 10 cm/s (Krist-
marked fluctuations in salinity and temperature, both in mannsson et al. 1989).
space and time (Stefánsson 1962, Swift and Aagaard 1981, The southward transport of deep water through the Den-
Malmberg 1984, Hopkins 1991, Buch et al. 1992 and 1996, mark Strait varies between 1 and 7 Svd, and these variations
Malmberg and Kristmannsson 1992). typically occur in bursts of one-day duration at intervals of
Below approximately 600 m, the deep water is mostly several days (Meincke 1983). The cause of these fluctuations
Norwegian Sea Deep Water (nearly 60% of the total vol- is not yet fully understood. A representative figure of 2.9 Svd
ume), but some Eurasian Basin Deep Water has also been is used as a mean value for the Denmark Strait overflow
detected near the Denmark Strait (Buch et al. 1992). Arctic (Dickson and Brown 1994). The overflow from the Den-
Intermediate Water accounts for 30%, and the remaining mark Strait increases rapidly by entrainment to 5.2 Svd at
10% is distributed amongst the surface water masses (Swift 160 km downstream from the sill. Its speed decreases to
and Aagaard 1981). The Arctic Intermediate Water is par- about 33 cm/s south of Dohrn Bank, and further decreases
tially formed locally in winter and comprises much of the to 25-30 cm/s from Angmagssalik to Cape Farewell. At
overflow water that enters the North Atlantic (Swift et al. Angmagssalik, the volume transport has increased to 10.7
1980). Svd by merging of the eastern deep waters and entrainment
of Labrador Sea Water. By the time the overflow current
Overflow waters reaches Cape Farewell, its transport has further increased
Below the Atlantic inflow to the Nordic Seas, intermediate to 13.3 Svd.
and deep waters flow out into the North Atlantic basin. Water masses associated with the different overflows over
Because of the ridge system in the southern part of the the Greenland-Scotland ridge are Arctic Intermediate Water
Nordic Seas, only a small amount of deep water, 10-15% and Arctic Deep Waters (Norwegian Sea Deep Water, Green-
of the circulation in the deep basins (>1500 m), will escape land Sea Deep Water, and Eurasian Basin Deep Water). The
to the North Atlantic. Flow through the deep channels characteristics of these overflow streams are that they are
(500-800 m) of the Greenland-Scotland Ridge consists fast (up to 60 cm/s long-term mean), thin (100-150 m), fol-
mainly of Arctic Intermediate Water (Swift et al. 1980, low topography with intense near-bottom shear, have maxi-
Blindheim 1990, Bönisch and Schlosser 1995), but also mum entrainment close to the source, and nonlinear deceler-
some deep waters. ation and deepening occurs downstream as a result of fric-
The Faeroe channels are the deepest connections across tional effects (Dickson and Brown 1994).
the Greenland-Scotland ridge system. They have complex Water from the Iceland-Scotland overflow undergoes
topography. Cold water (t < 3°C) leaves the Norwegian Sea stronger cross-isopycnal mixing near its origin than does
along the bottom of the Faeroe-Shetland channel and the Denmark Strait overflow. The North East Atlantic Deep
through the Faeroe Bank Channel in a relatively thin near- Water is therefore less dense, and thus is found above North
bottom flow, with mean velocities of 50 cm/s (Saunders West Atlantic Deep Water from the northern Irminger Sea
1990). Strong mixing and entrainment of North Atlantic to the continental margin of North America. The densest
Water occurs when the Overflow Water leaves the channel water does not pass southward around the Grand Banks
and starts sinking into the deep basins of the North Atlan- (Swift 1984).
tic. This is where production of North East Atlantic Deep
Water (mixing between North Atlantic Water and Overflow
3.5.4.4. The Labrador Sea, Baffin Bay, Canadian
Water) begins. The upper boundary of the North East At-
Archipelago, and Canadian inland seas
lantic Deep Water is found at about 2000 m all the way
from the Faeroe Bank Channel to the east coast of North Labrador Sea
America. The East Greenland Current, after passing through the Den-
The Iceland-Faeroe ridge, with a sill depth less than mark Strait, is joined by a branch of the Irminger Current.
500 m, is the shallowest portion of the Greenland-Scot- Intense mixing between the two water masses takes place
land ridge system. Intermittent overflow (Meincke 1983) during the subsequent southward flow (Buch 1985). Off
across the ridge joins the waters of the Faeroe Bank Chan- southeast Greenland, the current is augmented further by
nel outflow and undergoes the same modifications. How water recirculating from the North Atlantic Current. The
much North Atlantic Water is entrained into the Overflow West Greenland Current is the continuation of these water
Water is not easy to say, but it is thought that most of the masses and follows the west coast of Greenland to the north-
mixing takes place just south of the ridges. The combined ern part of Baffin Bay. Along its northward course, the cur-
southward flow of North East Atlantic Deep Water along rent branches several times westward into the Labrador Sea.
the eastern flank of the Reykjanes Ridge is estimated at Just south of the Davis Strait, the West Greenland Current
3.5-5.7 Svd (Dickson and Brown 1994, Price and O’Neil divides and sends a major branch westward, while the in-
Barringer 1994). The densest component of this water shore portion continues northward through the Davis Strait
mass remains in the northeastern basin, while the rest en- and into Baffin Bay (Figure 3·33).
ters the northwestern Atlantic through the Charlie Gibbs The hydrographic conditions in the upper layers on the
Fracture Zone. eastern side of the Labrador Sea depend on a variable mix-
The topography and current distribution in the Denmark ture of cold, low-salinity Polar Water and warmer, more
Strait is relatively simple. Cold southward currents are lo- saline Irminger Water. Beneath this upper water mass is a
cated along the Greenland slope, with Polar waters in the water mass that forms from a mixture of Irminger Water
near surface layers and the Arctic intermediate and deep and Labrador Sea Water (Lazier 1973, Clarke and Gascard
Pressure Pressure convection (see section 3.5.4.6) to as deep as 2000 m in

dbar dbar winter. Renewal occurs in the central Labrador Sea, where
Temperature (°C)
A 5.0
B the isopycnals rise to meet the sea surface because of the
0 0 -1.0 2.0 2.5 3.5 3.0 0
0
3.5 4.5 cyclonic circulation of the sea. Surface water cooled here
4.0
3.5 sinks to intermediate depths along isopycnal surfaces (Lee
and Ellett 1967, Lazier 1973, Talley and McCartney 1982,
1 000 1 000 Clarke and Gascard 1983, Gascard and Clarke 1983). At a
depth of 1500 m, the water has a temperature of about
Labrador 3.4°C and a salinity of about 34.9. Labrador Sea Water is
advected into the Irminger Sea, into the North Atlantic
2 000 3.0 2 000
Ocean, and southward along the western boundary of the
3.0
North Atlantic Ocean.
The Deep Western Boundary Current is found along the
3 000 2.5
3 000 lower continental slope. It flows north along the West
A B Greenland slope and continues around the entire Labrador
2.0
1.5
Greenland Sea until it exits north of Flemish Cap. Water masses associ-
ated with this current are the North East Atlantic Deep
0 500 1 000 km Water and the Denmark Strait Overflow Water (also called
Distance
the North West Atlantic Deep Water). The North East At-
Pressure Pressure lantic Deep Water, characterized by a salinity maximum be-
dbar
Salinity
dbar
tween values of 27.80 and 27.88, originates in the Ice-
A B
0 31 32 34.5
0 land-Scotland overflow (Clarke 1984). It circulates above
34
34.84
34.80
34.96 and seaward of the cold Denmark Strait Overflow Water
34.88
(Swift 1984, Lazier and Wright 1993).
34.86
34.86 Davis Strait sea ice formed in the fall is advected south-
1 000 1 000 ward in the Labrador Current where it is joined by floes ex-
34.88 34.84
ported from the Foxe Channel of the Hudson Strait. Car-
34.92 ried rapidly to the south, the ice reaches the eastern coast
34.90
2 000
Labrador
2 000
of Newfoundland by January-February. By March-April, it
has reached its southern extreme, approximately at the lim-
its of the current. In April, the marginal ice zone retreats
34.94
rapidly back to the north. In some years, Arctic ice of the
3 000 3 000 East Greenland Current rounds the southern tip of Green-
A B Greenland
land and makes its way northward in the West Greenland
34.92
34.90
Current.
0 500 1 000 km Baffin Bay

Distance
Baffin Bay is a relatively deep basin (> 2100 m), but com-
Figure 3·35. Temperature and salinity of the Labrador Sea. The cold munication with the Arctic Ocean and the Labrador Sea is
freshwater from the Arctic Ocean and Baffin Bay is seen above the con-
tinental shelf (Source: Stein and Wegner 1990). limited by sills (250 m in the north, 640 m in the south).
The overall circulation is cyclonic, and the two dominant
1983, Clarke 1984). Vertical distributions of temperature features of the circulation are the West Greenland Current
and salinity for the Labrador Sea are shown in Figure 3·35. and the Baffin Island Current, also called the Baffin Land
The Labrador Current (Figure 3·33) flows southward Current or the Canadian Current (Bailey 1957, Collin and
over the continental shelves and slopes of Labrador and Dunbar 1964, Muench 1971, Buch 1985, Rudels 1986b).
Newfoundland from the Hudson Strait around the Tail of The lower part of the West Greenland Current shows
the Grand Banks of Newfoundland and the coast of Nova clear influence of Atlantic Water as it enters Baffin Bay,
Scotia. The Labrador Current is comprised of water from with a temperature maximum at about 500 m, but it expe-
the Hudson Strait, the Baffin Island Current, and the branch riences a rapid temperature decrease as it moves northward
of the West Greenland Current that turns westward in the along the Greenland coast. In the northern parts of Baffin
Davis Strait. The Labrador Current transports 3.8 0.9 Svd, Bay, the influence of Polar Water flowing southward from
of which 85% occurs over the slope between the 400 and the Nares Strait is strong. When the ‘Atlantic’ layer exits
1200 m isobath and 15% occurs over the shelf (Lazier and through the Davis Strait (as part of the Baffin Island Cur-
Wright 1993). rent), it might contain as little as 1/6 water from the south
The upper layers of the western part of the Labrador Sea and as much as 5/6 water from the Arctic Ocean (Rudels
consist of cold, low-salinity water from the Hudson Strait 1986b).
and the Baffin Bay, as evidenced by temperatures as low as The Baffin Island Current, flowing south along the east
–1.5°C and salinities less than 34.0. This water often ex- coast of Canada, is much stronger and more voluminous
tends offshore to form the surface layer over most of the than the West Greenland Current. The reason for this is
sea (Lazier 1973). Offshore of the Labrador Current, one that the Baffin Island Current is augmented by inflowing
can frequently find Irminger Water at depths of 200 to currents from Smith Sound, Jones Sound, and Lancaster
1000 m that has been carried all the way around the Labra- Sound. South of the sounds, the Baffin Island Current
dor Sea. continues southward along the east coast of Baffin Island,
An important feature at intermediate depth is Labrador through the Davis Strait, and into the Labrador Sea, where
Sea Water, a low-salinity water mass, formed as a result of it becomes part of the Labrador Current. The Baffin Island
mixing and cooling surface water masses transported into Current generally has higher speeds toward the south, as it
this region. This intermediate water mass is renewed via accumulates water from the various inflows.
Three water masses can be distinguished in Baffin Bay: and ultimately passes through the Hudson Strait and into
the Labrador Sea. Water flowing through Jones and Lan-
1. the surface water (upper 200 m), characterized by cold
caster Sounds has a much higher Bering Sea Water compo-
temperatures of about –1.6°C and a salinity of about
nent than water entering Baffin Bay through Smith Sound
33.7; in summer a layer of fresher (30-32) and warmer
(Codispoti and Owens 1975, Jones and Coote 1980).
(2-5°C) water covers the surface. Winter convection is
An important feature of this region is the high velocity
thought to be limited to this layer, because of a strong
currents, especially in some passages, that run through the
thermocline below;
Arctic Islands. Maximum speeds can be more than 2.0 m/s
2. the Atlantic Water, which is warm (t > 0°C) and saline
and the energy available in these currents is sufficient to
(34.5) with a temperature maximum at about 500 m
bring warmer, more saline water from below the sill depth
(0.5-1.5°C); and,
up to the surface. The result is a reduced rate of sea-ice for-
3. the Baffin Bay Deep Water (below 1400 m), which is a
mation. Enhanced heat flow and mechanical removal of ice
very uniform water mass with a temperature of – 0.44°C
by these strong currents are in many places responsible for
and salinity of 34.4.
large open-water areas, polynyas, that are so common in
In the central portion of Baffin Bay, a depression (> 2200 m this area and which often remain ice-free even in midwinter
deep) is filled with Baffin Bay Bottom Water. The deep and (Topham et al. 1983, Melling et al. 1984).
bottom waters of Baffin Bay originate in the Arctic Ocean The new ice cover in the Archipelago has generally devel-
(from depths of about 250 m) and enter Baffin Bay mainly oped by the middle of October. Ice within the channels of
through the Nares Strait. Because of higher density, these the Archipelago is shorefast, typically reaching thicknesses
Arctic Ocean waters sink to the bottom (Bailey 1956, Tcher- of 1.7 to 2.3 m. It breaks up by the middle of July. The break-
nia 1980, Rudels 1986b, Bourke and Paquette 1991). up of sea ice proceeds into the Archipelago from Amundsen
The sea ice of this region is comprised mostly of locally Gulf. Break-up is not always complete, allowing multi-year
generated, first-year ice, with a small amount of multi-year ice to develop, in some areas reaching great thicknesses.
Arctic ice transported southward through the Nares Strait. Once break-up has occurred, there is a general easterly drift
of the locally formed ice, together with some ice advected
Nares Strait from the Arctic through the M’Clure and other straits.
The Nares Strait lies between the Canadian Archipelago
and Greenland. It connects the Smith Sound of Baffin Bay Foxe Basin
with the Lincoln Sea of the Arctic Ocean. The water in the Foxe Basin is a shallow inland sea with an average depth of
Nares Strait is advected almost unchanged from the Lincoln less than 100 m. Water from the Arctic Ocean enters Foxe
Sea. There are seasonal changes in the water structure due Basin through the Fury and Hecla Strait (Figure 3·36). Out-
to freezing and heating, but these are confined to the near- side the strait, the Arctic Water is stratified, but intense mix-
surface layer of 50 m. The most obvious feature of the wa- ing takes place as it passes the sill of the Fury and Hecla Strait
ter structure is a layer of cold water with a temperature and the water enters Foxe Basin as a homogeneous water
minimum (t = –1.4°C, 32 < S < 33.5) at a depth of about 75 mass. Very high current speeds, up to 3 m/s, have been ob-
m. Below the sill depth (250 m), the water is warmer (about served in the strait (Collin and Dunbar 1964, Sadler 1982).
0°C) and more saline (34.0 < S < 34.8) (Sadler 1976). The inflow is rather small, 0.04 Svd in the winter (Sadler 1982)
Canadian Archipelago Baffin

The Canadian Archipelago lies on the extensive polar conti- Bay
nental shelf of North America and constitutes a network of
shallow channels that forms a transition zone between the
waters of the Arctic Ocean and Baffin Bay. The water which
flows through the many straits and sounds of the Canadian
Foxe
Archipelago originates mainly in the surface layer of the Basin
Arctic Ocean, the net transport being from the Arctic Ocean.
The volume transport through the different channels is not
confidently known, but is thought to be about 1.7 Svd (Fissel Hudson
et al. 1988). These flows are likely to be important to the Strait
freshwater and contaminant budgets of the Arctic Ocean

since they consist mostly of Polar Surface Water. About half
the Polar Surface Water which ultimately discharges into
the Atlantic Ocean passes through these channels (Stige-
brandt 1981). Hudson
Bay
In the western part of the Archipelago, the water is iden-
tical in physical properties to that of the Canadian Basin,
the upper layer being characterized by Polar Surface Water
and the lower layer with Atlantic Water. Moving eastward,
the halocline is warmed up as a result of heat diffusion
from the underlying Atlantic water, and the Atlantic Water
found in the different basins of the western Archipelago is
therefore colder than Atlantic Water with the same salinity
in the Canadian Basin (Melling et al. 1984).
The water leaves the Archipelago by three main exits.
The two largest, Jones Sound (150 m 12 km) and Lan-
caster Sound (130 m 55 km) empty into Baffin Bay, while
the smaller, Fury and Hecla Strait, empties into Foxe Basin Figure 3·36. Surface currents in Hudson Bay, Hudson Strait and Foxe Basin.
and 0.1 Svd in the summer (Barber 1965), but even so, it adjacent seas is limited by the sill depths, the deepest (185 m)
fills the northern half of the Foxe Basin in one year, and has being the one between Hudson Bay and the Hudson Strait
a substantial influence on the circulation in the basin (Prin- (Prinsenberg 1986a).
senberg 1986a). Hudson Bay is almost completely covered with ice during
Arctic Water is lighter than the water in Foxe Basin and the winter, and is completely free of ice in the summer. With
flows southward, on top of it, along the coast of Melville the exception of some influx through Foxe Basin, the ice is
Peninsula. Northeast of Southampton Island, a branch moves local. It begins to form in November, and has almost com-
westward into the Frozen Strait, through Roes Welcome pletely covered the bay by December. The bay remains cov-
Sound and into the Hudson Bay. A second branch turns ered until at least May, when break-up begins. By June, the
northeasterly and joins water entering Foxe Basin from the ice retreats from the northeastern and northwestern coast-
Hudson Strait. Both branches enter a counterclockwise gyre lines, but it remains in the southwestern portion through
in the area south of Prince Charles Island. A third branch July, not disappearing totally until August.
continues through Foxe Channel and into the Hudson Strait. James Bay is extremely shallow, with only a few locations
In the eastern and central part of the Foxe Basin, the having depths over 50 m. In summer, the water in James Bay
large tides (up to 8 m) result in a well mixed and essentially is warmer and fresher than Hudson Bay water, and is there-
homogeneous water mass from top to bottom; whereas on fore easily identified as it flows northward along the west
the western side, the water is stratified (Campbell 1964, coast of Quebec. The large runoff rates (0.01 Svd) of the James
Prinsenberg 1986a). Seasonal changes in surface tempera- Bay region account for 61% of the freshwater brought into
ture are very small, from –1.7°C to about 3.0°C. Surface Hudson Bay (Prinsenberg 1986b). In James Bay, a cyclonic
salinity can occasionally fall far below 20.0, reflecting dilu- gyre exists, partly caused by wind stress and partly by fresh-
tion by meltwater and river runoff. By late summer and water runoff. Water from James Bay enters the Hudson Bay
early autumn, surface salinities are back to a level of 29.0- with velocities above 15 cm/s (Prinsenberg 1986c).
32.0. During winter, the density of the surface water in- The main hydrographic feature of the Hudson Strait is a
creases due to cooling and brine release. The cold, saline marked across-channel gradient in the surface temperature
bottom water (t = –1.97°C, S = 34.07) in deeper parts of and salinity, with higher temperature and lower salinity in
Foxe Basin and Foxe Channel is a result of dense water for- the south (Drinkwater 1986). The southeastward flow in the
mation over shallow shelves in the eastern basin during Hudson Strait originates in the outflows from Foxe Basin
winter (Campbell 1964). and Hudson Bay (Figure 3·36). As these waters proceed to-
A sill (185 m) separating Hudson Bay and Foxe Channel ward the east, they tend to be confined to a thinner surface
prevents the inflow of this cold, saline bottom water into layer and finally lose most of their identity as they reach the
Hudson Bay, except for intermittent occasions when this waters of the open ocean (Collin and Dunbar 1964). At the
water mass reaches a depth of 185 m or less (Campbell northern side of the strait, water from the Baffin Island Cur-
1964). Even so, the deep, homogeneous bottom layer in rent flows northwestward (Iselin 1927, LeBlond 1980), send-
Hudson Bay is assumed to be a result of intermittent over- ing out branches to the southern side that join the southeast-
flow, rather than the formation of dense water inside Hud- ern outflowing current. Water from Hudson Bay is modified
son Bay (Prinsenberg 1986a). through mixing with colder surface water from Foxe Basin
and some colder, higher-salinity deep water in the strait.
Hudson Bay, James Bay, and Hudson Strait When this water approaches the eastern entrance to the Hud-
Hudson Bay, which is a broad, shallow, semi-enclosed sea, son Strait, it undergoes intense vertical mixing and further
lies almost entirely south of the Arctic Circle. It does not in- modification.
teract directly with the Arctic Ocean; rather, it is externally
forced by the Atlantic Ocean via the Hudson Strait, and the
3.5.4.5. The Bering Sea
waters within the bay are modified by substantial runoff
(Prinsenberg 1991). The Bering Sea is the northernmost extension of the Pacific
Hudson Bay has an average depth of 125 m. Water col- Ocean, separated from the Pacific by the Aleutian Islands,
umn properties vary throughout the year because of melt- which form a partial barrier to water exchange between the
ing, freezing, and river runoff. Surface cooling and wind North Pacific Ocean and the Bering Sea. The water entering
mixing cause the seasonal pycnocline to deepen from 10 m the Bering Sea from the south is Alaskan Stream Water. The
in the summer to 40 m in the late fall. During winter, salt is Alaskan Stream is the northern boundary current of the Pa-
rejected from the developing sea-ice cover and the pycno- cific subarctic gyre and extends from the head of the Gulf of
cline deepens even more, to about 90 m. Alaska to the western Aleutian Islands (Figure 3·37). The
Circulation in Hudson Bay (Figure 3·36) is cyclonic with Alaskan Stream flows along the Aleutian Islands in a nar-
southward flow on the west side and northward flow on the row, high-speed and very stable westward current, generally
east side (Hachey 1935, Barber 1967). No return loop has seaward of the 1000 m isobath (Stabeno and Reed 1991).
been observed in the northern part of the bay, rather the Exchanges to the north (and south) are small, relative to the
surface water exits into the Hudson Strait (Prinsenberg total volume transport (15-20 Svd) of the Alaskan Stream
1983). Watershed contribution to Hudson Bay from its (Reed 1984). The maximum sill depths of the major openings
shores is large (0.02 Svd). As this water flows over the sur- are 430 m (Amukta Pass), 1155 m (Amchitka Pass), 2000 m
face of the bay, it is deflected to the right by the Coriolis (Near Strait), and 4420 m (Kamchatka Strait) (Favorite 1967).
force and adds to the cyclonic circulation (Prinsenberg The easternmost pass through which significant northward
1986b). The surface layer of Hudson Bay is strongly influ- transport (>1 Svd) of Alaskan Stream Water occurs is Amchit-
enced by the river runoff - salinities lie between 24.0 and ka Pass (Reed 1990, Stabeno and Reed 1994), while water
30.0 in summer. Surface temperatures in summer are be- transport through the passes east of 180° are intermittent
tween 5 and 9°C. and limited. The major inflow occurs through the Near Strait
Surface water entering from Foxe Basin and the Hudson (10 Svd), but is variable. In 1991, there was essentially no in-
Strait sinks below the surface layer in the Hudson Bay be- flow of Alaskan Stream Water (Stabeno and Reed 1992), but
cause of its higher density. Deep water exchange with the in 1992, there was a well developed northward flow (Reed
AG side of the Aleutian Islands. Inflows through the other passes

BSW /
NS join this flow eastward. The Bering Slope Current (5 Svd)
NS follows the continental slope along the Aleutian basin up to
IFA Cape Navarin where it bifurcates; a smaller part (about 10%)
West
Alaska C. of the flow follows the 60-70 m isobath around the Gulf of
BSW / CA
CW / BSW Anadyr, and the rest turns south and joins the Kamchatka
a C.
atk Current (Paluskiewicz and Niebauer 1984, Royer and Emery
ch Bering CA
1984, Kinder et al. 1986). The Kamchatka Current forms
Kam
Slope C. CW
BSW the western boundary current of the Bering Sea Gyre. Out-
Co
mm AS flow from the Bering Sea to the Pacific Ocean is confined to
ande AS/CW the Kamchatka Strait. The latest measurement of the outflow
r C.
from the Bering Sea through the Kamchatka Strait ranges
IFA from 6 Svd in 1991 (Stabeno and Reed 1992) to 12 Svd in
1990 (Verkhunov and Tkachenko 1992).
BSW The vertical density structure of the southwestern Bering
Alaskan Stream Sea can be represented by a surface layer, about 200 m thick,
a deeper transition layer near 500 m, and a bottom layer.
BSW: Bering Sea Water
IFA: Ice forming area
CW: Coastal Water
AG: Anadyr Gulf Water
The deep Bering Sea is a vast plain lying at a depth of 3800
AS: Alaskan Stream NS: Norton Sound Water m with occasional hollows up to 4150 m deep (Tsunogai et
CA: Convective area
al. 1979). The Bering Sea Deep Water has the highest silica
Figure 3·37. Schematic diagram of circulation and the extent of different concentration in the world ocean (Mantyla and Reid 1983).
water masses in the Bering Sea (Source: Takenouti and Ohtani 1974). The Bering Sea Shelf is very broad (500 km), unusually
flat, and featureless. The shelf is divided into three regimes
and Stabeno 1993). All deep-water inflow to the Bering Sea (Kinder and Schumacher 1981): 1) the inner shelf (< 50 m);
has to take place through the Kamchatka Strait. 2) the central or middle shelf (50-100 m); and, 3) the outer
The Bering Sea has a vast continental shelf that occupies shelf (> 100 m). The shelf break occurs near the 200 m isobath.
about 44% of the total area – 43% of the shelf is abyssal On the southeastern shelf there are two water masses:
with water depths generally > 3000 m, and 13% is conti- Alaskan Coastal Water and Central Shelf Water (Coachman
nental slope (Hood and Kelley 1974). Bering Sea Water is 1986). The Alaskan Coastal Water (CW in Figure 3·37) is a
comprised of Pacific Ocean water that enters through the mixture of seawater and freshwater runoff from land, formed
Aleutian passes and is modified by local processes. Figure in the coastal zone from the shore to the 50 m isobath. It
3·37 is a schematic diagram of circulation and the extent of tends to be vertically homogeneous because of mixing by
water masses and their characteristics. tidal currents and wind (as shown in Figure 3·38), except
Circulation within the Bering Sea basin is generally cy- near river mouths where the freshwater input is sufficient to
clonic. Most of the flow through the Near Strait turns east- create a two layered structure. Estimates of the mean annual
ward and follows the Commander Current along the north runoff are about 8.0-10.0 103 m3/s, with a minimum in
Shelf break Middle Inner
front front front
500 km
Oceanic Outer Middle Coastal
domain domain domain domain
0 0
50 50
Mixing energy : Generalized fluxes :

100 100
wind salt
150 salinity salinity tidal freshwater 150

tempera tempera
ture ture
salinity
200 tempera
ture
200
250 250
300 Depth 300

Depth Depth
350 350
Depth Depth
(m) (m)
Figure 3·38. The different domains of the Bering Sea continental shelf are separated by fronts. The coastal domain (depth <50 m) tends to be vertically
homogeneous, the middle domain (50-100 m) shows a clear two-layered structure, while the outer domain (100-170 m) is dominated by lateral interac-
tion between the shelf water and the Bering Sea Water (Source: Coachman and Walsh 1981).
spring and maximum in early summer (Roden 1967, Coach- winds are the main cause of these seasonal variations. When
man et al. 1975, Overland 1981). strong winds (> 8 m/s) from the northeast or east prevail,
The Central Shelf Water is formed between the 50 m and which is mostly a fall phenomenon, the flow through the
100 m isobaths from a mixture of seawater and freshwater Bering Strait reverses (Coachman and Aagaard 1988). Most
(from ice melting and precipitation). It has a tendency to- of the water that finally crosses the Bering Strait originates
ward a two-layered structure (Coachman and Walsh 1981, from the branch of the Bering Slope Current that enters the
Kinder and Schumacher 1981, Muench 1983) with a sur- Gulf of Anadyr (Anadyr Current) before going through the
face layer 10-40 m thick overlying a relatively homogene- Anadyr Strait or the Shpanberg Strait (Coachman et al. 1975).
ous deep layer (Figure 3·38). In winter, the shallower part Small amounts of ice may also be imported into the Arc-
of this area (< 70 m) becomes a cold, homogeneous water tic Ocean through the Bering Strait, but this can be safely ig-
column because of cooling and wind mixing. In spring, nored in first-order budgets, since it is much smaller that the
when ice starts to melt and the winter storms cease, the errors in the other terms. The calculations of property trans-
two-layered structure is re-established. Horizontal gradients port into the Arctic Ocean from the Pacific Ocean are rela-
in both salinity and temperature are small. The Central Shelf tively straightforward, provided contaminant data are avail-
Water is found in the convective areas (CA) and ice-forma- able, and can probably be better constrained than any of the
tion areas (IFA) (Figure 3·37). A strong inner front sepa- other ocean exchanges.
rates the Central Shelf Water from the Alaskan Coastal Wa- The flow of Bering Sea Water through the Bering Strait is
ter (Coachman and Walsh 1981) and an outer front also in- less than 5% of the total inflow into the Bering Sea and is
hibits lateral fluxes from the west. not considered an important factor in the water budget. How-
The outer continental shelf (>100 m isobath) is a zone of ever, this flow does permit the exit of a considerable amount
lateral water mass interaction (Figure 3·37) between Central of the freshwater runoff along the coast of the eastern Bering
Shelf Water and Bering Sea Water (BSW). A northwest current Sea, as well as some of the dilute water found off the coast
(1-10 cm/s) exists over the outer shelf, while the flow over the after the sea ice over the continental shelf melts in late spring.
middle shelf is insignificant (Schumacher and Kinder 1983). Contaminants delivered by rivers to the shelf will therefore
Mean salinity in July ranges from 31% near the Alaskan eventually end up in the Arctic Ocean.
coast to 33 at the 200 m isobath (Kinder et al. 1986). Be- The sea ice in the Bering Sea is comprised mostly of local-
cause of the low salinities in the surface layer on the conti- ly generated ice, although there is some imported ice south-
nental shelf, the depth of vertical convection due to ice for- ward through the Bering Strait from the Chukchi Sea. The
mation is only about 50 m. Particularly in the Gulf of Ana- first-year ice is thinner than on many of the northern shelf
dyr (Figure 3·37), where oceanic water enters at a lower seas, typically reaching a thickness of 20-40 cm. Formation
level, convection is restricted to the upper 30 m, and no occurs primarily in the northern Bering Sea, especially in
cold high-salinity water is found at the bottom (Takenouti coastal polynyas in Norton Sound and to the south of St.
and Ohtani 1974). Lawrence Island. The ice then circulates to the southwest in
Over the continental shelf, the West Alaska Current fol- the Bering Gyre, where it moves toward the marginal ice
lows the coastline from north of the Alaska Peninsula to the zone and is destroyed by melting. The marginal ice zone is
Bering Strait, entering the Arctic Ocean on the eastern side of located at the shelf break of the southeastern Bering Sea in
the strait. The current is influenced by water transports from winter, and retreats rapidly back to the north in mid-May.
the local rivers: the Kvichak, Nushagak, and Kuskokwim Riv-
ers, and most of all by the Yukon River farther north. Any
3.5.4.6. Vertical mixing
contaminant load brought by these rivers to the sea enters
the West Alaskan Current, and is thereby transported to the In the Arctic, near-surface seasonal variations can be ac-
north and into the Arctic Ocean. The current appears to be counted for by a combination of winter cooling, freezing,
weak (only a few cm/s), at least in the southern part (Kinder and convective mixing, and summer meltwater addition.
et al. 1986), transporting about 0.1 Svd to the Bering Strait. Different mechanisms for vertical mixing or a combination
The northern Bering Shelf (also called the Chirikov Ba- of these may result in formation of denser water masses. The
sin) lies north of St. Lawrence Island and south of the Be- density of the water produced determines to what depth the
ring Strait. Two water masses, Norton Sound water (NS) water sinks. Especially important is the process of deep con-
and Gulf of Anadyr water (AG), that are warm and have vection which helps ventilate the deep basins. Convection is
low salinity, are formed in the upper 10 m during summer, a result of unstable distribution of density in the water col-
overlying a bottom layer of colder and more saline water umn with heavy water above lighter water (Rudels 1993).
(Figure 3·37). This situation can arise from cooling and by evaporation or
Circulation in the Chirikov Basin is dominated by a freezing, which increase surface salt concentration.
northward net water transport into the Arctic Ocean. This There are two main types of deep convection (Killworth
flow is driven by a sea surface slope down to the north which 1983), shown schematically in Figure 3·39. The first is con-
results from the sea level difference between the North Pa- vection near an ocean boundary (shelf convection), and in-
cific and Arctic Ocean (Stigebrandt 1984, Aagaard et al. volves the formation of a dense water mass on the shelf,
1985, Overland and Roach 1987). The regional circulation which then descends down a continental slope and into the
is therefore normally northward in the Anadyr, Bering, and deep basin of the ocean (brine formation, left hand side of
Shpanberg Straits. Mean currents within these straits are the figure). The second is open ocean deep convection (mid-
about 25 cm/s in the Bering Strait, 15 cm/s within the Ana- gyre convection), occurring far from land, and especially in
dyr Strait, and 5 cm/s in the Shpanberg Strait (Aagaard et regions with a shallow thermocline or halocline (i.e., in the
al. 1985). Strong northerly winds coincide in general to center of cyclonic eddies). When the upper layer density is
weaken the northward currents. increased, local open ocean convection chimneys may be
The flow through the Bering Strait to the Arctic Ocean is formed (Killworth 1979).
subject to an annual cycle, with maximum transport (1.0-1.5 Cooling and the release of salt into the water from the
Svd) during the summer months and minimum transport (0.3- generation of sea ice are the driving factors in the process of
0.5 Sv) during the winter months (Coachman 1993). The vertical convection and formation of dense water in the Arc-
Russia
Barents
Sea
Chukchi
Sea
A
Iceland
Greenland
Baffin Bay
Atlantic
Ocean
Canada
B
Brine Mid-Gyre Mid-Gyre

