•11111111111 FI0200007

POSIVA 2001-05

Permafrost: occurrence and physicochemical processes

Lasse Ahonen Geological Survey of Finland

.33/15

October 2001

POSIVA OY Toolonkatu 4, FIN-00100 HELSINKI, FINLAND Phone (09) 2280 30 (nat.), ( + 358-9-) 2280 30 (int.) Fax (09) 2280 3719 (nat.). ( + 358-9-) 2280 3719 (int.) Posiva-raportti - Posiva Report Raportin tunnus - Report code POSIVA 2001-05 Posiva Oy Toolonkatu 4, FIN-00100 HELSINKI, FINLAND Julkaisuaika - Date Puh. (09) 2280 30 - Int. Tel. 4358 9 2280 30 October 2001

Tekija(t) - Author(s) Tbimeksiantaja(t) - Commissioned by

Lasse Ahonen Geological Survey of Finland Posiva Oy

Nimeke-Title

PERMAFROST: OCCURRENCE AND PHYSICOCHEMICAL PROCESSES

Tiivistelma-Abstract

Bedrock of the Northern Hemisphere areas to the north of about the 60th latitude are nowadays dominated by permafrost conditions. Fennoscandia is a major exception being characterised by temperate climate. In studying deep geological disposal of long-living nuclear waste, long-term climatic changes have to be taken into account. One of the scenarios to be studied is the extension of the deep permafrost conditions to the disposal site.

Quaternary climatic fluctuations and their possible reasons are discussed shortly. The author's conclusion is that future climatic changes cannot be undoubtedly derived from the past variations, mainly because of the current anthropogenic involvement and of the poorly known dynamics of the major climate-affecting factors like ocean currents, which cannot be treated in a deterministic way.

In low-porosity crystalline rocks permafrost may propagate to the depth of about 500 metres in some thousands to ten thousands of years. On the other hand, the major effects of permafrost are related to the freezing of water in the pores. Water expands about 9 percent in freezing, and the increasing stress may lead to pressure melting of ice. Dissolved salts in water do not accommodate into the solid ice, but they form saline water or brine segregations having freezing point of even less than minus ten degrees. A front of saline water may develop beneath the frozen bedrock. Pockets of saline water may also occur in ice, and unfrozen adsorption water may occur on the grain boundaries.

With respect to the radionuclide transport processes, permafrost as such is a barrier, while the unfrozen domains (taliks) beneath major lake and river systems are potential flow paths.

Avainsanat - Keywords Climatic variations, nuclear waste, freezing of water, cryopegs, gas hydrates, taliks

ISBN ISSN ISBN 951-652-106-1 ISSN 1239-3096

Sivumaara - Number of pages Kieli - Language 42 English Posiva-raportti - Posiva Report Raportin tunnus - Report code POSIVA 2001-05 Posiva Oy Töölönkatu 4, FIN-00100 HELSINKI, FINLAND Julkaisuaika - Date Puh. (09) 2280 30 - Int Tel. -(558 9 2280 30 Lokakuu 2001

Tekijä(t)-Author(s) Toimeksiantaja(t) - Commissioned by

Lasse Ahonen Geologian tutkimuskeskus Posiva Oy

Nimeke -1 itle

IKIROUTA: ESIINTYMINEN JA FYSIKAALIS-KEMIALLISET PROSESSIT

Tiivistelmä-Abstract

Pohjoisella pallonpuoliskolla likimain 60 leveyspiirin pohjoispuolisten alueiden kallioperä on tällä hetkellä suurelta osin ikijään vallassa. Fennoskandian alue muodostaa poikkeuksen kuuluen lauhkean ilmaston vyöhykkeeseen. Tutkittaessa pitkäikäisen ydinjätteen loppusijoitusta kallio- perään, mahdolliset pitkien aikavälien ilmastomuutokset on kuitenkin huomioitava. Yksi tutkittavista tulevaisuus-skenaarioista on syvälle kallioon ulottuvan ikiroudan leviäminen loppusijoituspaikalle.

Raportissa käsitellään lyhyesti Kvartäärikauden ilmastovaihtelua ja sen mahdollisia syitä. Kirjoittajan johtopäätös on, etta tulevia ilmastomuutoksia ei voida varmuudella johtaa menneistä ilmastovaihteluista, koska ihmisen vaikutusta ja ilmastoon ratkaisevasti vaikuttavien merivirtojen dynamiikkaa ei voida käsitellä ennalta määrättyinä prosesseina.

Vähähuokoisissa kiteisissä kivissä ikirouta voi edetä noin 500 metrin syvyyteen muutamien tuhansien - kymmenen tuhannen vuoden ajanjaksona. Toisaalta ikiroudan pääasialliset vaikutukset kallioperässä liittyvät huokostilassa olevan veden jäätymiseen. Vesi laajenee jäätyessäan noin 9 prosenttia ja tästä johtuvan jännityksen kasvu voi purkautua esimerkiksi painesulamisena. Veteen liuenneet suolai eivät jäädy, vaan ne muodostavat suolavesierkaumia, joiden jäätymispiste voi olla jopa alle miinus kymmenen astetta. Jäätyneen kallion alapuolelle voi muodostua suolaisen veden rintama. Myös jäan sisässä voi esiintyä suolataskuja ja raerajoilla voi olla jäätymätöntä adsorptiovettä.

Radionuklidien kulkeutumisen kannalta ikirouta on sinänsä este, kun tåas vesistöjen alapuolelle syntyvät jäätymättömät alueet (talikit) muodostavat mahdollisen kulkeutumisreitin.

Avainsanat - Keywords Ilmastonvaihtelu, ydinjäte, veden jäätyminen, suolavesierkaumat, kaasuhydraatit, talikit ISBN ISSN ISBN 951-652-106-1 ISSN 1239-3096 Sivumåärä - Number of pages Kieli - Language 42 Englanti PLEASE BE AWARE THAT ALL OF THE MISSING PAGES IN THIS DOCUMENT WERE ORIGINALLY BLANK PREFACE

This study is based on research contract between Posiva Oy and Geological Survey of Finland (GTK). Contact persons were Margit Snellman from Posiva Oy and Lasse Ahonen from GTK. The author wishes to thank Margit Snellman, Liisa Wikstrom and Aimo Hautojarvi for their comments. TABLE OF CONTENTS

1 INTRODUCTION 3

2 OCCURRENCE AND DISTRIBUTION OF PERMAFROST 4

3 FACTORS AFFECTING FORMATION OF PERMAFROST : 11

3.1 Depth and growth rate of permafrost 11

3.2 Climatic requirements and scenarios of permafrost growth 16

3.2.1 Climatic prerequisites of permafrost growth 16

3.3 Processes and scenarios related to the past and future climates 18

4 FLOW AND TRANSPORT IN PERMAFROST CONDITIONS 23

4.1 Physical and chemical processes related to ground freezing 23

4.1.1 Volume/pressure conditions in freezing process 23

4.1.2 Dissolved components in freezing 26

4.1.3 Formation and behaviour of gas hydrates 28

4.2 Flowpaths and transport processes in permafrost conditions 32

4.2.1 General 32

4.2.2 The presence of unfrozen water at the small scale 33

4.2.3 Cryopegs... 33

4.2.4 Taliks 34

4.3 Frost cracking and other periglacial processes 35

5 DISCUSSION AND CONCLUSIONS 36

5.1 Future climatic conditions 36

5.2 Effects of permafrost on a deep nuclear waste repository 37

6 REFERENCES 39 1 INTRODUCTION

Deep geological disposal of nuclear waste aims at permanent isolation of the hazardous material from the ground-surface biosphere. Long half-life of many radionuclides of the waste implies that the performance of the repository must be assessed in time perspective of hundreds of thousands years. During the geological history, especially the last two million years, Fennoscandia has been subjected to repeated glacial and periglacial periods. It is reasonable to assume that such climatic changes may occur also in the future.

