The environmental impacts of oil spills on land in the arctic regions

T H E E N V I R O N M E N T A L I M P A C T S O F O I L S P I L L S

O N L A N D I N T H E A R C T I C R E G I O N S

J. M. R A I S B E C K and M. F. M O H T A D I

Dept. of Chemical Engineering, University of Calgary, Calgary, Alberta, Canada

(Received 15 November, 1973; in final form 7 March, 1974)

Abstract. Some of the important aspects of oil spills which have significant impact on the Arctic evironment have been investigated. Simple theoretical models have been developed for the movement of oil on permeable and impermeable surfaces. The latter model has been verified by laboratory experiments. Vertical infiltration of oil in the unsaturated water zone of soil has also been studied both theoretically and experimentally. Areas requiring future investigation are discussed.

1. Introduct ion

The explo i ta t ion o f the oil and gas resources of the Arc t ic regions involves no t only

vast increases in the number o f dr i l l ing rigs, s torage facilities and pipelines, bu t also

cons t ruc t ion of a ne twork of roads , airstr ips, power lines, houses, sewage d isposa l and

water t r ea tment facilities. These projects have to be implemented in areas of dis-

cont inuous and cont inuous permafros t , and are cont ingent upon the deve lopment and

app l i ca t ion o f new knowledge.

Large scale deve lopment o f the N o r t h should take into account the result ing

env i ronmenta l consequences. A n i m p o r t a n t issue, which is the subject o f this paper ,

is the envi ronmenta l p rob lems which arise out of the spillage o f oil in the Arctic.

A l t h o u g h there are inadequa te da ta for forecast ing the a m o u n t o f crude oil which may

eventual ly enter the Arc t ic envi ronment , the potent ia l seriousness o f the p rob lem can

be real ized by reference to Table I.

In add i t i on to crude oil, there is the poten t ia l danger o f the spil lage of pe t ro leum

products , in par t icular , p roduc ts o f in te rmedia te molecu la r weight, which no rma l ly

TABLE I

Estimated reserves, production and spillage of crude oil in the arctic

Location Reserves Production Spillage 109 bbls (1980 onwards) bbls day -1

106 bbls day -1

Alaskan North Slope 11-21 a 2.0-2.8 e 700 a Arctic Islands and N.W.T. 28-70 b 0.5 a 150 a

Oil and Gas Journal, January 1973. b An Energy Policy for Canada, 1973. ° U.S. Cabinet Task Force, 1970. a Author's estimate.

Water, Air, and Soil Pollution 3 (1974) 195-208. All Rights Reserved Copyright © 1974 by D. Reidel Publishing Company, Dordreeht-Holland

196 J. M. RAISBECK AND M. F. MOHTADI

present the greatest risk of pollution (Dietz, 1971). Only recently (December, 1972) there was a spill of 30000 gal of diesel fuel from storage tanks in Holman, N.W.T.

Although there are many ways by which the Arctic environment may be contaminated by oil, the main causes fall into two categories:

(a) Technical breakdowns, and (b) Human and miscellaneous faults. Production, transportation and storage of oil can affect the thawing of the perma-

frost. In areas of ice rich soils this can lead to differential ground settlement causing serious damage to drilling rigs, pipelines and storage facilities. A typical example of leaks from storage tanks is shown in Figure 1. Other breakdowns may result from the

Fig. 1. Oil leak from storage tanks at Norman Wells.

embrittlement of steel, plastics, and synthetic rubber under extremely cold conditions. Furthermore, in the Arctic regions man's efficiency and work output are generally much less than normal and he is more susceptible to err. Oil spillages may also be caused by earthquakes (the proposed Trans-Alaska Pipeline System crosses an active seismic fault (Wellbaum, 1973)) or have a natural origin (Ebbley, 1944).

Whatever the cause, the immediate consequence of an oil spill on land is the creation of an aesthetically objectionable landscape. When Arctic soils, supporting vegetative growth, are contaminated with oil, the growing tissues are usually completely killed. In addition it appears that certain trees and plants, including mosses, lichens and sedge species show virtually no recovery (Wein andBliss 1973; Richard and Deneke, 1972).

THE ENVIRONMENTAL IMPACTS OF OIL SPILLS 197

It has even been stated (Brooks, 1970) that a large scale oil spill in the Arctic could disrupt the food chain.