A Formation Convection Convection B
0m
Surface Waters
σ 0 =27.9
Intermediate Waters
σ 1 =32.8 Atlantic
Iceland
Sea Ocean
Depth
Deep Waters
Canadian σ 2 =37.5 North
Basin Greenland
Sea
Eurasian
Basin
Bering Denmark
Strait Lomonosov Fram
Ridge Strait Strait
4 000 m
Greenland Gyre
Arctic Ocean Mid Gyre

outflow convection
Mixing
Figure 3·39. Circulation and water mass structure in the Arctic Ocean
and Nordic Seas. Mixing processes, such as brine formation, result in
denser water that is transported off the shelves and into the deep basin.
Open ocean convection takes place in the gyres of the Greenland and
Iceland Seas (Source: Aagaard et al. 1985).
tic (Killworth 1983, Carmack 1986, Carmack 1990, Rudels a wide range of densities is produced in this manner, depending
1993). The first type of convection takes place on the conti- on initial salinity, bathymetry, ice cover, and wind condition.
nental shelves surrounding the Arctic Ocean, whereas the On shelves with complicated topography, such as the Bar-
second type of convection is found in the Greenland, Ice- ents and Kara Seas, it is likely that bottom-water accumu-
land, and Labrador Seas. Because ice forms on open water lates in hollows and enclosed depressions (Midttun 1985).
at a greater rate than under existing ice, persistent diver- When the accumulation is sufficient to overtop and overflow
gence of the ice cover maximizes the formation of dense shelf depressions, and where shelf troughs extend across the
water (Melling et al. 1984). shelf to the edge, the bottom water runs off and sinks to its
density level. The flow from the shelves is thought to be both
Shelf convection in the Arctic localized and intermittent.
The waters on the shelves surrounding the Arctic Ocean are The higher the surface salinity, the less brine has to be in-
normally stratified during summer due to ice melt, freshwa- troduced to achieve the density required to produce bottom
ter inflow from rivers, and precipitation. During winter this water. In order to maximize bottom-water production, it is
stratification is broken down as a result of freezing and brine therefore important that the upper layers are preconditioned
release, and vertical mixing takes place (Aagaard 1994). If by removal of freshwater components prior to freezing, and
the water column becomes homogeneous down to the sea that inflow of river water during winter is blocked (Macdon-
floor, further brine release will result in a movement of the ald et al. 1989, Macdonald and Carmack 1991, Omstedt et
densest water (called bottom water) toward the edge and al. 1994, Melling and Moore 1995).
down the continental slope. Eventually these dense water Bathymetry is of great importance for this process, since
plumes detach from the slope and move into the interior of vertical convection reaches the bottom earlier in shallow
the Arctic Ocean at the appropriate density levels (Aagaard areas. Here the water column quickly homogenizes, and bot-
et al. 1985, Rudels 1993). Water at freezing temperature with tom-water formation starts.
The strong winds that are typical for the autumn, to- as the lower halocline originates in the Barents and Kara Seas,
gether with reduced river inflows, can result in well-mixed, but winter convection in the interior of the Arctic Ocean can
relatively saline nearshore water (Carmack et al. 1989) be- also be of importance in forming the Arctic Ocean halocline
fore ice starts to form. Also, upwelling of denser water in (Rudels et al. 1996).
periods with offshore winds may bring more saline water to The bottom waters that run off the banks and flow into
the surface. When ice starts to form, wind is important in neighboring depressions and eventually into the deeper ba-
moving the existing ice away, so that new ice can form in sins may transport with them suspended sediments and asso-
the same area. The reason for this is that most of the salt is ciated contaminants.
rejected back to the underlying water during the freezing
process (Gow and Tucker 1990) and faster ice growth in Open ocean convection in the Arctic
open water areas means that more brine will then be re- Open ocean convection requires cyclonic circulation in a gyre,
leased there than under thicker ice cover. a preconditioning phase to create low static stability within
Due to its high density, bottom-water flows along the sea the gyre, and intense and rapid surface forcing (e.g., by wind)
floor under the influence of the buoyancy, friction, and Co- (Killworth 1983). These conditions for open ocean convec-
riolis forces. Unfortunately, no concurrent observations of tion occur in the Greenland, Iceland, and Labrador Seas.
the temperature, salinity, and speed of such currents exist. The Greenland Sea has been recognized as a source for
Calculations with a stream-tube model indicate that drain- deep and bottom-water formation since the time of Nansen
age currents require a long time to cross a featureless shelf (1906). Water produced in the Greenland Sea Gyre, that is
(more than 40-60 days), chiefly because of the predomi- Greenland Sea Deep Water, is one of the world’s densest wa-
nance of the Coriolis force, which discourages down slope ter types. In the Greenland Sea, cold, fresh, less dense surface
flow. However, more rapid drainage is possible through water must sink through the warmer, more saline, denser
submarine canyons (Melling and Lewis 1982). water underneath. Since the surface salinity is too low to
Bottom water formed on the shallow banks around No- allow for deep-reaching convection even if the water is
vaya Zemlya (Midttun 1985) fills the western and eastern cooled to the freezing point, the freshwater has to be re-
Novaya Zemlya troughs. From here, it flows through the moved to create a water dense enough to convect into the
St. Anna Trough into the Arctic Basin. The topography of deeper layers. Formation of sea ice aids initiation of deep
the troughs inhibits strong advection and mixing and al- convection, since freshwater is extracted in the ice while
lows the salt to accumulate. The St. Anna Trough outflow is most of the salt remains in the water (Rudels 1993).
therefore dense and can penetrate into the deep water of the Earlier theories about the origin of bottom-water gener-
Eurasian Basin and contribute to its ventilation. In contrast, ally involved mixing the surface (Polar Water) and interme-
the bottom water of the Laptev Sea is less dense and will diate water (Atlantic Water) via diffusion, turbulence, eddies,
only intrude into the layer of Atlantic Water (Quadfasel et ice edge upwelling, or other processes (Carmack and Aagaard
al. 1993). Bottom water has also been observed in the Bear 1973, McDougall 1983, Clarke 1986, Rudels 1990). The
Island Channel, on the Svalbard Bank, and in Storfjordren- mixed water, which was then saltier, further cooled at the
na (Sarynina 1969, Anderson et al. 1988, Quadfasel et al. surface and sank. A more recent theory, also based on mix-
1988, Blindheim 1989). In the Chukchi Sea (Aagaard et al. ing of water masses, adds the ‘salinization’ mechanism of
1981), Beaufort Sea (Melling and Lewis 1982, Melling and brine rejection by early winter ice formation: ice forms at
Moore 1995), and Barrow Canyon, similar water has been the surface in early winter, rejecting salt and thus increasing
observed descending into the Canadian Basin and mixing the density of the surface water. Arctic winds then blow the
with existing subsurface water of the same density (Garri- ice away, exposing the ‘preconditioned’ surface water to
son and Becker 1976). cooling and mixing by the strong, cold winter winds. Deep
Much of the bottom water produced on the shelves is convection then occurs in a series of short, episodic bursts
not dense enough to sink into the deep basins, but enters (Pawlowicz et al. 1994).
the depths of the halocline in the Arctic Ocean. The signa- Quadfasel and Rudels (1990) monitored the buildup of a
ture of this shelf water can be seen over most of the Arctic salty layer in the top 5 m of the water column in the Green-
Ocean in the temperature-salinity structure of the halocline land Sea Gyre with a drifting bag equipped with tempera-
(Figure 3·40). Both physical and chemical properties indi- ture and conductivity sensors. The layer became unstable
cate that the origin of the upper halocline is in the Chukchi and ejected a plume draining this top layer. Such plumes en-
Sea (Aagaard et al. 1981, Jones and Anderson 1986), where- train ambient water and the upper Arctic layer will eventu-
Freezing and ally be enriched with salt, so that consequent plumes pene-
Temperature, °C
-2 0 2
brine release trate through the halocline and into the relatively warm in-
0 termediate layer (Rudels et al. 1989, Rudels 1990).
Low-salinity surface
mixed layer This warm intermediate water must rise to the surface so
Depth
that continuity can be preserved. Here it melts the ice and
(m)
Cold halocline layer
temporarily stops the convection, but after cooling of the
surface layer and new ice formation, the haline convection
250
cycle can start again. CTD (conductivity, temperature, depth)
profiles obtained in the central Greenland Sea Gyre during
several winter cruises apparently confirmed this concept of
Atlantic Water
convection cycles by showing that apparent convection depths
500 increase stepwise rather than continuously (Muench 1990).
The severity of the winter cooling influences the amount of
33 34 35 deep water that is produced in a given year, and to which
Salinity
depth the convection will take place (Aagaard 1968, Car-
Figure 3·40. Schematic representation of the temperature and salinity
structure of the upper Arctic Ocean and how the halocline layer is main-
mack and Aagaard 1973, Schott et al. 1993).
tained by brine-rich water produced on the shelves (Source: Aagaard et Water produced by convection in the Iceland Sea is an in-
al. 1981). termediate water mass nearly as cold as the deep water, but
slightly less saline, and therefore lying above the deep water Double diffusion
(Swift and Aagaard 1981). The key to the formation of this In areas where warm and salty (e.g., Atlantic) water under-
dense intermediate water in winter is seasonal heat loss. lies cold and fresher (e.g., Polar) water, a high density sub-
The salinity of the Arctic domain waters ultimately limits surface layer can be produced by double diffusive processes
the density of the water cooled at the sea surface in winter. at the interface. Heat is lost from the saline layer into the
Maintenance of relatively high salinities in the northern and upper layer, and from there it is expelled to the atmosphere
central Iceland Sea is important for the convective process. (Carmack and Aagaard 1973, McDougall 1983). The diffu-
The East Greenland Current, carrying relatively low-salinity sion of salinity into the upper layer is much, much slower
water, if spread over the Iceland Sea, could bring about suf- than the heat diffusion, so the lower layer will remain salty
ficient stratification to block winter formation of the dens- while it cools, and the density of the layer will increase.
est water masses. A salinity of 34.7 from the surface down- Both open ocean convection and shelf convection intro-
ward to about 100 m proved to be the critical value in this duce waters with different temperature-salinity characteris-
area. At salinities of 34.7 and lower, the surface water does tics into the water column. Open ocean convection leads to
not reach a density high enough to start a deep convection, large horizontal gradients which may be so strong that a
even at a freezing temperature of –1.8°C. Salinities of 34.8 slight disturbance results in vertical gradients capable of
or more, however, make this possible (Malmberg 1984). establishing double diffusive convection, which can drive
The part of the Iceland Sea with the highest surface layer horizontal mixing of the waters. Shelf convection results in
densities in winter is well away from the areas that in sum- large vertical gradients and inversions of temperature and
mer and fall show the strongest influence of the relatively salinity induce double diffusive processes, which act to re-
low-salinity surface water carried by the East Greenland move these anomalies (Rudels 1993).
Current and East Icelandic Current. The layer formed in
winter spreads laterally along isopycnal surfaces through- Vertical particle fluxes
out the Iceland Sea. The water produced constitutes a ma- Vertical mixing is limited in the Arctic Ocean by the stratifi-
jor proportion of the overflow of dense water into the North cation of the water column produced by freshwater inputs
Atlantic. of runoff and ice melt. Therefore, settling particles may form
The development of cyclonic circulation and concen- a significant component of the vertical transport of material
trated cooling and mixing within a cyclonic gyre, together which short circuits the sluggish diffusive transport. Vertical
with the presence of a source of warm, salty water (i.e., the particle fluxes in the Arctic Ocean vary widely among the
Irminger Sea Water) lead to deep convection in the Labra- marginal seas, where strong seasonal inputs of terrestrial
dor Sea (Clarke and Gascard 1983). This scenario will prob- and marine particles occur; along slopes where sediments
ably only occur in the Western Labrador Sea during winters are advected and redeposited by boundary currents; and, in
in which there has been a significant cold westerly wind dur- the interior ocean, in which surface water is isolated from
ing the early winter (January-February). Since the Labrador both terrestrial and atmospheric inputs, and where produc-
Shelf is normally ice-covered during winter, the Arctic air tivities are much lower (Subba Rao and Platt 1984, O’Brien
reaches out to the deep western Labrador Sea before it meets et al. 1991, Macdonald et al. 1993, Wheeler et al. 1996).
open water. Here the heat and vapor flux (evaporation) into On shelves, in the marginal ice zone, and under the Trans-
the air rises dramatically, resulting in intense cooling and a polar Drift, particle fluxes may be relatively large due to sup-
rise in salinity of the water in this area. This is the precon- ply from coastal sediments or from high biological produc-
ditioning phase of Labrador Sea Water renewal. tivity. Pfirman et al. (1990) proposed that sediment contri-
When a cyclonic gyre forms in the Western Labrador Sea buted from sea ice forming along the Siberian margin would
(scale ≈200 km), various water types (i.e., remnants of Irmin- result in elevated particle flux and sedimentation rates under
ger Water and Atlantic Water) present within this gyre are the Transpolar Drift. Stein et al. (1994) further proposed
mixed into the mixed layer and eventually cooled by the cold that oceanic and turbidity currents are even more important
and dry northwest winds blowing offshore from Labrador. than sea ice as sources of sediment to the Eurasian Basin,
It is believed that the gyre retains the developing deep mixed with active resuspension and redeposition along the sea bed.
layer in this general area long enough for the transformation One of the more important regions for vertical particle flux
to Labrador Sea Water to take place (Clarke and Gascard occurs in marginal ice zones, where algal mats falling off the
1983). The intensity of the convection varies from year to melting ice during break-up can result in separation of algal
year. Data suggest that significant renewal of Labrador Sea blooms from grazers, allowing phytodetritus to settle rela-
Water occurs on only 6 out of 10 years studied (Lazier 1980), tively undegraded to depth (Carey 1987, Hsiao 1987, Bru-
and it occurs to varying depths (Lazier 1973, Gascard and land et al. 1989, Hebbeln and Wefer 1991, Riebesell et al.
Clarke 1983). Convection in this region will not reach to 1991). These regions, which deliver carbon-rich material to
the bottom, however, because the North Atlantic Deep Wa- the benthos, are potentially important for the rapid trans-
ter is a denser water mass than the Labrador Sea Water. In- port of sediments and contaminants from the ocean surface
stead, the water produced will spread out over large parts of to the sea floor. Another kind of process produces pellets
the North Atlantic on top of the North Atlantic Deep Water. that are released where ice melts. During its transit, particle-
laden ice is reworked by freeze-thaw cycles which tend to
Upwelling move enclosed sediment to the surface of the ice. Surface ab-
The potential importance of upwelling in the Arctic in con- lation of the dirty ice can then form ‘cryoconites’, or pellets,
nection with deep-water formation is that it can bring deeper, which are effective scavengers of particle-reactive contami-
usually more saline water (Atlantic water) to shelf sites where nants, and settle quickly through the water column once the
the water can be cooled by surface exchange and returned ice has melted (Pfirman et al. 1990).
to the deep basins with increased density (Aagaard et al. In contrast to shelves and margins, permanently ice-covered
1981). How efficient this process is in producing dense wa- regions have very low particle fluxes (Fischer et al. 1988, Har-
ter is still uncertain, but the net fluxes associated with up- grave et al. 1994) because they are far from inorganic sedi-
welling seem to be small, except maybe in some of the can- ment sources and, although moderate primary production
yons (Aagaard and Roach 1990). may occur, much of the production is recycled in the surface
water (Wheeler et al. 1996). In support of this, Ku and lowing calculation of transport times (Dahlgaard et al.
Broecker (1967), Finkel et al. (1977), Moore and Smith 1986, Kautsky 1987). Livingston et al. (1984), using the
(1986), and Bacon et al. (1989) concluded from various same tracers, deduced that water could transit from Spits-
natural radionuclide inventories (230Th, 10Be, 210Pb) that bergen to the North Pole in about three years.
scavenging intensity, and hence particle production and Current velocities in the Greenland and Iceland Seas have
sinking, in the interior Arctic Ocean is low. Variations in been used to calculate the time it will take the Deep Water
particle fluxes occur between basins in the interior ocean as to flow from the Fram Strait to the Denmark Strait (Buch et
suggested by Scholten et al. (1995). Distributions of 230Th al. 1996). The result was about 700 days. Based on satellite-
and 231Pa suggest low flux in the deep waters of the Maka- tracked drifter observations taken in 1982-83, Royer and
rov Basin and higher flux in the Eurasian Basin. Emery (1984) found a transit time around the Bering Sea on
Particle-reactive contaminants, such as some of the arti- the order of one year.
ficial radionuclides, will be incorporated into the vertical
particle flux, and since sinking particles are often carbon- Ages and residence times
rich and contain biogenic components, bioaccumulating Estimates of ages and residence times of water masses in the
contaminants, such as OCs and mercury, are likely to be in- Arctic Ocean and adjacent seas have been based mostly on
volved too. For example, Cochran et al. (1995) found that distribution of transient tracers, including 14C, tritium, and
rapid ventilation and scavenging at the Nansen Basin mar- chlorofluoromethanes. The residence time is a measure of
gin resulted in the removal of particle-reactive radionuclides the average time a water parcel spends in a certain reservoir,
at rates comparable to those of the North Atlantic. Inor- while the age is defined as the mean time elapsed since the
ganic components of the particle flux will tend to be cou- water parcel left the surface. The different 14C radioactivities
pled to the biogenic flux because ‘indiscriminate packaging’ in the surface and the deep water can be used to calculate
by zooplankton of small particles produces larger aggre- the age of the deep water, since the only change in the amount
gates which then dominate the flux (Honjo 1990). of 14C in seawater is due to decay.
Due to logistical difficulties, there are few direct mea- The residence time for waters in the Arctic Ocean (Figure
surements of particle fluxes in the interior Arctic Ocean. 3·28) varies with location and depth. On the Eurasian shelf,
Hargrave et al. (1989, 1994) used sediment traps at the the residence time is one to three years (Codispoti and Low-
Canadian Ice Island to measure fluxes characteristic of man 1973, Hanzlick and Aagaard 1980, Östlund and Hut
regions under permanent ice cover. Peaks in particle sedi- 1984, Pavlov and Pfirman 1995). Mean residence time of
mentation were observed during late summer (August- river runoff on the shelves of the Siberian seas, however, has
September) and mid-winter (December-January). Average been estimated from salinity and tracer data (tritium, 3He,
particle fluxes were low (1.1 g/m2/y), compared to other and the 18O/16O ratio) to be 3.5 2 years (Schlosser et al.
year-round measurements made in marginal seas (3-120 1994b). On the Canadian shelf, the residence time for fresh-
g/m2/y), as summarized in Hargrave et al. (1994). Not water is less than a year (Thomas et al. 1986). Over the deep
enough sedimenting material was collected to analyze for basins, the residence time for the surface and halocline wa-
contaminant content, but the seasonal pattern of deposi- ters is about ten years (Aagaard and Coachman 1975, Öst-
tion suggests that scavenging of particle-associated com- lund 1982, Livingston et al. 1984), although surface water
pounds from the water column would be greatest during in the Transpolar Drift may take less than five years to cross
the late summer months. from the Siberian shelf to the Fram Strait (Schlosser et al.
In locations where there is seasonal ice cover, biological 1995a).
fluxes are usually found to undergo strong variability, with Atlantic Water has a shorter residence time in the Eurasian
maximum fluxes 10 to 1000 times the minimums (Atkinson Basin than in the Canadian Basin, about 25 and 30 years,
and Wacasey 1987, Honjo et al. 1988, Wefer et al. 1988, respectively (Aagaard and Greisman 1975, Wallace and
Wefer 1989, Honjo 1990). In winter, biogenic fluxes are Moore 1985). Eurasian Basin Deep Water has the shortest
extremely low, but near shelves these may be augmented by residence time of the deep waters of the Arctic Ocean, about
particles carried off the shelves in dense-water flows (Honjo 75 years, because of its connection with the Greenland-Nor-
et al. 1988, Wefer 1989). wegian Seas (Bönisch and Schlosser 1995). 14C and 39Ar data
show an estimated mean residence time of 250-300 years for
Eurasian Basin Bottom Water and 300 years for Canadian
3.5.4.7. Oceanic time scales
Basin Bottom Water (Schlosser et al. 1994b, Bönisch and
Understanding water mass characteristics and renewal rates Schlosser 1995). The Amundsen and Nansen Basins are
are important for the determination of the sensitivity of the treated as one reservoir, i.e., the Eurasian Basin, since the
Arctic waters to environmental changes. The northern wa- mean renewal times of the deep and bottom water of the
ter masses interact extensively with the surface ocean in two basins are fairly similar.
their formation regions, and hence have been tagged with In the Nordic Seas, the mean age of the waters between
anthropogenic substances that have entered the surface 50 to 500 m depth (i.e., Upper Arctic Intermediate Water;
layer of the ocean during the past several decades. Mea- Fig. 3·32) is five to ten years, while the mean age for the
surements of these tracers can be very useful in calculations upper 500 m in the Norwegian Atlantic Current is only one
of the time required for a water mass to travel from its to two years (Schlosser et al. 1995b). The less saline compo-
source region to another region. Salinity and temperature nent of Arctic Intermediate Water represents a rapid turnover
anomalies that are advected with the currents can also be of a small system, and the residence time is three to four
used to estimate transport times. years (Swift et al. 1980).
Atlantic Water flowing north from the Rockall Channel Measurable concentrations of the chlorofluoromethanes
will probably take two to three years to reach Bear Island F-11 and F-12 have penetrated to the deep basins of the Nor-
(Blindheim and Loeng 1981) and half a year from Bear Is- dic Seas. They show progressively lower concentrations with
land to the west coast of Novaya Zemlya. Using cesium and increasing depth, reflecting the greater isolation of deeper
strontium released from Sellafield on the Irish Sea as a tracer, waters from air-sea exchange. Calculations based on these
one can see that the isotopes follow the main currents, al- chlorofluoromethanes give a time scale for deep-water con-
vection in the Greenland Sea of 40 years and a time scale transport from temperate source regions to the Arctic and
for lateral mixing between the deep Greenland Sea and the the exchanges between the atmosphere and the land and
deep Norwegian Sea of 20 to 30 years (Bullister and Weiss ocean in the northern hemisphere have made significant
1983). The residence times for the Greenland Sea Deep progress toward this end. These models could perhaps be
Water and the Norwegian Sea Deep Water have been esti- enhanced through greater coupling with compartment and
mated from tracer balances to be 28 and 31 years, respec- subcompartment specific models, though whether or not the
tively (Bönisch and Schlosser 1995). increased parameterization would effectively enhance the
Tritium and 85Krypton have been used to estimate the predictive ability of the model is not known.
age of the Denmark Strait Overflow Water and the North
East Atlantic Deep Water just south of the Denmark Strait.
3.6.2. Atmospheric modeling
Denmark Strait Overflow Water was the densest water ob-
served south of the Denmark Strait, and there are two types, As mentioned in section 3.2, for most contaminants, the at-
a low-salinity type and a slightly higher salinity, more dense mosphere provides the fastest transport medium. Therefore,
type. Tritium and 85Krypton data reveal that the low-salinity in order to understand how contaminants are brought to the
type resided behind the Greenland-Iceland ridge for about Arctic, it is important to quantitatively calculate the atmos-
one year, before flowing into the Irminger Sea, compared to pheric transport in an explicit manner. Furthermore, having
about 15 years for the higher salinity type. North East At- established a measurement-validated modeling tool, the model
lantic Deep Water forms in the northeastern Atlantic from a can be used to estimate the relative importance of different
mixture of water flowing out of the Norwegian Sea and emission sources, to evaluate possible effects of new emis-
North Atlantic Water. About 70% of the tritium and 85Kryp- sions, and to increase our understanding of the relative im-
ton burden of North East Atlantic Deep Water comes from portance of various processes involved in the transport.
North Atlantic Water and 30% from Norwegian Sea water. There are two types of models used for contaminant path-
The average age of the North East Atlantic Deep Water just ways studies: 1) three-dimensional atmospheric models suit-
south of the Denmark Strait, relative to its formation in the able for studying the movement of one-hop compounds and,
northeast Atlantic, is approximately 7.5 years (range: 1-11.5 to a lesser extent, multi-hop compounds; and, 2) two-dimen-
years) (Smethie and Swift 1989). sional multi-compartmental models suitable for multi-hop
The ventilation time of the Labrador Sea Water depends compounds. Both types are usually run on a domain that is
on the production rate. With a production rate of 8 Svd global, although single-hop compounds can often be studied
(McCartney and Talley 1984), the ventilation time is nine effectively with a hemispheric scale model.
years; but if the production rate is lower, the ventilation
time is higher. Using the 2 Svd value calculated by Worthing-
3.6.2.1. Three-dimensional atmospheric models
ton (1976), a ventilation time of 36 years is derived (Talley
and McCartney 1982). Models for meteorological fields
In Baffin Bay, the upper layer (200 m) has a residence There are two different classes of models used to simulate the
time of about two years (Top et al. 1980, Rudels 1986b), air motions and dispersion of contaminants in the atmos-
while the warmer layer below has a residence time of 8 to phere. Physical models are physical analogues to the atmos-
20 years. From oxygen consumption, Top et al. (1980) cal- phere, consisting of tanks filled with liquids set in rotation in
culated the residence time of the bottom water in Baffin Bay order to reproduce the effects of the planetary rotation. Ex-
to be 500-1000 years. ternal forcings of motions can be controlled, and the result-
The normal residence time for water in the Chirikov Ba- ing motions and their transport capacities can be observed
sin (between St. Lawrence Island and the Bering Strait) is and measured. Physical models are no longer in wide use.
about two weeks, and varies little during the summer; but The second class of models are numerical models – math-
with the appearance of reversed flows in the fall, the resi- ematical tools which calculate the physical quantities utiliz-
dence times are two to four times longer (Coachman 1993). ing electronic computers. The elementary laws of nature
(mass conservation, Newton’s second law, the laws of ther-
modynamics) and the properties of the air (the equation of
3.6. Modeling state) and its composition are all formulated as mathemati-
cal differential equations. It is presumed that it is not neces-
3.6.1. Introduction
sary to calculate the motion of each molecule in air, but that
A complete review of modeling as it relates to the physical the physical laws can be formulated for separate air parcels.
and chemical fate of contaminants in the various compart- These are assumed small enough to be treated as homoge-
ments of the Arctic environment is beyond the scope of this neous and large enough to disregard molecular discontinu-
report. Consequently, only a summary of some of the more ities. In particular, molecular diffusion of total mass between
relevant models that have been adapted for northern condi- air parcels is neglected. Thus, the law of mass conservation
tions, or specifically applied to contaminant transport in the can be applied for each individual air parcel, and a unique
Arctic are reviewed here. Within the terrestrial/freshwater air velocity can be defined as the velocity of each parcel. The
compartment, only models related to the fluvial transport size of each air parcel is assumed to be small enough to be
and fate of contaminants in freshwater systems will be con- mathematically treated as an infinitesimal quantity. The math-
sidered. However, some of the atmospheric models do con- ematical equations contain derivatives of physical quantities
sider the soil surface in the context of sources and exchanges. (wind velocity, temperature, pressure, etc.) in all space direc-
It appears that data gaps and data inconsistencies among tions and in time. The time derivatives are of first order and,
different compartments remain a problem for these kinds in principle, the time-evolution of all the quantities in any
of linkages. point in the atmosphere can be found, provided proper ini-
To date, there has been no attempt to fully integrate all tial values are given, as well as boundary conditions describ-
of the compartments in a single model, due to the complexing the interactions with the Earth’s surface and the universe.
ity that would be required. However, the recent advances in In practice, however, the equations for these models are
atmospheric models directed at understanding contaminant generally impossible to solve in analytic fashion. They are
nonlinear and coupled, so that the distribution of one quan- Such initial conditions, called analyses, are made through a
tity determines the distribution of other quantities. Further- process called data assimilation. Modern data assimilation is
more, motions, which vary in time with a typical frequency a method which permits information from observations to
and have a typical spatial size, cannot be treated without propagate in space downwind of the observation points.
taking into account the influence from motions with other Thus, the influence of good observations can be seen as well
frequencies and sizes. For example, the development of cy- in data sparse regions, such as the Arctic. Fields of meteoro-
clones the size of 1000 km depends, among other things, on logical quantities from a numerical model which is run using
the structure and strength of turbulent eddies ranging in data assimilation can be taken out in all grid nodes with
size from 10-100 m. Initial values and boundary conditions equally spaced intervals. The resultant data set is considered
cannot be given accurately, but are subject to uncertainties, to be time resolved, in that it is the best available estimate of
due to measurement errors and insufficient coverage of the quantities in each grid node at the given time levels, re-
measurement networks. Research initiated by Lorenz (1963) flecting the scales of motion resolved by the grid. A model
has shown that the equations governing atmospheric mo- which is not using data assimilation, but only calculates the
tions are crucially dependent on the initial conditions. Thus, meteorological quantities with no corrections will produce
the time-development of two initial conditions which are fields which may resemble analyzed fields, but they cannot
arbitrarily close, but not exactly equal, will lead to com- be viewed as valid for specific dates and times. Such fields
pletely different states of the atmosphere within a limited can be taken as possible realizations of the given climate, de-
time span (the predictability limit). This difficulty is why all termined by the insolation and ground surface boundary
weather prediction is based on regular meteorological ob- conditions, but in many cases even this is not possible as the
servations which are disseminated in real-time to forecast ‘model-climate’ may deviate considerably from the real cli-
centers around the world. This is not the case, for example, mate due to factors such as weak subgrid parameterizations.
for predictions of tidal water. Atmospheric models can be solved for parts of the Earth
The mathematically complex nature of the equations for (limited area models, LAM), or for the whole globe (general
atmospheric state quantities provokes a need for approxi- circulation models, GCM). LAMs cannot be run without
mations. It is also clear from the chaotic properties of the lateral boundary conditions, presumably from a GCM run
atmosphere that observations must be incorporated in some with a coarser resolution (larger distance between the grid
way in order to be able to reproduce time-resolved features nodes) than the LAM. LAMs are run in national weather
similar to those observed. First of all, it must be accepted centers for the purpose of weather prediction over a few
that it is impossible, at least so far, to calculate motions and days, while GCMs are run at fewer centers (e.g., European
features on scales of variations from fractions of millimeters Center for Medium Range Weather Forecasts, ECMWF, or
and seconds to thousands of kilometers and months simul- National Meteorological Center, NMC in USA) for the whole
taneously. The equations are therefore subject to Reynolds- period of predictability, approximately five to ten days de-
averaging, i.e., separating turbulent and small-scale flows pending on flow type. GCMs are also run at certain loca-
from larger-scale flows. Nonlinear terms give rise to terms tions without data assimilation in order to study general cir-
in the averaged equations which describe the influence of culation and climate-related topics in a research model. Full
the turbulent flow on the larger-scale flow. Thus, the prob- climate simulations and climate predictions can only be made
lem of turbulence closure is encountered, an area of research using models that incorporate ocean circulations, sea ice,
which has produced a large amount of scientific literature. and land ice. Such model runs require enormous data re-
The Reynolds-averaged equations are still analytically un- sources and machinery to handle all the model output. They
solvable, and so must be further approximated by so-called are run at only a few locations around the world (see IPCC
discretization methods. For example, all derivatives in space (1996) for an overview).
and time can be replaced by finite differences over discrete
intervals in the time and space coordinates, but discretiza- Dispersion models
tions can also be made by, for example, using finite Fourier Atmospheric dispersion models are numerical models which
transforms (spectral discretization), or so-called finite ele- produce approximate numerical solutions to mass budget
ment methods. In any case, the method of discretization di- equations for a given set of contaminants. The contaminants
rectly or indirectly defines a grid of node-points, in which are in our case always considered to be gravitationally neu-
the state quantities are defined and known, but not the space tral (i.e., they follow the motion of the air precisely). The
and time in between. As for the Reynolds averaging, the only exception is contaminants that are captured by precipi-
procedure of discretization gives rise to new terms in the tation, in which case they follow the precipitation until they
equations which describe the influence of the scales that are hit the ground and deposit there, or they are reintroduced as
unresolved by the grid. These are called subgrid processes, air contaminants by evaporation of the precipitation element.
which must be parameterized as a function of the resolved Dispersion models need information about emission
scale quantities in order to close the system. sources, in addition to meteorological fields, which cause
These parameterizations are made in all routine meteoro- transport and precipitation scavenging. Many physical and
logical models, and they frequently involve crude assump- chemical processes are also linked to meteorological condi-
tions for parameter values, which are difficult to defend in- tions, in particular inside clouds. The modeling of contami-
dependently from observations and measurements. Such pa- nants seriously influenced by the presence of clouds (e.g., a
rameters must often be tuned in order to obtain reasonable major part of the production of sulfate from sulfur dioxide
model results (as compared with actual measurements). takes place inside clouds) is highly impacted by the uncer-
Clouds and precipitation processes are heavily involved in tainty related to clouds in meteorological models. In relation
subgrid parameterizations; thus, these are the most uncer- to dry deposition to the ground, data on the properties of
tain components of atmospheric models. vegetation, snowcover, ice cover, or open water may be re-
In order to produce quantitative results which can be quired, as well as the state of boundary layer turbulence,
compared with actual measurements taken at specific points which determines the aerodynamic resistance toward dry
and time-levels, calculations with the numerical models deposition. For some contaminants, the state of the bound-
have to use initial data which are based on measurements. ary layer may also be of crucial importance for natural
emissions from the sea, or for the re-emission of multi-hop age and used ECMWF data as input to produce dispersion
contaminants. calculations for three full years. These results showed good
For single-hop contaminants, a pure atmospheric disper- comparability with actual measurements.
sion model will be sufficient. As long as the emissions, de- Full three-dimensional global modeling of multi-hop con-
position processes, and the meteorological fields influencing taminants has so far not appeared in the literature. Strand
them are determined, the system is closed. To further evalu- and Hov (1993) used a two-dimensional, zonally-averaged
ate the fate of a single-hop contaminant being deposited, model to calculate the global distribution of the - and -
modeling of freshwater and/or oceanic transport is required. HCH isomers. The model also included the oceans and land/
But these models do not have to interact any further with soils as separate model compartments. Meteorological data
the atmospheric model. Dispersion models aimed at under- were taken from Plumb and Mahlman (1987) which are zo-
standing Arctic distribution of single-hop contaminants nally-averaged, based on a GCM without data assimilation.
should at least include coverage of emission sources in the There are several model results published which cover
major areas of the northern hemisphere. Arctic areas, but none are specific to Arctic contaminant be-
For multi-hop contaminants, distribution in all compart- havior. Langner and Rodhe (1991) calculated global distri-
ments of the environment must be carefully studied so that butions of sulfur compounds based on monthly-averaged
rates of revolatilization can be predicted. However, since meteorological data, with relatively coarse resolution. Ben-
many of the multi-hop contaminants (e.g., semivolatile or- kowitz et al. (1994) calculated sulfur distributions based on
ganic compounds (SOC) or mercury) have a large fraction ECMWF meteorological data for October and November
of their total mass in the other compartments (e.g., ocean, 1986, focusing on the North Atlantic Ocean.
terrestrial), and have been accumulating in these compart-
ments over decades, the day-to-day transport in the other On-line dispersion models
compartments is not crucial. A more reasonable approach The main benefit of running a dispersion model as an inte-
is to map the concentrations in the land, freshwater, and grated part of a meteorological model, in spite of the loss of
oceanic compartments. Modeling the whole accumulation actual time-resolution, is the possibility of including feed-
process over decades requires multi-compartmental models back effects between the contaminant and the meteorologi-
with full interaction (see section 3.6.2.2). Dispersion mod- cal variables determining the transport, chemistry, and depo-
els aimed at understanding Arctic distribution of multi-hop sition of the contaminant. For example, sulfur may change
contaminants should have global coverage. the atmospheric contents of cloud condensation nuclei, and
thus the lifetime of clouds and their optical properties. In-
Off-line dispersion models creased cloudiness may increase the efficiency of sulfur diox-
Dispersion models which are not part of models for meteo- ide oxidation to sulfate. Sulfate also directly causes an increase
rological fields, are called off-line models. This is a natural in clear air scattering of solar radiation and thus an increase
choice for modelers who do not run their own meteorologi- of planetary albedo. Soot, or black carbon, may likewise
cal model, or who want to produce time-resolved calcula- cause a decrease of the clear air albedo over regions with
tions of contaminants for comparison with measurements. very high natural surface albedo (e.g., in the Arctic). In total,
In principle, such modeling could be made on-line inside these processes may perturb the natural radiation balance in
the full meteorological model run with assimilation of me- a way which can cause changes in large-scale wind systems.
teorological observations. In practice, however, there are Another benefit of using on-line models is the availability
still large problems in the initial phase after inserting mea- of full meteorological data with the same accuracy as the
surements. These so called ‘spin-up’ problems reveal a seri- dispersion model. Thus, no interpolation is needed. The main
ous under-prediction of clouds and precipitation due to a drawback is that day-to-day comparisons with measurement
mismatch between the vertical motions produced in the mod- data is impossible. Agreements must be relied on in a more
el and the analyzed field of relative humidity. Therefore, un- statistical sense.
til better methods come in to use (e.g., the 4D-VAR assimi- Dastoor and Pudykiewicz (1996) recently made calcula-
lation method being developed at several numerical weather tions of Arctic sulfur transport in the Canadian global spec-
prediction centers), the ‘spin-up’ problem is avoided by using tral model. So far the model does not include feedbacks to
short-range forecasts (~ 6-12 hours) as input to the disper- the meteorological fields, although the results appear pro-
sion model. These data are not available for each time-step mising. With little focus on the Arctic, Taylor and Penner
in the dispersion model, thus, off-line modeling necessitates (1994) used the Lawrence Livermore National Laboratory
some kind of time interpolation between the inputs. version of the Community Climate Model 1 of the US Na-
Off-line modeling with the scope of understanding Arctic tional Center for Atmospheric Research to calculate the sul-
contaminants has so far been mainly concerned with Arctic fur distribution with sulfate radiative forcing and feedback
haze and sulfur compounds. Iversen (1989b) developed and on the meteorological fields. Similar calculations have also
applied a hemispheric-scale model for sulfur dioxide and be made by Pham et al. (1995) and Feichter et al. (1996).
particulate sulfate. Based on meteorological data from NMC, Joussaume (1990) modeled dust of crustal origins in a
the model was run for March 1983, and a period in June- GCM, although without any feedback mechanisms to at-
July 1983. Emissions were taken from Semb (1985). The mospheric processes. A minor, but still significant amount
model output was compared with observations taken at was calculated to be transported to the Greenland icecap
ground level and with aircraft in the Norwegian Arctic. through elevated routes (height 5-10 km above ground)
Seasonal contrasts as well as episodic behavior were rea- from the Sahara, Middle East, and desert areas in Asia. Sim-
sonably well reproduced. Later, the model was run for Oc- ilar calculations made for conditions during the last glacial
tober 1982 and January 1983, with a more comprehensive maximum (18 000 years before present) lead to larger bur-
model validation (Tarrason and Iversen 1992). Recently, the dens of crustal dust, but still underestimated compared with
model has been run for a full year (1988), with meteorolo- ice-core observations (Joussaume 1993). Genthon and Ar-
gical data from ECMWF (Tarrason and Iversen 1996). Re- mengaud (1995) have also calculated desert dust distribu-
sults can be found in chapter 8 of this report. Christensen tions, as well as sea salt and radioactive isotopes, with spe-
(1995) developed a model with almost hemispheric cover- cial emphasis on deposition in Greenland and Antarctica.
The Bergen Model is a global multi-compartmental model