Periglacial conditions are typical to the areas around glaciated terrains. Perennially frozen bedrock - permafrost - is a characteristic feature of the periglacial environment. Deep permafrost may affect the performance of a deep nuclear waste repository through different mechanisms and, consequently, features, events and processes related to the permafrost scenario must be examined carefully.

Nurmi (1985) made a comprehensive study of the possible long-term environmental changes in the future, effects of permafrost on the bedrock and groundwaters were discussed in the report. Recently, Gascoyne (2000a) studied the effects of permafrost on the hydrogeochemistry of bedrock.

The aim of the present report is to

• Evaluate factors and processes leading to the formation of permafrost.

• Study depth extent and growth rate of the freezing front.

• Study physical and chemical processes related to the ground freezing. 2 OCCURRENCE AND DISTRIBUTION OF PERMAFROST

Permafrost is defined as soil or rock that remains below 0°C throughout the year, and forms when the ground cools sufficiently in winter to produce a frozen layer that persists throughout the following summer.

Permafrost conditions prevail in predominant parts of the upper ground of the Northern Hemisphere, north of the 60th latitude (Figure 2.1). The main exceptions are Fennoscandia, Iceland and northwestern parts of Russia, which - due to the warming effects of Gulf Stream and predominant southwestern Atlantic winds - belong to the Temperate Zone. Greenland at the same latitudes is covered by thick continental ice. Mountain permafrost may exist at high altitudes also at low latitudes (e.g., in Alps). Mountain permafrost reaching the depth of about 60 metres is also known to occur in the northern side of the hill Yllas in Finnish Lapland (Kukkonen and Safanda 2001).

In the continuous permafrost zone, permafrost occurs everywhere beneath the ground surface except large bodies of water. The discontinuous permafrost zone is divided to the widespread permafrost zone, where permafrost underlies 50-90% of the land area, and to the sporadic permafrost zone, where it occurs mostly in peatlands and underlies 10-50% of the land area. The continuous permafrost area of the Northern Hemisphere is divided into the High Arctic (largely barren with some tundra vegetation), Middle Arctic (tundra with some bare ground) and Low Arctic (continuous tundra cover) (Gascoyne 2000a). The zone of discontinuous permafrost may be covered with boreal forests and wetlands and the snow cover persists for at least six months of the year.

Permafrost underlies an estimated 20%-25% of the world's land surface; it occupies the land surface of more than 50% of Russia and Canada, 82% of Alaska, 20% of China, and probably underlies parts of the Antarctic and Greenland glaciers. Generally, however, bedrock beneath continental ice sheets is not frozen, because of the thermal insulation of ice.

Figure 2.2. shows the distribution and depth of permafrost in Canada. Thick (up to 500 metres and more), continuous permafrost is characteristic to the most of the area of Northwestern Territories (areas north of 60th latitude), with the exception of the westernmost part (Yukon). Permafrost may reach at least the depth of 650 metres in Alaska and High Arctic islands of Canada (Brown et al. 1997).

The bedrock of the mid-continental Canada around Hudson Bay is mainly composed of Archean and Proterozoic granites and gneisses. Main part of the northern Islands and the area towards west is composed of unmetamorphosed Mesozoic sediments (in the vicinity of the red line of Figures 2.1 and 2.2) and subsequently of Cordillerian mountain belt.

A schematic cross section of the depth extension of permafrost is shown in Figure 2.3. It is based on field studies along the Norman Wells Pipeline, which is the first completely buried oil pipeline in permafrost terrain in Canada. sporadic widespread, discontinuous WM continuous |1|| glacier

Figure 2.1. Present distribution of permafrost in the Northern Hemisphere (IPA 1995).

Figure 2.4 shows the estimated distribution and thickness of permafrost in Russia. Permafrost reaching the depth of more than 1000 meters is common in the central Siberia between the rivers Yenisey and Lena. The area is characterised by continental arctic climate with low precipitation and very low annual mean temperature. To the east of river Lena, average depth of permafrost is lower. Thickness Permafrost Thickness (metres) • No Permafrost Present O < )0 O 10to <50 © 50to =500 •Glaciers and be Caps Provincial.and Territorial. Boundaries /V Water

Permafrost Map,of Canada Continuous Permafrost z_^i Extensive Discontinuous Permafrost Sporadic Discontinuous Permafrost Isolated Patches of Permafrost iiiUiJ Knomn.Subsea.Permafrost No Permafrost

Figure 2.2. Distribution and depth of permafrost in Canada (Heginbottom et al. 1995).

Figure 2.3. Schematic cross section of the permafrost along the Mackenzie Valley (line a - b in Figures 2.1 and 2.2). Modified from data of Geological Survey of Canada (2000). ^. j. >• limit of recent permafrost ML:.—

Totai thickness(m) of permafrost, including relict permafrost and permafrost with cryopegs, measured from its upper surface The narrow columns shov/ depths lo J 0-15 sn-ioo I~FMl 11 200---.020D--.o0a Ts~j -oo-riiic pT-i inn mo 300-1100 relict layer(s): a -up lo 100m 1 0-25 I ,7 | 50-150 41)0-tiOO ) 12 .] 200-500 ii§k!i ijiin-ono 900-I500 b - 100 - 200 m c - greater Irian 200 m ~2 0-50 I B I 100-200 |;.y|3;:;j 300-400 400-700 $8i$$\ 600-1000 IIOO-I300 IOO-I0OO j 25-50 [ a | 100-300 iFJl 300-500 500-700 | 700-900 ?|3 l0tl-l000aM-r.o:s "j 0-100 | 10 I 200-300 jr;.j5;;i| 300-600 20•'! 500-8DO S 800-1000 (in regions where thickness of permafrosl increases greatly from valleys lo peaks)

Figure 2.4. Distribution and depth of permafrost in Siberia after Yershov (1991).