Oil spilled on land eventually finds its way into river drainage systems with detri- mental effects to fresh water and fishery resources. The most serious aspect of water contamination by oil is undoubtedly the objectionable effect on the taste and odor. Some petroleum products can cause perceptible taste and odors even when present in concentrations well below 1 ppm. Organic sulphur compounds such as mercaptans

present in the oil add appreciably to the unpleasant smell and taste. In the Arctic regions

where there is a shortage of drinking water, odor and color contamination of water can pose serious problems.

Although a limited number of studies have been carried out in the Arctic, the bulk of information available on the environmental problems associated with oil spillage on land relate to the temperate conditions. It is important, therefore, to recognize

similarities between the temperate and Arctic situation and show how the information pertinent to temperate zones can be applied to the solution of Arctic problems.

2. The Nature of the Arctic Environment

Although the north often lives up to its popular conception of a desolate land mass,

intense cold, snow and strong winds, there are significant seasonal and regional varia- tions of climate in this region (see Table lI).

The northern topography is similar in many ways to that of the temperate zones.

However, the cold climatic conditions have resulted in certain ground features characteristic of the Arctic such as polygons, mounds, thermokarst lakes and palsas.

Polygons cover many tens of thousands of square kilometers of the northern region

and are particularly abundant in the southern part of the Mackenzie Delta. Mound s or frost hummocks usually comprise bunshaped deposits of fine-grained mineral soil. The

tops of the mound may be bare or turfy whereas the intermound depressions are covered usually by organic matter.

Virtually all the land mass north of the line of latitude 60 °N is covered with perma-

frost. Permafrost is usually classified as continuous or discontinuous, based on geo-

TABLE II Summary of long-term climatic characteristics at various settlements in the Arctic

Fairbanks ~ lnuvik b Norman Wells ° Resolute b

Mean annual precipitation, cm Mean annual snowfall, cm Mean daily temperature, °C Mean January temperature, °C Mean July temperature, °C

29.69 27.61 33.45 13.61 169.2 172.7 143.3 73.2 --3.3 --9.6 --6.3 --16.2

--24.0 --30.9 --23.7 --32.4 15.5 13.9 16.1 4.6

U.S. Weather Bureau. b Meteorological Branch, Department of Transport (Canada). ° Atmospheric Environment Service (Canada).

198 J .M. RAISBECK AND M, F. MOHTADI

GROUND ICE WEDGE-

TABULAR--

/•S• ROUND RFACE

UNFROZEN

{A)

Fig. 2.

----ACT,VELA.ER\ /GROOND PERELETO. (UNTHAWED f['f lj~r- PART OF s/ SURFACE 4 I I l~#ff/I ACTIVE

T ~ LAYER) --PERMAFROST TABLE~

Y.#3 VA ~TALI K ~//[ Y/l (SEASONAL

///////////////'///} ~J FROST //A v.j DOES NOT //A PENETRATE TO

PERMAFROST ~ ~ PERMAFROST /~ TABLE )

/ # f PERMAFROST .,___---~ ~ # ,SLANO ~ /

UNFROZEN /

(B)

Schematic diagrams of typical profiles in (A) continuous and (B) discontinuous permafrost.

graphic extent. In the continuous zone, permafrost exists everywhere below the natural ground surface. The depth of permafrost varies from several hundred feet in the south to several thousand feet in the north (Figure 2A). The discontinuous zone of permafrost comprises a mosaic of frozen and unfrozen masses and/or layers intermittently distributed below the ground surface (Figure 2B).

Hydrology of the north is greatly influenced by its cold climate. All surface water bodies are frozen over in the winter, small streams cease to flow and shallow lakes freeze to the bottom. Large rivers, however continue to flow although their flow rate is greatly reduced. The snow melt period results in the peak discharge of rivers and streams. It is not uncommon for as much as 80% of the annual flow of Arctic streams to occur in less than 3% of the year (Brown, 1972).

3. Present Work

The movement of oil on and within the Arctic soil is influenced by such factors as the configuration of the ground surface, the type and amount of vegetation, the nature of the soil, the extent of water saturation, duration of the spill, the ambient climatic conditions and the physical properties of the oil.