3.6.2.2. Multi-compartmental models
for HCHs with atmosphere-soil and atmosphere-ocean ex-
Global scale multi-compartmental models have the poten- change represented with a two-dimensional zonally-aver-
tial to: aged atmospheric model. An existing, detailed atmospheric
model was simplified (Strand and Hov 1993) by reducing
1. contribute significantly to our understanding of the ex-
the vertical and meridional resolution to six equally spaced
tent of transport of contaminants from sources in tropi-
latitude zones and four vertical layers. This was then com-
cal and temperate zones to the Arctic and Antarctic;
bined with a modified oceanic transport model (Siegenthaler
2. assist in the formulation and testing of hypotheses sur-
and Joos 1992) and a soil-atmosphere exchange model for
rounding the ‘cold condensation’ effect tending to con-
trace organics developed by Jury et al. (1983, 1984a, 1984b).
centrate substances in regions of cold climate (Wania and
In summary, the model includes the atmosphere, ocean wa-
Mackay 1993);
ter, cultivated and uncultivated soil, and the processes of at-
3. quantify fluxes of contaminants to the Arctic;
mospheric advection and convection, diffusive gas exchange
4. estimate the fraction of the total global use of each con-
between atmosphere and soil or water, wet deposition, and
taminant which reaches the Arctic; and,
chemical degradation. Seasonality is taken into account by
5. predict the time response to contamination and deconta-
defining specific temperatures, precipitation rates, and at-
mination (e.g., 5 years or 50 years).
mospheric transport parameters for four seasons.
Since the models necessarily treat several media including One weakness of this model is that no consideration is
air, soils, and ocean water, as well as a variety of biota, they given to particle-associated chemicals in atmosphere and
synthesize or integrate the components of the global and surface water, and, therefore, to particle-mediated transport
Arctic systems (Figure 3·1). The application to developing processes, such as dry particle deposition from the atmos-
rational international contaminant control strategies is ob- phere or contaminant scavenging by settling suspended
vious. For example, an early attempt to model the global matter in the surface ocean. This may be an acceptable sim-
DDT cycle with a four latitude-zone box model by Ostro- plification in the case of HCHs, which, due to relatively
mogil’skii et al. (1985) led to the conclusion that DDT con- high volatilities and water solubilities, occur mostly in the
tamination will persist for many decades and impact the gaseous or in the truly dissolved phase, but would require
Arctic much more than the Antarctic. After checking their adjustments for less volatile or less soluble chemicals. It is
model with observations in various environmental media, also not totally appropriate for Arctic precipitation which
they estimated that after a global ban of DDT, the rate of for much of the year is dominated by snowfall, an effective
decline in atmosphere, soil, and ocean will be a factor of scavenger, even of highly volatile contaminants at low tem-
10-20, 30, and 1-2% per decade, respectively. peratures (Gregor 1996, Hoff et al. 1995). Also, no land-to-
Although more comprehensive models are still in their ocean exchange is taken into account, which may be very
infancy, considerable progress in model development has important for river-transported chemicals. Furthermore,
been made in recent years. Existing global models for per- ocean-atmosphere exchange is assumed to occur at ocean
sistent organics cover a wide range of complexity. These water temperatures, which are considered to be constant
models are all based on a multi-compartmental global sys- throughout the year and which often differ considerably
tem, shown in Figure 3·1, with varying degrees of complex- from the temperature of the overlying atmospheric bound-
ity. For instance, the atmospheric compartment has been ary layer. Since a sensitivity analysis revealed the Henry’s
treated by Wania and Mackay (1995) and by Strand and Law constant as one of the most influential input parame-
Hov (1993). ters, this is potentially a very serious source of error.
Modeling studies are conducted in one of two ways as The Combined Model (Figure 3·41) is essentially an ex-
determined by the type of question to be answered and the pansion of a regional multimedia model developed by Mac-
type of model. These two are as follows: kay et al. (1992), designed to describe the fate of chemicals
in smaller regions of approximately 105 km2. It consists of
1. Long-term, low-spatial resolution model – starting with a
ten linked, latitudinally-determined climatic zones, with
‘clean’ global environment and simulating the entire time
each zone consisting of air, fresh and marine water, culti-
period a chemical has been used on a global scale/or as
vated and uncultivated soil, and freshwater sediments. The
long as it has been persistent in the environment. This is
the approach taken in the 2-D and 1-D models described N-Polar N-B
orea
l
below. At this time, this is impractical for a 3-D model, Figure 3·41. Schematic dia- N-
Te
gram of the Combined Model. m
as a result of the immense demand in computing time Ten climatic zones, each con-
pe
ra
te
and input parameters; and, taining well-mixed compart-
N-
2. Short-term, high-spatial resolution model – defining the ments, are connected by

Su
b tr
global contaminant loading in system compartments (Fig- meridional atmospheric and

op
oceanic exchange (Source:

ic
ure 3·1) at one point in time, and simulating a shorter Wania and Strand 1998).
time period (e.g., one year). This is the approach taken
N -T ro p
by the 3-D models.

icic
The Bergen and Toronto Models

S -T ro p
There have been two attempts at simulating the dispersion

of OCs on a zonally-averaged, global scale. Strand and Hov climate zone
p ic
(1996) modeled the distribution of HCHs using a 2-D, at-

tro
atmosphere
ub
mospheric model (Bergen Model), while Wania and Mackay (four layers)
S -S
developed a multimedia compartment model for OCs based cultivated uncultivated soils ra
te
pe
on the fugacity approach (Toronto Model) (Wania and Mac- soils freshwater system Te
m
S-
kay 1993, Wania 1994, Mackay and Wania 1995, Wania (water and sediment)
bpola
r
S-Su
and Mackay 1995). surface ocean S-Polar
zonally-averaged climatic bands are similar to the zones The Bergen Model is superior in its treatment of advec-
defined in the Bergen Model. Transport and exchange tive and diffusive transport processes in the global atmos-
processes between these compartments as well as degrada- phere, while the Toronto Model succeeds better in describ-
tion and export to the deep sea are parameterized. The ing the reversible climate-dependent exchange processes
model includes the phase partitioning of chemicals between between the atmosphere and the Earth’s surface. Recently,
air and aerosols, as well as between air and suspended mat- Wania and Strand (in prep.) have combined the best parts
ter. However, there is no vertical subdivision of the atmos- of both approaches by incorporating into the Toronto Mod-
pheric compartments, which makes the description of at- el a 2-D description of the atmosphere which adopts the
mospheric transport processes very simplistic. Seasonally vertical layering and the deduction of transport parameters
varying parameters, such as temperature and atmospheric from the Bergen Model. This ‘combined’ model is the first
exchange rates, are defined as sinusoidal functions. fugacity-based model which includes stacked atmospheric
“Bergen” Model “Toronto” Model “Combined” Model

Strand and Hov Wania and Mackay Wania and Strand
Stratosphere Stratosphere
Middle/upper troposphere Middle/upper troposphere

Atmosphere
Lower/middle troposphere Lower/middle troposphere
Atmospheric boundary layer Atmospheric boundary layer
Cultivated Uncultivated Agricultural Fresh Natural Agricultural Fresh Natural

soil soil Surface soil water soil soil water soil
ocean Surface Surface
ocean ocean
Deep Sediment Sediment
ocean
Figure 3·42. A comparison of the structures of the Bergen, Toronto, and Combined zonally-averaged global models (Source: Barrie et al. 1997).
Table 3·10. A comparison of characteristics of global zonally-averaged multi-compartmental toxics models.

–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Bergen Model Toronto Model Combined Model
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Zonal subdivision 6 zones, equally spaced 9 zones, based on climatic criteria 10 zones, based on climatic criteria
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Compartments in 4 atmospheric layers (stratosphere, 1 atmospheric layer (entire troposphere), 4 atmospheric layers (stratosphere,
each zonal middle/upper and lower/middle surface ocean, cultivated and unculti- middle/upper and lower/middle
‘building block’ troposphere, atmospheric boundary vated soil, freshwater, freshwater troposphere, atmospheric boundary
layer), surface ocean, deep ocean sediment layer), surface ocean, cultivated and
(subdivided into several layers), uncultivated soil, fresh water, freshwa-
cultivated and uncultivated soil ter sediment
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Treatment of atmos- Meridional and vertical exchange Meridional exchange described with Meridional and vertical exchange
pheric transport described with horizontal and vertical latitude-dependent horizontal eddy described with horizontal and
advection velocities and eddy diffusivities, lumping advective and vertical advection velocities and eddy
diffusion coefficients and diffusive processes diffusion coeficients
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Treatment of Modified High-Latitude-Exchange/ Meridional exchange described with Meridional exchange described with
oceanic transport Interior Diffusion-Advection Model estimated horizontal eddy diffusivi- estimated horizontal eddy diffusivi-
(HILDA) by Siegenthaler and Joos ties, lumping advective and diffusive ties, lumping advective and diffusive
(1992) processes processes
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Treatment of Diffusive gas exchange at ocean water Diffusive gas exchange (at different air Diffusive gas exchange (at different air
atmosphere–ocean temperature, dissolution in rain and ocean water temperatures), disso- and ocean water temperatures), disso-
exchange lution in rain, dry and wet particle lution in rain, dry and wet particle
deposition deposition
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Treatment of Description of soil–atmosphere Description of soil–atmosphere Description of soil–atmosphere
atmosphere–soil exchange based on Jury et al. exchange based on Mackay and Stiver exchange based on Mackay and Stiver
exchange (1983, 1984a, 1984b) (1991) (1991)
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Treatment of 4 seasonal values for air temperature, Sinusoidal functions for air and ocean 12 monthly values for air and ocean water
seasonality precipitation, evaporation, and water temperature, sea ice coverage, temperature, sea ice coverage, and
atmospheric transport parameters and atmospheric exchange parameters atmospheric exchange parameters
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Technicalities Zonally-averaged continuity equation, 54 linear mass balance equations in 90 linear mass balance equations in
step-wise solution using finite fugacity notation solved step-wise using fugacity notation solved step-wise using
difference approximation finite difference approximation finite difference approximation
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Test chemicals -HCH and -HCH -HCH and -HCH, DDT -HCH and -HCH
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Reference Strand and Hov (1996) Wania (1994), Wania and Mackay (1995) Wania and Strand (1998)
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
compartments of variable density, and special considera- Army Corps of Engineers – Hydraulic Engineering Center
tion is given to the treatment of vertical atmospheric trans- (HEC) or the Environmental Protection Agency (EPA).
port and wet deposition processes. Temperatures in ocean These models, listed below, are well established and thor-
water and the four atmospheric layers, as well as the verti- oughly tested, and were included in the RSS with as few
cal and horizontal atmospheric transport parameters are modifications as possible.
input as monthly averages. Figure 3·42 and Table 3·10
HEC-2 is used to compute water-surface profiles for
compare various characteristics of the three zonally-aver-
steady, gradually varying flow in both prismatic
aged models.
and nonprismatic channels. Both subcritical and
supercritical flow profiles can be estimated, as
3.6.3. Freshwater systems well as the effects of various obstructions, such
3.6.3.1. Introduction as bridges, culverts, weirs, and structures in the
overbank region.
In order to understand how contaminants are transported
DAMBRK is used for simulation of rapidly changing, un-
within and from the terrestrial/ freshwater compartment,
steady flow, for example from a dam breach or a
it is important to quantitatively link the processes that con-
flood wave.
trol transport. Such a modeling tool, once validated with
HEC-6 is used for computation of sediment transport,
measurements, can be used to increase our understanding
erosion, and deposition in rivers. Its hydraulic
of the relative importance of individual processes as well
capabilities are similar to those of HEC-2. In ad-
as to ultimately estimate contaminant deliveries. Although
dition, the model can simulate bed changes due
the different processes in river systems influence each other,
to erosion and deposition of sediments.
most models address only one or a few of these aspects,
RICE simulates hydraulics and ice processes in rivers.
and tend to focus on a specific question or location.
The most important ice simulation processes are
The river system as a whole, with all its processes, from
frazil ice production, icecover generation, and
watershed runoff and riverine processes to the discharge
decay and transport of ice.
via estuaries into the ocean, is not yet adequately repre-
FINNECO is a water quality model for lakes. The model is
sented by a single model or even coupled models. As well,
based on the EPAECO model and has subse-
unlike atmospheric and oceanic models, which are larger
quently been modified in Finland and Norway
in scale and not confined by national boundaries, freshwa-
to enable it to be used in ice-covered lakes. The
ter models tend to be country-specific and are presented
model can compute water quality and algal
as such below.
growth in a lake as a function of climatic con-
ditions and the water quality of the lake inflow.
3.6.3.2. Norway The model can also compute the vertical temper-
ature distribution in the lake.
A new computer model system has recently been developed
QUAL2E is a water quality model for rivers. The model is
in Norway. This River System Simulator (RSS) combines
based on the EPA model with the same name,
14 different models to describe river processes, with partic-
but has been extended to include benthic algae
ular emphasis on the environmental effects of river regula-
in the RSS version.
tions. The system integrates only well-tested models, and
AKVASS is a groundwater model that describes the inter-
makes use of one common database and a common user
action between river and groundwater in the
interface (Killingtveit and Harby 1994).
river plain. The model can compute effects, for
The RSS consists of three main parts: the main user in-
example, in the form of lowered groundwater
terface, the simulation models, and the database. The parts
level when streamflow is reduced due to hydro-
are briefly described separately below. The relational data-
power regulation, or dynamic fluctuations in
base is based on a common logical data model or database
groundwater level due to rapidly varying flow
scheme, which reflects as closely as possible the real physi-
downstream from peaking hydropower plants.
cal system, rather than the data structure used by the indi-
vidual models. The computations are performed by a
number of individual models, exchanging data through the Consequences for humans and ecosystems
database via a standard database query language (Struc- Most of the models in this group have been designed and
tured Query Language, SQL). implemented as part of the RSS project. The reason for this
The 14 models included in the River System Simulator was that there were few such models in existence, and they
can be grouped into four main categories. These are: were poorly documented and tested, or for other reasons
were regarded as unsuitable for inclusion in the RSS.
1. input from the watershed into the river system;
2. hydropower system simulation models; BIOLAKE estimates the potential fish harvest from a lake
3. physical, chemical, and biological processes in rivers based on simple relations to zooplankton and
and lakes; and, benthos. Fish harvests are estimated for pelagic
4. consequences for humans and ecosystems. and benthic fish species. If both types of species
exist in the same lake, the total harvest is re-
Only the latter two will be considered further, since they are
duced by competition. BIOLAKE also esti-
most relevant to contaminant issues.
mates the impacts of lake regulation on fish
Physical, chemical, and biological processes harvests.
in rivers and lakes BIORIV is a model used to describe biological condi-
tions in rivers, and the effect of changing phys-
This group contains the largest number of models in the
ical conditions. In the first version, the effect of
RSS, including models for river hydraulics, ice and water
water temperature is primarily considered.
temperature, sediment transport, and water quality. Most
Time from spawning to hatching, juvenile fish
models in this group were developed previously by the US
growth, and time to smoltification are all • improvements of the juvenile fish habitat;
strongly water temperature-related simula- • possible water temperature-related impacts on fish
tions done by the model. growth; and,
HABITAT simulates physical habitat (living conditions) • possibilities for canoeing/rafting.
in rivers, and how this is affected by changes
Although the RSS has proven very useful in its applications,
in streamflow. The model is useful in studies
there has been no consideration, apparently, of using this
of how the habitat of different fish species is
model to predict the effect of these changes on contaminant
affected by river regulation.
transport within the system, though this project seems quite
RECREATE is a model used to compute how a proposed
feasible. More effort should be put into applying the RSS to
change in stream flow will affect the recreatio-
help understand contaminant transport and fate in Arctic
nal value of the river.
river systems.
Examples of the use of River System Simulator
Testing has been an important issue in the complete devel-
3.6.3.3. Canada
opment of the RSS. Three Norwegian rivers were chosen as
reference rivers, each posing different problems and pos- Much of the modeling work in Canada has focused on the
sessing different biological, hydrological, and geographical Mackenzie River, especially hydrology and ice. See for ex-
conditions. They are the Gjengedal River, Halden River, ample the following sources for:
and Stjørdal River. Only the Gjengedal and Stjørdal River
• 1-D hydrodynamic modeling for the Mackenzie Delta
systems are discussed here.
(J. Kerr, pers. comm., Studies Engineer, Environment Can-
Gjengedal River system ada, Yellowknife, NWT, 1996);
• a multi-channel suspended sediment transport model for
The Gjengedal River is situated in western Norway, in the
the Mackenzie Delta (S. Fassnacht, pers. comm., Depart-
county of Sogn og Fjordane. The catchment area is 32 km
ment of Civil Engineering, University of Waterloo, Water-
long and includes 171 km2 at the outlet to the Hyen Fjord.
loo, Ontario, Canada, 1996) also based on the 1-D hy-
The catchment area rises from sea level to 1077 m.a.s.l. The
drodynamic model;
largest lake is Storevatnet, with an area of 31 km2. Hydro-
• river ice (Beltaos et al. 1993, Martinson et al. 1993,
logically, the river is typical of rivers in the middle part of
Chambers et al. 1994); and,
western Norway, with a distinct spring flood, several rain
• hydrology of small systems (Cassell and Pangburn 1991,
storms in the fall, and very low runoff during the winter
Munro 1991).
period. The mean annual runoff is 11.6 m3/s.
The RSS was used to study possible impacts of a planned Other modeling efforts have looked at evapotranspiration
hydropower development. The lake Storevatnet was pro- from Arctic wetlands, snowfall, and fresh snow and snow-
posed as a storage reservoir for the hydropower plant. Alter- pack processes, but there is as yet no attempt at integrating
native locations of the power station outlet were investi- these on a larger scale.
gated. The most significant impacts on the environment Data from an intensive study of the processes control-
were predicted in the river reaches between the power sta- ling contaminant fate at Amituk Lake, NWT (Diamond
tion outlet and the original lake outlet at the dam site. et al. 1996) has been used to attempt to fully integrate
The RSS was also used to analyze the economic potential physical process data with contaminant fate data. These
of different locations of the hydropower plant with alterna- modelers built a general, whole lake, mass balance model
tive restrictions of minimum flow and reservoir filling level. based on the QWASI model (Quantitative Water Air Sedi-
The model ENMAG was set up to simulate these alterna- ment Interaction), developed originally by Mackay and co-
tives. The regulation of the lake Storevatnet will affect the workers (Mackay 1991). The model uses equivalence as
growth and stock of brown trout, the only fish in the lake. the equilibrium criterion, which is suitable for most chem-
The model BIOLAKE was used to simulate the expected icals, rather than fugacity, which is suitable for VOCs only.
changes in fish harvest. The planned regulation height is A multi-species time-dependent model was developed to
28 m, which probably will lead to approximately an 80% represent the large proportion of meltwater and chemical
decrease in harvest. This model is based on empirical data, loadings that flow through the lake without mixing with
and may be too conservative for the Storevatnet site, since the water column, because of the rapid and dominant spring
brown trout is the only species of fish. freshet. The model also accounts for ice cover for up to
46 weeks of the year, during which no air-water exchange
Stjørdal River system of chemicals occurs, and contaminant concentrations in
A 70 year old hydropower system is about to be refurbished the water column may increase as they are excluded from
in the upper part of the Stjørdal River system in the middle the ice layer. The application of this model to understand-
part of Norway. The major impacts will be in the upper part ing the fate of contaminants will be considered further in
of the river system, with pronounced changes in the flow chapter 6.
regime. Different types of changes and impacts will also Models have been developed for a portion of the Yukon
take place all the way downstream to the outlet in the fjord. River system (Diamond et al. 1996, Barrie et al. 1997), how-
Altogether, eight models were tested for both the upper and ever, since these are still in the early stage of development,
lower part of this river system. The RSS was used to study they will not be considered here.
topics, such as:
• the effects of floods and changes in the flood and flow 3.6.3.4. United States
regime;
Model development in the United States parallels that in
• ice conditions and water temperature changes;
Canada. Emphasis has been on developing models that are
• economic revenue of hydropower production;
concerned with one or several physical transport processes
• possible groundwater level changes in the river valley;
specific to Arctic conditions, but little effort has been directed
at fully integrating these processes to provide, for example,

3.6.4.2. Types of models
estimates of contaminant delivery, even seasonally, from
large river systems. Some examples of these models include Numerical modeling is a tool that has proven useful and
Kane et al. (1993) for energy-related modeling of snowmelt will continue to prove itself useful in understanding and
and Cassell and Pangburn (1991) who modified the Stream- simulating the Arctic Ocean and, indeed, the whole cli-
flow Synthesis and Reservoir Regulation (SSARR) model to mate system. There has been evolutionary progress to-
account for cold region effects. ward completely coupled, ice-ocean numerical models.
At first, decoupled models were being used. Semtner
(1976a) performed an ocean modeling study of the Arc-
3.6.3.5. Russia
tic Basin wherein ice properties were prescribed, and the
A deterministic, mathematical model has been developed general circulation was calculated by the model. On the
for runoff, called the HYDROGRAPH SHI-96. It covers other hand, there have been many more modeling studies
all types of runoff and may be applied to any physiogra- where ice was explicitly modeled and where oceanic para-
phic region and basins of any size. The design interval of meters were prescribed.
the model is daily or shorter and the model input includes Maykut and Untersteiner (1971) developed a one-dimen-
precipitation amount and duration, air temperature, and air sional, thermodynamic ice model which was simplified and
humidity deficit. The model generates a continuous hydro- shown by Semtner (1976b) to perform well, even if ice is
graph for the design interval. Other information required represented by a low-resolution vertical grid. Parkinson and
by the model includes: Washington (1979) used Semtner’s model and a simplified
ice dynamics model to simulate the yearly ice cycle in the
• formation and melting of snowcover;
Arctic and Antarctic. Hibler (1979) developed a horizontal,
• evapotranspiration;
two-dimensional, transport model of the Arctic Basin which
• infiltration and surface runoff;
exhibited realistic properties. Ice growth rates were pre-
• dynamics of soil moisture and drainage water;
scribed a priori as a function of ice thickness and time of the
• formation of underground runoff;
year. A one-dimensional, bulk mixed-layer model has been
• runoff transformation as part of overland flow and
developed by Lemke (1987) and coupled to Semtner’s ice
within the channel; and,
model. Mellor and Kantha (1989) coupled a one-dimen-
• runoff at the outlet.
sional, turbulence closure model to an ice model. The ice
This model is part of the ‘Runoff – Erosion – Contamina- model also explicitly recognized the importance of ice to
tion’ System model (V. Vuglinsky, pers. comm., State Hy- the ocean heat transfer role of open leads.
drological Institute, St. Petersburg, Russia, 1996). The papers by Hibler and Bryan (1987) and Semtner
In conclusion, it is clear that additional effort should (1987) brought forth simulation studies of the Arctic Basin
be directed at more comprehensive, circumpolar model- using a full three-dimensional, coupled ice-ocean model. It is
ing activities to quantify delivery of contaminants from apparent from these papers that oceanic heat distribution
the land surface to the fluvial system. Subsequently, exist- does play an important role in determining the location of
ing river models, such as the RSS, the SSARR, or the HY- the marginal ice zone and the other icecover properties, such
DROGRAPH SHI-96, need to be tested and calibrated for as mean ice thickness and concentration. These models did
estimating contaminant transport and river-mouth load- not incorporate mixed layer physics even though, as re-
ings to the marine environment. Testing of these models marked by Hibler and Bryan, the mixed layer ought to be
with existing data will help determine information needs considered the essential coupling medium between ice and
for their refinement. International coordination and coop- ocean. A three-dimensional, coupled ice-ocean model which
eration in this effort is required to enhance information incorporated coupling with the mixed-layer physics was de-
compatibility. veloped by Mellor and Kantha (1989).
Until the last decade or two, modeling studies have been
somewhat idealized and lacking in adequate resolution. Due
3.6.4. Marine system modeling
to the availability of increasingly affordable and powerful
3.6.4.1. Objectives of modeling
computer resources, numerical modeling is now entering an
One of the latest attempts in marine system modeling has age where ever more realistic simulations of oceans are to be
been to define the major Arctic Ocean transport pathways expected.
for contaminants through application of a coupled dynam-
ic–thermodynamic ice-ocean model to the geographical Ice models
area encompassing the Arctic Basin and adjacent seas. Nu- Häkkinen and Mellor (1992) developed a coupled dynamic
merical models must utilize both historical and newly ac- model that will predict ice motion, which is primarily driven
quired field data to provide information on physical trans- by winds. The model is built on governing ice equations.
port. Concurrent analyses of field data will contribute to The dynamic prognostic variables are ice concentration,
understanding the model output within the context of perti- mass, and velocity. The ice mass equations provide an accu-
nent physical transport processes. rate balance between advection of ice mass and sources and
Models can provide a quantitative, climatological picture sinks of ice. There are some empirical coefficients in the ice
of water mass and contaminant transports everywhere in concentration equations which deal with how to convert ice
the Arctic. The pathways for both bottom-released and growth in open water to an increase in fractional ice cover,
river-origin contaminants can be described by the model. and ice melt to a decrease in fractional ice cover. The ice mo-
Since prediction of the ice drift is included, the pathways mentum balance consists of Coriolis force, advection of mo-
for contaminants in the ice from riverine sediments can be mentum, wind stress, interfacial ice-ocean stress, and forcing
simulated. All of these modeling components can provide a from internal stresses. Away from the coastal areas, the
basis for the design of a long-term monitoring plan for the main balance is between wind stress and ice-ocean stress.
interpretation of measurements and for the design of future Near the coasts, the ice internal stresses become important
modeling scenarios. (Hibler 1979, Häkkinen and Mellor 1992).
Thermodynamics It should be noted that, whereas the models include
The pioneering study of ice thermodynamics was done by some vertical variability for the thermodynamics, the equa-
Maykut and Untersteiner (1971), in which vertical structure tions for momentum are invariably slab models. They do
of sea ice was highly resolved, but ice concentration and not allow for vertical variability in velocity, a reasonable
horizontal variability were neglected. The ice models that simplification.
will be discussed below are essentially slab models in which
vertical structure is simplified to minimally model the con- Requirements and constraints
duction process (Semtner 1976b). The following conduc- This is a review of sea ice-ocean models which will be/are
tion description used in Häkkinen and Mellor (1992) re- used as components of global coupled atmosphere-ocean-
sembles the model of Semtner (1976b), with modifications land-ice models. Using numerical ocean models, the re-
to account for leads. A schematic of the thermodynamics is search goal is to understand the numerous interactions be-
shown in Figure 3·43. tween lateral and vertical exchanges of heat and salt that
QAI dominate the high latitude ice-ocean system. Exactly the
hs snow Qs
T3
QAO
same processes of advection and mixing which govern tem-
T2
hI
ice QI2
T1
perature and salinity also govern the advection and diffu-
FT
ice QI0
T0 sion of contaminants.
FT This discussion is limited to a review of models and mod-
eling studies of the Arctic and its peripheral seas. Further-
WAO
more, discussion is restricted to three-dimensional, primitive
WAI equation, numerical ocean models, which acknowledge tem-
WRO
perature and salinity variations and which have been used or
WIO are currently being developed for Arctic studies in a coupled
Figure 3·43. A schematic of the three-level thermodynamic snow ice sys- ice-ocean model. These types of ocean models are the focus
tem model (Source: Häkkinen and Mellor 1992). because the inclusion of thermodynamics and salinity effects
are important for both short-term and long-term simula-
In the model, the conductive heat flux at the sea-ice sur- tions. Secondly, any contaminants can be modeled as scalar
face is balanced by the atmospheric heat flux, including ra- variables, just like temperature and salinity, but they will
diative, sensible, and latent components, if surface tempera- have different surface boundary conditions and sources and
ture is below freezing. If there is a net heat gain at the sur- sinks. Finally, some examples on how models can be used to
face and the surface is at or above freezing, the heat gain is describe different scenarios for transport and spreading of
used to melt the snow and ice. At the ice/water interface, contaminants are given.
any imbalance between the conductive heat flux from the
ice slab and the sensible heat flux from the water column Quality and availability of data
results in either freezing or melting. The surface tempera- Initialization of realistic three-dimensional ocean models re-
ture over snow or ice, T3, is determined iteratively (a single quires good quality hydrographic data. This will consider-
iteration for monthly winds, but several iterations for daily ably shorten the spin-up time of the model and improve its
varying winds) from the surface energy balance, where the accuracy in anticipated one- to five-year simulations. The
atmospheric flux, QAI, has to equal the conductive heat flux first available historical data set was compiled by Levitus
through snow, Qs. However, the heat capacity of the snow (1982), and was based on a sparse set of hydrocasts. An im-
is neglected, so that conductive flux through ice, QI2, equals proved version was compiled by Levitus and Boyer (1994)
Qs. QAI consists of the solar radiation and longwave back and Levitus et al. (1994). Prognostic model simulations re-
radiation formulas, which are adapted from Parkinson and quire forcing fields consisting of atmospheric data (winds,
Washington (1979), and turbulent sensible and latent heat air temperatures, humidity, short and long wave radiation
fluxes modeled using bulk formulas. The surface albedo is fields, cloudiness, precipitation) and lateral transports of
assumed to be a function of surface temperature (Robock water at the open boundaries. The analyzed atmospheric
1980). The ice cover is considered as a single slab, rather data set most useful for this work is available from ECMWF.
than dividing it into several thickness classes, each of which The transports through the Norwegian Sea and the Den-
would have their separate heat balance. Up to 10 cm of mark and Bering Straits are derived from available transport
water is allowed to accumulate at the surface; excess water estimates. There are several sets of river runoff values avail-
can drain to the ocean. In the fall, this surface water freezes able. However, the accuracy of total river input is question-
before bottom accretion begins. able because of the large year-to-year variability, and because
A common conclusion from the ice modeling studies is many smaller rivers are usually not included.
that the snow-ice system has its largest sensitivity to the
surface albedo. In effect, all other variability in surface forc- Availability of models
ing components, such as cloudiness and sensible and latent A coupled ice-ocean model using multi-layer sigma coordi-
heat flux, are secondary compared to the albedo effect. nates accommodating both shallow shelf regions, such as the
However, once one defines the albedo model, in which the Kara and Barents Seas, and the deep basins is now available.
Arctic sea ice will not vanish during summer in the present This model traces the flow of contaminants in surface/sub-
climate, the other forcing components, i.e., oceanic heat surface layers and their (possible) sinking to the deep ocean.
flux, ice divergence, snowfall, cloudiness, and air tempera-
tures (and winds), also show strong sensitivity in determin-
3.6.4.3. Ocean Models
ing the Arctic ice mass. Considering all the uncertainties in
all of the forcing components listed above, the ice models Detailed discussion here is restricted to three-dimensional,
appear to be relatively stable in giving an average Arctic primitive equation numerical ocean models which include
sea-ice thickness of about 2.5-3.3 m when climatological description of temperature and salinity variations and which
radiative and turbulent heat fluxes, snowfall, and oceanic have been used in large-scale Arctic studies in a coupled ice-
heat fluxes representative of the central Arctic are used. ocean model. These models have the potential to describe con-
taminant transport. Exceptions will be identified. For meso- pled Arctic model: a good description of mixing (which
scale ice-ocean models and ice-ocean process studies, the poses further demands on the surface forcing) and good esti-
reader is referred to a review by Häkkinen and Mellor mates for river runoff, lateral inflows and outflows, and
(1990). their spatial distribution. Either of these elements can be cir-
cumvented by using a diagnostic description for parts of the
Specific ocean models ocean model, such as relaxation to climatology, however,
The first numerical ocean model and the present day stan- this limits a model’s usefulness.
dard is the Bryan-Cox Model (Bryan 1969, Cox 1984),
which was later numerically improved by Semtner (1976a) Partially diagnostic models
so that it is sometimes called the Bryan-Cox-Semtner Hibler and Bryan (1987) were the first to present results
Model. It is conceptually the simplest of the models, an im- from a coupled ice-ocean model for the Arctic. The ocean
mediate advantage, in that it uses a conventional z-level model was partially diagnostic for the deeper ocean, in that
vertical coordinate and spherical coordinates. It is used by a Newtonian damping factor forced the temperatures and
many large-scale ocean modelers (Bryan and Holland 1989, salinities toward climatological values, while the upper
Semtner and Chervin 1992). ocean could adjust prognostically to the surface forcing de-
The Princeton Ocean model (POM) (Blumberg and Mel- termined by ice freezing and melt. The main result from the
lor 1980, Mellor 1993) was initially developed for applica- simulations was to show the importance of the northward
tion to estuaries and coastal oceans, although it is currently heat transport by the Norwegian Atlantic Current, which is
being applied to ocean basins. In estuarine applications, the responsible for determining the ice extent in the Greenland
sigma coordinate system, together with a free surface and and Barents Seas. This same heat source is responsible for
the turbulence closure submodel (Mellor 1973, Mellor and year round ice melt in the Greenland Sea. The coupling also
Yamada 1974, Mellor and Yamada 1982), provided a bot- intensified the oceanic Beaufort Gyre and the East Green-
tom boundary which converted tidal energy into turbulence land Current, which the authors described as a readjustment
and mixing. Comparisons with current meter measure- to the forcing because the initial salinity and temperature
ments, tide gauge data, and salinity intrusion into estuaries fields were smoothed estimates of the observations.
were favorable. The model’s horizontal grid is orthogonal The climatology of the Arctic ice thickness field supports
curvilinear, which supports rectilinear or spherical coordi- high ice thicknesses north of Greenland due to the mechani-
nates as special cases. cal pileup of ice transported by the Transpolar Drift to the
A variation on the theme of sigma-coordinate models is vicinity. However, the model, using surface forcing from
offered by the Spectral, Primitive Equation Model (SPEM) 1979, produced a highly anomalous ice thickness field with
developed by Haidvogel (Hedstrom 1990, Haidvogel et al. a large ice buildup along the East Siberian coast and a weak
1991). It is similar in some respects to the POM model, ex- build up north of Greenland.
cept that it has a rigid lid and is distinguished by the fact This Hibler-Bryan Model has been implemented for opera-
that variables on the sigma coordinate are expanded verti- tional use at the US Navy’s Fleet Numerical Oceanography
cally in series of Chebyshev polynomials (i.e., it is a spectral Center (Riedlinger and Preller 1991). Forecasted seasonal ice
model in the vertical, instead of a level model). It has been trends are in good overall agreement with observations for
used in a number of process studies, but not yet for coupled growth and decay. However, the ice cover in the Barents and
ice-ocean studies, although such a coupled version is under Greenland Seas appears to be somewhat excessive, which
development. the authors attribute to poor model resolution to describe
The forerunner of the isopycnal layer models is the Bleck narrow currents like the West Spitsbergen Current.
and Boudra Model (Bleck and Boudra 1986, Bleck et al. Another coupled ice-ocean model by Piacsek et al. (1991)
1992). This particular model has not yet been coupled to is also partially diagnostic. The Hibler ice model is coupled
sea ice, but the isopycnic formulation was adopted by to a high resolution mixed layer, where turbulence is calcu-
Oberhuber (1993a) in his version of the isopycnal model lated according to the level 2.5, Mellor and Yamada turbu-
with a sea-ice cover using Hibler’s viscous-plastic rheology lence closure model (Mellor and Yamada 1982). However,
and Semtner’s thermodynamics. An advantage of isopycnal the deeper ocean is diagnostic with a geostrophic velocity
coordinates is that increased resolution is automatically ob- field determined from Levitus’ (1982) climatology. Corre-
tained in regions with strong density gradients. The code is sponding to a perpetual year, 1986 forcing, the model pro-
complicated due to the need to cope with vanishing isopyc- duces a very realistic seasonal variability, with the exception
nal layer thicknesses near the ocean surface and bottom. of the wintertime ice extent in the Barents Sea. They con-
sider the inclusion of the mixed layer dynamics to give a su-
Coupled ice - ocean, Arctic modeling studies perior ice thickness field compared to coupled models with-
In the Arctic Ocean and the peripheral seas, the Greenland, out mixed layer dynamics. The mean oceanic heat flux in the
Iceland, Norwegian, and Barents Seas, the overall stratifica- model varies in the ice-covered area from 5 to 15 W/m2, at-
tion structure is determined by both river runoff and At- taining even larger values northeast of Spitsbergen in the
lantic and Pacific inflow, and by dynamic and thermody- area of the submerged Atlantic waters. These values are
namic interactions with the ice cover. As a result of heat ex- on the high side compared to the traditional view of about
change with the atmosphere, deep waters are formed in this 2 W/m2, and to even lower values, less than 0.5 W/m2, as
region, a process which is augmented by brine rejection due suggested by other one-dimensional modeling studies (Mel-
to ice formation. Alternatively, since density is strongly lor and Kantha 1989). Comparison of observed buoy tracks
salinity-dependent, freshwater fluxes at the ocean surface and simulated drift tracks are in reasonably good agreement
are especially important because of their stabilizing effect. considering the coarse resolution of the model.
In the Arctic Ocean, the freshwater layer prevents heat ex-
change between upper and deeper parts of the water col- Prognostic models
umn, and in the Greenland Sea, an excess freshwater cap in The first comprehensive prognostic ocean model for the Arc-
the form of ice can prohibit the renewal of deep waters. tic Ocean was described by Semtner (1976a) using the Bryan-
These processes prescribe the needs of a limited-area cou- Cox-Semtner (BCS) ocean model. The model was later ex-
panded by Semtner (1987) to include a dynamic-thermody- (1993). Their paper presents a set of sensitivity studies of near-
namic ice cover comprised of a three-level snow-ice system ly all possible forcing components. They find that cloud cover
(Semtner 1976b) and an ice rheology simplified from the is one of the least known, but most sensitive components,
model by Hibler (1979). This ice-ocean system was driven while snow fall (up to 2 m/y) did not appear to be crucial.
by monthly surface forcing and specified inflow-outflow
fluxes at the boundaries as in the Semtner (1976a) model. Evaluation and verification of model results
Overall, the results from the ice model component There now exist fairly sophisticated, numerical models for
showed agreement with the observed ice extent, but ice ice cover and oceans. They appear to give fairly realistic re-
thicknesses were much lower compared to the generally ac- sults even though they are mostly coarse resolution models.
cepted average values of 2.5-3 m. The model predicted a The basic criterion of evaluating an equilibrium ocean model
modest ice growth in the Beaufort Sea Gyre and much is the question of how well it can reproduce the seasonal hy-
larger growth occurring on the Eurasia coast. The model drography. The comparison is inadvertently limited to scalar
also showed that ice melt occurred in the Greenland Sea variables instead of velocity fields of which we know very
even in mid-winter. The simulated oceanic circulation pro- little except for coarse features, such as the cyclonic flow of
duced the main circulation features in the upper ocean, such the Atlantic origin waters around the Arctic Basin. Remote
as the Beaufort Gyre, Transpolar Drift, East Greenland Cur- sensing gives an ideal way to validate ice model results, at
rent, Norwegian Atlantic Current, and Barents Shelf circula- least for ice concentration fields. Also the Arctic Buoy Pro-
tion. Surface salinities were reasonably well reproduced in gram can provide information on ice drift for validation
the model, except that their gradients were not strong purposes. The least known quantity observationally is the
enough in the central Arctic. ice thickness. To date, the best means of measuring ice thick-
This model has been applied to the study of inter-an- ness is through submarine sonar observations, but this type
nual ice variability by Fleming and Semtner (1991) for the of data is classified, and thus is largely unavailable for moni-
period of 1971-1980. Their main conclusion is that using toring and validation.
inter-annually varying forcing produces much improved
sea-ice cover variability for the annual cycle, compared to
3.6.4.4. Modeling of oceanic contaminant transport
the model forced by mean monthly climatology. The inter-
in Arctic seas
annual forcing produced large variability in ice thickness
fields, with much less degree of variability in sea-ice ex- The starting point for modeling the transport of contami-
tent. The oceanic heat flux variability has the strongest in- nants in Arctic seas, is a coupled ice-ocean model as de-
fluence in determining the monthly average ice edge posi- scribed above. In addition, a transport model is needed, ei-
tions. However, overly thin ice thickness is still a problem ther as an extension integrated into the ocean model or as a
in the model which is, according to the authors, probably separate model driven by the results from the ocean model.
a result of an excessive melt-freeze cycle. Also, the authors Depending on the contaminant in question, the transport
consider that not having an explicit treatment of mixed model can be more or less complicated. If the contaminant is
layer dynamics produces inadequacies in the vertical mix- dissolved in seawater, it is treated as a passive tracer. This
ing processes. can be done in two ways. In the Eulerian approach, an ad-
In particular, the model is used to test the conjecture of vection-diffusion equation is solved for the development of
Aagaard and Carmack (1989) that the Great Salinity the tracer concentration in a similar way to the equations
Anomaly was a consequence of an anomalously large ice for salinity and heat in the ocean model itself. An alternative
export in 1968. The model simulations explored the high is the Lagrangian, or particle-tracking approach. Here, tra-
latitude, ice-ocean circulation changes due to wind field jectories of passive particles representing the contaminant
changes while other forcing components, such as air tem- are calculated from the current field of the ocean model. Both
peratures, cloudiness, snowfall, and river runoff, were cli- methods are well developed, and have well-known good and
matological. The simulated ice extent in the Greenland Sea bad properties.
increased during the 1960s, reaching a maximum in 1968, If the contaminant is bound to particles, the situation is a
as observed. The maxima in ice extent coincided with large lot more complex. The particle may go into the sediments,
pulses of ice export through the Fram Strait. The ice export be resuspended, captured in ice, etc. The physical, chemical
from the Arctic Ocean to the Greenland Sea in 1968 was and geological processes are not known well enough. Realis-
the largest in the simulation, being about twice as large as tic models are not available at present. Further process stud-
the simulated average for 1955-1975 and corresponding to ies and modeling are needed to gain knowledge before these
1600 km3 of excess freshwater. The simulated upper water processes can be incorporated or parameterized into re-
column in the Greenland Sea has a salinity minimum in the gional or large-scale pollution transport models.
fall of 1968, followed by very low winter salinities. In addi-
tion to the above average ice export to the Greenland Sea, Examples of dispersion modeling
there was also a supply of freshwater from the Arctic There has been relatively little activity on modeling of dis-
Ocean. The total simulated freshwater input of 2500 km3 persion of oceanic contaminants in Arctic areas. Recently,
to the Greenland Sea compared well with the estimated the activity has increased somewhat, but most of the results
total freshwater excess of the Great Salinity Anomaly of are so fresh that they have not yet reached the reviewed lit-
about 2200 km3, as it passed through the Labrador Sea erature. This increased activity is mostly concerned with the
(Dickson et al. 1988). potential threats of nuclear contamination from former So-
A paper by Oberhuber (1993b) used the OBH Model for viet sources, in particular in the Kara Sea.
most of the Atlantic Ocean from 30°S to 90°N. Thus, a sec- The largest modeling effort has been in the USA under
tor of the Arctic Ocean was included in the model domain. the Arctic Nuclear Waste Assessment Program (ANWAP).
The ice model was as described in section 3.6.4.2; however, The model used is the coupled ice-ocean model of the US
the sensible heat storage in the ice was neglected. The model Naval Research Laboratory. This model consists of a Bryan-
resolution is 2° by 2°. This model has been adopted into Cox ocean model coupled with a Hibler ice model. The
a limited area Arctic–Nordic seas model by Holland et al. modeling work is reported in Preller and Edson (1995).
to passive transport with the currents, a simple formulation