Permafrost in Siberia has been interpreted to be very old, even dating back to the Pliocene epoch and beginning of the Pleistocene epoch (more than two million years ago), but the process of ground freezing has continued also during the Holocene (Figure 2.5). Due to the arid continental conditions, the area of Central Siberia having very deep permafrost has not been covered by ice sheet during the Pleistocene. CRYOGENIC AGE (TIME OF INITIAL CONTINUING EXISTENCE OF GROUND IN THE FROZEN STATE )

Pliocene-Pleistocene and Pleistocene; where alas complex occurs: f~ 1 Late Holocene to present Pleistocene to late Holocene I I Present {seasonally freezing ground) Pleistocene; permafrost found only at some depth from the |-*--=- -*| Southern limit of permafrost in the Pleistocene layer of seasonal freezing Pleistocene; at depths of 5 - 80m: late Holocene j'^••'•••i.l Possible southern limit of permafrost in the Pliocene Limit of present day permafrost Late Holocene; lower frozen layer: Pleistocene Late Holocene; lower frozen layer possibly Pleistocene Boundary between frozen ground of different ages TYPE OF CRYOGENESIS PERMAFROST: SYNCRYOGENIC AND POLYCRYOGENIC (MIXED PERMAFROST: EPICRYOGENIC SYNCRYOGENIC AND EPICRYOGENIC) r*' f. '• Syncryogenic deposits with large ice wedges (to 40 m and more) - the "icy" or "yedoma" In glacial, marine and other deposits extending from the surface, ^——! complex, vol. ice cont, 0.6-0.8; polycryogenic deposits with large ice wedges {to 20m} - the vol. ice conL 0.2-O.4; developed at depth (relict) vol. ice cont. to 0.2 "alas" complex and coastal-marine deposits), vol. ice cont. 0.3-O.S Sedimentary, effusive and intrusive formations of the platform, vol. ice f~ ~ Syncryogenic alluvial and glacial deposits with various inclusions of ice from time of initial cont 0.03-O.2 ^ formation, vol. ice cont. 0.2-0.6; polycryogenic fluvioglacial deposits, vol. ice cont. 0.1-0.4 Intrusive and metamorphic shield formations, vol. ice cont. 0.02-0.15 Sedimentary and effusive, intrusive and other formations in mountain "' Syncryogentc alluvial deposits with ice wedges, vol. ic& conL 0.3-0.6; potycryogenic slope systems deposits with weak ice tensing and ice wedges, vol. ice cont. 0.2-0.5 I I Boundary of cryogenic type

Figure 2.5, Age of the permafrost in Siberia after Yershov (1991). Geologically, the most extensively frozen part of the Siberian area belongs to the Central Siberian Craton, which forms the oldest consolidated element of Northern Asia (Khain 1985). The Archean basement is exposed in two major areas, the Anabar massif in the north and the Aldan Shield in the south (Figure 2.6). Western part of the shield is covered by the extensive Upper Permian - Lower Triassic plateau-basalt, while sedimentary cover (up to 10 - 12 km thick) predominates in the eastern part. The floods Yenisey in the west and Lena in the east mark the limits of the permafrost system.

A distinct feature in the permafrost map is the very deep permafrost within the crystalline Archean Anabar massif compared to the thinner permafrost of the surrounding sedimentary formations. Topographically the Anabar massif do not differ from its surroundings, being classified as highland with elevations between about 400 - 800 m. Other areas of very thick permafrost (600 - 1500 m), situated southeast and east of the Anabar massif are hosted by sedimentary formations. These areas are well studied due to the diamond mining from Kimberlites, while direct drilling information from the Anabar massif is less comprehensive. Very deep permafrost (700 - 900 m) has been observed in the northeast of the lake Baikal at the latitudes around 55°. However, this area is also topographically high with elevations up to more than 3000 m. Other topographically high area in the western part of the Central Siberian Craton, the Putorana mountains (elevations up to 2000 m) is also characterised by deep permafrost reaching the depth of 1000 metres. Very deep permafrost is also typical to the Taymyr area in the north.

In the western Siberia, the area between Novaja Zemlja and river Yenisey, depth of permafrost is between 300 and 500 metres. In the Southwestern Siberia (between Ural mountains and river Yenisey) permafrost is relict in nature, but may reach the depth of more than 200 metres. The patchy nature of the permafrost distribution of Eastern Siberia may be attributed to the high topographic differences and to the effect of the Pacific Sea. 10

Figure 2.6. Geology of the Siberian Craton (Khain 1985). 1. Early Precambrian basement; 2-4. Fold framings delimiting the craton; 5-7. Sedimentary platform cover, isobases indicate the thickness of the formation, 8. Plateau basalt, 9. Folds of platform cover, 10. Salt domes, 11. Ultrabasic-alkaline intrusions and kimberlite pipes, 12. Faults. 11

3 FACTORS AFFECTING FORMATION OF PERMAFROST

3.1 Depth and growth rate of permafrost

Permanent freezing of water in the geological medium requires that the annual mean temperature of the ground surface must be below zero. Heat from the earth's interior is transported mainly by conduction to the surface and a downward-increasing temperature gradient (g) tends to form when thermal equilibrium conditions are approached. Typical values of temperature gradient in Fennoscandia are around 10-20 mK/m.

Ground temperature variations are strongly affected by the variations of the surface temperature. Kukkonen (1986) calculated the depth extent of cyclic (sinusoidal) temperature variations on the ground: daily variations reach the depth of about 10-20 cm, annual variations reach the depth of several metres, while long-term climatic variations (100 ky) may be observed even at the depth of 1000 metres. Thickness of the seasonally freezing and thawing active layer may vary from some metres to tens of metres (Yershov 1998), depending for example on the amplitude of temperature variations (difference between temperature of warm period and cold period).

Thermal conductivity of ice is higher that that of liquid water. Cooling of ground during frozen period may thus be more effective than warming during summer period. Depending on the moisture content of the ground, this may affect the near-surface temperature gradient.

Formation of permafrost implies the transformation of liquid water to solid form, leading to the liberation of latent heat of freezing (6 U/mol). Same amount of energy is consumed in melting of ice. Heat capacity (Cp, unit J/kg-K) of the rock is the other factor counteracting the change of thermal conditions in rock. Heat capacities of crystalline rocks are around 2000 kJ/m3K (assuming average density of granitic rock), while latent heat of freezing of water is 334 kJ/dm3 (density 1 kg/dm3). Consequently, moisture content (porosity, p) of the rock becomes important thermal parameter at porosity values higher than about 1 %. Conductive heat flow is dependent on the parameters thermal diffusivity (ar) and thermal conductivity (AT), which are related by:

XT =p-Cp-a 3.1.

Values of these parameters are of the order: Xj « 1-5 Js'^m^-K"1, and aj^l-lO"6 m2-s'\ Heat flux across a unit area is calculated as the product of thermal conductivity and thermal gradient (Fourier's law):

0>Q = AT-g 3.2. Modelling of the propagation and degradation of permafrost is based on the solution of the heat conduction equation in appropriate boundary conditions. Complicated systems may be modelled by numerical solutions, but analytical solutions of simple systems or rough simplifications have also been presented.