3.1. SPREAD OF OIL ON THE SURFACE

3.1.1. Impermeable Media

The forces which cause the spread of oil on an impermeable surface are gravity and pressure. The retarding forces are inertia and viscous. Surface tension may assist or resist spreading depending on the circumstances.

Oil spilled onto a flat impermeable surface (Figure 3) spreads in all directions and the seepage area varies as a function of time. If we assume a constant slick thickness,


r+dr I AT t _ ~

RATE OF SPI LLAGE ,Q(t) AI R RADIUS.r-'~"I

OIL SLICK AT TIME,t I OF CONSTANT T - - - THICKNESS ,h 2 OIL -'~ - ' ' x

J

Fig. 3. Schematic diagram of the spread of oil on an impermeable medium.

h then

and

d r Q (t) = 27crh - - (1)

dt

t

0

If oil is added at an average rate, (~, then

( 0 ~ 1/2 r = t 1/2 \ ~ ] . (3)

The value of h will be influenced by the roughness of the ground surface and the physical properties of the oil, in particular the viscosity. For a smooth solid surface, provided the viscosity of the oil remains constant, the equilibrium value of h will be determined primarily by surface tension and gravity froces. The height of the oil film can now be predicted from the following equation proposed by Davies and Rideal (1963):

k/2~ (1 -cosO) h = (4)

Pg

Warm oil spilled on very cold frozen soil or on ice will inevitably solidify and cease to flow. The value of h under such circumstances will be highly temperature dependent.

We have carried out a number of experiments to simulate the spreading of oil on a fiat ice surface in a specially constructed rectangular trough to which we have given the name 'Arctic Simulation Rig' (Figure 4). Figure 5 shows the results of a typical ex- periment using Norman Wells crude oil and it demonstrates that the thickness of the oil film is essentially constant during a steady, continuous oil leak.

To verify the validity of Equation (3) for large spills on a snow surface, experimental results obtained by McMinn and Golden (1973) are reproduced in Figure 6, and

2 0 0 J. M. RAISBECK A N D M. F. MOIcITADI

10 2 --

OJ 0J

CAMERA

THERMO'BULB

H E A T EXCHANGER

. THERMO EXPANSION

1' I I i |

I I

E L E C T R I C A L ~ ' ~ ! [ L ~ I I t I ~ 1 POWER [ f , - - ~ r ~ l i I I [ / I

MOTOR PRESSURE COMPRESSOR RECEIVER CONTROL CONDENSER

Fig. 4. Structural details of the Arctic simulation rig.

EVAPORATOR PRESSURE REGULATOR VALVE

DATA STOPPED - - ADDING O,L I J OIL: NORMAN WELLS CRUDE OIL , / TEMP. OF OIL PRIOR TO ADDITION,=C=5 ~ - o o AMBIENT TEMR. °C =-2 ~ "~'o °~ VALUE OF h(DETERMINED 66 min AFTER ~ STOPPING OIL ADDITION),cm =0.096 ~ , , "o~- AVERAGE FLOW RATE.Q.cm 3/rain = 4 DURATION OF SPILL. rain = 50

, /

I I I l i l l l I I I I l l l l l r i i ~ l l = f l r L i i i r l l l I I I I r l r l l I I 0 102 I 0 5 104

T I M E , t ( i n s a c )

Fig. 5. Spread of oil on ice in the Arctic simulation rig.

T H E E N V I R O N M E N T A L I M P A C T S O F O I L S P I L L S 2 0 1

10 3

Fig. 6.

u

v

l0 2

DATA

OIL:PRUOHOE BAY CRUDE OIL TEMR OF OIL PRIOR TO ADDITION,°C=I5 ESTIMATED AMBIENT TEMR, °C = - 5 0 VALUE OF h(OBTAINED IN SPILL(2) WHEN OIL HAD LEAKED ONTO SNOW SURFACE FOR 281 SEC.) crn =1.50 AVERAGE FLOW RATE, Q, crn 5/see = 4870

o

10 I I I ~ I i q l ] u I I t i i i 1 [ I I i I r l l l l 0.1 I I0 10 2

T I M E , t ( i n sec)

Spread of oil on snow under Arctic winter conditions (McMinn and Golden, 1973).

compared with Equation (3). The thickness of the oil film was calculated from these authors' experimental data corresponding to a lapse time of 28.1 s. Equation (3) is seen to describe the oil spreading rate fairly well - in fact much more accurately than the theory proposed by McMinn and Golden. The discrepancy between the absolute values of h is attributable to the differences in surface roughness and the viscosity of oils used.