Russia Ob is used to remove a fraction of the concentration from water
Yenisey
Pechora into the sediments. He concludes both for 137Cs and 239Pu
that a serious contamination of adjacent seas is unlikely. Fig-
ure 3·45 shows the modeled concentration of 137Cs six years
after release east of Novaya Zemlya.
Canada’s Institute of Ocean Science (IOS) uses a prognostic
Sellafield ocean model coupled with a sea-ice model to investi-gate how
Alaska
a tracer is transported within the Arctic Ocean. A novel fea-
Greenland ture of the IOS model is the representation of the subgridscale
eddies as a driving force in the mean circulation rather than as
traditional eddy viscosity. This eddy force (‘neptune’) was de-
Canada scribed by Holloway (1992) and implemented in models by
Alvarez et al. (1994). Other features of the IOS model include
the use of flux corrected transport (FCT) after Gerdes et al.
(1991). The model includes inflow at the Bering Strait, pas-
sages through the Barrow and Nares Straits in the Canadian
Archipelago, and an open boundary in the Greenland Sea.
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 PCi/L Forcing by seasonal wind and buoyancy is applied for 120
years, at which time a tracer source is ‘turned on’ in the Nor-
wegian Atlantic Current, then followed for 15 years. This rep-
Figure 3·44. Distribution of radioactivity (PCi / L) in the surface layer resents a simplified European radionuclide source. Subsequent
of the ocean after ten years of simulation of Preller’s model. Source loca- evolution can be compared with the tracer observations.
tions are the Ob, Yenisey. and Pechora Rivers, as well as Sellafield.
(Source: Preller and Edson 1995). Profound differences between the models are seen in col-
umn-integrated tracer burdens (Figure 3·46) and in vertical
The model area contains the Arctic Ocean, the Nordic sections (Figure 3·47). Even though diffusion was made as
Seas, and the subpolar gyre in the North Atlantic and parts small as possible (while not allowing large negative concen-
of the North Pacific as shown in Figure 3·44. The main con- trations), usual modeling (Figure 3·46) produces a diffuse
taminant sources are the rivers Ob and Yenisey in the Kara picture in an unrealistic circulation. In particular, Fram
Sea, Pechora in the Barents Sea, and the Sellafield plant in Strait inflow is too weak along the Eurasian slope and trans-
the Irish Sea. The contaminants are transported as passive port north of Greenland is westward, counter to observed
tracers in an Eulerian way. A Lagrangian approach is used circulation (Rudels et al. 1989, Schlosser et al. 1995b). In
to look at contaminant transport by the ice. For the Nordic contrast, when the model includes neptune effect and FCT
and Barents Seas, the modeled contribution from Sellafield (Figure 3·46), there is persistent eastward flow all along the
agrees with observed radiation levels. For the Kara Sea, the Eurasian slope, returning flow along the Lomonosov Ridge,
local river sources must be added to obtain high enough ra- cyclonic circulation in the Makarov Basin and an anticyclo-
diation levels. Figure 3·44 shows the levels of radiation in nic flow around the Chukchi Plateau/Northwind Ridge.
the surface with both Sellafield and the river sources after North from Greenland, mid-depth flow is eastward. (Here, a
10 years. caution is warranted in that these remarks are based upon
More regional work for the Kara and Barents Seas has studies from the particular IOS model using geopotential co-
been done by Harms (1997). He uses the Hamburg Shelf ordinates. Another model, using isopycnal layers (Holland et
Ocean Model (HamSOM) documented by Backhaus (1985), al. 1995) would be expected to behave differently. Careful
together with an ice model based on Hibler’s model. The model intercomparisons remain to be done).
Eulerian transport model is used to study the dispersion The tracer distribution shown in Figure 3·46 is more con-
from waste dump sites east of Novaya Zemlya. In addition sistent with flows inferred from current meter data (Aagaard
80°
1989) and water properties (Rudels et al. 1994, McLaughlin
et al. 1995, Schlosser et al. 1995b). Ice motion (not shown)
resembles the synthesis by Rigor (1992). Perhaps most im-
portantly, Figure 3·46 is consistent with what is known
about the entry of European reprocessing radionuclides into
the Arctic Ocean (Livingston 1988) and their distributions
Novaya
Zemlya as revealed by recent submarine and icebreaker sections
75° (Ellis et al. 1995, Smith et al. 1995).
Tsivolky Bay Figure 3·47 shows a vertical section through the North
Pole along Y = 0 in Figure 3·46. Striking differences between
Novaya Zemlya Trough usual modeling and the model with FCT/neptune (Figure 3·47)
Stepovogo Bay
are seen in stronger tracer inflow from the Fram Strait and a
Abrosimov Bay
return flow along the Lomonosov Ridge. A tracer core (Fig-
ure 3·47) has circulated around the Laptev slope and is re-
70° turning along the Lomonosov Ridge.
Velocities, normal to Y = 0, are shown in Figure 3·47.
Bq/m3 Both models produce westward flow over the Eurasian Ba-
sin. However, the FCT/neptune model exhibits two impor-
50° 60° 70° 80° 90° 100°
tant differences. A strong eastward flow along the Eurasian
Figure 3·45. Yearly and depth-averaged concentrations (Bq / m3) of 137Cs
after 6 years from Harms’ model. The sources are instantaneous releases
slope displaces westward flow to the interior, and flow along
of 1 PBq from the dump sites in Abrosimov Bay, Stepovogo Bay, Tsivolky the Amundsen flank of the Lomonosov Ridge is evident
Bay, and Novaya Zemlya Trough (Source: Harms 1997). throughout the water column.
a (a)
500
200
10 1000
2000
2000
0
20
500
3000
Y
200
4000
0
2000
3000
20
102 0
0000
0
200
200
0
50
-10
2000
00
200
30
-15 0 15
b (b)
500
200
10 1000
2000
2000
0
20
500
3000
Y
200
4000
0
2000
3000
20
102 0
0000
0
200
200
0
50
-10
2000
00
200
30
-15 0 15
X
0 3.0·10 -13 6.0·10 -13 9.0·10 -13 1.2·10 -12 1.5·10 -12
Figure 3·46. Vertically integrated tracer burden shown after 15 years. Units are burden per unit surface area, normalized by cumulative release. Bottom
topography is in meters; a) no neptune, centered difference, and b) neptune, flux-corrected transport (Source: Alvarez et al. 1994).
Level Depth Level Depth

(m) (m)
0 0
5 122.5 5 122.5
10 497.5 10 497.5
15 1122.5 15 1122.5
20 1997.5 20 1997.5
25 3122.5 25 3122.5
(a) (c)
30 4487.5 30 4487.5
0 10 20 0 10 20
Level Depth Level Depth
(m) (m)
0 0
5 122.5 5 122.5
10 497.5 10 497.5
15 1122.5 15 1122.5
20 1997.5 20 1997.5
25 3122.5 25 3122.5
(b) (d)
30 4487.5 30 4487.5
0 10 20 0 10 20
X out X in
0.0 0.2 0.4 0.6 0.8 1.0 -8.0 -5.6 -3.2 -0.8 1.6 4.0
Tracer concentration Velocity in cm/s
Figure 3·47. Tracer concentration on a vertical section (along Y = 0 on Figure 3·46) after 15 years; a) no neptune, centered difference; b) neptune, flux-
corrected transport; c,d) velocity normal to the section with positive ‘into’ the page, negative ‘out’ of the page; c) no neptune, centered difference; and,
d) neptune, flux-corrected transport (Source: Alvarez et al. 1994)
3.7. Summary and information gaps The most accurate models for studying atmospheric
transport pathways are those using globally or hemispheri-
3.7.1. Atmospheric pathways
cally gridded meteorological fields based on climatic obser-
Atmospheric pathways are important in the overall context vations. The density of meteorological observations in the
of the delivery of contaminants to the Arctic. The atmosphere Arctic is lower than elsewhere, due to the remoteness and
is the most important pathway relative to ocean and terres- high cost of operating these stations, and consequently, the
trial/freshwater pathways for one-hop compounds (acids, accuracy of the predictions is reduced. These models have
metals except mercury, involatile organics). Given the cur- proven invaluable in that they can mathematically describe
rent configurations of anthropogenic sources at mid-latitudes, physical/chemical processes that help explain observed at-
the most favored pathway into the Arctic is from the Eura- mospheric levels of contaminants and inter-pathway ex-
sian continent from November to May. The relative impor- changes.
tance of atmospheric transport compared to marine and ter- Physical/chemical processes in the polar atmosphere are
restrial/freshwater is very contaminant-specific for the multi- unique, and not as well understood or documented as those
hop compounds, such as OC pesticides, Hg, and PCBs. While for temperate regions. Some of the processes that need to be
all of the compartments (Fig. 3·1) play a role in transporting understood better to improve the modeling capability and
these contaminants, the speed of transfer through the at- the general understanding of contaminant transport include:
mosphere suggests that this compartment is particularly im-
• cloud, fog, and precipitation scavenging of contaminants
portant in the global cycling of these types of compounds.
in the Arctic during the summer when there is extensive
The airshed for the Arctic is northern hemisphere for one-
low-level cloud and light precipitation under conditions of
hop contaminants, but global for multi-hop contaminants.
continuous daylight;
Models exist that quantitatively simulate atmospheric
• chemical transformation of contaminants in air under
pathways. These have been derived from general circulation
Arctic conditions of temperature and sunlight;
models, which have required extensive international coop-
• scavenging of contaminants by Arctic precipitation under
eration, resulting in models that cover a large geographic
low temperature conditions, especially the role of large
area, including the Arctic. Thus, it has been a natural exten-
surface area snow crystals that descend slowly through
sion to apply these models to understanding contaminant
the air column;
transport from source regions to the Arctic.
• air-surface exchange of contaminants to better under- ported to and within the freshwater system. At present, dis-
stand the recycling possibilities of multi-hop compounds charge and sediment concentration in freshwater systems are
under Arctic conditions; and, reasonably well documented. Estimates of contaminant
• gas-particle phase partitioning within the atmosphere, loads from rivers to the oceans, on the other hand, are quite
and the role of temperature and other conditions in con- crude, being dependent upon a very limited set of observa-
trolling this process. tions and lacking a large-scale modeling capability. The pau-
city of site-specific and larger-scale studies in the Arctic lim-
Arctic atmospheric pathway studies suffer from a paucity of
its the understanding of processes and determination of
observations, both spatially and temporally. This is especi-
mass transport. Site-specific investigations of contaminant
ally a problem with respect to multi-phase studies, such as
transport and fate in Arctic freshwater systems are required
the simultaneous measurement at the surface (land, fresh-
to understand the basic processes and to evaluate and cali-
water, and marine, including ice surfaces for both lakes and
brate site-specific and larger-scale models. For example,
oceans) and through the air column. This deficiency can
studies are needed on freeze and thaw effects on snowpack
only be addressed through greater use of fully integrated
release; permafrost effect on infiltration and on ground and
observational platforms (i.e., surface-based stations com-
drainage waters; humic acids and their influence on adsorp-
bined with aircraft and ships). In some cases, these observa-
tion of hydrophobic compounds; and, inorganic particles
tions would need to be augmented by surface-level routine
and sorption processes. Larger-scale studies and monitoring
observations from additional locations using observatories
are required to more precisely measure the total mass of
and buoys. In addition, the intensive observations would
contaminants delivered to oceans and the temporal distribu-
have to be supplemented by remote sensing techniques (e.g.,
tion of this delivery, as well as to characterize the particulate
satellite imagery, lidars, etc.) as this is the only means of
and dissolved phase contributions. This information is nec-
providing broad spatial coverage cost-effectively.
essary in order to improve the estimation of the transport of
hydrophobic and soluble contaminants from the land sur-
face to headwater systems using large-scale mass transport
3.7.2. Terrestrial/freshwater pathways models (e.g., the River System Simulator), which have not
The deposition of contaminants from the atmosphere through been used extensively for the Arctic.
snow, rain, and dry fallout is a major regional-scale source Our knowledge of transport and levels of contamination
of contamination in the Arctic. However, the processes con- of surface waters in the Arctic is restricted, but even less is
trolling scavenging, fate, and release are poorly understood. known about processes and levels in drainage waters and
Since the atmosphere serves as a relatively homogeneous their discharge to Arctic rivers. The transport and fate of
source, spatial variability of contaminant deposition seems contaminants in drainage waters are probably partly con-
to be largely controlled by the precipitation fluxes to the trolled by similar variables as those controlling surface wa-
land and ocean surfaces. Precipitation measurements are ters (e.g., organic matter, pH, ions), but these processes, as
spatially limited in the Arctic, especially over the ocean, and well as the extent or potential for contamination of Arctic
remotely sensed data are currently imprecise. Consequently, groundwater, have hardly been investigated. Such knowl-
mass fluxes of contaminants to the land and ocean surfaces edge would be most useful for areas where local contamina-
are difficult to estimate at present. Improved sampling de- tion sources are a concern.
vices for contaminants in precipitation, developed for Arc- There has been no investigation of sediment resuspension
tic climatic conditions, together with better accounting of in Arctic lakes in the context of contaminants transport. As
precipitation accumulation in remote areas, would enhance well, there are few investigations of shallow lakes (2-3 m
the accuracy of deposition information. deep), which are the most common in the Arctic. Their heat
Just as the processes that control scavenging from the at- budget (warmer and not stratified during the summer, but
mosphere are poorly understood and quantified, so too are often totally frozen in the winter) is likely to influence the
the physical and chemical processes that control the exchange fate of contaminants, but this has not been investigated.
between snowpacks and the atmosphere. For example, only Ice cover is an important phenomenon in the Arctic. It
recently has it been recognized that there is a large differ- plays a role in freshwater systems, especially with respect to
ence between the air/snowpack exchange of shallow and oxygen exchange and consumption, which in turn could in-
deep snowpacks. The fate of contaminants in cold and fluence the fate of contaminants in the environment. Al-
metamorphosing snowpacks is largely controlled by snow- though not studied to any great extent, it is clear that ice
melt and refreezing, solar radiation, changing snow particle will form a reasonably impermeable barrier between the at-
size and surface area, and exchange with the atmosphere mosphere and the water for the majority of the year. This
and surface soil. These processes have only begun to be in- obviously limits the exchange of atmospheric pollutants
vestigated. Field studies provide evidence that up to 80% of with the water, but also retains any pollutants introduced
the total solute load in the snowpack is released with the below the ice within the limited water volume. Sources of
first 20-30% of meltwater. Laboratory studies indicate that, contaminants that could be a concern include contaminated
whereas dissolved pollutants (organic compounds with high groundwater drainage, industrial and municipal discharges
water solubility) are rapidly dispersed into the water at the below the ice, and natural seeps of hydrocarbons such as
onset of melting, those that are adsorbed (to particles) are occur in the Mackenzie River. Also, the role of ice in remo-
more recalcitrant, and only leave the snow with the final bilizing contaminants in rivers and in delivering contami-
meltwater. An understanding of the rates of these processes nants to the ocean has received minimal attention. The role
and the factors controlling them is necessary in order to of ice in controlling exchanges and the magnitude of conta-
model the delivery of contaminants to the aquatic system minant loadings by ice should be evaluated to determine if
from land and ice runoff. This is a clear gap in our under- more extensive work is required.
standing of contaminant delivery to the oceans from the Hydrophobic contaminants sorb to particles and their
terrestrial/freshwater compartment. transport is strongly related to the quantity and nature of
The contaminants delivered to the Arctic and deposited suspended loads in rivers. Areas with reduced water velocity,
to the terrestrial/freshwater environment are readily trans- such as lakes, floodplains, and freshwater deltas, are possi-
ble deposition zones for contaminants. The quantities and relatively little dilution over remarkably large distances in
physical aspects of sediment transport are relatively well the Arctic – more so than in any other ocean basin. Our un-
known. This is important for determining the relative con- derstanding of the role of this redistributed river water in
tribution of the different phases of contaminants (dissolved long-range contaminant transport is limited, as is the influ-
or associated with particles and ice) transported in the river. ence of variable riverine discharge on ocean circulation.
However, the concentrations of contaminants in the differ- The Arctic is not in a steady-state, but how it oscillates is
ent phases and their spatial and temporal variability remain still uncertain. Because most data are from the summer and
largely unknown. have been obtained relatively recently, there is less known
Much emphasis has been placed on the pathways for about winter conditions and interannual variability. The Beau-
long-range transport of contaminants, nevertheless, the fort Gyre seems to shift between a small gyre and a larger
contribution of local Arctic sources of contaminants to the one, but the causes of these changes and their effects are un-
total contaminant transport budget requires further investi- known. The Atlantic layer seems to have warmed by about
gation. In Canada, this seems to be only a local concern, 0.5-1°C during recent years. The invasion of warm water,
whereas in Russia, Arctic sources may be a major compo- following boundary currents around the Eurasian and Cana-
nent of the total budget. Many large river basins draining dian Basins, is accompanied by a shift in the front separating
to the Arctic extend well into temperate latitudes in both the Atlantic and Pacific water masses in the upper part of
Canada and Russia. Investigations in Canada have shown the Arctic Ocean. The fate of this displaced Pacific water, in-
that there is little industrial and agricultural impact on Can- cluding its pathway out of the Arctic Ocean, is not known.
adian river systems flowing to the north; however, in Rus- Transport pathways and mechanisms are not the same for
sia, there is much greater potential for both industrial and all contaminants. Some contaminants tend to become bound
agricultural discharges to directly affect the quality of wa- to particles while others remain in the dissolved phase. Since
ter. Some information is available for industrial and munici- particles tend to sediment and therefore take different path-
pal sources, but little seems to be known about the agricul- ways than dissolved constituents, it is crucial to understand
tural impacts on these river systems. both the phase geochemistry of the contaminant and the
ocean processes of particle production and vertical flux. Pre-
sently, there are few data with which to estimate the relative
3.7.3. Ocean pathways importance of particles in the transport of contaminants and
Our schematic knowledge of the general circulation in the the movement of contaminants from surface water and
Arctic Ocean has improved over the past decade due espe- coastal zones to deeper parts of the ocean.
cially to tracer studies carried out during basin transects. Sea ice acts as a lid on the surface of the ocean, hindering
However, there are still large gaps in our knowledge of the ocean-atmosphere exchange of contaminants. Clearly, ice
current structure in the Makarov and Canada Basins, of the controls the seasonal cycle of semivolatile constituents in the
exchange between the ocean margins and the interior, of the ocean surface, but measurements are limited.
coupling between basins, and of the relative importance of Recent studies indicate that sea ice also plays a role in
the Canadian Arctic Archipelago and the Fram Strait for transport and redistribution of contaminants, although little
surface water exiting the Arctic. Tracer fields often give a concrete information is available on the quantitative signifi-
direct measure of the rate of propagation of a property and, cance of this pathway, as detailed below. The process of re-
in the most widely cited example, radionuclide tracers from taining particles/contaminants during transport and releas-
the European reprocessing plants have shown dramatically ing them at the sea surface far away makes sea ice a unique
that transport times are much faster than earlier believed. transport mechanism.
But tracer fields often do not provide good estimates of vol- Many dissolved pollutants are excluded with salt during
ume flows. Poorly constrained volumetric flows, together the freezing process, thus, sea ice without incorporated se-
with a very poor knowledge of the contaminant concentra- diments or organic material may be less contaminated than
tions (particulate and dissolved), mean that, at present, only the water from which it was formed. However, much of the
approximate guesses of quantitative transport rates are pos- ice formed in shallow regions of the Siberian seas entrains
sible within the Arctic Ocean. Although volumetric flows resuspended sediments and organic material, and may there-
are better constrained at the entrances and exits from the fore incorporate associated contaminants. Some of this ice is
Arctic Ocean, there is still considerable uncertainty and exported from the shelf and transported over thousands of
variance, and the contaminant concentrations are, as yet, kilometers. The relative importance of atmospheric deposi-
poorly determined or not determined at all. Therefore, we tion versus marine and river sources of contaminants to sea
need better documentation of volume transports of ice and ice is not clear. As well, information is lacking on the signifi-
water, and their contaminant burdens into, out of, and cance of contaminated ice being exported from a large re-
within the Arctic. gion where it is formed, into a smaller region where it dissi-
Due to ice cover, circulation in the central Arctic Ocean pates and/or is exported.
has not been as well documented as in the adjacent seas. Due to continued atmospheric deposition of aerosols and
For example, the amount of water exchanged between the condensates during transport, and perhaps entrainment from
Beaufort Gyre and the Transpolar Drift is not well quanti- the surface ocean microlayer, contaminant loads may in-
fied. This is important since the Beaufort Gyre can retain a crease in older sea ice. The fate of contaminants deposited to
contaminant and recirculate it for many years, while conta- sea ice in snow as the snow melts is unknown. Some portion
minants in the Transpolar Drift exit the Arctic Ocean in may be lost in meltwater runoff, some may volatize, and
just a few years. Similarly, water leaving the Beaufort Gyre some may adsorb to sediments in the ice, but the distribu-
through the Canadian Archipelago could be important for tion among these is unknown.
contaminant flux out of the Arctic. Most particles and associated contaminants are probably
The Arctic Ocean is more dynamically coupled to its sur- released at the sea surface along the marginal ice zone. Be-
rounding seas than believed earlier. Little is known about cause of biological activity in this region, it is possible that
the variability in the different seas and how they affect the contaminants released here enter the food chain. Sea-ice
main system. For example, river runoff is transported with transport could provide a link between pollutant source
areas and distant wildlife. The biological communities that flux studies. Through the use of multi-media models, a
seem to be most at risk from long-range pollutant transport great deal could be achieved efficiently by fully integrat-
by sea ice are those along the marginal ice zones, where ing processes, observed levels, and trends. In other words,
large amounts of ice melt. Such areas are found in the Bar- integrating the observed environmental measurements
ents and Greenland Seas, the Iceland Sea, the west coast of with the pathways and sources information, so as to pro-
Greenland and Baffin Bay, and to a lesser degree the other vide assessment and feedback on the design and imple-
marginal seas. The degree to which sea-ice transport of con- mentation of future monitoring and process research ac-
taminants contributes to body burdens in these regions is tivities, as well as management and/or mitigation mea-
not known. sures. Models are also essential in assessing the sensitivity
Presently, models of ocean and ice circulation are often of individual or linked processes, which in turn can be in-
of too low resolution to resolve boundary currents and ed- structive in assigning priorities to often complex research
dies, both of which are likely to be the main modes of cont- questions. In particular, models have not been used to
aminant transport. Furthermore, contaminants enter the their fullest in the area of terrestrial/freshwater pathways.
models only as passive tracers, and future modeling efforts There is also a need for integrated models to help assess
need to incorporate more realistic contaminant properties complex questions, including evaluation of the relative
including uptake by biota, particle scavenging, and kinetic importance of processes, estimation of transport fluxes,
losses. and assessment of remedial measures.
3.7.4. Conclusions 3.7.5. Recommendations

Following from the summary and information gaps discus- The overriding conclusion from this work is the general in-
sed above, it can be concluded that a number of steps are ability to quantify fluxes and rates on a circumpolar scale.
required to better understand the pathways of contami- Consequently, these recommendations are intended to en-
nants into and within the Arctic. These are presented below hance our understanding of the inter- and intra-compart-
as general conclusions which are relevant to all compartmental delivery of contaminants by providing guidance for
ments: future system-based monitoring and research activities. This
can be approached in two ways, either through the evalua-
1. Tracers are important in determining and confirming
tion of pathways to determine 1) ‘contaminant focusing
pathways. To date, these have been used in the ocean
zones’ or 2) ‘zones of influence’ of known source regions.
pathway studies (e.g., radionuclides released from Sell-
These concepts are illustrated in Figure 3·48a and b (next
afield), but greater use needs to be made of natural and
page), respectively.
anthropogenic tracers in all compartments. Emphasis
should be placed on tracers that mimic the major conta- 1. Accordingly, it is recommended that contaminant path-
minants, or are representative of groups of contaminants, ways be evaluated in a comprehensive and systematic way
thereby enhancing pathway studies, as well as perhaps to determine contaminant focusing zones as a result of long-
reducing their cost. Selected PAHs, PCB congeners, sta- range transboundary transport through one or more envi-
ble isotopes, and a range of other contaminants could ronmental compartments. Specific focusing zones include:
prove to be very useful in this regard and need to be ex- • atmospheric and ocean conduits and deposition areas;
amined more fully. • marginal ice zones and polynyas; and,
2. Arctic processes are largely understood qualitatively, • zones of increased precipitation (e.g., coastal and moun-
however, the ability to quantify them is inadequate. In tainous areas).
particular, it is essential that transport processes and
their relative importance or magnitude within and be- This approach will help to account for existing areas of rela-
tween compartments (air, terrestrial, water, ice, sedi- tively elevated contaminant levels and to identify other po-
ments, biota) be determined. There is a need for prioriti- tential zones of impact based on the physical/chemical data.
zation to ensure that the most important processes are Some examples to illustrate what is meant by contami-
investigated within the context of contaminant delivery nant focusing zones include:
to the Arctic environment. • Atmospheric pathway convergence and terrestrial/fresh-
3. In general, there is a qualitative understanding of the water focusing: The mountain range on the west coast of
transformation and fate of different contaminants or North America focuses the deposition of atmospheric
groups of contaminants under varying Arctic conditions. contaminants through precipitation. These are then car-
However, detailed information and the ability to quan- ried through runoff into sensitive drainage systems result-
tify these processes is lacking. While a considerable ing in accumulation in lakes;
amount of information is available on this subject for • Estuaries: In productive environments, estuaries act as a
many of the contaminants under temperate conditions, trap for sediments and contaminants, especially where
there is still little or no understanding as to how, for ex- nearby sources contribute contaminants to river or ocean
ample, temperature and light affect the rate of these ice and water; and,
processes. In order to be able to fully assess the magni- • Marine convergence: This may be an important factor for
tude and direction of pathways, a better understanding a number of areas, as suggested by both pathway and con-
of transformation and fate of pollutants in the Arctic is taminant residue information. For example, in the Sval-
required. bard area, air masses, precipitation scavenging, ocean cur-
4. Ice is a dominant multi-compartment medium of the rents, and sea ice may all interact to focus contaminants.
Arctic. However, the role of ice in modulating contami- In the Canadian Arctic Archipelago, the flux of water and
nant fate and controlling pathways is poorly understood. contaminants out of the Arctic Ocean takes place almost
This is an important gap. entirely in the upper 100 m and is consequently coupled
5. There is ample evidence of the limited effectiveness of with air exchange and freshwater inputs which together
sporadic and poorly integrated process, pathway, and could uniquely impact upon the local marine ecology.
• the Murman area, for which the long-range zone of influ-