The most straightforward method to estimate the depth extent of permafrost is to calculate the asymptotic steady state depth of zero temperature from an assumed 12 average surface temperature and assumed temperature gradient (g). By assuming thermal equilibrium of the system, depth-extent (X) of permafrost is simply

X = T/g 3.3.

Where T is the (long-term) average surface temperature. However, ground-surface temperatures are known to vary in different time scales for many reasons, and many of the variations are cyclic in nature. McEwen and de Marsily (1991) calculated the variation of the ground temperature field by using the analytical solution of heat conduction equation in the case of harmonically varying surface temperature. Climatic oscillation periods of 10000, 40000 and 120000 years were adopted from the ACLIN climatic model (Kukla et al. 1981). Sinusoidal variation of annual mean temperature in Sweden was assumed to be between +5° and -15° C in Sweden during next 150000 years. The results indicated maximum permafrost thickness of about 600 metres after about 60 thousand years (Figure 3.1), but the heat generation of the waste prevents freezing of the repository. The contribution of latent heat was not included in this model, which may slightly overestimate the calculated depth of permafrost.

Temperature ("Q

100 200 30D 400 500 600 700 800 ftOO 1000 HOC 1200 Depth (m)

5000 10000 20000 40000 60000 80000 60000 100000

Figure 3.1. Temperature versus depth and time (years after present) assuming surface temperature variation according to the ACLIN model. Geothermal gradient 10 mK/m (McEwen and de Marsily 1991). 13

Kukkonen and Safanda (2001) modelled the permafrost conditions of northern Fennoscandia during holosene using numerical model including latent heat, which allows the effect of the water-filled porosity to be accounted. The model demonstrates that the propagation (or decay) of permafrost front in low-porosity rock may be of the order of about 0.1 m per year, if the change of surface temperature is about 1.5 degrees (Figure 3.2).

Delisle (1998) modelled the rate of permafrost growth using finite difference method. Latent heat liberation was included into the modelling scheme. In the model, annual mean temperature on the surface was allowed to decrease in linear way (1 degree in 1000 years) from initial value of 0° to -5° or -8°, respectively and to remain constant thereafter. Modelling results (Figure 3.3) indicate relatively linear increase of permafrost depth with linearly decreasing temperature. When a constant surface temperature is attained, permafrost propagates at decreasing rate, displaying an approximately parabolic relationship between time and depth. In these simulations maximum permafrost depth was about 250 metres or less during a cold period of 40 ky, depending on the choice of parameters. However the chosen porosity values were relatively high, thus being not representative for crystalline bedrock.

-70

- -120 O m - -170 I Permafrost depth

- -220

-270 -10000-8000 -6000 -4000 -2000 2000 YEARS A.D.

Figure 3.2 Variation of permafrost thickness during the Holocene, results of 1- dimensional transient modelling (Kukkonen and Safanda (2001). Thermal conductivity of thawed rock, Xr = 3.7 m'1 K"1, diffusivity ax = 1.37-10"6 rn^s"1, and basal geothermal gradient at 6 km, 8 mKm"1. Porosity values and the used temperature values indicated in the picture. 14

Initial terrestrial heat flow value in these simulations was 60 mWm'2, thus the initial thermal gradients were 27 mKm'1 and 35 mKm"1 corresponding to thermal conductivity values of 2.2 Wm^K'1 and 1.7 Wm^K'1, respectively. Using the lower gradient values, final equilibrium depths of permafrost are about 300 m and 185 m for Figures 3.2 a) and b), respectively, if final equilibrium gradients equal the initial gradients.

180 -»#••! (Ul)S S 160. _^—•*—• 140- 120-

Time (1000 years)

250

£ 200 i

% 150

1 100- •10%;2.2WAnK • 30%;1.7WAnK . 30%;2.2WftnK | 50' Tmin= -

0 10 15 20 25 30 35 40 Time (1000 years)

Figure 3.3. Simulation of permafrost growth in geological medium, assuming four different combinations of porosity and thermal conductivity. Linear temperature decrease from 0 to - -5 or from 0 - -8 was assumed in a) and b), respectively, followed by a constant surface temperature (Delisle 1998). 15

Rate of permafrost growth may also be estimated using semi-empirical and simplified formulas. Gascoyne (2000a) presents a general formula to describe permafrost growth as a function of time:

t. = X2/4D 3.4.

Where tc is the time required by subsurface temperature to adjust for change in climate, X is the depth, D is the thermal diffusivity. This expression describes time of freezing as a quadratic function of depth. If the temperature of water is initially close to the freezing point (within a few degrees), the analytical solution to the heat-conduction equation may be approximated by a simplified formula (Carslaw and Jaeger, 1959):

AT • t/(LHp) 3.5. where X is depth, Xr is thermal conductivity, AT is difference between initial temperature of water and freezing temperature, t is time elapsed, and LH is the latent heat of fusion in the formation of ice. Figure 3.4 demonstrates the permafrost propagation rates calculated using these equations with different values of parameters. It must be noted that the actual equilibrium depth of permafrost cannot be inferred from this type of calculations.

500 -i

2000 4000 6000 8000 10000 Time (years)

Figure 3.4. Simplified models for permafrost growth as a function of time. Solid lines represent simplified Neumans solution to problem of Stefan (Carslaw and Jaeger 1959) 1 assuming AT = 1.5° K in each case: a) p = 1 %, XT = 3.7 Wrn^K' , b) p = 5 %, XT = 3.7 Wm'K'1, c) p = 10 %, Xj - 2.2 Wm^K"1. Hatched line represents model of Osterkamp 2 1 and Gosink (1991): tc = X /4D (Gascoyne 2000a), assuming D = M0" mV. 16

3.2 Climatic requirements and scenarios of permafrost growth 3.2.1 Climatic prerequisites of permafrost growth

Heat transfer through the atmosphere - ground surface boundary is a complex process affected both by the properties of the ground and characteristics of the climate. It has been observed that annual mean temperature of air is generally slightly lower than the annual mean ground temperature, and that the difference increases with decreasing temperature (Kukkonen 1986). The reason is that ground is effectively heated during warm season, while - due to the isolating effect of snow cover - escape of heat during cold period is restricted. On the other hand, this effect may be counteracted by high moisture content of the ground: frozen ground allows more heat to be transported out than the seasonally thawed ground allows be transported in. Due to this phenomenon, permafrost may develop and be preserved beneath the seasonally freezing and thawing layer of mires, even though the annual mean temperature would be above zero.