3.1.2. Permeable Media

Oil spilled on a permeable surface will also spread as a film of constant thickness. For a continuous steady leak

dA h d t = 0 - Q ~ ' (5)

where Q~ is the quantity of oil infiltrated into the porous medium. In the capillary model of flow in porous media, with negligible h, the surface tension

and gravity forces can be related to the viscous drag forces by:

d x 2zcrcy cos 0 + rcr~gAQx = 8rc#x dt " (6)

In the case of oil flow within a porous soil, the surface tension forces are assumed to be far more predominant than the gravity forces. The expression obtained by neglecting the gravitational force is known as the Rideal-Washburn equation (Davies and Rideal,

202 S. M. RA~BECK AND M.F. MOHTADI

1963) : d x

rot cos 0 = 4#x ~-. (7)

Integration of Equation (7) gives

x/roy cos Ot x = 2/z

from which dx ~/] ,

Ui - d t - Ks t

where

K1 = ~// 'e]) COS 0

8#

If the area dA is formed during the interval of time • to ~ + d~ then

A

QI = f Ui dA 0

and after a change of variable

t f dA dr

e , = .

0

(8)

(9)

(10)

(11)

(12)

Substituting Equation (12) in Equation (6) and solving using the Laplace Transform, we obtain:

t

: exp t S U ) e r f c t ~ ) at (13) 0

Equation (13) may be integrated to give

A = K ~ erfc(Y) e x p ( Y 2 ) + ~ Y - I , (14)

where

for

and

Y > I A ~ 1(2 ~/~ t 1/2

r ~ K2 t 1/4,

(15)

(16)


where K2 is a constant and which has the value

rc7 cos 0 (17) K2 = 0 ~/ 2#re

It is of interest to compare the relationships for the radius of an oil slick under isothermal conditions on a flat surface for permeable and impermeable surfaces:

for impermeable medium ratl / 2

for permeable medium rat 1/4.

3.2. MOVEMENT OF OIL WITHIN THE SOIL

In general, the movement of oil within the soil is the result of the interaction of gravity, capillary and pressure forces. The complicated nature of Arctic soils makes the predic- tion of the movement of oil extremely difficult.

3.2.1. Development of a Simple Model

Fluid dynamic principles may be employed to predict the movement of oil in a homogeneous unsaturated soil. One can consider the one-dimensional movement of oil when there is no free oil remaining on the soil surface and assume that there is uniform saturation at all points throughout the oil zone, i.e., S is independent of depth x, but varies with time.

If S = S O and x = x 0 at t = t o

then

Sox o s - ( 1 8 )

Noting that for a capillary tube model

23' cos 0 Pc - - - (19)

rc

and that the relationship between permeability and porosity according to Marshall (1968) is

84/3F 2 k - (20)

8

We can combine Equations (7), (19) and (20) to obtain

Pc k dx - - - = x - - . ( 2 1 ) /34/3# dt

The capillary pressure-saturation relationship of the drainage cycle for water-soil


systems has been

where

described by Brooks and Corey (1966) as:

S * = (JPd~ ~ (22)

S* - S - Sr~. (23) 1 - Srl

Equations (22) and (23) can be applied to the oil-water-soil system by assuming that the water and oil form a single wetting phase and that S - Sr~ is the excess saturation over the residual saturation.

A variety of expressions has been used for the wetting phase relative permeability relationships, one of the most common being

k - S * ~ . ( 2 4 )

ko

Hence, by substituting Equations (22) and (24) in (21)

dx P~k° S *~-(1/~) - x (25) e4/3/2 dt '

where under isothermal conditions

Pdko 83/4/2 = K 3 (a constant).