(a)
ence is not well known, despite having large industrial
Iceshed and municipal atmospheric emissions and wastewater dis-
Watershed charges with extensive local and regional effects;
• the fate and impact of contaminants from regional
sources upon the Kara Sea; and,
• the extent of impact from point sources within the Arctic
(e.g., DEW Line sites) relative to long-range sources.
Pollutant source In order to fulfill the intent of these recommendations, inves-
region tigations of both contaminant focusing zones and the zones
of influence of source regions will have to consider the use-
Bioshed fulness of tracers, quantify the main processes, assess trans-
formation and fate in the specific situation, consider the
unique conditions of the Arctic, such as ice, and fully inte-
grate the use of multi-media models. Development of com-
prehensive investigations to address these recommendations
will require further definition and development within the
context of the information provided on contaminant levels
and the effects of specific contaminants as outlined in subse-
quent chapters.
Airshed
Acknowledgments
Editors
Dennis J. Gregor, Harald Loeng, Len Barrie.
(b)
Authors
Iceshed Watershed Len Barrie, Eva Falck, Dennis J. Gregor, Trond Iversen,
Harald Loeng, Rob Macdonald, Stephanie Pfirman,
Trond Skotvold, Elleke Wartena.
Contributors
B. Ådlandsvik, A.C. Brown, M. Diamond, P. Fellin,
V. Gordeev, S. Häkkinen, W. Hart, G. Holloway, V. Kim-
Area of
contaminant stach, L. Lockhart, H. Martin, K. Puckett, W. Schroeder,
focusing G. Stephens, W. Strachan, H. Svendsen, V. Vuglinski,
Bioshed B. Welch, A. Zhulidov.
Reviewers
Oceans: K. Aagaard, B. Hansen, S. A. Malmberg, T. Vinje,
G. Björk, S. Carlberg, E. Reimnitz, T. Rossby, T. Tucker
Atmosphere: W. Schroeder
Terrestrial/Freshwater: T. Prowse, M. Jansson
Entire chapter: D. Landers, M. Bewers.
References
Airshed
Aagaard, K., 1968. Temperature variations in the Greenland Sea deep
water. Deep-Sea Res. 15: 281-296.
Aagaard, K., 1981. On the deep circulation of the Arctic Ocean. Deep-Sea
Res. 82: 251-268.
Figure 3·48. Illustration of the concept of a) the ‘zone of influence’ Aagaard, K., 1984. The Beaufort undercurrent. In: P.W. Barnes, D.M.
through major contaminant pathways from a major source region, and b) Schell and E. Reimnitz (eds.). The Alaskan Beaufort Sea, ecosystems
the ‘contaminant focusing zone’ resulting from the convergence of path- and environments, pp. 47-71. Academic Press, Orlando.
ways carrying contaminants from a variety of sources to a single zone. Aagaard, K., 1989. A synthesis of the Arctic Ocean circulation. Rapp. P.-
v. Réun. Cons. int. Explor. Mer 188: 11-22.
Aagaard, K., 1994. Contamination of the Arctic: Ocean processes. Arct.
2. Subsequently, it is also recommended that the zones of Res. U.S. 8: 21-33.
influence of known regional or local sources be investigated Aagaard, K. and E.C. Carmack, 1989. The role of sea ice and other fresh
water in the arctic circulation. J. Geophys. Res. 94(C10): 14485-14489.
based upon an understanding of processes, pathways, and Aagaard, K. and E.C. Carmack, 1994. The Arctic Ocean and climate: a
flux rates. This will lead to the identification of local, or at perspective. The polar oceans and their role in shaping the global envi-
most regional zones of influence for specific sources of con- ronment. Am. Geophys. Un. 85: 5-20.
Aagaard, K. and L.K. Coachman, 1975. Toward an ice-free Arctic Ocean.
taminants. EOS 56: 484-486.
The first priority in this exercise would be to understand Aagaard, K. and P. Greisman, 1975. Towards new mass and heat budgets
the zone of influence of sources locally, regionally, and over for the Arctic Ocean. J. Geophys. Res. 80: 382l-3827.
Aagaard, K. and S.-A. Malmberg, 1978. Low frequency characteristics of
long distances, pointing to areas for which impact studies the Denmark Strait overflow. ICES C.M. 1978 C: 47. International
may be required. Second and subsequent priorities would Council for the Exploration of the Sea, Copenhagen.
have to consider information needs with respect to pathways. Aagaard, K. and A.T. Roach, 1990. Arctic Ocean-shelf exchange: Mea-
surements in Barrow Canyon. J. Geophys. Res. 95(C10): 18163-18175.
Some examples which illustrate the applicability of determin- Aagaard, K., L.K. Coachman and E.C. Carmack, 1981. On the halocline
ing zones of influence of contaminant sources include: of the Arctic Ocean. Deep-Sea Res. 28: 529-545.
Aagaard, K., J.H. Swift and E.C. Carmack, 1985. Thermohaline circula- position of glaciers. Ann. Glaciol. 7: 100-108.
tion in the Arctic Mediterranean seas. J. Geophys. Res. 90: 4833-4846. Barrie, L.A., 1986. Arctic air pollution flow of current knowledge. Atmos.
Aagaard, K., E. Fahrbach, J. Meincke and J.H. Swift, 1991. Saline out- Environ. 20: 643-663.
flow from the Arctic Ocean: Its contribution to the deep waters of the Barrie, L.A., 1991. Snow formation and process in the atmosphere that
Greenland, Norwegian, and Iceland seas. J. Geophys. Res. 96: 20433- influence its chemical composition. In: T.D. Davis, M. Tranter and
20441. H.G. Jones (eds.). Seasonal Snowpacks. NATO ASI Series, Volume 28,
Abelmann, A., 1992. Diatom assemblages in Arctic sea ice – indicator for pp. 1-20. Springer-Verlag, Berlin and Heidelberg, Germany.
ice drift pathways. Deep-Sea Res. Special Issue 39: S525-S538. Barrie, L.A., 1992. Scavenging ratios: black magic or a useful scientific
Ackermann, N.L., H.T. Shen and B. Sanders, 1994. Experimental studies tool. Precip. Scavenging Atmos.-Surface Exchange 1: 403-415.
of sediment enrichment of Arctic ice covers due to wave action and Barrie, L.A., 1995. Arctic aerosols: composition, sources and transport in
frazil enrichment. J. Geophys. Res. 99(C4): 7761-7770. ice core studies of global biogeochemical cycles. In: R.J. Delmas (ed.).
Ackley, S.F., 1982. Ice scavenging and nucleation: Two mechanisms for Global Environmental Change. NATO ASI Series, Volume 30, pp. 1-22.
incorporation of algae into newly-forming ice. EOS Trans. Am. Geo- Springer-Verlag, Heidelberg, Germany.
phys. Un. 63: 54-55. Barrie, L.A., 1996b. Occurrence and trends of pollution in the Arctic tro-
Ahlnäs, K. and G.R. Garrison, 1984. Satellite and oceanographic obser- posphere. In: E. Wolff and R.C. Bales (eds.). Chemical exchange be-
vations of a warm coastal current in the Chukchi Sea. Arctic 37: 244- tween the atmosphere and polar snow. Global Environmental Change.
254. NATO ASI Series I, Volume 43, pp. 93-130. Springer-Verlag, Heidel-
Alabyan, A.M., D.B. Babich and A.L. Bogomolov, 1991. Recent process- berg, Germany.
es of delta formation and history of the Yenisey Delta development. Barrie, L.A., 1996a. Personal communication. Atmospheric Environment
VINIT1, Moscow. (In Russian). 151p. Service, Environment canada, Downsview, Ontario, Canada.
Alabyan, A.M., R.S. Chalov, V.N. Korotaev, A.Y. Sidorchuk and A.A. Barrie, L.A. and R.S. Schemenauer, 1986. Pollutant wet deposition mecha-
Zaitsev, 1996. Natural and technogenic components of water and nisms in precipitation and fog water. Water Air Soil Pollut. 30: 91-104.
sediment supply to the Laptev sea. Ber. Polarforsch. (in press). Barrie, L.A. and R.S. Schemenauer, 1989. Wet deposition of heavy metals.
Albright, L.J., K.V. Masuda, G.L. Ennis and H. Schreier, 1980. Microbial In: J.M. Pacyna and B. Ottar (eds.). Control and fate of atmospheric
dynamics of two sub-Arctic Canadian rivers. Water Resour. Res. 14(9): trace metals. NATO ASI Series C 268, pp. 203-231. Kluwer Academic
1353-1362. Publishers, Dordrecht, The Netherlands.
Allan, R.J., 1979. Heavy metals in bottom sediments of Great Slave Lake Barrie, L.A., M.P. Olson and K.K. Oikawa, 1989b. The flux of anthropo-
(Canada): A reconnaissance. Environ. Geol. 3: 49-58. genic sulphur into the Arctic from mid-latitudes in 1979/80. Atmos.
Allen, J.R.L., 1970. Physical processes of sedimentation: An introduction. Environ. 23: 2505-2512.
In: J. Sutton and J.V. Watson (eds.). Earth Science series No.1, 248p. Barrie, L.A., D. Gregor, B. Hargrave, R. Lake, D. Muir, R. Shearer, B.
Unwin University Books. George Allen and Unwin Ltd., London, U.K. Tracey and T. Bidleman, 1992. Arctic contaminants: sources, occur-
Altshuler, V.M., Y.V. Sustavov and Y.M. Kazakova, 1970. On the energy rence and pathways. Sci. Total Environ. 122: 1-74.
and water exchange between the White and Barents Seas. Trudy Barrie, L.A., R. Macdonald, T. Bidleman, M. Diamond, D. Gregor, R.
Polyarn. Naucho-Issled. Inst. Morsk. Ryb. Khoz. Oceanogr. 27: 114- Semkin, W. Strachan, M. Alaee, S. Backus, M. Bewers, C. Gobeil, C.
118. (In Russian). Halsall, J. Hoff, A. Li, L. Lockhart, D. Mackay, D. Muir, J. Pudykie-
Alvarez, A., J. Tintore, G. Holloway, M. Eby and J.M. Beckers, 1994. Ef- wicz, K. Reimer, J. Smith, G. Stern, W. Schroeder, R. Wagemarm, F.
fect of topographic stress on the circulation in the western Mediter- Wania and M. Yunker, 1997. Chapter 2. Sources, occurrence and
ranean. J. Geophys. Res. 99: 16053-16064. pathways. pp 25-182. In: J. Jensen, K. Adare and R. Shearer (eds.).
Anderson, L.G. and E.P. Jones, 1986. Water masses and their chemical Canadian Arctic Contaminants Assessment Report. Indian and North-
constituents in the western Nansen Basin of the Arctic Ocean. Oceanol. ern Affairs Canada, Ottawa. 460p.
Acta 9: 277-283. Baskaran, M. and A.S. Naidu, 1995. 210Pb-derived chronology and the
Anderson, L.G., E.P. Jones, R. Lindegren, B. Rudels and P.-I. Sehlstedt, fluxes of 210Pb and 137Cs isotopes into continental shelf sediments, East
1988. On the chemistry of the cold, high salinity bottom waters of Chukchi Sea, Alaskan Arctic. Geochim. Cosmochim. Acta 59: 4435-
the Arctic Ocean shelves. Cont. Shelf Res. 8: 1345-1355. 4444.
Anderson, L.G., G. Björk, O. Holby, E.P. Jones, G. Kattner, K.P. Kolter- Beltaos, S., D.J. Calkins, L.W. Gatto, T.D. Prowse, S. Reedyk, G.J. Scrim-
mann, B. Liljeblad, R. Lindegren, B. Rudels and J. Swift, 1994. Water geour and S.P. Wilkins, 1993. Physical effects of river ice. In: T.D. Prowse
masses and circulation in the Eurasian Basin: results from the Oden and N.C. Gridley (eds.). Environmental aspects of river ice, pp. 3-74.
91 expedition. J. Geophys. Res. 99: 3273-3283. Ministry of Supply and Services, National Hydrology Research Insti-
Antonov, V.S., 1970. Siberian rivers and Arctic seas. Probl. Arktiki Ant- tute, Saskatoon, Saskatchewan.
arktiki 36/37: 142-152. Bengtsson, L., 1994. Northern hydrology in Sweden. In: T.D. Prowse,
Arnesen, R.T. and E.R. Iversen, 1995. Transport of heavy metals from C.S.L. Ommaney and L.E. Watson (eds.). Northern hydrology: Inter-
Norwegian mines. NIVA Report O-94021. NIVA (Norwegian Insti- national perspectives, pp. 109-127. Environment Canada, Saskatoon,
tute for Water Research), Oslo, Norway. (In Norwegian). Saskatchewan, NHRI Science Report No. 3.
Asplin, L., 1995. Examination of local circulation in a wide, stratified Bengtsson, L., T. Hellström and L. Rakoczi, 1990. Redistribution of sedi-
fjord and exchange of water with the adjacent ocean due to constant ments in three Swedish lakes. Hydrobiologia 192: 167-181.
local up-fjord wind. In: H.R. Skjoldal, C. Hopkins, K.E. Erikstad and Benkowitz, C.M., C.M. Berkowitz, R.C. Easter, S. Nemesure, R. Wagener
H.P. Leinaas (eds.). Ecology of fjords and coastal waters. Elsevier and S.E. Schwartz, 1994. Sulfate over the North Atlantic and adjacent
Science B.V., Amsterdam. continental regions: Evaluation for October and November 1986 using
Atkinson, E.G. and J.W. Wacasey, 1987. Sedimentation in arctic Canada: a three-dimensional model driven by observation-derived meteorology.
Parcticulate organic carbon flux to a shallow marine benthic commu- J. Geophys. Res. 99: 20725-20756.
nity in Frobisher Bay. Polar Biol. 8: 3-7. Benson, C.S., 1982. Reassessment of winter preciptiation on Alaska’s Arc-
Backhaus, J.O. 1985. A three-dimensional model for the simulation of tic Slope and measurments on the flux of windblown snow. University
shelf dynamics. Dt. Hydrogr. Z. 38: 165-187. of Alaska, Geophysical Institute, Fairbanks, Alaska, Report No. 288, 26p.
Bacon, M.P., C.-A. Huh and R.M. Moore, 1989. Vertical profiles of some Berner, H. and G. Wefer, 1990. Physiographic and biological factors con-
natural radionuclides over the Alpha Ridge, Arctic Ocean. Earth Pla- trolling surface sediment distribution in the Fram Strait. In: U. Bleil
net. Sci. Lett. 95: 15-22. and J. Thiede (eds.). Geological history of the polar oceans: Arctic ver-
Bailey, W.B., 1956. On the origin of Baffin Bay deep water. J. Fish. Res. sus Antarctic, pp. 317-335. Kluwer Academic Publishers, Dordrecht,
Bd Can. 13: 303-308. The Netherlands.
Bailey, W.B., 1957. Oceanographic features of the Canadian archipelago. Bidleman, T.F., 1988. Atmospheric processes. Environ. Sci. Technol. 22(4):
J. Fish. Res. Bd Can. 14: 731-769. 361-367.
Ballschmiter, K. and M. Zell, 1980. Baseline studies of the global pollu- Bidleman, T.F. and L. McConnell, 1995. A review of field experiments to
tion – l. Occurence of organohalogens in pristine European and Ant- determine air-water gas exchange of persistent organic pollutants. Sci.
arctic aquatic environments. Int. J. Environ. analyt. Chem. 8: 15-35. Total Environ. 159: 101-117.
Barber, F.G., 1965. Current observations in Fury and Hecla Strait. J. Fish. Bleck, R. and D.B. Boudra, 1986. Wind-driven spin-up in eddy resolving
Res. Bd Can. 22: 225-229. ocean models formulated in isopycnic and isobaric coordinates. J.
Barber, F.G., 1967. A contribution to the oceanography of Hudson Bay. Geophys. Res. 91: 7611-7621.
Mar. Sci. Br., Dept. of Mines and Technical Surveys, Ottawa, Ms. Bleck, R., C. Rooth, D. Hu and L.T. Smith, 1992. Salinity-driven thermo-
Rep. Ser., No. 4, 69p. cline transients in a wind-and thermohaline-forced isopycnic coordi-
Barnes, P.W. and E. Reimnitz, 1974. Sedimentary processes on Arctic nate model of the North Atlantic. J. Phys. Oceanogr. 22: 1486-1505.
shelves off the northern coast of Alaska. In: J.C. Reed and J.E. Sater Blindheim, J., 1989. Cascading of Barents Sea bottom water into the
(eds.). The coast and shelf of the Beaufort Sea, pp. 439-476. The Arc- Norwegian Sea. Rapp. P.-v Réun. Cons. int. Explor. Mer 188: 49-58.
tic Institute of North America, Washington D.C. Blindheim, J., 1990. Arctic intermediate water in the Norwegian Sea.
Barnes, P.W., E.M. Kempema and E. Reimnitz, 1990. Source, characteris- Deep-Sea Res. 37: 1475-1489.
tics and significance of sediment pellets formed on the sea ice of the Blindheim, J., 1993. Seasonal variations in the Atlantic inflow to the Nor-
Arctic Basin. In: S.F. Ackley and W.F. Weeks (eds.). Sea ice properties dic Seas. ICES C.M. 1993 C: 39. International Council for the Explo-
and processes: Proceedings of the W.F. Weeks Sea Ice Symposium, pp. ration of the Sea, Copenhagen.
106-108. Cold Regions Research and Engineering Laboratory, Han- Blindheim, J. and H. Loeng, 1981. On the variability of Atlantic influence
over, New Hampshire, CRREL Monograph 90-1. in the Norwegian and Barents Seas. Fiskeridir. Skr. Ser. Havunders. 17:
Barrie, L.A., 1985. Atmospheric particles: their physical and chemical 161-189.
characteristics and deposition processes relevant to the chemical com- Bloom, N. and S.W. Effler, 1990. Seasonal variations in the mercury speci-
ation of Onodoga Lake (New York). Water Air Soil Pollut. 53: 251-265. 1995, Vancouver, B.C., Canada.
Blumberg, A.F. and G.L. Mellor, 1980. A coastal ocean numerical model. Carlson, T.N., 1981. Speculations on the movement of polluted air to the
In: J. Sunderman and K.-P. Holz (eds.). Mathematical modelling of Arctic. Atmos. Environ. 15: 1473-1477.
estuarine physics. Proc. Int. Symp., Hamburg, 1978, pp. 203-214. Carmack, E.C., 1986. Circulation and mixing in ice-covered waters. In:
Springer-Verlag, Berlin. N. Untersteiner (ed.). The geophysics of sea ice, pp. 641-712. Plenum,
Bönisch, G. and P. Schlosser, 1995. Deep water formation and exchange New York.
rates in the Greenland/Norwegian Seas and the Eurasian Basin of the Carmack, E.C., 1990. Large scale physical oceanography of polar oceans.
Arctic Ocean derived from tracer balances. Prog. Oceanogr. 35: 29-52. In: W.O. Smith (ed.). Polar oceanography, pp. 171-222. Academic
Borys, R.D., E.E. Hamilton and P.J. DeMott, 1988. The chemical frac- Press, San Diego, California.
tionation of atmospheric aerosol as a result of snow crystal fromation Carmack, E.C. and K. Aagaard, 1973. On the deep water of the Greenland
and growth. J. Atmos. Chem. 7: 213-239. Sea. Deep-Sea Res. 20: 687-715.
Borys, R.D., D. Del Vecchio, J.-L. Jaffrezo, C.I. Davidson and D.L. Mitchell, Carmack, E.C., R.W. Macdonald and J.E. Papadakis, 1989. Water mass
1993. Assessment of ice parcticle growth processes at Dye-3, Green- structure and boundaries in the Mackenzie shelf estuary. J. Geophys.
land. Atmos. Environ. 27A: 2815-2822. Res. 94(C12): 18043-18055.
Bourke, R.H. and R.G. Paquette, 1991. Formation of Baffin Bay bottom Carter C.W. and I.H. Suffet, 1982. Binding of DDT to dissolved humic
and deep waters. In: P.C. Chu and J.C. Gascard (eds.). Deep convec- materials. Environ. Sci. Technol. 16(11): 735-740.
tion and deep water formation in the ocean. Elsevier, Amsterdam. Cassell, E.A. and T. Pangburn, 1991. Interactive modelling of cold regions
Bourke, R.H., A.M. Weigel and R.G. Paquette, 1988. The westward watersheds. In: T.D. Prowse and C.S.L. Ommanney (eds.). Northern
turning branch of the West Spitsbergen Current. J. Geophys. Res. 93: hydrology: Selected perspectives. Proceedings of the Northern Hydro-
14065-14077. logy Symposium, July 10-12, 1990, pp. 363-377. Environment Cana-
Boyle, E.A., J.M. Edmond and E.R. Sholkovitz, 1977. The mechanism of da, National Hydrology Research Institute, Saskatoon, Saskatchewan,
iron removal in estuaries. Geochim. Cosmochim. Acta 41: 1313-1324. Canada.
Brimblecombe, P., M. Tranter, P.W. Abrahams, I. Blackwood, T.D. Davies Cavalieri, D.J. and S. Martin, 1994. The contribution of Alaskan, Siberian
and C.E. Vincent, 1985. Relocation and preferential elution of acidic and Canadian coastal polynyas to the cold halocline layer of the Arctic
solute through the snowpack of a small, remote, high-altitude Scottish Ocean. J. Geophys. Res. 99: 18343-18362.
catchment. Ann. Glaciol. 7: 141-147. CCN, 1986. PCBs in perspective. Can. Chem. News 38: 15-17.
Brimblecombe, P., S.L. Clegg, T.D. Davies, D. Shooter and M. Tranter, CCREM (Canadian Council of Resource and Environment Ministers),
1987. Observations of the preferential loss of major ions from melt- 1987. Canadian water quality guidelines, prepared for the Task Force
ing snow and laboratory ice. Water Resour. Res. 21: 1279-1286. on Water Quality Guidelines of the Canadian Council of Resource and
Broecker, W.S., T.-H. Peng and T. Takahashi, 1980. A strategy for the use Environment Ministers by Water Quality Objectives Division. Water
of bomb-produced radiocarbon as a tracer for the transport of fossil Quality Branch, Inland Waters Directorate, Environment Canada,
fuel CO2 into the deep sea source regions. Earth Planet. Sci. Lett. Ottawa, Ontario.
49: 463-468. CGEIC (Canadian Global Emission Inventory Centre), 1996. Atmospheric
Bruland, K. W., P.K. Bienfang, J.K.B. Bishop, G. Eglinton, V.A.W. Ittek- Environment Service, Environment canada, Downsview, Ontario.
kot, R. Lampitt, M. Sarnthein, J. Thiede, J.J. Walsh and G. Wefer, Unpublished.
1989. Group report: Flux to the seafloor. In: W.H. Berger, V.S. Sme- Chambers, P., G.J. Scrimgeour, A. Pietroniro, J.M. Culp and I. Loughran,
tacek and G. Wefer (eds.). Productivity of the ocean: Present and past. 1994. Oxygen modelling under river ice covers. In: T.D. Prowse (ed.).
Report of the Dahlem workshop on productivity of the ocean: Present Proceedings of the workshop on environmental aspects of river ice,
and past, Berlin, April 24-29, 1988. pp. 193-215. Wiley, Chichester. August 18-20, 1993, pp. 235-260. National Hydrology Research Insti-
471p. tute, Saskatoon, Saskatchewan, Canada.
Brunskill, G.J., 1986. Environmental features of the Mackenzie System. Cheng, H., S. Leppinen and G. Whitley, 1993. Effects of river ice on chem-
In: B.R. Davies and K.F. Walker (eds.). The ecology of river systems, ical processes. In: T.D. Prowse and N.C. Gridley (eds.). Environmental
pp. 435-471. Dr. W. Junk, Dordrecht, The Netherlands. aspects of river ice. National Hydrology Research Institute, Saskatoon,
Bryan, K., 1969. A numerical model for the study of circulation of the Canada, NHRI Science Report No. 5.
world oceans. J. comput. Phys. 4: 347-376. Chernyak, S.M., C.P. Rice and L.L. McConnell, 1996. Evidence of cur-
Bryan, F. and W.R. Holland, 1989. A high resolution simulation of the rently used pesticides in air, ice, fog, seawater and surface microlayer
wind and thermohaline-driven circulation of the North Atlantic. Taha in the Bering and Chukchi Seas. Mar. Pollut. Bull. 32: 410-419.
Hulikoua, Proc. of the Hawaiian Winter Workshop, pp. 99-116. Christensen, J.H., 1995. Transport of air pollution in the troposphere to
University of Hawaii. the Arctic. PhD thesis, National Environmental Research Institute,
Buch, E., 1985. Seasonal and year to year variations of West Greenland Roskilde, Denmark, 377p.
waters in recent years. Rapp. P.-v. Réun. Cons. int. Explor. Mer. 185: Church, M., 1974. Hydrology and permafrost with reference to northern
141-151. North America. In: Proceedings Workshop Seminar on Permafrost
Buch, E., S.-A. Malmberg and S.S. Kristmannsson, 1996. Arctic Ocean Hydrology, pp. 7-20. Canadian National Committee for IHD, Ottawa.
deep water masses in the western Iceland Sea. J. Geophys. Res. 101: CIP, 1995. Natural environment in Russia. A brief survey. ECOS J.: 1-119.
11965-11973. Claridge, F.B. and A.M. Miira, 1981. Erosion control along transporta-
Bullister, J.L. and R.F. Weiss, 1983. Anthropogenic chlorofluoromethanes tion routes in northern climates. Arctic 34(2): 147-157.
in the Greenland and Norwegian Seas. Science 221: 265-268. Clarke, R.A., 1984. Transport through the Cape Farewell-Flemish Cap
Bursey, G.G., T.W.D. Edwards and S.K. Frape, 1991. Water balance and section. Rapp. P.-v. Réun. Cons. int. Explor. Mer 185: 120-130.
geochemistry studies in a tundra watershed, district of Keewatin, Clarke, R.A., 1986. The formation of Greenland Sea deep water. ICES
Northwest Territories. In: T.D. Prowse and C.S.L. Ommanney (eds.). C.M. 1986 C: 2. International Council for the Exploration of the Sea,
Northern hydrology: Selected perspectives. Proceedings of the North- Copenhagen.
ern Hydrology Symposium, 10-12 July 1990, Saskatoon, Saskatche- Clarke, R.A. and J.-C. Gascard, 1983. The formation of Labrador Sea
wan, pp. 17-31. Environment Canada, Saskatoon, Saskatchewan, water, Part I: Large-scale processes. J. Phys. Oceanogr. 13: 1764-1778.
NHRI Symposium No. 6. Clarke, R.A., J.L. Reid Jr. and J.H. Swift, 1986. The Greenland Sea in
Buzov, A.Y., 1991. Natural factors and their influence on transit sailing winter. ICES C.M. 1986 C: 32. International Council for the Explor-
on the Northern sea Route - Part I. The Northern Sea Route Project - ation of the Sea, Copenhagen.
Pilot Studies Report. The Fridtjof Nansen Institute, Oslo. Coachman, L.K., 1986. Circulation, water masses, and fluxes on the south-
Cadle, S.H., J.M. Dasch and N.E. Grossnickle, 1984a. Northern Michi- eastern Bering Sea shelf. Cont. Shelf Res. 5: 23-108.
gan snowpack - a study of acid stability and release. Atmos. Environ. Coachman, L.K., 1993. On the flow field in the Chirikov Basin. Cont. Shelf
18: 807-816. Res. 13: 481-508.
Cadle, S.H., J.M. Dasch and N.E. Grossnickle, 1984b. Retention and Coachman, L.K. and K. Aagaard, 1974. Physical oceanography of Arctic
release of chemical species by a northern Michigan snowpack. Water and subarctic seas. In: Y. Herman (ed.). Marine geology and oceano-
Air Soil Pollut. 22: 303-319. graphy of the Arctic Seas, pp. 1-72. Springer-Verlag, New York.
CAFF (Conservation of Arctic Flora and Fauna), 1994. The state of pro- Coachman, L.K. and K. Aagaard, 1988. Transports through Bering Strait:
tected areas in the circumpolar Arctic 1994. Directorate for Nature Annual and inter-annual variability. J. Geophys. Res. 93(C12): 15535-
Management, Trondheim, CAFF Habitat Conservation Report No. 1., 15539.
163p. Coachman, L.K. and C.A. Barnes, 1961. The contribution of Bering Sea
Calles, B., 1977. The river Elvegårdselv and its delta; a study of fluvial water to the Arctic Ocean. Arctic 14: 147-161.
transport and topographical changes. University of Uppsala, Dept. of Coachman, L.K. and C.A. Barnes, 1963. The movement of Atlantic water
Physical geography, UNGI report 45. in the Arctic Ocean. Arctic 16: 9-16.
Campbell, N.J., 1964. The origin of cold high-salinity water in Foxe Ba- Coachman, L.K. and J.J. Walsh, 1981. A diffusion model of cross-shelf
sin. J. Fish. Res. Bd Can. 21: 45-55. exchange of nutrients in the southeastern Bering Sea. Deep-Sea Res.
Cannon, G.A. and J.R. Holbrook, 1981. Wind induced seasonal interac- 28A(8): 819-846.
tion between coastal and fjord circulation. In: R. Sætre and M. Mork Coachman, L.K., K. Aagaard and R.B. Tripp, 1975. Bering Strait: the reg-
(eds.). Norwegian coastal current, pp. 131-151. University of Bergen. ional physical oceanography. University of Washington Press, Seattle.
Carey, A.G., 1987. Parcticle flux beneath fast ice in the shallow southwest- Cochran, J.K., D.J. Hirschberg, H.D. Linvingston, K.O. Buesseler and
ern Beaufort Sea, Arctic Ocean. Mar. Ecol. Prog. Ser. 40: 247-257. R.M. Key, 1995. Natural and anthropogenic radionuclide distributions
Carey, A.G., 1995. Importance of site-specific factors in influencing cont- in the Nansen Basin, Arctic Ocean: Scavenging rates and circulation
aminant distributions in ecosystems. In: (Anon. ed.) Abstract book. timescales. Deep-Sea Res. II 42(6): 1495-1517.
2nd. SETAC World Congress (16th Annual Meeting), 5-9 November Codispoti, L. and D. Lowman, 1973. A reactive silicate budget for the
Arctic Ocean. Limnol. Oceanogr. 18: 448-456. ECMP, 1995. Information prepared by the Environmental Committee of
Codispoti, L.A. and T.G. Owens, 1975. Nutrient transports through Murmansk province for the AMAP expert group. In: Proposals for
Lancaster Sound in relation to the Arctic Ocean’s reactive silicate environmentally sound investment projects in the Russian part of the
budget and the outflow of Bering Strait waters. Limnol. Oceanogr. Barents region, Volume 1: Non-radioactive contamination. The Nordic
20: 115-119. Environment Finance Corporation (NEFCO), Barents Region Environ-
COGLA, 1989. Database listing of onshore wells, N.W.T. and Yukon. mental Programme, Phase I report. AMAP, Oslo. 147p.
Canada oil and gas lands administration, Ottawa (Computer database). Eisma, D., 1993. Suspended matter in the aquatic environment. Springer-
Collett Jr., J.L., B. Oberholzer and J. Staehlin, 1993. Cloud chemistry at Verlag, Berlin-Heidelberg, 315p.
Mt. Rigi, Switzerland: dependence on drop size and relationship to Eitzer, B.D. and R.A. Hites, 1989. Atmospheric transport and deposition
precipitation chemistry. Atmos. Environ. 27A: 33-42. of polychlorinated dibenzo-r-dioxins and dibenzofurans. Environ. Sci.
Collin, A.E. and M.J. Dunbar, 1964. Physical oceanography in Arctic Technol. 23: 1396-1401.
Canada. In: H. Barnes (ed.). pp. 45-75. Ocean. Mar. Biol. Ann. Rev. Elíasson, J., 1994. Northern hydrology in Iceland. In: T.D. Prowse, C.S.L.
2, George Allen and Unwin, London. Ommaney and L.E. Watson (eds.). Northern hydrology: International
Colony, R. and A.S. Thorndike, 1985. Sea ice motion as a drinkard’s perspectives, pp. 41-80. Environment Canada, Saskatoon, Saskatche-
walk. J. Geophys. Res. 90: 965-974. wan, NHRI Science Report No. 3.
Cooper, L.W., C.R. Olsen, D.K. Solomon, I.L. Larsen and R.B. Cook, Ellis, K.M., J.N. Smith, R.P. Nelson, L. Kilius, R. Macdonald, E. Carmack
1991. Stable isotopes of oxygen and natural and fallout radionuclides and S.B. Moran, 1995. Distribution of artificial radionuclides in the
used for tracing runoff during snowmelt in Arctic watershed. Water Arctic Oceans from the 1994 Arctic Ocean Section. In: P. Strand and
Resour. Res. 27(9): 2171-2179. A. Cooke (eds.). Environmental Radioactivity in the Arctic. Proceed-
Cooper, L.W., C. Solis, D.L. Kane and L.D. Hinzman, 1993. Application ings of the second International Conference on Environmental Radio-
of oxygen-18 tracer techniques to arctic hydrological processes. Arct. activity in the Arctic, Oslo, August 21-25, 1995, pp. 204-207. Nor-
alp. Res. 25(3): 247-255. wegian Radiation Protection Authority, Østerås, Norway. 416p.
Coquery, M., D. Cossa and J.M. Martin, 1995. The distribution of dis- Elverhøi, A., S.L. Pfirman, A. Solheim and B.B. Larssen, 1989. Glacio-
solved and particulate mercury in three Siberian estuaries and adja- marine sedimentation on epicontinental seas – exemplified by the
cent arctic coastal waters. Water Air Soil Pollut. 80(1/4): 653-664. northern Barents Sea. Mar. Geol. 85: 225-250.
Cotham, W.E. and T.F. Bidleman, 1991. Estimating the atmospheric Encyclopedia Britannica, 1990. The New Encyclopedia Britannica, 15th
deposition of organochlorine contaminants to the arctic. Chemo- ed., vol. 14, pp. 288-292. Encyclopedia Britannica Inc., Chicago.
sphere 22(1-2): 165-188. English, M.C., M.A. Stone, B. Hill, P.M. Wolfe and R. Ormson, 1996.
Cowardin, L.M., V. Carter, F.C. Golet and E.T. LaRoe, 1979. Classifi- Assessment of impacts on the Slave River Delta of Peace River im-
cation of wetlands and deepwater habitats of the United States. US poundment at Hudson Hope. Northern River Basins Study, Edmonton,
Department of the Interior, Fish and Wildlife Service, Washington Alberta, Project Report No. 74, p. 91.
D.C., 103p. ENVIRODAT, 1996. Water Quality data Base. Environment Canada, Ot-
Cox, M.D., 1984. A primitive equation, 3-dimensional model of the tawa, Canada.
ocean. Princeton University, GFDL, GFDL Ocean Group Tech. Rep. Evans, S.M., 1996. Biomagnification of persistent organic contaminants in
No.1, 141p. Great Slave Lake. In: J.L. Murray, R.G. Shearer and S.L. Han (eds.).
D’Asaro, E.A., 1988. Observations of small eddies in the Beaufort Sea. J. Synopsis of research conducted under the 1994/95 Northern Contami-
Geophys. Res. 93: 6669-6684. nants Program, pp. 215-220. Indian Affairs and Northern Develop-
Dahlgaard, H., A. Aarkrog, L. Hallstadius, E. Holm and J. Rioseco, ment, Ottawa, Canada, Environmental Studies No. 73.
1986. Radiocaesium transport from the Irish Sea via the North Sea Favorite, F., 1967. The Alaskan Stream. Int. N. Pac. Fish. Com. Bull. 21: 20.
and the Norwegian Coastal Current to East Greenland. Rapp. P.-v. Feichter, J., E. Kjellstrom, H. Rodhe, F. Dententer, J. Lelieveld and G.-J.
Réun. Cons. int. Explor. Mer 186: 70-79. Roelfs, 1996. Simulation of the tropospheric sulfur cycle in a global
Dai, M.-H. and J.-M. Martin, 1995. First data on trace metal level and climate model. Atmos. Environ. 30: 1693-1707.
behaviour in two major Arctic river-estuarine systems (Ob and Yeni- Fellin, P., L.A. Barrie, D. Dougherty, D. Toom, D. Muir, N. Grift, L. Lock-
sey) and in the adjacent Kara Sea, Russia. Earth Planet. Sci. Lett. hart and B. Billeck, 1996. Air monitoring in the Arctic; results for se-
131: 127-141. lected persistent organic pollutants for 1992. Environ. Toxicol. Chem.
Dastoor, A.P. and J. Pudykiewicz, 1996. A numerical global meteorologi- 15: 253-261.
cal sulphur transport model and its application to Arctic air pollution. Finkel, R., S. Krishnaswamit and D.L. Clark, 1977. Be-10 in Arctic Ocean
Atmos. Environ. 30: 1501-1522. sediments. Earth Planet. Sci. Lett. 35: 199-204.
Davidson, C.I., 1989. Mechanisms of wet and dry deposition of atmos- Fischer, G., D. Fütterer, R. Gersonde, S. Honjo, D. Ostermann and G. We-
pheric contaminants to snow surfaces. In: H. Oeschger and C.C. fer, 1988. Seasonal variability of parcticle flux in the Weddell Sea and
Langway (eds.). The environmental record in glaciers and ice sheets. its relation to ice cover. Nature 335: 426-428.
Report of the Dahlem workshop on the environmental record in gla- Fissel, D.B., J.R. Birch, H. Melling and R.A. Lake, 1988. Non-tidal flows
ciers and ice sheets, Berlin, March 13-18, 1988. pp.29-51. Wiley, in the Northwest Passage. Canadian Institute of Ocean Science, Sidney,
Chichester. 400p. British Columbia, Canada, Tech. Rep. Hydrogr. Ocean Sci. 98, 157p.
Davies, T.D., C.E. Vincent and P. Brimblecombe, 1982. Preferential elu- Fleming, G.H. and A.J. Semtner, 1991. A numerical study of interannual
tion of strong acids from a Norwegian ice cap. Nature 300: 161-163. ocean forcing on Arctic ice. J. Geophys. Res. 96: 4589-4603.
Dean, K.G., W.J. Stringer, K. Ahlnäs, C. Searcy and T. Weingartner, Foldvik, A., K. Aagaard and T. Torresen, 1988. On the velocity field of
1994. The influence of river discharge on the thawing of sea ice, the east Greenland Current. Deep-Sea Res. 35: 1335-1354.
Mackenzie River Delta: albedo and temperature analyses. Polar Res. Franz, T.P., D.J. Gregor and S.J. Eisenreich, (in press). Snow deposition of
12: 93-94. atmospheric semivolatile organic chemicals. In: J.E. Baker (ed.). At-
Diamond, M.L., H. Freitas and M. Kawai, 1996. Modelling inorganic mospheric deposition of contaminants to the Great Lakes and coastal
and organic contaminants in arctic freshwater lakes. In: J.L. Murray, waters. Society for Environmental Toxicology and Chemistry (SETAC).
R.G. Shearer and S.L. Han (eds.) Synopsis of research conducted Freitas, H., M. Diamond, R. Semkin and D. Gregor, 1997. Contaminant
under the 1994/95 Northern Contaminants Program, pp. 119-130. fate in High Arctic lakes: Development and application of a mass bal-
Department of Indian and Northern Affairs, Ottawa, Canada, Envi- anced model. Sci. Total Environ. (submitted)
ronmental Studies No. 73. French, H.M., 1978. Sump studies I. Indian and Northern Affairs Canada,
Dickson, R.R. and J. Brown, 1994. The production of North Atlantic Ottawa, Environmental Studies No. 6, 52p.
deep water: Sources, rates, and pathways. J. Geophys. Res. 99: 12319- Friedman, G.M. and J.E. Sanders, 1978. Principles of sedimentology. John
12341. Wiley and Sons Inc., New York, 792p.
Dickson, R.R., J. Meinke, S.A. Malmberg and A.J. Lee, 1988. The “great Futsaeter, G., G. Eidnes, G. Halmo, S. Johansen, H.P. Mannvik, L.K. Syd-
salinity anomaly” in the northern North Atlantic. Prog. Oceanogr. nes and U. Witte, 1991. Report on oil pollution. The State of the Arc-
20: 103-151. tic Environment Reports, University of Lapland, Arctic Centre, Rova-
Doskey, P.V. and A.W. Andren, 1981. Modeling the flux of atmospheric niemi, pp. 270-334.
polychlorinated biphenyls across the air/water interface. Environ. Sci. Garland, T.R., R.G. Riley, E. Snyder-Conn and D.F. Woodward, 1988.
Technol. 15: 705-711. Drilling fluids and the Arctic tundra of Alaska: Assessing contamina-
Douglas, B.C., 1991. Global sea level rise. J. Geophys. Res. 96: 6891-6992. tion of wetlands habitat and the toxicity to aquatic invertebrates and
Drinkwater, K.F., 1986. Physical oceanography of Hudson strait and fish. Arch. Environ. Contam. Toxaphene 17(5): 638-697.
Ungava Bay. In: I.P. Martini (ed.). Canadian inland seas, pp. 237-264. Garrison, G.R. and P. Becker, 1976. The Barrow submarine canyon: a
Elsevier Oceanography Series 44, Elsevier, New York. drain for the Chukchi Sea. J. Geophys. Res. 81: 4445-4453.
Driscoll, C.T., C.L. Schofield, R. Munson and J. Holsapple, 1994. The Garrison, D.L., S.A. Ackley, K.R. Buck, 1983. A physical mechanism for
mercury cycle and fish in Adirondack lakes. Environ. Sci. Technol. establishing algal populations in frazil ice. Nature 306: 363-365.
28(3): 136-143. Gascard, J.C. and R.A. Clarke, 1983. The formation of Labrador sea
Duce, R.A., V.A. Mohnen, P.R. Zimmerman, D. Grosjean, W. Cautreels, water. Part II: Mesoscale and smaller scale processes. J. Phys. Oceanogr.
R. Chatfield, R. Jaenicke, J.A. Ogren, E.D. Pellizzare and G.T. Wal- 13: 1779-1797.
lace, 1983. Organic material in the global troposhere. Rev. Geophys. Gaul, H., 1989. Organochlorine compounds in water and sea ice of the
Space Res. 21: 921-952. European Arctic Sea. In: R. Shearer and A. Bartonova (eds.). Global
Duxbury, A.C. and A.B. Duxbury, 1994. An introduction to the world’s significance of the transport and accumulation of polychlorinated hy-
oceans. Wm.C. Brown Publisher, Dubuque, Iowa, 472p. drocarbons in the Arctic, The 8th international conference of Comité
Dynesius, M. and C. Nilsson, 1994. Fragmentation and flow regulation of Arctique International, Oslo, 18-22 September 1989.
river systems in the northern third of the world. Science 266: 753-762. Genthon, C. and A. Armengaud, 1995. GCM simulations of atmospheric
tracers in the polar latitudes: South Pole (Antarctica) and Summit dependent solute redistribution at the ice/water phase boundary. II.
(Greenland) cases. Sci. Total Environ. 160/161: 101-116. Experimental investigation. J. Chem. Phys. 62: 3085-3092.
Gerdes, R., C. Koberle and J. Willebrand, 1991. The influence of numeri- Gubala, C.P., D.H. Landers, M. Monetti, M. Heit, T. Wade, B. Lasorsa
cal advection schemes on the results of ocean general circulation mod- and S. Allen-Gil, 1995. The rates of accumulation and chronologies of
els. Climate Dynamics 5: 211-226. atmospherically derived pollutants in Arctic Alaska, USA. Sci. Total
GESAMP, 1985. Interchange of pollutants between the atmosphere and Environ. 160/161: 347-361.
the ocean. IMO/FAO/UNESCO/WMO/WHO/IAEA/UN/UNEP. Haas, J., 1991. Estimates of evaporation in the Yukon and Northwest
GESAMP Reports and Studies No. 23, 55p. Territories. In: T.D. Prowse and C.S.L. Ommanney (eds.). Northern
GESAMP, 1993. Anthopogenic influences on sediment discharge to the hydrology: Selected perspectives, pp. 189-203. Proceedings of the
coastal zone and environmental consequences. IMO/FAO/UNESCO. Northern Hydrology Symposium, 10-12 July, 1990, Saskatoon,
GESAMP Reports and Studies No. 52, 66p. Saskatachewan. Environment Canada, Saskatoon, Saskatchewan,
Giorgi, F., 1986. A particle dry-deposition parameterization scheme for NHRI Symposium No.6.
use in tracer transport models. J. Geophys. Res. 91(D9): 9794-9806. Hachey, H.B., 1935. The circulation of Hudson Bay water as indicated by
Giovando, L.F. and R.H. Herlinveaux, 1982/1 (1982 or 1981?). A discus- drift bottles. Science 82: 275-276.
sion of factors influencing the dispersion of pollutants in the Beaufort Haidvogel, D.B., J.L. Wilkin and R.E. Young, 1991. A semi-spectral prim-
Sea. Pacif. mar. Sci. Rep. 81(4): 198. itive equation ocean circulation model using vertical sigma and orthog-
Gloersen, P., 1995. Modulation of hemispheric sea-ice cover by ENSO onal curvilinear horizontal coordinates. J. comput. Phys. 94: 151-185.
events. Nature 373: 503-506. Håkanson, L., 1994. A model to predict gross sedimentation in small gla-
Gloersen, P., W.J. Campbell, D.J. Cavalieri, J.C. Comisco, C.L. Parkinso- cial lakes. Hydrobiologia 284: 19-42.
na and H.J. Zwally, 1992. Arctic and Antarctic-1987, Satellite pas- Håkanson, L. and M. Jansson, 1983. Principles of lake sedimentology.
sive-microwave observations and analysis. NASA SP-511, 290p. Springer-Verlag, Berlin and Heidelberg, Germany. 316p.
Goldberg, E.D., 1975. Synthetic organohalides in the sea. Proc. R. Soc. Häkkinen, S. and G.L. Mellor, 1990. One hundred years of Arctic ice
Lond. Ser.B 189: 277-289. cover variations as simulated by a one dimensional coupled ice-ocean
Goldberg, E.D., 1976. The health of the oceans. Unesco, Paris, 172p. model. J. Geophys. Res. 95: 15959-15969.
Goldschmidt, P.M., S.L. Pfirman, I. Wollenburg and R. Henrich, 1992. Häkkinen, S. and G.L. Mellor, 1992. Modeling the seasonal variability of
Origin of sediment pellets on the Arctic sea floor: sea ice or icebergs? the coupled Arctic ice-ocean system. J. Geophys. Res. 97: 20285-20304.
Deep-Sea Res. Special Issue 39: S539-S565. Hansen, A.D.A. and H. Rosen, 1984. Vertical distribution of particulate
Gordeev, V.V. and I.S. Sidorov, 1993. Concentrations of major elements carbon, sulfur and bromine in the Arctic haze and comparison with
and their outflow into the Laptev Sea by the Lena River. Mar. Chem. ground-level measurements at Barrow, Alaska. Geophys. Res. Lett. 11:
43(1-4): 33-46. 381-384.
Gordeev, V.V., J.-M. Martin, I.S. Sidorova and M.V. Sidorova, 1995. A Hanson, A.M., 1965. Studies of the mass budget of Arctic pack-ice floes.
reassessment of the Eurasian river input of water, sediment, major ele- J. Glaciol. 5: 701-709.
ments and nutrients to the Arctic Ocean. Am. J. Sci. (accepted) Hanzlick, D. and K. Aagaard, 1980. Freshwater and Atlantic water in the
Gosselink, J.G. and R.E. Turner, 1978. The role of hydrology in freshwa- Kara Sea. J. Geophys. Res. 85: 4937-4942.
ter wetland ecosystems. In: R.E. Hook, D.F. Whigham and R.L. Simp- Hardisty, P., V. Schilder, T. Dabrowski and J. Wells, 1991. Yukon and
son (eds.). Freshwater wetlands. Ecological processes and manage- Northwest Territories ground-water data base. In: T.D. Prowse and
ment potential. New York, 378p. C.S.L. Ommanney (eds.). Northern hydrology : Selected perspectives,
Gould, W.J., J. Loynes and J. Backhaus, 1985. Seasonality in slope cur- pp. 465-482. Proceedings of the Northern Hydrology Symposium, 10-
rent transports N.W. of Shetland. ICES C.M. 1985 C: 7. International 12 July, 1990, Saskatoon, Saskatachewan. Environment Canada, Saska-
Council for the Exploration of the Sea, Copenhagen. toon, Saskatchewan, NHRI Symposium No. 6.
Gow, A.J. and W.B. Tucker III, 1990. Sea ice in the polar regions. In: W.O. Hargrave, B.T., W.P. Vass, P.E. Erickson and B.R. Fowler, 1988. Atmos-
Smith Jr. (ed.). Polar oceanography. Part A, Physical science. Academic pheric transport of organochlorines to the Arctic Ocean. Tellus 40B:
Press Inc., San Diego, California. 480-493.
Graedel, T.E. and J.P. Franey, 1975. Field measurements of submicron Hargrave, B.T., B. von Bodungen, R.J. Conover, A.J. Fraser, G. Phillips
aerosol washout by snow. Geophys. Res. Lett. 2: 325-328. and W.P. Vass, 1989. Seasonal changes in sedimentation of particulate
Graf Pannatier, E. and D.J. Gregor, 1996. Mackenzie delta overbank and matter and lipid content of zooplankton collected by sediment trap in
lake sedimentation and contaminant accumulation study. IWD-NWT the Arctic Ocean off Axel Heiberg Island. Polar Biol. 9: 467-475.
NOGAP Progress Report #3, May 1996, 41p. (Unpublished report). Hargrave, B.T., B. von Bodungen, P. Stoffyn-Egli and P.J. Mudie, 1994.
Gray, C.B., 1994. Lake Laberge. In: R.J. Allan, M. Dickman, C.B. Gray Seasonal variability in particle sedimentation under permanent ice
and V. Cromie (eds.). The book of Canadian lakes, pp. 474-484. The cover in the Arctic Ocean. Cont. Shelf Res. 14: 279-293.
Canadian Association on Water Quality, Gloucester, Ontario, Canada. Harms, I.H., 1997. Modeling the dispersion of 137Cs and 239Pu released
Monograph Series #33. from dumped waste in the Kara Sea. J. Mar. Sys. 13(1997): 1-13.
Gray, C.B., M. Dickman, B. Krushelnicki and V. Cromie, 1992. A survey Harper, J.R., 1990. Morphology of the Canadian Beaufort Sea coast. Mar.
of Canadian lakes. National Water Research Institute, Environment Geol. 91: 75-91.
Canada, Burlington, Ontario, Canada. NWRI Contribution #91-18 Hart, K.M., J. Tremp, E. Molnar and W. Giger, 1993. The occurrence and
(unpublished). the fate of organic pollutants in the atmosphere. Water Air Soil Pollut.
Gray, D.M., D.I. Norum and J.M. Wigham, 1973. Infiltration and phys- 68: 91-112.
ics of flow of water through porous media. In: D.M. Gray (ed.). Hebbeln, D. and G. Wefer, 1991. Effects of ice coverage and ice-rafted
Handbook on the principles of hydrology, pp. 5.1-5.41. Water Infor- material on sedimentation in the Fram Strait. Nature 350(4): 409-411.
mation Center Inc., New York. Hecky, R.E., R.W. Newbury, R.A. Bodaly, K. Patalas and D.M.
Gregor, D.J., 1990. Deposition and accumulation of selected agricultural Rosenberg, 1984. Environmental impact prediction and assessment:
pesticides in Canadian Arctic snow., In: D.A. Kurtz (ed.). Long range the southern Indian Lake experience. Can. J. Fish. aquat. Sci. 41: 720-
transport of pesticides, pp. 373-386. Lewis Publishers, Chelsea, 732.
Michigan. Hedstrom, K.S., 1990. User manual for a semi-spectral primitive equation
Gregor, D.J., 1991. Organic micropollutants in arctic snow and firn. In: regional ocean circulation model. Version 3.0. Institute for Naval
M.-K. Woo and D.J. Gregor (eds.). Arctic environment: Past, present Oceanography, Technical Note FY 90-2, 90p.
and future, p. 164. Proceedings of a symposium held at McMaster Hequette, A., M.-H. Ruz and P.R. Hill, 1995. The effects of the Holocene
University, Nov. 14-15, 1991. McMaster University, Department of sea level rise on the evolution of the southeastern coast of the Cana-
Geography, Hamilton, Ontario. dian Beaufort Sea. J. coast. Res. 11(2): 494-507.
Gregor, D.J., 1996. Snow, ice and temperature as determinants of organic Hibler III, W.D., 1979. A dynamic sea ice model. J. Phys. Oceanogr. 9: 815-
chemical fate in northern ecosystems. In: J.L. Murray and R.G. Shearer 846.
(eds.) Synopsis of research conducted under the 1994/95 Northern Hibler III, W.D. and K. Bryan, 1987. A diagnostic ice-ocean model. J. Phys.
Contaminants Program, pp. 57-59. Ottawa, Canada, Environmental Oceanogr. 17: 987-1015.
Studies No. 73. Hill, P.R. and O.C. Nadeau, 1989. Storm-dominated sedimentation on the
Gregor, D. and J. Eamer, 1993. Base flow water quality and contaminant inner shelf of the Canadian Beaufort Sea. J. Sedimentary Petrology 59:
levels in the upper Yukon River system. In: J.L Murray and R.G. 455-468.
Shearer (eds.). Synopsis of research conducted under the 1992/93 Hobbie, J.E., 1973. Arctic limnology: a review. In: M.E. Button (ed.). Ala-
Northern Contaminants Program, pp. 69-77. Department of Indian skan Arctic tundra. Arctic Institute of North America, Washington
and Northern Affairs, Ottawa, Canada, Environmental Studies No. D.C., Technical Paper No. 25.
70. Hoekstra, P. and R.D. Miller, 1967. On the mobility of water molecules in
Gregor, D.J. and W. Gummer, 1989. Evidence of atmospheric transport the transition layer between ice and a solid surface. J. Colloid Interface
and deposition of organochlorine pesticides and PCB in Canadian Sci. 25: 155-173.
Arctic snow. Environ Sci. Technol. 23(5): 561-565. Hoff, J.T., F. Wania, D. Mackay and R. Gilham, 1995. Sorption of nonpo-
Grigoriev, V.Y. and B.L. Sokolov, 1994. Northern hydrology in the For- lar organic vapours by ice and snow. Environ. Sci. Technol. 27: 2174-
mer Soviet Union (FSU). In: T.D. Prowse, C.S.L Ommaney and L.E. 2180.
Watson (eds.). Northern hydrology: International perspective, pp. Holland, D.M., L.A. Mysak, D.K. Manak and J.M. Oberhuber, 1993.
147-179. Environment Canada, Saskatoon, Saskatchewan, Science Sensitivity study of a dynamic thermodynamic sea ice model. J. Geo-
report No.3. phys. Res. 98: 2561-2586.
Groen, P., 1969. The waters of the sea. Van Norstrand, New York. Holland, D.M., L.A. Mysak and J.M. Oberhuber, 1995. An investigation
Gross, G.W., C.-H. Wu, L. Bryant and C. McKee, 1975. Concentration of the general circulation of the Arctic Ocean using an isopycnal ocean
model. Tellus 48A: 138-157. Arctic Ocean into the channels of the Canadian Arctic archipelago: the
Holloway, G., 1992. Representing topographic stress for large scale ocean history of Hobson’s Choice Ice Island. Polar Record 29(171): 305-312.
models. J. Phys. Oceanogr. 22: 1033-1046. Jeffries, M.O., H.R. Krouse, W.M. Sackinger and H.V. Serson, 1989.
Holtan, G., D. Berge, H. Holtan and T. Hopen, 1994. Annual report on Stable-isotope (18O/16O) tracing of fresh, brackish, and sea ice in multi-
direct and riverine inputs to Norwegian coastal waters during the year land-fast sea ice, Ellesmere Island, Canada. J. Glaciol. 35: 9-16.
year 1993 – Paris Convention. SFT (Norwegian State Pollution Con- Johannesson, M. and A. Hendrien, 1978. Chemistry of snow meltwater:
trol Authority), Oslo. SFT Report 580194 (NIVA-report 090001), changes in concentration during melting. Water Resour. Res. 14(4):
138p.. 815-619.
Holtan, H. and S.O. Åstebøl, 1991. Handbook for collection of data on Johansen, B.E. and H. Tømmervik, 1992. Finnmarksvidda, mapping of
pollutant inputs to fjords and water courses. Revised edition – No- vegetation. NORUT Information Technology, Tromsø, Norway.
vember 1991. SFT (Norwegian State Pollution Control Authority), NORUT Information Technology report 2020. (In Norwegian).
Oslo. SFT Report No. 91:10. (In Norwegian). Johnson, L., 1975a. Physical and chemical characteristics of Great Bear
Honjo, S., 1990. Particle fluxes and modern sedimentation in the Polar Lake, Northwest Territories. J. Fish. Res. Bd Can. 32(11): 1971-1987.
Oceans. In: W.O. Smith (ed.). Polar oceanography, Part B: Chemistry, Johnson, L., 1975b. The Great Bear Lake: its place in history. Arctic
biology and geology, pp. 687-739. Academic Press, San Diego, Cali- 28(4): 230-243.
fornia. Johnson, L., 1994a. Great Bear Lake. In: R.J. Allan, M. Dickman, C.B.
Honjo, S., S.J. Manganini and G. Wefer, 1988. Annual parcticle flux and Gray and V. Cromie (eds.). The book of Canadian lakes, pp. 549-559.
a winter outburst of sedimentation in the northern Norwegian Sea. The Canadian Association on Water Quality, Gloucester, Ontario,
Deep-Sea Res. 35(8): 1223-1234. Canada. Monograph Series #33.
Hood, D.W. and E.J. Kelley, 1974. Summary of early exploration and Jones, E.P. and L.G. Anderson, 1986. On the origin of the chemical prop-
oceanographic features. In: D.W. Hood and E.J. Kelley (eds.). Oceano- erties of the Arctic Ocean halocline. J. Geophys. Res. 91: 10759-10767.
graphy of the Bering Sea with emphasis on renewable resources, pp. Jones, E.P. and A.R. Coote, 1980. Nutrient distribution in the Canadian
xv-xxi. University of Alaska, Institute of Marine Science, Fairbanks, archipelago: indicators of summer water mass and flow characteristics.
Alaska. Can. J. Fish. aquat. Sci. 37: 589-599.
Hope, D., M.F. Billett and M.S. Cresser, 1994. A review of the export of Jónsson, S. and A. Foldvik, 1992. The transport and circulation in Fram
carbon in river water: Fluxes and processes. Environ. Pollut. 84: 301- Strait. ICES C.M. 1992. C: 10. International Council for the Explo-
324. ration of the Sea, Copenhagen.
Hopkins, T.S., 1991. The GIN Sea – A synthesis of its physical oceanog- Jopling, A.V., 1960. An experimental study on the mechanics of bedding.
raphy and literature review 1972-1985. Earth-Sci. Rev. 30: 175-318. PhD dissertation, Harvard University, Cambridge, Massachusetts.
Hsiao, S.I.C., 1987. Sedimentation in Arctic Canada: species composition Joussaume, S., 1990. Three-dimensional simulations of the atmospheric
and biomass of phytoplankton contributed to the marine sediments of cycle of desert dust particles using a General Circulation Model. J.
Frobisher Bay. Polar Biol. 7: 245-251. Geophys. Res. 95: 1909-1941.
Hunter, J.G., 1960. Hazen Lake. Fish. Res. Board Canada, Arctic Unit Joussaume, S., 1993. Paleoclimatic tracers: An investigation using an
Annual Report and Investigators Summaries, April 1, 1959 to March atmospheric general circulation model under ice age conditions 1.
31, 1960. Desert dust. J. Geophys. Res. 98: 2767-2805.
HYDAT, 1994. Hydrology Data Base, version 4.94. Environment cana- Jury, W.A., W.F. Spencer and W.J. Farmer, 1983. Behavior assessment
da, Ottawa, Canada. model for trace organics in soil. I. Model description. J. Environ.
Ibrahim, M., L.A. Barrie and F. Fanaki, 1983. An experimental and theo- Qual. 12: 558-566.
retical investigation of the dry deposition of particles to snow, pine Jury, W.A., W.F. Spencer and W.J. Farmer, 1984a. Behavior assessment
trees and artificial collectors. Atmos. Environ. 17: 781-788. model for trace organics in soil. II. Chemical classification and para-
Igamberdiev, V.M., V.N. Lystsov and V.M. Makeev, 1995. Identification meter sensitivity. J. Environ. Qual. 13: 567-572.
and assessment of land-based activities in Russian Federation that Jury, W.A., W.F. Spencer and W.J. Farmer, 1984a. Behavior assessment
contribute to the degradation of the Arctic marine environment. Re- model for trace organics in soil. IV. Review of experimental evidence.
port prepared by Arctic workshop of ACOPS (Advisory Committee J. Environ. Qual. 13: 580-586.
on Protection of the Sea), Moscow. (In Russian). 192p. Jury, W.A., W.F. Spencer and W.J. Farmer, 1984c. Behaviour assessment
Ingram, R.G., 1981. Characteristics of the Great Whale River plume. J. model for trace organics in soil: III. Applications of screening model. J.
Geophys. Res. 86: 2017-2023. Environ. Qual. 13: 573-579.
Ingram, R.G. and P. Larouche, 1987. Variability of an under-ice river Kane, D.L., 1994. Northern hydrology in Alaska. In: T.D. Prowse, C.S.L.
plume in Hudson Bay. J. Geophys. Res. 92: 9541-9547. Ommaney and L.E. Watson (eds.). Northern hydrology: International
IPCC, 1996. Climate change 1995. The science of climate change. Cam- perspectives, pp. 5-20. Environment Canada, Saskatoon, Saskatche-
bridge University Press, Cambridge, 572p. wan, NHRI Science Report No.3.
Iribarne, J.V. and L.A. Barrie, 1995. The oxidation of S(IV) during rim- Kane, D.L., R.E. Gieck and L.D. Hinzman, 1990. Evaporation from a
ing by cloud droplets. J. Atmos. Chem. 21: 97-114. small Alaskan arctic watershed. Nordic Hydrology 12l(4/5): 253-272.
Iribarne, J.V. and T. Pyshnove, 1990. The effect of freezing on the com- Kane, D.L., L.D. Hinzman, C.S. Benson and G.E. Liston, 1991. Snow
position of supercooled droplets - I. Retention of HCl, HNO3, NH3, hydrology of a headwater Arctic Basin: I. Physical measurements and
and H2O2. Atmos. Environ. 24A: 383-387. process studies. Water Resour. Res. 27(6): 1099-1109.
Iribarne, J.V., L.A. Barrie and A. Iribarne, 1983. Effect of freezing on sul- Kane, D.L., R.E. Gieck, G. Wendler and L.D. Hinzman, 1993. Snowmelt
phur dioxide dissolved in supercooled droplets. Atmos. Environ. 17: at a small Alaskan watershed, 2. Energy related modeling results. In:
1047-1050. T.D. Prowse, C.S.L. Ommaney and K.E. Ulmer (eds.). Proceedings of
Iselin, C.O., 1927. A study of the northern part of the Labrador Current. the Ninth International Northern Research Basins
Bull. Nat. Res. Council 61: 217-222. Symposium/Workshop, pp. 227-248, August 14-22, 1992. Whitehorse,
Iversen, T., 1984. On the atmospheric transport of pollution to the Arc- Dawson city, Eagle Plains, Yukon and Inuvik, Northwest Territories.
tic. Geophys. Res. Lett. 11: 457-460. National Hydrology Research Institute, Environment Canada,
Iversen, T., 1989a. Some statistical properties of ground level air pollu- Saskatoon, Saskatchewan, Canada.
tion at Norwegian Arctic stations and their relation to large scale Kautsky, H., 1987. Investigation on the Distribution of 137Cs, 134Cs and
atmospheric flow systems. Atmos. Environ. 23: 2451-2462. 90Sr and the water mass transport times in the Northern North Atlan-
Iversen, T., 1989b. Numerical modelling of the long range atmospheric tic and the North Sea. Dt. Hydrogr. Z. 40: 49-69.
transport of sulphur dioxide and particulate sulphate to the Arctic. Kempema, E.W., E. Reimnitz and P.W. Barnes, 1989. Sea ice sediment
Atmos. Environ. 23: 2571-2595. entrainment and rafting in the Arctic. J. Sed. Petrol. 59: 308-317.
Iversen, T., 1996. Atmospheric transport pathways for the Arctic. In: E. Kidd, K.A., D.W. Schindler, D.C.G. Muir and R.H. Hesslein, 1996. The
Wolff and R.C. Bales (eds.). Chemical exchange between the atmos- biomagnification of organochlorines through the food web of Lake
phere and polar snow. Global Environmental Change. NATO ASI Laberge and other Yukon Lakes. In: J.L. Murray, R.G. Shearer and
Series I, Volume 43, pp. 71-92. Springer-Verlag, Berlin and Heidel- S.L. Han (eds.). Synopsis of research conducted under the 1994/95
berg, Germany. Northern Contaminants Program, pp. 209-214. Indian Affairs and
Jaffrezo, J.L., M.P. Clain and P. Masclet, 1994. Polycyclic aromatic hy- Northern Development, Ottawa, Canada, Environmental Studies no. 73.
drocarbons in the polar ice of Greenland. Geochemical use of these Killingtveit, Å., 1994. Northern hydrology in Norway. In: T.D. Prowse,
atmospheric tracers. Atmos. Environ. 28: 1139-1145. C.S.L. Ommaney and L.E. Watson (eds.). Northern hydrology: Inter-
Jantunen, L.M. and T.F. Bidleman, 1995. Reversal of the air-water gas national perspectives, pp. 81-107. Environment Canada, Saskatoon,
exchange direction of hexachlorocyclohexanes in the Bering and Saskatchewan, NHRI Science Report No.3.
Chukchi Seas: 1993 versus 1988. Environ. Sci. Technol. 29: 1081- Killingtveit, Å. and A. Harby, 1994. Multi purpose planning with the
1089. “River System Simulator” – A decision support system for water
Jaworowski, Z., 1989. Pollution of the Norwegian Arctic: A review. resource planning and operation. Presented at Environmental model-
Norsk Polarinsitutt Rapportserie 55. Norwegian Polar Institute, Oslo, ling seminar, Trondheim 17-19 August 1994.
Norway. 93p. Killingtveit, Å. and K. Sand, 1991. On aerial distribution of snowcover in
Jeffries, M.O., 1987a. The growth, structure and disintegration of Arctic a mountainous area. In: T.D. Prowse and C.S.L. Ornmanney (eds.).
ice shelves. Polar Record 23: 631-649. Northern hydrology: Selected perspectives, pp. 189-203. Proceedings
Jeffries, M.O., 1987b. Stucture and growth of Arctc ice shelves and ice of the Northern Hydrology Symposium, 10-l2 July, 1990, Saskatoon,
islands. NRC of Canada Associate Committee on Geotechnical Saskatachewan. Environment Canada, Saskatoon, Saskatchewan,
Research, Ottawa, Technical Memorandum 41, pp. 39-48. NHRI Symposium No. 6.
Jeffries, M.O. and M.A. Shaw, 1993. The drift of ice islands from the Killworth, P.D., 1979. On chimney formations in the oceans. J. Phys.
Oceanogr. 9: 531-554. Washington D.C., Publ. 13, 173p.