Snow is an effective thermal insulator. The effect of the snow depth on the annual frost depth is depicted in Figure 3.5. In snow-free conditions, depth of the annual frost may reach the depth of more than 1.5 metres even in the relatively mild climatic conditions of southwestern Finland. On the other hand, in the snowy northeastern Finland (Kuusamo region) snow cover is normally more than 50 cm thick, and the depth of annual frost is only about 10 - 30 cm (National Board of Survey 1988). In the present conditions of Finland, only sporadic permafrost occurs in mires and tops of hills of northern Lapland, where the annual mean temperature is between 0° and -2° C, corresponding to frost sums of about 1400 - 1600 day0 C.

Development of deep continuous permafrost covering most of the area of Finland would require a considerable decrease both in annual precipitation and in annual mean temperature. If the annual temperature would decrease clearly below zero, but precipitation would not decrease significantly, accumulation of snow and formation of an ice cover would probably be the result. However, due to the strong albedo of the formed ice sheet, further cooling and increasing average atmospheric pressure around the glacier might result to a decreasing precipitation also around the glacier fore land. 17

10 cm

400 500 800 1200 1400 d *C

PiMcastmraa • KShtannnu > Frost aim

Al« • Cmridl I • Am AkxV.OmridcV.AmV 0- 0

200- 200

400- 400

600 \- a / 500 \ V •"I \ / v / 1000- 1000 V / 200- 1200 1960/61 s 400- 1400 1858-64 IEO0 \r ISS2/63 \ 1800 i 1 0 0-

-^ 40- 40 •I \\ Sj 80- 80 120- 120 1 \ ) ! I! 160- 161 %.

•"•' it 200- 20C-

240' XI XII I II III IV V VI XI XII I II III IV V VI

Figure 3.5. a) Frost depth in open ground in natural state, for different snow depths and frost sums (degree-days below zero), b) Annual frost depth in bare soil as a function of time and frost sum on the southwestern coast (area I) and in Lapland (area V) during a mild winter and severe winter (hatched lines). Solid line shows the average of the observation period of six years (National Board of Survey 1988). 18

3.3 Processes and scenarios related to the past and future climates

Climate cooling during the late Tertiary led to gradual formation of glaciers on the circum-Arctic continents and freezing of the Arctic Sea for 2 - 3 million years ago, marking the beginning of the Quaternary Ice Age. During the early Quaternary, climatic fluctuations gradually formed a more and more regular pattern, which during about the last one million years developed to regular oscillations between cold glacial and warm interglacial periods having successions of about 120 000 years. The climate in Fennoscandia has varied between conditions of thick continental ice sheet and warm climate corresponding to present conditions in Central Europe. Periglacial system with permafrost is a typical transitional environment near the ice margin in dry cold conditions.

Future climatic conditions cannot be known, but predictions based on different assumptions and models may be presented and have been presented. The most straightforward prediction would be that, in the future, climatic oscillation between cold and warm periods would continue in the same manner as in the past. However, without having an insight on the reasons of the climatic oscillations, a prediction of this type is of no quantitative value.

In principle, Quaternary climatic oscillations have been explained by:

• Some external/internal forcing, which launches (triggers) a sequence of events leading deterministic to a particular result. Many scientists have supported the Milankovich' hypothesis, according to which climatic oscillations are due to oscillations in Earth's spinning axis and orbital eccentricity (e.g., Eronen 1992).

• Stochastic events, which - due to their internal dynamics - organise themselves to oscillating patterns. A well-known example is the phenomenon known as El Nino- Southern Oscillation (ENSO) having frequency of about ten years, but climatic oscillations in other times scales may also take place (affecting for example the ocean currents).

• A composite of the two first, for example: a cyclic process caused by its internal dynamics, but synchronised by some external or internal forcing acting as 'pacemaker' (Berger andLoutre 1997).

The most popular theory during the last decades has been the 'Astronomic theory' of climatic fluctuations. It dates back to the 19th century, when the small variations in the Earth's orbit could be discerned and calculated. At the same time first information about the past glaciations was also available. Yugoslavian astronomer Milankovitch adopted and promoted the still valid physical reasoning of the interaction between orbital variations and climatic variations: During periods of decreased summer insolation of the Northern Hemisphere (65° N), length of the snowy season tends to increase leading to increasing albedo. This, in turn, promotes the reflection of the coining radiation, and the positive feedback effect gradually leads to the formation of persistent ice sheets (e.g., Berger and Loutre 1997). 19

Earth's orbital parameters and their effect on the summer solstice radiation are summarised in Figure 3.6. Earth's orbit around Sun deviates slightly from a perfect circle with an amount described by the parameter eccentricity. The e-value 0.04 means that the distance between Earth and Sun varies about 7 - 8 % during the year, while e- value zero means that Earth's orbit is circular.

Earth's spinning axis and the normal of the orbital plane (ecliptica) form an angle, value of which being at present 23.5°. Because Earth's axis is turning around the ecliptica (process called precession) once in about 20 000 years, also the season during which each hemisphere is closest to the sun (i.e. in periheli position) is changing within the same cycle (climatic precession). At present the distance between Earth and Sun is closest during midwinter, while during summer the Sun is furthest away. When similar conditions appear again after 25 000 years, Earth's orbit is almost circular. With respect to the Milankovitch theory (low summer insolation) the most favourable conditions for glaciation are those during which periheli is during northern winter (as now) and the orbital eccentricity is high (as was for 120 000 years ago).

The most important factor affecting the summer radiation of each hemisphere is the obliquity, meaning the angle between ecliptica and Earth's axis, which oscillates between about 22° and 25° within a cycle of about 40 000 years. The present value of obliquity, 23.5° lies thus close to the average. Changing obliquity manifests in the moving position of the Arctic Circle: During next 10000 years Arctic Circle will move about 150 km towards the north, then during next 20000 years about 300 km southwards and so on. Northern Hemisphere mid-summer radiation is at it's minimum when the obliquity is at minimum (i.e., Polar Circle has shifted to it's northernmost position).

The combination of these orbital parameters is a time- and latitude-bound solar energy flux, shown in Fig. 3.6 for the latitude 65N during June. It can be estimated that the variation from the present value is at maximum about 2% toward higher and about 8 % toward lower midsummer radiation within the next 100 000 years. 20

200 -50 -100 -150 -200 (kyrBP)

Figure 3.6. Variations of earth's orbital parameters and their effect on northern hemisphere (65° N) summer solstice radiation (Wm"2) within past and future 200 ky calculated by Berger (1978).

The original Milankovich's model with the simple linear (albedo) feedback could not reproduce satisfactorily the climatic variations observed in proxy record of past climate. Consequently, later development of the astronomical climate models focused on the model 'tuning' by using more complex feedback parameters, e.g. higher order feedback terms, lag-terms etc. Best known of these models are the ACLIN (Astronomical Climate Index) and Imbrie & Imbrie models (e.g., Forsstrom 1999). Both these models use 65° N midsummer radiation as the climatic forcing function, but are tuned to match known climatic history derived from geological records. Output of these models for the next 100 000 years is compared with the 'Milankovitch' forcing in Figure 3.7. The models predict a cold period after 60 ky, which is also next time when the northern hemisphere summer insolation value is lower than at present.