Also from Equations (18) and (23) we have

and

S* - S°x° 1 _ K4 (26) X (1 - - Srl ) x

Qv = K 6 t - o , (27)

where Qv is the rate of increase in soil volume inundated with oil at time, t and

- (1/4) + 1 q~ - (28)

q - (1/)t) + 2

For many soils, 3 ~< q ~< 4 (Brooks and Corey, 1966), and for the snow-water-air system t/= 2 (Colbeck, 1971). Taking 1 ~< 2 ~ 2, we find from Equation (28):

0.75 ~< q5 ~< 0.82 for soil, (29)

0.67 ~< ~b ~< 0.71. for snow. (30)

3.2.2. Experimental Verification

Raisbeck (1972) studied the effect of 'rainfall' on the mode of migration of oil through uniformly packed beds of soil. Mixtures of 10~ soil and 90~ sand were used as the


,=

t¢3 E t)

> O m- if) <Z w c r 0 Z

O >

w Z O NI c3 W

Z

¢ Z O (D

._1 5 la- O W

Z < W

I 0 - -

IO-I

10-2

10-3

a,

Ist RAIN ADDITION TO iOLUMN 12

QV = 19f -0.81 ~ ' ~

:

1 0 - 4 ~ , , I I0 I0 z

, , I , , , , I , , , , I , , , , I I03 I04 105 106

TIME, m i n I i I I I I I I I I 2 3 I0 20 30 IO0 200 300

Fig. 7.

TIME,t(in days) Mean rate of the increase in the volume of oil contaminated

zone as a function of time (Raisbeck, 1972).

soil model. Kerosine was injected at the soil surface and water was added intermittently over a period of up to 46 weeks to simulate natural precipitation. The experimental results o f Raisbeck (Figure 7) confirm the validity of Equation (27). Equation (27) has been shown to be valid also for the movement of kerosine in sandy-loam soils.

Field studies are being pursued by researchers from the University of Toronto (MacKay and Phillips, 1973) to investigate the mode of oil movement in soil at test plots at Norman Wells and Tuktoyaktuk. There has been no publication so far describ- ing this work and it is therefore not possible to make comparisons between theoretical, laboratory and field studies.

4. Future Studies

Important areas where future investigations should be directed include: (a) The lateral spreading of oil over terrain features common to the Arctic;


(b) The movement of oil through snow, ice, frozen ground and the saturated water zone;

(c) The effect of an oil layer on the energy transfer through an air-soil interface; (d) The intrusion of water-soluble oil fraction into natural water resources; and (e) The role of chemical and biological degradation in the fate of spilled oil. Some work has already been done on the movement of oil over northern terrain

features and also on snow and ice. It is already known that the oil spilled on land, if not contained entirely in the unsaturated water zone, will reach the capillary fringe where it will move parallel with and above the groundwater table, as depicted in Figure 8. However, more information is needed to relate the area of the pancake layer to the most permeable part of the saturated subsoil and the groundwater gradient.

OIL Z /-VERTICAL SEEPAGE

.~j~-- LATERAL SPREADING

/--JA2ER RIVER

-7~7--~7~z~ NCAKE " ~ PERMAFROST "'"

t 2 ~

Fig. 8. Schematic d iagram of the m o v e m e n t of oil within Arctic soil.

The existence of a film of oil at the soil surface increases the rate of energy absorption by the soil. Whether the extra energy absorbed causes significant thawing of the permafrost is uncertain. One factor complicating the problem is that the oil inundated soil has a lower thermal conductivity than the non-contaminated soil, thus reducing the heat transfer rate.

Of the possible routes by which water-soluble components of oil can reach an Arctic water resource, sub-surface flow is of particular interest. In natural soil the rate of movement of the extractables is reduced by adsorption and biodegradation. Although the effectiveness of these latter processes has been highlighted by Raisbeck (1972), further work is required in this area.

The role of chemical oxidation in the degradation of oil in the Arctic environment is not clear. More research is needed to determine the success of fertilizer addition, the use of oil-tolerant grasses and bacterial seeding in the acceleration of natural soil reclamation.

5. Summary and Conclusions

(1) The exploitation of the oil and gas resources of the Arctic is contingent upon the development and application of new engineering knowledge relevant to the particular


environmental conditions of the north. Any large scale development of these resources will inevitably create new environmental problems.

(2) Oil spillages in the Arctic could result from technical breakdowns, embrittlement of materials, mechanical stresses on oil bearing facilities by thermokarst, human errors, and possibly earthquakes.

(3) Harmful consequences of oil spillages in the Arctic include destruction of vegetation and subsequent prevention or stunting of growth, disruption of animal life and pollution of natural water resources.