Killworth, P.D., 1983. Deep convection in the world ocean. Rev. Geophys. Levitus, S. and T.P. Boyer, 1994. World ocean atlas. Volume 4: Tempera-
Space Phys. 21: 1-26. ture. NOAA Atlas NESDIS 4. National Oceanic and Atmospheric Ad-
Kinder, T.H. and J.H. Schumacher, 1981. Hydrographic structure over ministration (NOAA), Washington D.C.
the continental shelf of the southeastern Bering Sea. In: D.W. Hood Levitus, S., R. Burgett and T.P. Boyer, 1994. World ocean atlas. Volume 3:
and J.A. Calder (eds.). The Eastern Bering Sea shelf: Oceanography Salinity. NOAA Atlas NESDIS 3. National Oceanic and Atmospheric
and resources, pp. 31-52. Office of Marine Pollution Assessment, Administration (NOAA), Washington D.C.
National Oceanic and Atmospheric Administration. University of Lewis, E.L., 1982. The Arctic Ocean: water masses and energy exchanges.
Washington Press, Seattle. In: L. Rey and B. Stonehouse (eds.). The Arctic Ocean, pp. 43-68.
Kinder, T.H., D.C. Chapman and J.A. Whitehead Jr., 1986. Westward Macmilan Press, London.
intensification of the mean circulation on the Bering Sea shelf. J. Phys. Lewis, P., 1991. Sedimentation in the Mackenzie Delta. In: P. Marsh and
Oceanogr. 16: 1217-1229. C.S.L. Ommanney (eds.). Mackenzie Delta: Environmental interactions
Kirkland, R.A. and C.B.J. Gray, 1986. Reconnaissance of the chemical and implications of development, pp. 37-38. Proceedings of the Work-
and biological limnology in four large lakes of the Yukon River basin. shop on the Mackenzie Delta 17-18 October, 1989, Saskatoon, Saskat-
Environment Canada, National Hydrology Research Institute, Saska- chewan. Environment Canada, Saskatoon, Saskatchewan, National
toon, Saskatchewan, Canada, Paper No. 33, 23p. Hydrology Research Institute Symposium No.4.
Kislalioglu, M. and F. Berkes, 1994. La Grande 2 (LG2 Reservoir). In: Li, S.M., R.W. Talbot, L.A. Barrie, R.C. Harriss, C.I. Davidson and J.-L.
R.J. Allan, M. Dickman, C.B. Gray and V. Cromie (eds.). The book Jaffrezo, 1993. Seasonal and geographical variations of methane sul-
of Canadian lakes, pp. 188-208. The Canadian Association on Water phanic acid in the Arctic troposphere. Atmos. Environ. 27A: 3011-3024.
Quality, Gloucester, Ontario, Canada. Monograph Series #33. Lindsey, G.C., K. Patalas, R.A. Bodaly and C.P. Archibald, 1981. Glaci-
Klinck, J.M., J.J. O’Brien and H. Svendsen, 1981. A simple model of ation and physical, chemical and biological limnology of Yukon Lakes.
fjord and coastal circulation interaction. J. Phys. Oceanogr. 11: 1612- Fisheries and Oceans Canada, Winnipeg, Manitoba, Canada, Can.
1626. Tech. Rep. Fisheries and Aquatic Sciences No. 966, 37p.
Knezovich, J.P., F.L. Harrison and R.G. Wilhelm, 1987. The bioavailabil- Lisitsin, A.P., 1988. Avalanche sedimentation and interuptions in sediment
ity of sediment-sorbed organic chemicals: A review. Water Air Soil accumulation in seasonal oceans. Nauka, Moscow, 308p. (In Russian).
Pollut. 32: 233-245. Liss, P.S. and P.G. Slater, 1974. Flux of gases across the air-sea interface.
Kotchetov, S.V., I.Y. Kulakov, V.K. Kurajov, L.A. Timokhov and Y.A. Nature 247: 181-184.
Vanda, 1994. Hydrometeoroligical regime of the Laptev Sea. Federal Livingston, H.D., 1988. The use of Cs and Sr isotopes as tracers in the Arc-
Service of Russia for Hydrometeorology and Monitoring of the Envi- tic Mediterranean Seas. Phil. Trans. R. Soc. London A325: 161-176.
ronment, The Arctic and Antarctic Research Institute, St. Petersburg, Livingston, H.D., S.L. Kupferman, V.T. Bowen and R.M. Moore, 1984.
Russia. manuscript, 85p. Vertical profile of artificial radionuclide concentrations in the central
Kovacs, A., 1972. Ice scoring marks floor of Arctic shelf. The Oil and Arctic Ocean. Geochim. Cosmochim. Acta 48: 2195-2203.
Gas Journal, October 23, 1972, vol. 70, no. 43. Lockhart, W.L., 1996. Depositional trends – Lake and marine sediments.
Krishnappan, B.G., R. Stephens, J.A. Kraft and B.H. Moore, 1995. Size In: J.L. Murray, R.G. Shearer and S.L. Han (eds.). Synopsis of research
distribution and transport of suspended particles, Athabasca River, conducted under the 1994/95 Northern Contaminants Program, pp.
February and September, 1993. Northern River Basins Study, Edmon- 61-69. Indian Affairs and Northern Development, Ottawa, Canada,
ton, Alberta, June, 1995, Project Report no. 51, 10p. Environmental Studies no. 73.
Kristmannsson, S.S., S.-A. Malmberg and J. Briem, 1989. Inflow of warm Lockhart, W.L. and D. Muir, 1996. Food chain accumulation, biological
Atlantic water to the subarctic Iceland Sea. Rapp. P.-v. Réun. Cons. effects and sediment contamination in Lake Laberge and other Yukon
int. Explor. Mer. 188: 74. Lakes. In: J.L. Murray, R.G. Shearer and S.L. Han (eds.). Synopsis of
Krysell, M. and D.W.R. Wallace, 1988. Arctic ocean ventilation studied research conducted under the 1994/95 Northern Contaminants Pro-
with a suite of anthropogenic halocarbon tracers. Science 242: 746-748. gram, pp. 199-207. Indian Affairs and Northern Development, Otta-
Ku, T.-L. and W.S. Broecker, 1967. Rates of sedimentation in the Arctic wa, Canada, Environmental Studies no. 73.
Ocean. Prog. Oceanogr. 4: 95-104. Lockhart, W.L., R. Wagemann, B. Tracey, D. Sutherland and D.J. Thomas,
Kuusisto, E., 1994. North hydrology in Finland. In: T.D. Prowse, C.S.L. 1992. Presence and implications of chemical contaminants in the fresh-
Ommaney and L.E. Watson (eds.). Northern hydrology: International waters of the Canadian arctic. Sci. Total Environ. 122(1-2): 165-243.
perspectives, pp. 129-146. Environment Canada, Saskatoon, Saskat- Loeng, H., 1989. Ecological features of the Barents Sea. In: L. Rey and V.
chewan, NHRI Science Report No.3. Alexander (eds.). Proceedings of the sixth Conference of the Comité
Kuznetsov, I.M., 1983. About the ice exchange through Karskiye Vorota Arctique International, 13-15 May 1985, Leiden. E.J. Brill, Leiden,
Strait. Proc. AARI 380: 123-128. (In Russian). The Netherlands.
Kvambekk, A.S. and T. Vinje, 1993. The ice transport through Fram Loeng, H., 1991. Features of the physical oceanographic conditions of the
Strait. Paper presented at the Nansen Centennial Symposium, Bergen, Barents Sea. Polar Res. 10: 5-18.
Norway, 21-25 June, 1993. (Unpublished). Loeng, H. and T.E. Vinje, 1979. On the sea ice conditions in the Green-
LaChapelle, E.R., 1969. Field guide to snow crystals. University of Wash- land and Barents seas. In: (Anon. ed.). Proceedings of the fifth Inter-
ington Press, Seattle. national Conference on Port and Ocean Engineering under Arctic
Lamb, D. and R. Blumenstein, 1987. Measurement of the area of entrap- Conditions. POAC Vol. 1. NTH, Trondheim.
ment of sulphur dioxide by rimed ice. Atmos. Environ. 21: 1765-1772. Loeng, H., V. Ozhigin, B. Ådlandsvik and H. Sagen, 1997: Water fluxes
Langner, J. and H. Rodhe, 1991. A global three-dimensional model of through the Barents Sea. ICES J. Mar. Sci. 54: 310-317.
the tropospheric sulphur cycle. J. Atmos. Chem. 13: 225-263. Lorenz, E.N., 1963. Deterministic non-periodic flow. J. Atmos. Sci. 20:
Larson, L.W. and E.L. Peck, 1974. Accuracy of precipitation measure- 130-141.
ments for hydrologic modelling. Water Resour. Res. 10(4): 857-863. Macdonald, R.W. and J.M. Bewers, 1996. Contaminants in the arctic
Lauff, G.H. (ed.). 1967. Estuaries. American Association for the marine environment: priorities for protection. ICES J. mar. Sci. 53:
Advancement of Science, Washington D.C., 757p. 537-563.
Laws, E.A., 1993. Aquatic pollution: An introductory text. John Wiley Macdonald, R.W. and E.C. Carmack, 1991. The role of large-scale under-
and Sons Inc., New York, 611p. ice topography in separating estuary and ocean on an Arctic shelf.
Lazier, J.R.N., 1973. The renewal of Labrador Sea water. Deep-Sea Res. Atmosphere-Ocean 29(1): 37-53.
20: 341-353. Macdonald, R.W. and D.J. Thomas, 1991. Chemical interactions and sedi-
Lazier, J.R.N., 1980. Oceanographic conditions at O.W.S. Bravo 1964- ments: The western Canadian Arctic shelf. Cont. Shelf Res. 11: 843-864.
1974. Atmosphere-Ocean 18: 227-238. Macdonald, R.W., E.C. Carmack, F.A. McLaughlin, K. Iseki, D.M. Mac-
Lazier, J.R.N. and D.G. Wright, 1993. Annual velocity variations in the donald and M.C. Obrien, 1989. Composition and modification of
Labrador Current. J. Phys. Oceanogr. 23: 659-678. water masses in the Mackenzie shelf estuary. J. Geophys. Res. 94:
LeBlond, P.H., 1980. On the surface circulation in some channels of the 18057-18070.
Arctic archipelago. Arctic 33: 189-197. Macdonald, R.W., E.C. Carmack and D.W.R. Wallace, 1993. Tritium and
Lee, A. and D. Ellett, 1967. On the water masses of the Northwest At- radiocarbon dating of Canada Basin deep waters. Science 259: 103-104.
lantic Ocean. Deep-Sea Res. 14: 183-190. Macdonald, R.W., D.W. Paton, E.C. Carmack and A. Omstedt, 1995. The
Leister, D.L. and J.E. Baker, in press. Influence of non-filterable particles freshwater budget and under-ice spreading of Mackenzie River water
on the scavenging of organic contaminants by rain. Environ. Sci. Tech- in the Canadian Beaufort Sea based on salinity and 18O/16O measure-
nol. 29: (in press). ments in water and ice. J. Geophys. Res. 100: 895-919.
Lemke, P., 1987. A coupled one-dimensional sea ice-ocean model. J. Geo- Macdonald, R.W., S.M. Solomon, R.E. Cranston, H.E. Welch, M.B. Yun-
phys. Res. 92: 13164-13172. ker and C. Gobeil, 1995a. A sediment and organic carbon budget for
Lesack, L.F.W., P. Marsh and R.E. Hecky, 1991. Ice-cover growth and the Canadian Beaufort shelf. Mar. Geol. (submitted)
freeze-out of solutes in a Mackenzie Delta lake. In: T.D. Prowse and Mackay, D., 1991. Multimedia environmental models: the Fugacity ap-
C.S.L. Ommanney, (eds.). Northern hydrology: Selected perspectives, proach. Lewis Publishers, Chelsea, Michigan.
pp. 219-236. Canada Ministry of Supply and Services, Ottawa, NHRI Mackay, D. and W. Stiver, 1991. Predictability and environmental chem-
Symposium No.6. istry. In: R. Grover and A.J. Cessna (eds.). Environmental chemistry of
Leuenberger, C., J. Czuczwa, E. Heyerdahl and W. Giger, 1988. Aliphatic herbicides, Vol. II, pp. 281-297. CRC Press, Boca Raton, Florida.
and polycyclic aromatic hydrocarbons in urban rain, snow and fog. Mackay, D. and F. Wania, 1995. Transport of contaminants to the Arctic:
Atmos. Environ. 22(4): 695-705. partitioning, processes and models. Sci. Total Environ. 160/161: 25-38.
Levitus, S., 1982. Climatological atlas of the world ocean. National Mackay, D., S. Paterson and W.Y. Shiu, 1992. Generic models for evalu-
Oceanic and Atmospheric Administration, U.S.Dept. of Commerce, ating the regional fate of chemicals. Chemosphere 24: 695-717.
Malmberg, S.-A., 1984. Hydrographic conditions in the east Icelandic Arctic Centre, Rovaniemi, The State of the Arctic Environment Re-
Current and sea ice in north Icelandic waters, 1970-1980. Rapp. P.-v. ports 2, pp. 82-153.
Réun. Cons. int. Explor. Mer. 185: 170-178. Melnikov, I.A. and G.L. Pavlov, 1978. Characteristics of organic carbon
Malmberg, S.-A. and S. Jónsson, 1997. Timing of deep convection in the distribution in the waters and ice of the Arctic Basin. Oceanology
Greenland and Iceland Seas. ICES J. Mar. Sci. 54(3): 300-309. 18(2): 163-167.
Malmberg, S.-A. and S.S. Kristmannsson, 1992. Hydrographic conditions Meybeck, M., 1993. C, N, P and S in rivers: from sources to global inputs.
in Icelandic waters 1980-1989. ICES mar. Sci. Symp. 195: 76-92. In: R. Wollast, F.T. Mackenzie and L. Chou (eds.). Interactions of C,
Malmberg, S.-A., S.S. Kristmannsson and E. Buch, 1990. Greenland Sea N, P and S. Biogeochemical Cycles and Global Change. NATO ASI
Project in the western part of the Iceland Sea from Jan Mayen to Den- Series, Volume 14, pp. 163-193. Springer-Verlag Berlin-Heidelberg.
mark Strait. ICES C.M. 1990 C: 27. International Council for the Ex- Meybeck, M. and A. Ragu, 1995. GEMS/WATER contribution to the
ploration of the Sea, Copenhagen. Global Register of River Inputs (GLORI). Provisional Final Report.
Manley, T.O., K.L. Hunkins and R.D. Meunch, 1987. Current regimes UNEP/WHO/UNESCO.
across the east Greenland polar front at 78o 40’ north latitude during Meybeck, M., D.V. Chapman and R. Helmer (eds.). 1989. Global envi-
summer 1984. J. Geophys. Res. 92: 6741-6753. ronment monitoring system, global freshwater quality: A first assess-
Mantyla, A.W. and J.L. Reid, 1983. Abyssal characteristics of the world ment. World Health Organization and United Nations Environment
ocean waters. Deep-Sea Res. 30: 805-833. Programme. Basil Balckwell Ltd., Oxford, United Kingdom, 306p.
Martin, J.J., P.K. Wang and H.R. Pruppacher, 1980. A theoretical deter- Meybeck, M.G., G. Friedrich, R. Thornas and D. Chapman, 1992. Rivers
mination of the efficiency with which aerosol particles are collected In: D. Chapman (ed.). Water quality assessments, pp. 239-316. Chap-
by simple ice crystal plates. J. Atmos. Sci. 37: 1628-1638. man and Hall, London.
Martin, J.M., D.M. Guan, F. Elbaz-Poulichet, A.J. Thomas and V.V. Gor- Midttun, L., 1985. Formation of dense bottom water in the Barents Sea.
deev, 1993. Preliminary assessment of the distribution of some trace Deep-Sea Res. 32: 1233-1241.
elements (As, Cd, Cu, Fe, Ni, Pb and Zn) in a pristine aquatic envi- Midttun, L. and H. Loeng, 1987. Climatic variations in the Barents Sea.
ronment: the Lena River estuary (Russia). Mar. Chem. 43: 185-199. In: H. Loeng (ed.). The effect of oceanographic conditions on distribu-
Martin, S., 1981. Frazil ice in rivers and oceans. Ann. Rev. Fluid Mech. tion and population dynamics of commercial fish stocks in the Barents
13: 379-397. Sea. Proc. 3rd Soviet-Norwegian Symp. Murmansk, 26-28 May 1986.
Martinson, K., M. Sydor, N. Marcotte and S. Beltaos, 1993. RIVICE Institute of Marine Research, Bergen, Norway.
model update. In: T.D. Prowse (ed.). Proceedings of the Workshop on Milburn, D. and T.D. Prowse, 1996. The effect of river-ice break-up on
Environmental Aspects of River Ice, August 18-20, 1993, pp. 127- suspended sediment and selected trace-element fluxes. Nordic Hydro-
144. National Hydrology Research Institute, Saskatoon, Saskatche- logy 27: 69-84.
wan, Canada. Miller, N.L. and P.K. Wang, 1991. A theoretical determination of the col-
Mason, R.P. and W.E. Fitzgerald, 1991. Mercury speciation in open ocean lection rates of aerosol particles by falling ice crystal plates and columns.
waters. Water Air Soil Pollut. 56: 779-789. Atmos. Environ. 25A: 2593-2606.
Maykut, G.A. and N. Untersteiner, 1971. Some results from a time-depen- Milliman, J.D. and J.P.M. Syvitski, 1992. Geomorphic/tectonic control of
dent thermodynamic model of sea ice. J. Geophys. Res. 76: 1550-1575. sediment discharge to the ocean: the importance of small mountainous
McCarthy, L.H., T.G. Williams, G.R. Stephens, J. Peddle, K. Robertson rivers. J. Geol. 100: 525-544.
and D. Gregor, 1996. Baseline studies in the Slave River, NWT, 1990- Mitchell. M., 1956. Early identification of Arctic haze – reference to the
1994: Part I. Evaluation of the chemical quality of water and suspended Alaskan Arctic. J. Atmos. Terrestrial Phys. Special Suppl.: 195-211.
sediment from the Slave River (NWT). Sci. Total Environ. (in press). Mitra, S.K., U. Barth and H.R. Pruppacher, 1990. A laboratory study of
McCartney, M.S. and L.D. Talley, 1984. Warm-to-cold water conversion the efficiency with which aerosol particles are scavenged by snow
in the northern North Atlantic Ocean. J. Phys. Oceanogr. 14: 922-935. flakes. Atmos. Environ. 24A: 1247-1254.
McCrea, R.C. and J.D. Fischer, 1986. Heavy metal and organochlorine Moore, R.M. and J.N. Smith, 1986. Disequilibria between Ra, 21OPb and
contaminants in the five major Ontario rivers of the Hudson Bay 210Po in the Arctic Ocean and the implications for chemical modifica-
Lowland. Water Pollut. Res. J. Can. 21: 225-234. tion of the Pacific water inflow. Earth Planet. Sci. Lett. 77: 285-292.
McCrea, R.C. and G.M. Wickare, 1986. Organochlorine contaminants in Moss, A.J. and P.H. Walker, 1978. Particle transport by continental water
peatlands and selected estuary sites of the Moose River. Water Pollut. flows in relation to erosion, deposition, soils and human activities.
Res. J. Can. 21(2): 251. Sedim. Geol. 20: 81-139.
McDougall, T.J., 1983. Greenland Sea bottom water formation: a balance Mudroch, A., R.J. Allan and S.R. Joshi, 1992. Geochemistry and organic
between advection and double-diffusion. Deep-Sea Res. 30: 1109-1117. contaminants in the sediments of Great Slave Lake, Northwest Terri-
McLaren, I.A., 1964. Zooplankton of Lake Hazen, Ellesmere Island and tories, Canada. Arctic 45(1): 10-19.
a nearby pond with special reference to the copepod Cyclops scutifer Muench, R., 1983. Mesoscale oceanographic features associated with the
Sars. Can. J. Zool. 42: 613-629. central Bering Sea ice edge: February-March 1981. J. Geophys. Res.
McLaughlin, F.A., E.C. Carmack, R.W. Macdonald and J.K.B. Bishop, 88: 2715-2722.
1996. Physical and geochemical properties across the Atlantic/Pacific Muench, R.D., 1971. The physical oceanography of the northern Baffin
water mass front in the southern Canadian Basin. J. Geophys. Res. Bay region. In: Baffin Bay north water project science, Report I. Arctic
101: 1183-1197. Institute of North America, Montreal, Canada.
McVetty, B.D. and R.A. Hites, 1988. Atmospheric deposition of poly- Muench, R.D., 1990. Greenland Sea Project. EOS 71: 750-755.
cyclic aromatic hydrocarbons to water surfaces: A mass balance Müller, F., A. Ohmura and R. Braithwaite, 1976. On the climatic influ-
approach. Atmos. Environ. 22: 511-536. ence of North Water. In: International Geographical Congress: Sympo-
Means, J.C., J.J. Hassett, S.G. Wood and W.L. Banwart, 1979. Sorption sium on the geography of polar countries (extended summaries), pp.
properties of energy related pollutants and sediments. In: P.W. Jones 55-58. Gidrometeoizdat, Leningrad.
and P. Leber (eds.). Polynuclear aromatic hydrocarbons, pp. 327-340. Munro, D.S., 1991. On modelling surface meltwater discharge from
Third international symposium on Chernisigy and biology. Carcinogen- Arctic and Alpine glaciers. In: T.D. Prowse and C.S.L. Ommanney
esis and miaagenesis. Ann Arbor Science Publishers Inc., Ann Arbor. (eds.). Northern hydrology: Selected perspectives, pp. 253-262.
Meincke, J., 1983. The modern current regime across the Greenland- Proceedings of the Northern Hydrology Symposium, July 10-12, 1990.
Scotland ridge. In: A. Bott et al. (eds.). Structure and development of Environment Canada, National Hydrology Research Institute, Saska-
the Greenland-Scotland Ridge, pp. 637-650. Plenum, New York. toon, Saskatchewan, Canada.
Melling, H. and E.L. Lewis, 1982. Shelf drainage Beaufort Sea and their Murakami, M., T. Kimura, C. Magono and K. Kikuchi, 1983. Observa-
effect on the Arctic Ocean pycnocline. Deep-Sea Res. 29: 967-986. tions of precipitation scavenging for water-soluble particles. J. Met.
Melling, H. and R.M. Moore, 1995. Modification of halocline source Soc. Japan 61: 346-357.
waters during freezing on the Beaufort Sea shelf: evidence from oxy- Murakami, M., C. Magono and K. Kikuchi, 1985. Experiments on aerosol
gen isotopes and dissolved nutrients. Cont. Shelf Res. 15(1): 89-113. scavenging by natural snow crystals Part III. The effect of snow crystal
Melling, H. and D.A. Riedel, 1996. Development of seasonal pack ice in charge on collection efficiency. J. Met. Soc. Japan 63: 1127-1137.
the Beaufort Sea during the winter of 1991-92: A view from below. J. Murakami, M., K. Kikuchi and C. Magono, 1985. Experiments on aero-
Geophys. Res. 101(C5): 11975-11991. sol scavenging by natural snow crystals Part II. Attachment rate of 0.1
Melling, H., R.A. Lake, D.R. Topham and D.B. Fissel, 1984. Oceanic µm Dia. particles to stationary snow crystals. J. Met. Soc. Japan 63:
thermal structure in the western Canadian Arctic. Cont. Shelf Res. 130-135.
3: 233-258. Murphy, T.J., 1995. Comment on: “Seasonal variations in air-water ex-
Mellor, G.L., 1973. Analytic prediction of the properties of stratified change of PCBs in Lake Superior”. Environ. Sci. Technol. 29: 846-847.
planetary surface layers. J. Atmos. Sci. 30: 1061-1069. Mysak, L.A. and D.K. Manak, 1989. Arctic sea-ice extent and anomalies,
Mellor, G.L., 1993. Users guide for a three-dimensional, primitive equa- 1953-1984. Atmosphere-Ocean 27(2): 376-405.
tion, numerical ocean model. AOS Program, Princeton University, Nakamura, N. and A.H. Oort, 1988. Atmospheric heat budgets of the
Princeton Report NJ 08540. polar regions. J. Geophys. Res. 93D: 9510-9524.
Mellor, G.L. and L.H. Kantha, 1989. An ice-ocean coupled model. J. Nansen, F., 1906. Northern waters: Captain Roald Amundsen’s oceano-
Geophys. Res. 94: 10937-10954. graphic observations in the Arctic seas in 1901. Videnskabs-Selskabets
Mellor, G.L. and T. Yamada, 1974. A hierarchy of turbulence closure Skrifter. I. Mathematisk-Naturvitenskapelig Kalsee, 1906, Dybwad,
models for planetary boundary layers. J. Atmos. Sci. 31: 1791-1806. Christiania, 1, 145p.
Mellor, G.L. and T. Yamada, 1982. Development of a turbulence closure NEFCO, 1995. Proposals for environmentally sound investment projects
model for geophysical fluid problems. Rev. Geophys. Space Phys. in the Russian part of the Barents Region. V.1 Non-radioactive conta-
20: 851-875. mination. AMAP Report, AMAP Secretariat, Oslo, 147p.
Melnikov, S.A., 1991. Report on heavy metals. University of Lapland, Newbury, R.W., G.K. McCullough and R.E. Hecky, 1984. The Southern
Indian Lake impoundment and Churchill River diversion. Can. J. sequences of dumping radioactive waste into Arctic Seas, Oslo, Nor-
Fish. Aquat. Sci. 41(4): 548-557. way, 1-5 February, 1993.
Newton, J.L. and L.K. Coachman, 1974. Atlantic water circulation in the Pavlov, V.K., L.A. Timohov, G.A. Baskakov, M.Y. Kulakov, V.K. Kurazhov,
Canada Basin. Arctic 21: 297-303. P.V. Pavlov, S.V. Pivovarov and V.V. Stanovoy, 1994. Hydrometeorolo-
Nicholson, K.W., J.R. Branson and P. Geiss, 1991. Field measurements of gical regime of the Kara, Laptev and East Siberian Seas. Arctic and
the below-cloud scavenging of particulate material. Atmos. Environ. Arntarctic Research Institute, St. Petersburg, Russia. manuscript, 179p.
25A: 771-777. Pawlowicz, R.A., J.F. Lynch, W.B. Owens and P.F. Worcester, 1994. Deep
Norton, S.A., 1986. A review of the chemical record in lake sediments of water formation in the Greenland Sea. Oceanus 37:2, 9-11.
energy related air pollution and its effects on lakes. Water Air Soil Pearce, C.M., 1993. Overbank sedimentation patterns on the Mackenzie
Pollut. 39: 331-345. Delta, NWT. Report to the NWT Programs Branch, Environment
Nurnberg, D., D.K. Futterer, F. Niessen, N. Norgaard-Pedersen, C.J. Canada, Yellowknife, NWT.
Schubert, R.F. Spielhagen and M. Wahsner, 1995. The depositional Peulv, Ú.S., I. Broyelle, M.-A. Sicrer, I. Bouloubassi, A. Lorre, A. Saliot,
environment of the Laptev Sea continental margin: Preliminary results J.W. de Leeuw and M. Baas, 1993. Characterization of the organic
from the R/V Polarstern ARK IX-4 cruise. Polar Res. 14: 43-53. matter in an Arctic delta (Lena River) using biomarkers and macro-
O’Brien, M.C., K. Iseki, R.W. Macdonald, J.R. Forbes, Y. Liangfeng and molecular indicators. Organic Geochemistry, Stavanger, Falch Hurtig-
D. McCullough, 1991. Sediment trap data collected in the Beaufort trykk, pp. 393-397.
Sea, March 1987 - March 1988. Can. Data Rep. Hydrogr. Ocean Sci. Pfirman, S., M.A. Lange, I. Wollenburg and P. Schlosser, 1990. Sea ice
60(8): 237. characteristics and the role of sediment inclusions in deep-sea deposi-
Oberhuber, J.M., 1993a. Simulation of the Atlantic circulation with a tion: Arctic-Antarctic comparisons. In: U. Bleil and J. Thiede (eds.).
coupled sea ice-mixed layer-isopycnal general circulation model. Part Geological history of the polar oceans: Arctic versus Antarctic, pp.
I: Model description. J. Phys. Oceanogr. 23: 808-829. 187-211. Kluwer Academic Publishers, Dordrecht, The Netherlands.
Oberhuber, J.M., 1993b. Simulation of the Atlantic circulation with a Pfirman, S.L., D. Bauch and T. Gammelsrød, 1994. The northwestern
coupled sea ice-mixed layer-isopycnal general circulation model. Part Barents Sea: summer water mass distribution and circulation. In: O.M.
II: Model experiment. J. Phys. Oceanogr. 23: 830-845. Johannessen, R.D. Muench and J. E. Overland (eds.). The role of the
Oehme, M., 1991. Transport of polychlorinated aromatics and pesticides Polar Oceans in shaping the global environment. Geophysical Mono-
from North America and Eurasia to the Arctic. In: R. Shearer (ed.). graph Series, AGU 85, AGU, Washington D.C. 525p
Proc. 8th Int. Conf. on global significance of the transport and accu- Pfirman, S.L., H. Eicken, D. Bauch and W. Weeks, 1995a. The potential
mulation of polychlorinated hydrocarbons in the Arctic. Comite Arc- transport of pollutants by Arctic sea ice. Sci. Total Environ. 159: 129-
tique International, Oslo, Norway, Sept. 18-22, 1989. 146.
Oehme, M., J.-E. Haugen and M. Schlabach, 1995. Ambient air levels of Pfirman, S.L., J. Koegler and B. Anselme, 1995b. Coastal environments of
persistent organchlorines in spring 1992 at Spitsbergen and Norwe- the western Kara and eastern Barents Seas. Deep-Sea Res. 42: 1391-1412.
gian mainland: comparison with 1984 result and quality control mea- Pham, M., J.-F. Muller, G.P. Brasseur, C. Granier and G. Megie, 1995. A
sures. Sci. Total Environ. 160/161: 139-152. three-dimensional study of the tropospheric sulfur cycle. J. Geophys.
Omstedt, A., E.C. Carmack and R.W. Macdonald, 1994. Modeling the Res. 100: 26061-26092.
seasonal cycle of salinity in the Mackenzie shelf/estuary. J. Geophys. Piacsek, S., R. Allard and A. Warn-Varnas, 1991. Studies of the Arctic ice
Res. 99: 10011-10021. cover and upper ocean with a coupled ice-ocean model. J. Geophys.
Östlund, H.G., 1982. The residence time of the freshwater component in Res. 96: 4631-4650.
the Arctic Ocean. J. Geophys. Res. 87: 2035-2043. Pirazzoli, P.A., 1991. World atlas of Holocene sea-level changes. Elsevier
Östlund, H.G., 1993. Transport patterns of Siberian river water in the Oceanography Series 58, Elsevier, New York, 300p.
Arctic Basin. In: P. Strand and E. Holm (eds.). Environmental radio- Plumb, R.A. and J.D. Mahlman, 1987. The zonally averaged transport
activity in the Arctic and Antarctic, pp. 151-155. Norwegian Radia- characteristics of the GFDL circulation/transport model. J. Atmos. Sci.
tion Protection Authority, Østerås, Norway. 44: 298-327.
Östlund, H.G. and G. Hut, 1984. Arctic Ocean water mass balance from Pomeroy, J.W. and D.M. Gray, 1995. Snowcover accumulation, relocation
isotope data. J. Geophys. Res. 89C: 6373-6381. and management. National Hydrology Research Institute Saskatoon,
Östlund, H.G., G. Possnert and J.H. Swift, 1987. Ventilation rate of the Canada, Science Report no. 7, 144p.
deep Arctic Ocean from carbon 14 data. J. Geophys. Res. 92C: 3769- Pomeroy, J. and H.G. Jones, 1996. Windblown snow: sublimation, trans-
3777. port and changes to polar snow. In: E. Wolff and R.C. Bales (eds.).
Ostromogil’skii, A.K., A.O. Kokorin and M.I. Afanas’ev., 1985. Model- Chemical exchange between the atmosphere and polar snow. Global
ling the global DDT cycle. Met. Gidrologiya 2: 37-45. Environmental Change. NATO ASI Series I, Volume 43, pp. 453-490.
Ottar, B., 1989. Arctic air pollution: a Norwegian perspective. Atmos. Springer-Verlag, Heidelberg, Germany.
Environ. 23: 2349-2356. Pomeroy, J.W., T.D. Davies and M. Tranter, 1991. The impact of blowing
Overland, J.E., 1981. Marine climatology of the Bering Sea. In: D.W. snow on snow chemistry. In: T.D. Davies, M. Tranter and H.G. Jones
Hood and J.A. Calder (eds.). The Eastern Bering Sea shelf: Oceano- (eds.). Seasonal snowpacks – Processes of compositional change.
graphy and resources, pp. 15-27. Office of Marine Pollution Assess- NATO ASI Series G, pp. 71-113. Springer-Verlag, Berlin, Germany.
ment, National Oceanic and Atmospheric Administration. University Potanin, V.A. and S.V. Korotov, 1988. Seasonal variability of the main
of Washington Press, Seattle. currents in the southern Barents Sea and water exchange with the
Overland, J.E. and A.T. Roach, 1987. Northward flow in the Bering and adjacent areas. Geological and geographical problems of natural re-
Chukchi Seas. J. Geophys. Res. 92: 7097-7105. sourses exploitation in the northern seas. Murmansk, 1988, pp. 81-90
Pacyna, J.M., 1995. The origin of Arctic air pollutants: lessons learned (In Russian).
and future research. Sci. Total Environ. 160/161: 39-53. Preller, R.H. and R. Edson (eds.), 1995. Proceedings of the ONR/NRL
Pacyna, J.M. and B. Ottar, 1988. Vertical distribution of aerosols in the Workshop on Modeling the Dispersion of Nuclear Contaminants in
Norwegian Arctic. Atmos. Environ. 22: 2213-2222. the Arctic Seas. October 18-19, 1994. Monterey, California. Naval
Paluskiewicz, T. and H.J. Niebauer, 1984. Satelitte observations of circu- Research Laboratory, Stennis Space Center, Monterey, California,
lation in the eastern Bering Sea. J. Geophys. Res. 89: 3663-3678. U.S.A. Report no.: NRL/MR/7322—95-7584.
Paquette, R.G. and R.H. Bourke, 1974. Observations on the coastal cur- Price, J.F. and M. O’Neil Baringer, 1994. Outflows and deep water pro-
rent of Arctic Alaska. J. mar. Res. 32: 195-207. duction by marginal seas. Prog. Oceanogr. 33: 161-200.
Paquette, R.G. and R.H. Bourke, 1981. Ocean circulation and fronts as Pritchard, D.W., 1967. Observations of circulation in coastal plain estuar-
related to ice melt-back in the Chukchi Sea. J. Geophys. Res. 86: 4215- ies, In: G.H. Lauff (ed.). Estuaries. American Association for the
4230. Advancement of Science, Washington D.C., Publ. V. 83, pp. 37-44.
Paquette, R.G., R.H. Bourke, J.F. Newton and W.F. Perdue, 1985. The Prinsenberg, S.J., 1983. Time variability of physical oceanographic para-
east Greenland Polar Front in autumn. J. Geophys. Res. 90: 4866-4882. meters in Hudson Bay. Le Naturaliste Canadian 109: 685-700.
Parkinson, C.L. and W.M. Washington, 1979. A large-scale numerical Prinsenberg, S.J., 1986a. On the physical oceanography of Foxe Basin. In:
model of sea ice. J. Geophys. Res. 84(C): 311-337. I.P. Martini (ed.). Canadian inland seas, pp. 217-236. Elsevier Science,
Parkinson, C.L., J.C. Comiso, H.J. Zwally, D.J. Cavalieri, P. Gloersen New York.
and W.J. Campbell, 1987. Arctic Sea ice, 1973-1976: Satellite passive- Prinsenberg, S.J., 1986b. Salinity and temperature distributions of Hudson
microwave observations. NASA, Washington D.C., 296p. Bay and James Bay. In: I.P. Martini (ed.). Inland coastal seas, pp 163-
Parungo, F., C. Nagamoto and R. Madel, 1987. A study of mechanisms 186. Elsevier Press, New York.
of acid rain formation. J. Atmos. Sci. 44: 3162-3174. Prinsenberg, S.J., 1986c. The circulation pattern and current structure of
Pavlov, V.K., 1993. Formation features of the structure and variability of Hudson Bay. In: I.P. Martini (ed.). Canadian inland seas, pp. 187-204.
the hydrological processes of the Kara Sea area. Proceedings of the Elsevier Science, New York.
Conference on Radioactivity and Environmental Security in the Prinsenberg, S.J., 1991. Effects of hydro-electric projects on Hudson Bay’s
Oceans: New research and policy priorities in the Arctic and North marine and ice environments. James Bay Publication Series, No. 2, 7p.
Atlantic, Woods Hole, Massachusetts, U.S.A., 7-9 June, 1993. Woods Proehl, J.A. and M. Rattray Jr., 1984. Low-frequency response of wide
Hole Institute of Oceanography, Woods Hole, Mass., U.S.A. deep estuaries to non-local atmospheric forcing. J. Phys. Oceanogr.
Pavlov, V. and S.L. Pfirman, 1995. Hydrographic structure and variability 14: 904-921.
of the Kara Sea: Implications for pollutant distribution. Deep-Sea Res. Prowse, T.D., 1994. Environmental significance of ice to stream flow in
42: 1369-1390. cold regions. Freshw. Biol. 32: 241-259.
Pavlov, V.K. and V.A. Volkov, 1993. Numerical modeling of the oceano- Prowse, T.D. and M. Conly, 1996. Inpacts of flow regulation on the aquat-
graphic fields and transport of radioactive contaminants in the Kara ic ecosystem of the Peace and Slave Rivers. Northern River Basins Study,
seas. International Meeting on assessment of actual and potential con- Edmonston, Alberta, Synthesis Report No. 1, 168p.
Prowse, T.D. and R.L. Stephenson, 1986. The relationship between win- water properties through Bering Strait. J. Geophys. Res. 100(C9):
ter lake cover, radiation receipts and the oxygen deficit in temperate 18443-18457.
lakes. Atmosphere-Ocean 24(4): 386-403. Robock, A., 1980. Ice and snow feedbacks and the latitudinal and season-
Pruppacher, H.R., 1981. The microstructure of atmospheric clouds and al distribution of climate sensitivity. J. Atmos. Sci. 40: 986-997.
precipitation. In: P.V. Hobbs and A. Deepak (eds.). Clouds: Their for- Roden, G.I., 1967. On river discharge into the northeastern Pacific Ocean
mation, optical properties and effects, pp. 93-186. Academic Press and Bering Sea. J. Geophys. Res. 72: 5613-5629.
Inc., New York. Rognerud, S. and E. Fjeld, 1993. Regional survey of heavy metals in lake
Quadfasel, D. and J. Meincke, 1987. Note on the thermal structure of sediments in Norway. Ambio 22(4): 206-212.
the Greenland Sea gyres. Deep-Sea Res. 34: 1883-1888. Rognerud, S., S.A. Norton and V. Duvalter, 1993. Heavy metal pollution
Quadfasel, D. and B. Rudels, 1990. Some new observational evidence for in lake sediments in the border areas between Russia and Norway. SFT
salt induced convection in the Greenland Sea. Inst. Meereskunde, (Norwegian State Pollution Control Authority), Oslo, Norway. SFT
Hamburg, Techn. Rep. 4-90, 30p. Report 522/93, 18p.
Quadfasel, D., B. Rudels and K. Kurz, 1988. Outflow of dense water from Romanov, I.P., 1992. The ice cover of the Arctic Ocean. AARI, St. Peters-
a Svalbard fjord into the Fram Strait. Deep-Sea Res. 35: 1143-1150. burg, 211p.
Quadfasel, D., B. Rudels and S. Selchow, 1992. The Central Bank vortex Roots, E.F., 1982. The changing Arctic marine environment: Some basic
in the Barents Sea: watermass transformation and circulation. ICES considerations. In: L. Rey (ed.). The Arctic Ocean. The hydrographical
mar. Sci. Symp. 195: 40-51. environment and the fate of pollutants. Comite Arctique International,
Quadfasel, D., A. Sy and B. Rudels, 1993. A ship of opportunity section Macmillan Press, London.
to the North Pole: upper ocean temperature observations. Deep-Sea Rosenberg, D.M., R.A. Bodaly, R.E. Hecky, R.W. Newbury and K. Pata-
Res. 40: 777-789. las, 1985. Hydroelectric development in northern Manitoba: the
Radke, L.F., J.H. Lyons, D.A. Hegg, P.V. Hobbs and I.H. Bailey, 1984. Churchill-Nelson River diversion and flooding of southern Indian
Airborne observations of Arctic aerosols. 1. Characteristics of Arctic Lake. Bull. Can. Soc. Environ. Biol. 42(2): 31-42.
haze. Geophys. Res. Lett. 11: 393-396. Ross, C.K., 1978. Overflow variability in Denmark Strait. ICES C.M.
Rahn, K.A. and G. Shaw, 1982. Sources and transport of Arctic air pol- 1978 C: 21. International Council for the Exploration of the Sea,
lution aerosol: a chronicle of six years of ONR research. Naval Res. Copenhagen.
Rev. March (3): 3-26. Royer, T.C. and W.J. Emery, 1984. Circulation in the Bering Sea 1982-
Rasch, M. and N. Nielsen, 1995. Coastal morpho-stratigraphy and Holo- 1983, based on satelitte tracked drifter observations. J. Phys. Oceanogr.
cene relative sea level changes at Tuapaat, southeastern Disko Island, 14: 1914-1920.
central West Greenland. Polar Res. 14: 277-289. Rozanski, K., 1979. Krypton-85 in the atmosphere 1950-1977. A data re-
Rawson, D.S, 1950. The physical limnology of Great Slave Lake. J. Fish. view. Environ. Int. 2: 139-143.
Res. Bd Can. 8: 3-66. Rudd, J.W.M., 1995. Sources of methyl mercury to freshwater ecosystems:
Rawson, D.S., 1947. Northwest Canadian fisheries surveys in 1944-1945, A review. Water Air Soil Pollut. 80: 697-713.
Great Slave Lake Bull. Fish. Res. Bd Can. LXXII: 45-68. Rudels, B., 1986a. The tetha-S relations in the northern seas: Implications
Raynor, G.S. and J.V. Hayes, 1983. Differential rain and snow scaveng- for the deep circulation. Polar Res. 4: 133-159.
ing efficiency implied by ionic concentration differences in winter pre- Rudels, B., 1986b. The outflow of polar water through the Arctic archi-
cipitation. In: H.R. Pruppacher, R.G. Semonin and W.G.N. Slinn (eds.). pelago and the oceanographic conditions in Baffin Bay. Polar Res. 4:
Precipitation scavenging, dry deposition and resuspension, pp. 249- 161-180.
264. Elsevier Science Publishing, New York. Rudels, B., 1987a. On the exchange of water masses between the Polar
Redkin, Y.N., 1973. Washout of atmospheric aerosols by snow in the sur- Ocean and the North Atlantic. Ph.D. thesis, University of Gothenburg.
face layer. In: V.M. Voloshchuk and Y.S. Sedunov (eds.). Hydrody- Rudels, B., 1987b. On the mass balance of the Polar Ocean with special
namics and thermodynamics of aerosols. John Wiley and Sons, New emphasis on the Fram Strait. Norsk Polarinst. Skr. 188: 1-53.
York. Rudels, B., 1990. Haline convection in the Greenland Sea. Deep-Sea Res.
Reed, R.K., 1984. Flow of the Alaskan Stream and its variations. Deep- 37: 1491-1511.
Sea Res. 31: 369-386. Rudels, B., 1993. High latitude ocean convection. In: D.B. Stone and S.K.
Reed, R.K., 1990. A year-long observation of water exchange between the Runcorn (eds.). Flow and creep in the solar system: Observations,
North Pacific and the Bering Sea. Limnol. Oceanogr. 35: 1604-1609. modeling and theory, pp. 323-356. Kluwer Academic Publishers, The
Reed, R.K. and P.J. Stabeno, 1993. The recent return of the Alaskan Netherlands.
Stream to Near Strait. J. mar. Res. 51: 515-527. Rudels, B. and D. Quadfasel, 1991. The Arctic Ocean component in the
Regnell, O., 1994. The effect of pH and dissolved oxygen levels on meth- Greenland-Scotland overflow. ICES C.M. 1991 C: 30. International
ylation and partitioning of mercury in freshwater model systems. En- Council for the Exploration of the Sea, Copenhagen.
viron. Pollut. 84: 7-13. Rudels, B., D. Quadfasel, H. Friedrich and M.-N. Houssais, 1989. Green-
Regnell, O. and A. Tunlid, 1991. Laboratory study of chemical specia- land Sea convection in the winter of 1987-1988. J. Geophys. Res. 94:
tion of mercury in lake sediment and water under aerobic and anaero- 3223-3227.
bic conditions. Appl. Environ. Microbiol. 57: 789-795. Rudels, B., E.P. Jones and L.G. Anderson, 1994. On the intermediate depth
Rehkopf, J., M. Newiger and H. Grassl, 1984. A 2-D model of global waters of the Arctic Ocean. In: O.M. Johannessen, R.D. Muench and
aerosol transport. Atmos. Environ. 18: 2745-1752. J.E. Overland, (eds.). The Polar Oceans and their role in shaping the
Reimnitz, E. and K.F. Bruder, 1972. River discharge into an ice-covered global environment: the Nansen Centennial Volume, pp. 33-46. Ameri-
ocean and related sediment dispersal, Beaufort Sea, coast of Alaska. can Geophysical Union, Washington D.C. Geophysical Monograph 85.
Geol. Soc. Am. Bull. 83: 861-866. Rudels, B., L.G. Anderson and E.P. Jones, 1996. Formation and evolution
Reimnitz, E. and E.W. Kempema, 1987. Field observations of slush ice of the surface mixed layer and halocline of the Arctic Ocean. J. Geo-
generated during freeze-up in arctic coastal waters. Mar. Geol. 77: phys. Res. 101(C4): 8807-8821.
219-231. Runnells, D.D., T.A. Shepherd and E.E. Angino. 1992. Metals in water,
Reimnitz, E., S.M. Graves and P.W. Barnes, 1988. Map showing Beaufort determining natural background concentrations in mineralized areas.
Sea coastal erosion, sediment flux, shoreline evolution, and erosional Environ. Sci. Technol. 62(12): 2316-2323.
shelf profile. U.S. Department of the Interior, U.S. Geological Survey, Sackinger, W.M., W.M. Serson, M.O. Jeffries, H.D. Sohoemake and M.-H.
Miscellaneous Investigations Series, 22p. + map. Yan, 1985. Ice islands as hazards to arctic offshore production struc-
Reimnitz, E., E.W. Kempema, W.S. Wefer, J.R. Clayton and J.R. Payne, tures. Proceedings of the Offshore Technology Conference, pp. 399-
1990. Suspended-matter scavenging by rising frazil ice. In: S.F. Ackley 408. May 6-9, 1985. Houston, Texas.
and W.F. Weeks (eds.). Sea ice properties and processes, pp. 97-100. Sadler, H.E., 1976. Water, heat and salt transports through Nares Strait,
Proceedings of the W.F. Weeks Sea Ice Symposium. Cold Regions Re- Ellesmere Island. J. Fish. Res. Bd Can. 33: 2286-2295.
search and Engineering Laboratory, Hanover, New Hampshire, CRREL Sadler, H.E., 1982. Water flow into Foxe Basin through Fury and Hecla
Monograph 90-1. Strait. Le Naturaliste Canadian 109: 701-707.
Reimnitz, E., L. Marincovich Jr., M. McCormick and W.M. Briggs, 1992. Sælen, O.H., 1983. Exchange of bottom water between the Greenland
Suspension freezing of bottom sediment and biota in the Northwest and Norwegian Seas. Paper presented at 18th General Assembly. Int.
Passage and implications for Arctic Ocean sedimentation. Can. J. Union of Geod. and Geophys., Hamburg, Germany.
Earth Sci. 29: 693-703. Sarynina, R.N., 1969. Conditions of origin of cold deep-sea waters in the
Reimnitz, E., M. McCormick, K. McDougall and B. Brouwers, 1993. Bear Island channel. ICES C.M. 1969 Symp:28, 1-8. International
Sediment export by ice rafting from a coastal polynya, Arctic Alaska, Council for the Exploration of the Sea, Copenhagen.
U.S.A. Arct. alp. Res. 23: 83-98. Saunders, P.M., 1990. Cold outflow from the Fareo Bank Channel. J. Phys.
Riebesell, U., I. Schloss and V. Smetacek, 1991. Aggregation of algae Oceanogr. 20: 29-43.
released from melting sea ice: implications for seeding and sedimenta- Sauter, D.P. and P.K. Wang, 1989. An experimental study of the scaveng-
tion. Polar Biol. 11: 239-248. ing of aerosol particles by natural snow crystals. J. Atmos. Sci. 46:
Riedlinger, S.H. and R.H. Preller, 1991. The development of a coupled 1650-1655.
ice-ocean model for forecasting ice conditions in the Arctic. J. Geo- Schindler, D.W., H.E. Welch, J. Kalff, G.J. Brunskill and N. Kritsch, 1974.
phys. Res. 96: 16955-16977. Physical and chemical limnology of Char Lake, Cornwallis Island,
Rigor, I., 1992. Arctic Ocean buoy program. Argos Newsl. 44: 1-3. (75°N lat.). J. Fish. Res. Bd Can. 31(5): 585-607.
Rigor, I. and R. Colony, 1997. Sea ice production and transport of pollu- Schlosser, P., G. Bönisch, B. Kromer, K.O. Münnich and K.P. Koltermann,
tants in the Laptev Sea, 1979-1993. Sci. Total Environ. 202: 89-110. 1990. Ventilation rates of the waters in the Nansen Basin of the Arctic
Roach, A.T., K. Aagaard, C.H. Pease, S.A. Salo, T. Weingartner, V. Pav- Ocean derived by a multi-tracer approach. J. Geophys. Res. 95: 3265-
lov and M. Kulakov, 1995. Direct measurements of transport and 3272.
Schlosser, P., G. Bönisch, M. Rhein and R. Bayer, 1991. Reduction of Slinn, W.G.N., 1983. Air-to-sea transfer of particles. In: P.S. Liss and
deepwater formation in the Greenland Sea during the 1980s: Evidence W.G.N. Slinn (eds.). Air-sea exchanges of gases and particles, pp. 299-
from tracer data. Science 251: 1054-1056. 405. D. Reidel Publishing Co., Boston.
Schlosser, P., D. Bauch, R. Fairbanks and G. Bönisch, 1994a. Arctic river- Smethie, W.M. Jr. and J.H. Swift, 1989. The tritium/krypton-85 age of
runoff: mean residence time on the shelves and in the halocline. Deep Denmark Strait overflow water and Gibbs Fracture Zone water just
Sea Res. 41(7): 1053-1068. south of Denmark Strait. J. Geophys. Res. 94: 8265-8275.
Schlosser, P., B. Kromer, G. Östlund, B. Ekwurzel, G. Bönisch, H.H. Loos- Smethie, W.M. Jr., D.W. Chipman, J.H. Swift and K.P. Koltermann, 1988.
li and R. Furtschert, 1994b. On the 14C and 39Ar distribution in the Chlorofluoromethanes in the Arctic Mediterranean Seas: evidence for
central Arctic Ocean: implications for the deep water formation. formation of bottom water in the Eurasian Basin and deep water ex-
Radiocarbon 36(3): 327-345. change through Fram Strait. Deep-Sea Res. 35: 347-369.
Schlosser, P., J.H. Swift, D. Lewis and S.L. Pfirman, 1995a. The role of Smith, D.W. and T.D.W. James, 1979. Sump studies III - biological changes
large-scale Arctic Ocean circulation in the transport of contaminants. in permafrost terrain adjacent to high Arctic oil and gas wellsites.
Deep-Sea Res. 42: 1341-1367. Indian and Northern Affairs Canada, Ottawa, Environmental Studies
Schlosser, P., G. Bönisch, B. Kromer, H.H. Loosli, B. Bühler, R. Bayer, G. No. 16, 150p.
Bonani and K.P. Koltermann, 1995b. Mid 1980s distribution of 3H, Smith, J.N., K.M. Ellis, S. Forman, L. Polyak, G. Ivanov, D. Matishov,
3He, 14C, and 39Ar in the Greenland/Norwegian seas and the Nansen L.Kilius and S. Dahle, 1995. Radionuclide sources in the Barents and
Basin of the Arctic Ocean. Prog. Oceanogr. 35: 1-28. Kara Seas. In: P. Strand and A. Cooke (eds.). Proceedings of interna-
Schnell, R.C. and W.E. Raatz, 1984. Vertical and horizontal characteris- tional conference on environmental radioactivity in the Arctic, pp.
tics of Arctic haze during AGASP: Alaskan Arctic. Geophys. Res. 179-185. Oslo, Norway, Aug. 21-25, 1995. Norwegian Radiation
Lett. 11: 369-372. Protection Authority, Østerås, Norway.
Scholton, J.C., M.M. Rutgers van der Loeff and A. Michel, 1995. Distri- Solheim, A. and S.L. Pfirman, 1985. Sea-floor morphology outside a
bution of 230Th and 231Pa in the water column in relation to the venti- grounded, surging glacier: Bråsvellbreen, Svalbard. Mar. Geol. 65:
lation of the deep Arctic basins. Deep-Sea Res. II 42(6): 1519-1531. 127-143.
Schöndorf, T. and R. Herrmann, 1987. Transport and chemodynamics of Sparmacher, H., K. Fülber and H. Bonka, 1993. Below-cloud scavenging
organic micropollutants and ions during snowmelt. Nordic Hydrology of aerosol particles: particle bound radionuclides-experimental. Atmos.
18: 259-278. Environ. 24A: 1247-1254.
Schott, F., M. Visbeck and J. Fischer, 1993. Observations of vertical cur- Stabeno, P.J. and R.K. Reed, 1991. Recent Lagrangian measurements
rents and convection in the central Greenland Sea during the winter along the Alaskan Stream. Deep-Sea Res. 38: 289-296.
of 1988-1989. J. Geophys. Res. 98: 14401-14421. Stabeno, P.J. and R.K. Reed, 1992. A major circulation anomaly in the
Schreier, H., W. Erlebach and L. Albright, 1980. Variations in water qual- western Bering Sea. Geophys. Res. Lett. 19: 1671-1674.
ity during winter in two Yukon rivers with emphasis on dissolved Stabeno, P.J. and R.K. Reed, 1994. Circulation in the Bering Sea Basin
oxygen concentration. Water Res. 14(9): 1345-1351. observed by satellite-tracked drifters: 1986-1993. J. Phys. Oceanogr.
Schumacher, J.D. and T.H. Kinder, 1983. Low-frequency current regimes 24: 848-854.
over the Bering Sea shelf. J. Phys. Oceanogr. 13: 607-623. Staehelin, J., A. Waldvogel, J.L. Collett Jr., R. Dixon, R. Heimgartner,
Schumann, T., B. Zinder and A. Waldvogel, 1988. Aerosol and hydrome- W. Henrich, C. Hsu, L. Li, L. Mosimann, B. Oberholzer, A.S.H.
teor concentrations and their chemical composition during winter pre- Prevot, W. Schmid, T. Schumann, M. Steiner, M. Volken and B. Zinder,
cipitation along a mountain slope-I: Temporal evolution of the aero- 1993. Scientific goals and experiments of the project “Winter precipi-
sol, microphysical and meteorological conditions. Atmos. Environ. tation at Mt. Rigi”: an overview. Water Air Soil Pollut. 68: 1-14.
22: 1443-1459. Stefánsson, U., 1962. North Icelandic waters. Rit Fiskideildar 3: 1-269.
Scott, B.C., 1981. Sulphate washout ratios in winter storms. J. appl. Met. Stein, M. and G. Wegner, 1990. Thermohaline observation on waters off
20: 619-625. west Greenland. Nafo-Sci. Counc. Stud. 14: 29-37.
Scrimgeour, G.J., T.D. Prowse, J.M. Culp and P.A. Chambers, 1994. Eco- Stein, R., H. Grobe and M. Wahsner, 1994. Organic carbon, carbonate,
logical effects of river ice breakup: a review and perspective. Freshw. and clay mineral distributions in eastern central Arctic Ocean surface
Biol. 32: 261-275. sediments. Mar. Geol. 119: 269-285.
Semb, A., 1985. A circumpolar SO2 emission survey. Norwegian Institute Stigebrandt, A., 1981. A model of the thickness and salinity of the upper
of Air Research, Kjeller, Norway, NILU OR 69/85. layer in the Arctic Ocean and the relationship between the ice thick-
Semkin, R., 1996. Processes and fluxes of contaminants in aquatic sys- ness and some external parameters. J. Phys. Oceanogr. 11: 1407-1422.
tems - 1994/95. In: J.L. Murray, R.G. Shearer and S.L. Han (eds.). Stigebrandt, A., 1984. The North Pacific: A global-scale estuary. J. Phys.
Synopsis of research conducted under the 1994/95 Northern Conta- Oceanogr. 14: 464-470.
minants Program, pp. 105-118. Indian Affairs and Northern Devel- Stigebrandt, A., 1990. On the response of the horizontal mean vertical
opment, Ottawa, Canada, Environmental Studies no. 73. density distribution in a fjord to low-frequency density fluctuations in
Semtner, A.J., 1976a. Numerical simulation of the Arctic Ocean circula- the coastal water. Tellus 42A: 605-614.
tion. J. Phys. Oceanogr. 6: 409-425. Stirling, I.S. and H. Cleator, 1981. Polynyas in the Canadian Arctic. Cana-
Semtner, A.J., 1976b. A model for the thermodynamic growth of sea ice dian Wildlife Service.
in numerical investigation of climate. J. Phys. Oceanogr. 6: 379-389. Strand, A. and Ø. Hov, 1993. A two dimensional zonally averaged trans-
Semtner, A.J., 1987. A numerical study of sea ice and ocean circulation in port model including convection motions and a new strategy for the
the Arctic. J. Phys. Oceanogr. 17: 1077-1099. numerical solution. J. Geophys. Res. 98: 9023-9037.
Semtner, A.J. and R.M. Chervin, 1992. Ocean general circulation from a Strand, A. and Ø. Hov, 1996. A model strategy for the simulation of chlo-
global eddy-resolving model. J. Geophys. Res. 97: 5493-5550. rinated hydrocarbon distribution in the global environment. Water Soil
Shaw, G.E., 1995. The arctic haze phenomenon. Bull. Am. Met. Soc. 76: Air Pollut. 86: 283-316.
2403-2413. Stumm, W. and J.J. Morgan, 1981. Aquatic chemistry. John Wiley and
Shen, G., E.R. Sholkovitz and D.R. Mann, 1983. The P coagulation of Sons, New York.
dissolved 239,240Pu in estuaries as determined from a mixing experi- Sturgess, D.L., 1986. Precipitation measured by dual gauges, Wyoming-
ment. Earth Planet. Sci. Lett. 64: 437-444. shielded gauges, and in a forested opening. In: D.L. Kane (ed.). Pro-
Shen, H.T. and N.L. Ackermann, 1990. Wave-induced sediment enrich- ceedings of the Symposium: Cold regions hydrology, pp. 387-396. 22-
ment in coastal ice covers. In: S.F. Ackley and W.F. Weeks (eds.). Sea 25 July, 1986, Fairbanks, Alaska. American Water Resources Associa-
ice properties and processes: Proceedings of the W.F. Weeks Sea Ice tion, Bethesda, Maryland, AWRA Technical Publication Series No.
Symposium, pp. 100-102. Cold Regions Research and Engineering TPS-86-l.
Laboratory, Hanover, New Hampshire, CRREL Monograph 90-1. Subba Rao, D.V. and T. Platt, 1984. Primary production of Arctic waters.
Shiklomanov, I.A. and B.G. Skakalsky, 1994. Studying water, sediment Polar Biol. 3: 191-201.
and contaminant runoff of Siberian Rivers. In: B. Molnia and K.B. Sugden, D., 1982. Arctic and Antarctic – A modern geographical synthe-
Taylor (eds.). Proceedings of the Workshop on Arctic Contamination, sis. Blackwell, Oxford, 472p.
Interagency Arctic Research Policy Committee, May 2-7, 1993, Anchor- Svendsen, H., 1977. A study of the circulation in a sill fjord on the west
age, Alaska. Arct. Res. U.S. 8: 295-306. coast of Norway. Mar. Sci. Comm. 3: 151-209.
Sholkovitz, E.R., 1976. The flocculation of dissolved organic and inor- Svendsen, H., 1981. A study of circulation and exchange processes in the
ganic matter during the mixing of river water and seawater. Geochim. Ryfylkefjords. Ph.D. Thesis, University of Bergen, Geophysical Insti-
Cosmochim. Acta 40: 831-845. tute, Report No 55..
Sholkovitz, E.R., 1983. The geochemistry of plutonium in fresh and ma- Svendsen, H., 1995. Physical oceanography of coupled fjord-coast systems
rine water environments. Earth-Sci. Rev. 19: 95-161. in northern Norway with special focus on frontal dynamics and tides.
Sholkovitz, E.R., E.A. Boyle and N.B. Price, 1978. Removal of dissolved In: H.R. Skjoldal, C. Hopkins, K.E. Erikstad and H.P. Leinaas (eds.).
humic acid and iron during estuarine mixing. Earth Planet. Sci. Lett. Ecology of fjords and coastal waters. Elsevier Science B.V., Amster-
40: 130-136. dam, The Netherlands.
Shortreed, K.S. and J.G. Stockner, 1986. Trophic status of 19 sub-arctic Swift, J.H., 1984. The circulation of the Denmark Strait and Iceland-Scot-
lakes in the Yukon Territories. Can. J. Fish. aquat. Sci. 43: 797-805. land overflow waters in the north Atlantic. Deep-Sea Res. 31: 1339-
Shpaikher, A.O. and Z.S. Yankina, 1971. Relationship of mean annual 1355.
levels in the Arctic seas with hydrometeorlogical factors. Probl. Arctic Swift, J.H., 1986. The Arctic waters. In: B.G. Hurdle (ed.). The Nordic
Antarctic 38: 29-42. Seas, pp. 129-154. Springer-Verlag, New York.
Siegenthaler, U. and F. Joos, 1992. Use of a simple model for studying Swift, J.H. and K. Aagaard, 1981. Seasonal transitions and water mass
oceanic tracer distribution and the global carbon cycle. Tellus 44B: formation in the Iceland and Greenland Seas. Deep Sea Res. 28: 1107-
186-207. 1129.
Swift, J.H. and K.P. Koltermann, 1988. The origin of Norwegian Sea deposition of oxides of sulphur and nitrogen with emphasis on long
deep water. J. Geophys. Res. 93: 3563-3569. range transport modelling in North America. Atmos. Environ. 20:
Swift, J.H., K. Aagard and S.-A. Malmberg, 1980. The contribution of 2101-2123.
the Denmark Strait overflow to the deep North Atlantic. Deep-Sea Volken, M. and T. Schulmann, 1993. A critical review of below-cloud aero-
Res. 27: 29-42. sol scavenging results on Mt. Rigi. Water Air Soil Pollut. 68: 15-28.
Swift, J.H., T. Takahashi and H.D. Livingston, 1983. The contribution of Wadhams, P., 1986a. The seasonal sea ice zone. In: N. Untersteiner (ed.).
the Greenland and Barents Seas to the deep water of the Arctic Ocean. Geophysics of sea ice. NATO ASI Series B, Physics, 146, pp. 825-991.
J. Geophys. Res. 88: 5981-5986. Plenum Press, New York.
Tait, J.B., 1957. Hydrography of the Faroe Shetland Channel 1927-1952. Wadhams, P., 1986b. The ice cover. In: B.G. Hurdle (ed.). The Nordic Seas,
Mar. Res. Scot. 1957, No.2, 309p. pp. 21-85. Springer-Verlag, New York.
Takahashi, T., 1963. Chemical composition of snow in relation to their Wadhams, P. and R.J. Horne, 1980. An analysis of ice profiles obtained
crystal shapes. J. Met. Soc. Japan 41: 327-336. by submarine sonar in the Beaufort Sea. J. Glaciol. 25: 401-423.
Takenouti, A.Y. and K. Ohtani, 1974. Currents and water masses in the Wallace, D.W.R. and R.M. Moore, 1985. Vertical profiles of CCl3F (F-11)
Bering Sea: A review of Japanese work. In: D.W. Hood and E.J. Kelly and CCl2F2 (F-12) in the central Arctic Ocean basin. J. Geophys. Res.
(eds.). Oceanograpy of the Bering Sea with emphasis on renewable 90(C1): 1155-1166.
resources, pp. 39-57. University of Alaska, Institute of Marine Science, Wallace, D.W.R., R.M. Moore and E.P. Jones, 1987. Ventilation of the
Fairbanks. Arctic Ocean cold halocline: rates of diapycnal and isopycrial trans-
Talley, L.D. and M.S. McCartney, 1982. Distribution and circulation of port, oxygen utilization and primary production inferred using chloro-
Labrador Sea water. J. Phys. Oceanogr. 12: 1189-1205. fluormethane distributions. Deep Sea Res. 34: 1957-1979.
Tarrason, L. and T. Iversen, 1992. The influence of North American sul- Walsh, J.E. and C.M. Johnson, 1979. An analysis of arctic sea ice fluctua-
phur emissions over western Europe. Tellus 44B: 114-132. tions, 1953-77. J. Phys. Oceanogr. 9: 580-591.
Tarrason, L. and T. Iversen, 1996. Sulphur dispersion in the northern Wania, F., 1994. Temperate and chemical behavior in the environment –
hemisphere, simulations with a 3-dimensional time-resolved model. towards an understanding of the global fate of persistent organic
Tellus. (submitted) chemicals. Ph.D. Thesis, University of Toronto.
Taylor, K.E. and J.E. Penner, 1994. Response of the climate system to Wania, F. and D. Mackay, 1993. Modeling the global distribution of toxa-
atmospheric aerosols and greenhouse gases. Nature 369: 734-737. phene: a discussion of feasibility and desirability. Chemosphere 27: 2079-
Tchernia, P., 1980. Descriptive regional oceanography. Pergammon Ma- 2094.
rine Series Volume 3, New York, 253p. Wania, F. and D. Mackay, 1995. A global distribution model for persistent
Thomas, D.J., R.W. Macdonald and A.B. Cornford, 1986. Geochemical organic chemicals. Sci. Total Environ. 160/161: 211-232.
mass-balance calculations for the coastal Beaufort Sea, NWT, Can- Wania, F. and A. Strand, 1998. Introducing atmospheric layering in a glo-
ada. Rapp. P.-v. Réun. Cons. int. Explor. Mer 186: 165-184. bal distribution model for persistent organic pollutants based on the
Thomas, D.J., F. Noone, A. Blyth and B.D. Smiley, 1990. Arctic data fugacity approach. Submitted to Environ. Toxicol. Chem. February
compilation and appraisal, vol.20, Beaufort Sea: Chemical oceanogra- 1998.
phy - hydrocarbons, metals, pigments, nutrients, oxygen and others. Weeks, W., 1994. Possible roles of sea ice in the transport of hazardous
Can. Data Rep. Hydrogr. Ocean Sci. 5(20): 51. material. In: B. Molnia and K.B. Taylor (eds.). Proceedings of the
Thomas, D.J., B. Tracey, H. Marshall and R.J. Norstrom, 1991. Arctic Workshop on Arctic Contamination, Interagency Arctic Research Pol-
terrestrial ecosystem contamination. Sci. Total Environ. 122: 135- icy Committee, May 2-7, 1993, Anchorage, Alaska. Arct. Res. U.S. 8:
164. 34-52.
Thomas, R. and M. Meybeck, 1992. The use of particulate material. In: Weeks, W.F. and S.F. Ackley, 1986. The growth, structure, and properties
D. Chapman (ed.). Water quality management, pp. 121-170. Chap- of sea ice. In: N. Untersteiner (ed.). Geophysics of sea ice. NATO ASI
man and Hall, London. Series B, Physics, 146, pp. 9-164. Plenum Press, New York.
Thomsen, T., 1994. Northern hydrology in Greenland. In: T.D. Prowse, Wefer, G., 1989. Particle flux in the ocean: effects of episodic production.
C.S.L. Ommaney and L.E. Watson (eds.). Northern hydrology: Inter- In: W. H. Berger, V.S. Smetacek and G. Wefer (eds.). Productivity of
national perspectives, pp. 21-39. Environment Canada, Saskatoon, the ocean: Present and past, pp. 139-153. John Wiley and Sons, New
Saskatchewan, NHRI Science Report No.3. York.
Thorndike, A.F., 1986. Kinematics of sea ice. In: N. Untersteiner (ed.). Wefer, G., G. Fischer, D. Fuetterer and R. Gersonde, 1988. Seasonal parti-
Geophysics of sea ice. NATO ASI Series B, Physics, 146, pp. 489-550. cle flux in the Bransfield Strait, Antarctica. Deep-Sea Res. 35(6): 891-898.
Plenum Press, New York. Welch, H.E., 1974. Metabolic rates of arctic lakes. Limnol. Oceanogr.
Timofeyev, V.T., 1961. The movement of Atlantic water and heat into the 19(1): 65-73
Arctic Basin. Okeanologiya 1: 407-411. Welch, H.E., 1985. Introduction to limnology research at Sagvagjuac,
Timofeyev, V.T., 1963. Interaction of the Arctic ocean waters with Atlan- northern Hudson Bay. Can. J. Fish. aquat. Sci. 42(3): 494-505.
tic and Pacific waters. Okeanologiya 3: 569-578. Welch, H.E., 1994. Char Lake. In: R.J. Allan, M. Dickman, C.B. Gray
Timokhov, L.A., 1994. Regional characteristics of the Laptev and the east and V. Cromie (eds.). The book of Canadian lakes, pp. 563-570. The
Siberian Seas: Climate, topography, ice phases, thermohaline regime, Canadian Association on Water Quality, Gloucester, Ontario, Canada.
circulation. Ber. Polarforsch. 144: 15-31. Monograph Series #33.
Top, Z., W.B. Clarke, W.C. Eismont and E.P. Jones, 1980. Radiogenic he- Welch, H.E. and M.A. Bergmann, 1985a. Water circulation in small arctic
lium in the Baffin Bay bottom water. J. mar. Res. 38: 435-455. lakes in winter. Can. J. Fish. aquat. Sci. 42(3): 506-520.
Topham, D.R., R.G. Perkin, S.D. Smith, R.J. Anderson and G. den Har- Welch, H.E. and M.A. Bergmann, 1985b. Winter respiration of lakes at
tog, 1983. An investigation of a polynya in the Canadian archipelago: Sagvagjuac, N.W.T. Can J. Fish. aquat. Sci. 42(3): 521-528.
1. Introduction and oceanography. J. Geophys. Res. 88C: 2888-2916. Welch, H.E. and J.A. Legault, 1986. Precipitation chemistry and chemical
Tranter, M., P. Brimblecombe, T.D. Davies, C.E. Vincent, P.W. Abrahams limnology of fertilized and natural lakes at Sagvagjuac, N.W.T. Can. J.
and I. Blackwood, 1986. The composition of snowfall, snowpack and Fish. aquat. Sci. 43(6): 1104-1134.
meltwater in the Scottish Highlands – evidence for preferential elu- Welch, H.E., D.C.G. Muir, B.N. Billeck, W.L. Lockhart, G.J. Brunskill,
tion. Atmos. Environ. 20: 517-525. H.J. Kling, M.P. Olson and R.M. Lemoine, 1991. Brown snow: a long-
Tsiouris, S., C.E. Vincent, T.D. Davies and P. Brimblecombe, 1985. The range transport event in the Canadian Arctic. Environ. Sci. Technol.
elution of ions through field and laboratory snowpacks. Ann. Glaciol. 25: 280-286.
7: 196-201. Wendland, W.M. and R.A. Bryson, 1981. Northern hemispheris airstream
Tsunogai, S., M. Kusakabe, H. Iizumi, I. Koike and A. Hattori, 1979. regions. Monthly Weather Rev. 109: 255-270.
Hydrographic feature of the deep water of the Bering Sea - The Sea of Weschler, C.J., 1981. Identification of selected organics in the arctic aero-
Silica. Deep-Sea Res. 26: 641- 659. sol. Atmos. Environ. 15: 1365-1369.
Uhlmann, D.R., B.Chalmers and K.A. Jackson, 1964. Interaction between Wheeler, P.A., M. Gosselin, E. Sherr, D. Thibault, D.L. Kirchman, R. Ben-
particles and a solid-liquid interface. J. appl. Phys. 35: 298-299. ner and T.E. Whitledge, 1996. Active cycling of organic carbon in the
Uralov, N.S., 1960. On the advective component of heat balance in the central Arctic Ocean. Nature 380: 697-699.
southern Barents Sea. Trudy Gos. Oceanogr. Inst. 55: 3-20. (In Rus- Whitfield, P.H. and B. McNaughton, 1986. Dissolved oxygen depressions
sian). under ice cover in two Yukon rivers. Water Resour. Res. 22: 1675-1679.
van Everdingen, R.O., 1990. Ground-water hydrology. In: T.D. Prowse WHO/UNEP, in press. Regional former Soviet Union water quality assess-
and C.S.L. Ommanney (eds.). Northern hydrology, Canadian perspec- ment. Ed. by V. Kimstach and M. Meybeck.
tives, pp. 77-101. National Hydrology Research Institute, Environ- Wiens, L.H., 1991. The effect of Peace River flow regulation on inflows to
ment Canada, Saskatoon, NHRI Science Report No. 1. the Mackenzie Delta. In: P. Marsh and C.S.L. Ommanney (eds.). Mac-
Verkhunov, A.V. and Y.Y. Tkachenko, 1992. Recent observations of var- kenzie Delta: Environmental interactions and implications of develop-
iability in the western Bering Sea current system. J. Geophys. Res. 97: ment, pp. 189-191. Proceedings of the Workshop on the Mackenzie
14369-14376. Delta. 17-18 October, 1989, Saskatoon, Sask. National Hydrology
Vinje, T., 1985. Physical environment western Barents Sea: Drift, compo- Research Institute, Environment Canada, Saskatoon, Saskatchewan,
sition, morphology and distribution of the sea ice field in the Barents Symposium No.4.
Sea. Norsk Polarinst. Skr. 179C: 1-26. Williams, J., 1979. Introduction to marine pollution control. John Wiley
Vinje, T., 1987. Morphology and dynamics of the Barents Sea ice field. & Sons, New York, 173p.
POAC 77, University of Alaska, Fairbanks. Williams, M.W. and J.M. Melack, 1991. Solute chemistry of snowmelt
Vinje, T. and O. Finnekåsa, 1986. The ice transport through Fram Strait. and runoff in an alpine basin, Sierra Nevada. Water Resour. Res. 27:
Norsk Polarinst. Skr. 186: 4-39. 1575-1588.
Voldner, E.C., L.A. Barrie and A. Sirois, 1986. A literature review of dry Wollast, R. and J.C. Duincer, 1982. General methodology and sampling
strategy for studies on behaviour of chemicals in estuaries. Thalassis