During 1990's a new type of climate model - LLN 2D - has been under development in Louvain-la-Neuve, Belgium. In this model, both solar insolation and atmospheric CO2 have been adopted as forcing factors. Feedback terms are physically plausible, including interactions between air, ocean, ice sheet and continents. The model is two-dimensional along the longitudes. Model is described in more details by Gallee et al. (1992). 21

Warm ^ A Forcing A ^~ Imbrie /A rx /A —ACLIN

a 60 \ o \ J20 M trv /\ / V/ XX // / "VT °° V Cold time (ky)

Figure 3.7. Comparison of the astronomical forcing function (65° N summer solstice radiation) and the course of climate calculated by Imbrie and Imbrie (1980) model and ACLIN model during the next 100 ky. Scale of the insolation curve is 480 ±60 Wm'2, the climate model parameters are not in scale otherwise than with the respect to the fixed starting point at t = 0.

The CO2 cycle and interactions were not coupled as feedback processes in the LLN 2D model, but used as an input factor. The CO2 forcing function was derived from the variation of the atmospheric CO2 during the last glaciation (Jouzel et al. 1993), which was about 290 ppm during Eem interglacial and decreased towards the end of the glaciation to about 200 ppm. Termination of the three last glaciations has been accompanied by a rapid increase of atmospheric CO2 concentration from about 200 ppm to about 290 - 300 ppm (Fisher et al. 1999).

The results of LLN 2D model indicated that no glaciations will appear during next 150 ky, if atmospheric CO2 concentration will remain at least at the same concentration level as that of the Eem interglacial (290 ppm). Present (year 2000) atmospheric CO2 concentration is 370 ppm, increasing with more than 1 ppm each year. The results of this model are thus clearly different from the earlier using only solar radiation as forcing factor, and point out the importance of the atmospheric greenhouse gases as climate- affecting factors.

Climatic instability is a characteristic property of the Quaternary period. Ice core records have revealed that abrupt climate changes of at least regional scale have taken place both during glacial and interglacial conditions. Dansgaard et al. (1993) suggested that climate in North Atlantic region is able to reorganise itself to a new state even 22

within a few decades. The 'flickering' nature of the Late Pleistocene climate is well established by detailed studies of the Greenland ice cores (Taylor et al. 1993)

Adams et al. (2001) present a comprehensive study of sudden climate variations during the Quaternary, concluding that "All the evidence indicates that most long-term climate change occurs in sudden jumps rather than incremental changes". Sudden changes are a tangible manifestation of instability of a system. Small initial disturbances may cause a cascade of events amplifying the total process.

In northern Europe, disturbances in the circulation of northern Atlantic Ocean probably play a major role in either triggering or amplifying rapid climatic fluctuations. A key factor affecting the climate in northern Europe and northwestern Siberia is the warm salty Gulf Stream, which carries heat to the northern latitudes and sends mild air masses across the north Europe. Upon cooling and reaching the Barents Sea, the salty Gulf Stream water sinks and the 'pull' exerted by the sinking water helps to maintain the flow. The sinking water mass forms a major constituent of North Atlantic deep water. Studies of deep-sea sedimentary records have provided evidence that the cold periods during the end of the last glaciation (e.g., Heinrich Events) were associated with a decline of the warm surface inflow of the Gulf Stream (Rasmussen et al. 1997). Fading of the Gulf Stream or shifting of its position would cause major changes in the climate of northern Europe towards colder and drier conditions. 23

4 FLOW AND TRANSPORT IN PERMAFROST CONDITIONS

4.1 Physical and chemical processes related to ground freezing

4.1.1 Volume/pressure conditions in freezing process

The most pronounced physical change associated with the freezing of water in the rock fractures is the volume increase due to the water-ice phase change. Contrary to most compounds, specific volume of the solid phase (ice) is about 9 % higher than that of the liquid phase. Volume changes in the water/ice system are illustrated in Figure 4.1 within the temperature interval -10 - +10 °C. Construction of the figure is based on the following data and considerations.

Even though the sudden volume change of phase transition is the most important factor, volumes of the both phases are also dependent on pressure and temperature as approximated by:

• Isothermal compressibility describing the effect of pressure on the volume at constant temperature:

p V dp

• Thermal expansion coefficient describing the effect of temperature on the volume at constant pressure:

B =1^ 42 PT VdT Compressibility value of liquid water is well known through a wide temperature range. For liquid water at T » 0 °C, pwater = 5-10"10 Pa"1 (CRC 1994), being very close to the 10 1 corresponding value for ice, piCe « 3 - 4-10' Pa' (Dorsey 1940), which is, however, less exactly known. Compressibility values of crystalline rocks vary between about MO"8 and MO"11 Pa"1 depending on the fracturing (Freeze and Cherry 1979).

Temperature-volume data of liquid water at standard conditions is well known and published (e.g., CRC 1994). Thermal expansion data of ice from different original studies were reviewed by Dorsey (1940), indicating that the value is close to: Pice = 120 • 10"6 K"1. This is clearly higher than that of common crystalline rocks, varying 6 1 between pr0Ck = 5 - 30 • 10" K" (from the linear expansion data of Mesimaki (1998).

In addition to the salinity (e.g., Gascoyne 2000a), freezing point of water is also dependent on the pressure: increase of pressure by 100 bars decreases the freezing point by about one degree (e.g., CRC 1994). 24

1.090 1.005

K. E •o

•2 1.085 -

E

1.080 -10 -5 0 5 Temperature °C

Figure 4.1. Volume changes in the phase transition ice/water at 1 bar (solid line) and 50 bars (hatched line). Note the different scales for ice and liquid water. The figure is constructed according to the data presented in text.

As indicated by Figure 4.1, relative volume change in freezing is not notably dependent on the pressure within the relevant pressure range (i.e., to the hydrostatic 'depth' down to about 500 metres). Total water-filled porosity of crystalline rock is relatively low, being here assumed to be 1 % or less. The volume-pressure change during downward- advancing freezing of the bedrock/groundwater system may either:

• Be dispersed and dissipated into the non-frozen rock below, if there is interconnected porosity available and the freezing process is slow enough.

• If the mass flow and pressure equalisation is excluded, pressure of the freezing front will increase. If the pressure increase is evenly distributed into the rock mass, the increasing volume due to the freezing of water (1 % v/v) increases the total volume by about 0.09% causing and additional confined pressure of about 20 bar to the bedrock (from eq. 4.1). In reality, however, water in the bedrock is not evenly distributed, but it is mainly concentrated to a few fractures. Consequently, local pressure peaks may develop. 25

A schematic phase diagram for water is shown in Figure 4.2, indicating that at least five stable forms of ice are known (e.g., Dorsey 1940). However, only the specific volume of "ordinary" ice (ice-I) is higher than that of water. The dense forms of ice (Ice-II to Ice-VI) can only be formed at pressures higher than about 2 kbars, which is thus the maximum pressure able to be created by freezing of water in a completely rigid environment (temperature of about -20 °C or lower is also required).