(4) The spreading of oil spilled on land is influenced, in the main, by the nature of the surface, the permeability of the soil, the type and amount of vegetation, the physical properties of the oil and the prevailing climatic conditions. For the case of a continuous steady oil leak over an impermeable media under isothermal conditions the radius of the oil slick may be correlated with time by the simple expression

r : t I / 2 .

For an oil spill on a fiat porous surface, theoretical examination yields

r ~- K 2 t 114.

(5) Oil spilled on land in warm weather will move vertically in the unsaturated water zone and laterally in the capillary fringe. In cold weather, however, the principal mode of oil migration is lateral spreading. On the basis of the capillary model of flow in porous media the increase in the volume of the soil inundated with oil in the unsaturated zone may be related to time by a simple correlation

Qv = K t - 4 .

Both theory and laboratory experiments indicate that q5 has a value of approximately 0.8.

(6) Very little basic information has been published on the environmental problems of oil spills in the Arctic. In particular, little is known as to the intrusion of water- soluble oil components through Arctic soils into natural water resources. In view of the relative scarcity of potable water supplies in the north, this topic is of particular significance and requires careful study.

Acknowledgments

The authors wish to express their gratitude to the University of Calgary, The Boreal Institute of Northern Studies, Alberta Environment and the National Research Council of Canada for their financial support of this work. Part of the research re- ported was done whilst one of the authors (J.M.R.) was working for a doctorate degree at Birmingham University. The help and encouragement of Stichting CONCAWE and Prof. P. J. Garner in this connection are also gratefully acknowledged.

208 J . M . RAISBECK AND M. F. MOHTADI

Notation

A K1./(2 etc. Pc Pa Q Q QI

Qv S So Srl Uj g h k ko r

re

t to X

XO

AO 0 rh 2

0

T

erfc exp

area of oil slick at soil surface (seepage area) constants capillary pressure displacement pressure rate of spillage average rate of spillage rate of infiltration of oil into porous medium rate of increase in the volume of oil contaminated zone saturation initial saturation residual liquid saturation velocity of oil infiltration into porous medium acceleration due to gravity thickness of oil slick permeability to wetting phase total permeability radius of the oil slick radius of the capillaries time initial time depth of vertical penetration of oil into porous medium initial depth density difference between oil and air contact angle between oil and solid constants for particular soil viscosity of oil density of oil surface tension at the oil-air interphase time complementary error function exponential function

References

Brooks, J. W. : 1970, 'Environmental Influences of Oil and Gas Development with Reference to the Arctic Slope and Beaufort Sea', Bureau of Sport Fisheries and Wildlife.

Brooks, R. H. and Corey, A. T. : 1966, Proc. Am. Soc. Civil Engrs, Irrigation and Drainage Div. 94, 61. Brown, R. J. E. : 1972, Geomorph. Neue Folge Suppl. 13, 102. Colbeck, S. C. : 1971, 'One-Dimensional Water Flow Through Snow', U.S. Army Cold Regions

Research and Engineering Laboratory. Davies, J. T. and Rideal, E. K. : 1963, InterfacialPhenomena, Academic Press. Dietz, D. N. : 1971, 'Pollution of Permeable Strata by Oil Components' , Water Pollution by Oil, The

Institute of Petroleum. Ebbley, N. : 1944, Mineral Metall. 25, 415. MacKay, D. and Phillips, C. R. : 1973, Private communication, University of Toronto. Marshall, T. J. : 1968, Y. Soil. Sci. 9, 1. McMinn, T. J. and Golden, P. : 1973, 'Behavorial Characteristics and Clean- up Techniques of North

Slope Crude Oil in an Arctic Winter Environment' , Prevention and Control of Oil Spills, American Petroleum Institute.

Raisbeck, J. M. : 1972, Ph.D. Thesis, University of Birmingham. Rickard, W. E. and Deneke, F. : 1972, 'Preliminary Investigations of Petroleum Spillage, Haines-

Fairbanks Military Pipeline, Alaska',U.S. Army Cold Regions Research and Engineering Laboratory. Wein, R. W. and Bliss, L. C. : 1973, 'Experimental Crude Oil Spills on Arctic Plant Communities',

to be published. Wellbaum, E. W. : 1973, 'Oil Spill Prevention Measures for the Trans-Alaska Pipeline System', Pre-

vention and Control of Oil Spills, American Petroleum Institute.

Documents

The environmental impacts of oil spills on land in the arctic regions