Jugoslavia 18: 471-491.
Woo, M-k., 1992. Arctic streamflow. In: M-k. Woo and D. J. Gregor
(eds.). Arctic environment: Past, present and future. pp. 105-111.
Proceedings of a Symposium held at McMaster Universty, Nov. 14-
15, 1991. McMaster University, Department of Geography, Hamilton,
Ontario.
Woo, M-k. and P. Marsh, 1977. Determination of snow storage for small
eastern high Arctic basins. In: Proceedings of the Thirty-Fourth Annual
Eastern Snow Conference, Belleville, Ontario, Feb. 3-4, 1977. pp. 147-
162.
Woo, M-k. and J. Sauriol, 1980. Channel development in snow filled val-
leys, Resolute, N.W.T., Canada. Geografiska Annaler 62A: 37-56.
Woo, M-k., R. Heron, P. Marsh and P. Steer, 1983. Comparison of weath-
er station snowfall with winter snow accumulation in High Arctic
basins. Atmosphere-Ocean 21(3): 312-325.
Worthington, L.V., 1976. On the North Atlantic circulation. The Johns
Hopkins Oceangr. Stud. No. 6, 110p.
Yablokov, A.V., V.K. Karasev, V.M. Rumyantsev, M. Kokeyev, O.I. Petrov,
V.N. Lystov and A.F. Melyanenkov, 1993a. Russian Federation state
report on burial of radioactive wastes. Admin. Press. Russ. Fed., Mos-
cow, 108p.
Yablokov, A.V., V.K. Karasev, V.M. Rumyantsev, M.Y. Kokeyev, O.I. Pe-
trov, V.N. Lytsov, A.F. Yemelyanenkov and P.M. Rubtsov, 1993b.
Facts and problems related to radioactive waste disposal in seas adja-
cent to the territory of the Russian federation. Office of the President
of the Russian Federation, Moscow (also known as “The Yablokov
Report”), 72p.
Yunker, M.B., W.J. Cretney, B.R. Fowler, R.W. Macdonald, F.A McLaugh-
lin and B.G. Whitehouse, 1991a. On the distribution of dissolver hy-
drocarbons in natural water. Org. Geochem. 17(3): 301-307.
Yunker, M.B., R.W. Macdonald, B.R. Fowler, W.J. Cretney, S.R. Dalli-
more and F.A. McLaughlin. 1991b. Geochemistry and fluxes of hy-
drocarbons to the Beaufort Sea shelf: A multivariate comparison of
fluvial inputs and coastal erosion of peat using principal components
analysis. Geochim. Cosmochim. Acta 55: 255-273.
Yunker, M.B., R.W. Macdonald, W.J. Cretney, B.R. Fowler and F.A.
McLaughlin, 1993. Alkane, terpene, and polycyclic aromatic hydro-
carbon geochemistry of the Mackenzie River and Mackenzie shelf:
Riverine contributions to Beaufort Sea coastal sediment. Geochim.
Cosmochim. Acta 57: 3041-3061.
Zakharov, V.F., 1976. Cooling of the arctic and ice cover of the Arctic
seas. Trudy Trans. 337: 96.
Zakharov, V.F., 1994. On the character of cause and effect relationships
between sea ice and thermal conditions of the atmosphere. Ber. Polar-
forsch. 144: 33-43.
Zenkovitch, V.P., 1985. Arctic USSR. In: E.C.F. Bird and M.L. Schwarz
(eds.). The world’s coastline, pp. 863-871. Van Nostrand Reinhold
Company, New York.
Zhulidov, A.V., J.V. Headley, R.D. Robarts, A.M. Nikanorov and A.A.
Ischenko, 1977. Atlas of Russian wetlands: biogeography and metal
concentrations. National Hydrology Research Institute, Environment
Canada, Saskatoon, SA., Canada. 312p.
Zinder, B., T. Schumann and A. Waldvogel, 1988. Aerosol and hydro-
meteor concentrations and their chemical composition during winter
precipitation along a mountain slope-II. Enhancement of below cloud
scavenging in a stably stratified atmosphere. Atmos. Environ. 22:
2742-2750.
Zubakin, G.K., 1987. About the ice exchange in the Barents Sea. Arctic
and Antarctic Research Institute, St. Petersburg, Proceedings Vol. 410,
pp. 133-117. (In Russian).
Zubov, N.N., 1943. Arctic ice. Translation No.103, The U.S. Navy
Oceanographic Office and the American Meteorological Society. U.S.
Navy Elect. Lab., San Diego, California, U.S.A.