As can be seen in the phase diagram, increasing pressure leads to the melting of ice (Ice-I) in the temperature range from about -20 °C to 0 °C. Pressure melting of ice is well known from different environments, including ice beneath a glacier, as well as ice beneath ice skates. Pressure melting is thus also a potential mechanism releasing the overpressure due to freezing.

0 1000 2000 3000 4000 5000 6000 7000 Pressure (bar)

Figure 4.2. A simplified phase diagram for ice and water at temperatures below zero and pressures up to 7 kbar (Simplified from Dorsey 1940 with Ice-I /Liquid - boundary approximated from data of CRC 1994). 26

4.1.2 Dissolved components in freezing

Dissolved components of water are not able to accommodate into the ice lattice. Consequently, a water phase with increasing concentration of dissolved components will be separated in the freezing process. Due to the freezing point depression with increasing salinity of water, saline water or brine phase becomes stable at certain concentration - temperature values (Figure 4.3).

The amount of total dissolved solids (TDS) in fresh near-surface waters in Finland is of the order of 0.1 - 0.3 g/l (Lahermo et al. 1990), but the average salinity increases downward in the bedrock. At the depths between 200 and 400 metres, groundwater TDS-value is typically around 5 - 10 g/l, and below 400 metres more than 10 g/l (Ruotsalainen et al. 2000) occasionally reaching concentrations up to 100 g/l and more (Ahonen 1999, and references therein).

If assumed that the original salinity of groundwater in the bedrock is about 10 g/l, which is a typical value at the depth of about 400 - 500 metres. If 90 % of the water freezes slowly allowing the residual fluid to accommodate the dissolved salts, salinity of the residual fluid will be about 100 g/l. It can be estimated that the freezing temperature of such solution is about -5 °C. Freezing point depression due to the salinity of water is thus more pronounced than that due to the pressure increase (Figure 4.2).

BRINE and NaCI

BRINE and

NaCI* 2 H20 | 2 ICE and BRINE

•20

ICE and NaCf*2H20

100 200 300 NaCI concentration

Figure 43. Phase diagram for the system NaCI - H2O (Biggar and Sego 1993). 27

Increasing salinity of remnant water phase also leads to a supersaturation with respect to mineral phases. Minerals reported to form in cryogenic processes are calcite (CaCCh), silicates and oxides of iron and manganese (Vogt 1991), but - in many cases - formation of the reported precipitates may be attributed to the evaporation or sublimation in cold aridic conditions (e.g., Cailleux 1965). However, even the very soluble salts (e.g., mirabilite NaSC>4-10H2O) have been proposed to precipitate in cryogenic conditions (Yershov 1998). Mirabilite has also been reported to precipitate in temperate climate during the winter months when humidity is low and to disappear as the humidity rises, (http://www.mammothcave.com/glossary.htm).

Calcite is a relatively reactive mineral, being usually at least close to equilibrium with groundwater. The solubility equation for calcite can be written as

+ 2+ CaCO3 + H = Ca + HCO3', logK= 1.85, AH = -26 kJ/mol 4.3

Negative enthalpy value (exothermic reaction) indicates that the solubility of calcite increases with decreasing temperature. At T = 0 °C, stability constant of the dissolution reaction is log K = 2.27, and the solubility product of calcite can be written in logarithmic form as:

2+ Log[Ca ] + log [HC03] + pH = 2.27 4.4

2+ At pH 8.27, T = 0 °C, equilibrium concentrations of Ca and HC03" are thus about 1 mol/1 of each at a temperature of 0 °C. During freezing, the residual fluid phase becomes readily supersaturated with respect to calcite, which may precipitate out. This in turn depletes the solution with respect to calcium and bicarbonate, and the relative portion of other main ions, Na+, Cl", and possibly SO/ increases.

Solubility of mirabilite is very high, being defined by the equation:

+ = Na2SO410H2O = 2 Na + S04 + 10H2O,

log K =-1.1, AH = 79kJ/mol 4.5

Thus, at T = 0 °C, value of the stability constant is: log K = -2.37, and the logarithmic form of solubility product can be written as (assuming [H20] « 1):

+ = 2 log[Na ] + log[SO4 ] = -2.37 4.6

This solubility relationship is visualised in Fig. 4.4 indicating that mirabilite will precipitate out only in near-complete freezing conditions or through evaporation of the solution. 28

100

80 -.

60 -3 o> Mirabilite

dissolved salt 20

4 6 10 Na (g/t)

Figure 4.4. Stability field of mirabilite (Na2SO4-10H2O) as a function of sodium and sulphate concentration at T = 0° C.

4.1.3 Formation and behaviour of gas hydrates

Gas hydrates - clathrates - are ice-like crystalline compounds of water and gas molecules (Figure 4.5 a). The gas molecules are entrapped in a cage formed by a group of water molecules. The uncharged gas molecules supporting the cavity are bound by Van der Waals forces to the water molecules bound by hydrogen bonds to form the cages. The structures of gas hydrates are different from any of the known forms of ice, although they are also solids consisting mostly of water molecules. So far three different structures have been identified: structure I, structure II and structure H. The three- dimensional crystal structures of hydrates are quite complicated (Figure 4.5 b).

Type I structure can only hold small gas molecules, no larger than 5.2 A in diameter. These include methane, ethane, carbon dioxide, and hydrogen sulfide. Type II hydrate formers include larger molecules propane and isobutane. However, nitrogen, a relatively small molecule, also forms a Type II hydrate. Type H hydrates are formed by larger molecules but only in the presence of a smaller molecule, such as methane. Type H formers include: 2-methylbutane, methylcyclopentane, methylcyclohexane, and cyclo- octane. 29

a)

Figure 4.5. The habit (a) and crystal structure (b) of a typical gas hydrate: Structure I with two unit cells (http://web.mit.edu/caozt/www/Research/Research.html).

Gas hydrates occur abundantly in nature, both in Arctic regions and in marine sediments. Gas hydrates form when water and gases meet at high pressure and low temperature. These conditions exist naturally in subsurface permafrost regions on land, and in sediments under the oceans. Figure 4.6 displays the phase diagram of pure methane hydrate together with assumed thermal gradients in the two regions of the world, where hydrates are stable in the subsurface: permafrost regions (referred to as continental hydrate) and deep sea sediments (oceanic hydrates). The figure indicates that the stability of hydrates is considerably more pressure dependent than that of ice: the minimum pressure of formation at T = 0 corresponds to the hydrostatic depth of about 200 metres. The temperature-stability field increases with increasing pressure up to temperatures around 15° C at the depth of 1000 metres.