AAR-Ch03 Influence of Physical and Chemical Processes On Comtaminant Transport Into and Within The Arctic

Uploaded by

Copyright:

Available Formats

AAR-Ch03 Influence of Physical and Chemical Processes On Comtaminant Transport Into and Within The Arctic

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

AAR-Ch03 Influence of Physical and Chemical Processes On Comtaminant Transport Into and Within The Arctic

Uploaded by

Copyright:

Available Formats

Go back to opening screen 25

Contents 3.5.3.1. Ice and icebergs . . . . . . . . . . . . . . . . . . . . . . . . . . 62

3.2. Atmosphere 3.2.2. Atmospheric transport

mospheric winds deliver contaminants to the north. This is W/m 2 (a)

Precipitation, water equivalent (mm)

Figure 3·8. Conceptual model of the terrestrial/freshwater compartment

perate areas can result in an increase in sedimentation of

pean part and in western and central Siberia. It is only re-

aquatic media. There are no other Norwegian industrial ac-

north from agricultural use areas upstream in the drainage

% of total % of total % of total

Hydrographs of Canadian Arctic rivers

and lakes (Albright et al. 1980, Schreier et al. 1980, Prowse

Melt M el t 3.3.4.3. Suspended sediments, sedimentation,

Sediment Suspended sediments

Norway northern Norway to the sea is in excess of 150 103 ton-

ocean. In the middle part of the estuary, salt water and

River runoff Freshwater

Depth (m) Rivers

3.5.3. Ice Atlantic

over the ice, depositing its sediment load on the surface

1 000 Canada Basin Arctic Deep Water Eurasian Basin

son et al. 1994).

Arctic deep water is divided into Canadian Basin Deep Wa-

Sea Deep Sea Deep

3.5.4.3. The Nordic Seas

The Norwegian Coastal Current

The Norwegian Atlantic Current follows the Norwegian 3.7 26 6.3

Pressure Pressure convection (see section 3.5.4.6) to as deep as 2000 m in

0 500 1 000 km Baffin Bay

Canadian Archipelago Baffin

freshwater and contaminant budgets of the Arctic Ocean

AG side of the Aleutian Islands. Inflows through the other passes

Mixing energy : Generalized fluxes :

150 salinity salinity tidal freshwater 150

300 Depth 300

Brine Mid-Gyre Mid-Gyre

Arctic Ocean Mid Gyre

The Bergen Model is a global multi-compartmental model

2. Short-term, high-spatial resolution model – defining the ments, are connected by

global contaminant loading in system compartments (Fig- meridional atmospheric and

oceanic exchange (Source:

by the 3-D models.

The Bergen and Toronto Models

There have been two attempts at simulating the dispersion

(1996) modeled the distribution of HCHs using a 2-D, at-

“Bergen” Model “Toronto” Model “Combined” Model

Middle/upper troposphere Middle/upper troposphere

Atmospheric boundary layer Atmospheric boundary layer

Cultivated Uncultivated Agricultural Fresh Natural Agricultural Fresh Natural

Table 3·10. A comparison of characteristics of global zonally-averaged multi-compartmental toxics models.

at fully integrating these processes to provide, for example,

to passive transport with the currents, a simple formulation

Level Depth Level Depth

3.7.4. Conclusions 3.7.5. Recommendations

• the Murman area, for which the long-range zone of influ-

Oceanogr. 9: 531-554. Washington D.C., Publ. 13, 173p.

strategy for studies on behaviour of chemicals in estuaries. Thalassis

You might also like