Adding other trace gases or ions to the system yields a shift of the phase boundary. Increasing salinity shifts the phase boundary towards lower temperatures, whereas admixture of CO2, H2S and higher-molecular-weight hydrocarbons cause a shift into the opposite direction (e.g., Kvenvolden, 1993). The addition of a small amount of propane, which causes a change from structure I to II, leads to a pronounced shift of the hydrate stability field towards higher temperatures: Less than half the pressure is required at a given temperature to stabilize structure II hydrates than structure I hydrates (Sloan, 1990). 30

PERMAFROST DEEP SEA

phase/^NJ gas boundary |\ SQO depth of permafrost hydrate Q. \ 1000 hydrate seafloor

1SO0 thermal gradient BSR \J _J I - | N I -20 20 -20 0 20 temperature (°C) temperature (°C)

Figure 4.6. Simplified phase diagram of hydrates in a pure water/methane-system adapted to typical situations expected in permafrost (A) and in deep sea sediments (B). The diagram demonstrates that hydrates are stable at higher temperature than ice, but the stability field also extends to lower temperatures than the thermal gradient defines. From Sloan (1990) after Kvenvolden (1988).

Significant quantities of naturally occurring gas hydrates have been detected in many regions of the Arctic including Siberia (Makogon and others, 1972), the Mackenzie River Delta (Bily and Dick, 1974), and the North Slope of Alaska (Collett, 1983; Collett et al 1988). Permafrost may act as a seal, preventing gases escaping through faults from rising higher, and gases in permafrost are bound in solid clathrate structures. However, due to the high hydrostatic pressure, gases beneath the permafrost are probably also entrapped by solid hydrates down to about 1000 metres at average thermal gradient values (Figure 4.6 A).

At standard conditions (STP), one volume of saturated methane hydrate may contain as much as 164 volumes of methane gas (Davidson et al. 1978), while dissolved methane concentrations in deep groundwaters are of the order of 1 vol/vol at STP or less (Gascoyne 2000b, Sherwood Lollar et al. 1993). Because of this large gas storage capacity, gas hydrates may represent an important commercial source of natural gas.

The source of methane in clathrates may be either biogenic or (abiotic) thermogenic. Biogenic origin is typical for the sea-bottom gas hydrates, which may form relatively slowly over a wide area, as bacteria digest detritus that has sunk to the seafloor over a wide area. Thermogenic hydrates are formed when methane and other 31

gases vent through a fault and emerge on the seafloor. They are likely to be more concentrated and have smaller areal extent than biogenic gas hydrates.

Known permafrost gas hydrates in Alaska are stratigraphically associated with oil and gas occurrences. These accumulations include the Prudhoe Bay and Kuparuk River oil fields, and the heavy-oil and tar deposits within the West Sak and Ugnu sands (Werner, 1987). Distribution of gas hydrates in permafrost areas of northern Europe and Asia is less well known, but oil and gas occurrences are common in these areas too. 32

4.2 Flowpaths and transport processes in permafrost conditions 4.2.1 General

Transport of ionic or molecular compounds in permafrost conditions may only take place via:

• Diffusion through the permafrost

• Transport with flowing water

In general, diffusion can be considered to be a slow process even in the presence of a continuous liquid phase. In frozen rock, the main pathway for diffusion is the intergranular space between rock and ice or between individual ice crystals. The space may contain small amounts of unfrozen water sorbed on the surfaces or as intergranular salt-water rejections.

Groundwater flow requires: a) hydraulic gradient; b) hydraulically continuous flow paths (interconnected porosity). Frozen ground together with the cold aridic conditions required by the ground freezing may decrease the average hydraulic gradient by two reasons:

• low precipitation does not facilitate formation of uneven groundwater table

• hydraulic head distribution and differences cannot be transmitted by solid ice

Flowpaths may develop, if continuous bodies of liquid water are available. Formation of water bodies may be due to:

• freezing-point depression due to salinity (cryopegs)

• thermal conditions (taliks)

• pressure melting of ice

Freezing/thawing -related physical and chemical processes are best known from near- surface conditions. Mechanical effects include (Yershov 1998):

• frost cracking of the soil

• mechanical fracturing of rock due to freezing

• formation of frost heaves on the ground surface

• ground patterns due to frost-assisted movements

• solifluction 33

4.2.2 The presence of unfrozen water at the small scale

In the freezing process, liquid water crystallises to a more ordered crystalline form. Water molecules in liquid water mainly interact with each others by forming loose hydrogen bonds, while the molecules very close to mineral surfaces may have molecular-scale interactions with the atoms of the mineral surface. If the molecular interactions are strong enough, water molecules near the mineral surface segregate from the liquid phase and sorb onto the mineral surface by various chemical mechanisms. The sorbed water molecules may not either take part into the freezing process, if the 'sorption forces' are strong enough. The process results to the formation of electrolyte- rich adsorption layer of water molecules on the grain surfaces. Thickness of the layer may be some hundred nanometers (Mohamed et al. 1997).

Experimental studies (Anderson and Morgenstern 1973) have shown that ice-water interfaces develop during freezing and a thin liquid film probably exists at contact surfaces because the ice does not bind strongly to the soil. The ice crystal is then able to grow away from the particle surface into the pore space and reject solutes, which then concentrate within the void space as brine 'islands' (Hivon and Sego 1995). Depending on the grain size of the sediments, these brines may then migrate away from the freezing surface driven by density and capillary forces, and coalesce to form separate and possibly extensive saline water bodies (cryopegs).

Mohamed et al. (1997) modelled the diffusion processes in frozen soil by a 'dual flux concept', which considers the continuous adsorbed moisture as the diffusion pathway. The results indicated that, due to multi-flux effect, metal ion transport in the unfrozen boundary layer in frozen soil is faster than that in free water. However, the work did not consider diffusion in low-porosity crystalline rock, but only in loose soil systems.

4.2.3 Cryopegs

In the freezing process, a water phase with increasing concentration of dissolved components will be separated. Due to the freezing point depression with increasing salinity of water, saline water or brine phase becomes stable at certain concentration/ temperature values. These cryogenic exclusions - cryopegs - may contain hundreds of grams of total dissolved solids, and the freezing point of the brine may be -20° C or lower (Figure 4.3).

Saline water/brine body may develop below an advancing freezing front, being called subpermafrost cryopegs. Slow freezing favours crystallisation of a pure ice phase and quantitative segregation of the brine. The thickness of subpermafrost cryopeg layer may reach several hundreds of meters (Yershov 1998). Salinity increase of subpermafrost cryopeg may be opposed by:

• Diffusion of ions towards less saline water tends to level off the concentration gradient

• Increasing hydrostatic pressure due the volume expansion in freezing causes increasing hydraulic head and, consequently, water flow 34

Intrapermafrost cryopegs form if the segregating water phase is entrapped into the ice- rock matrix. Rapid freezing probably contributes to the process, as indicated by the sea- water freezing. Salt content of freshly formed sea ice is about 4 - 5 %o, but the amount decreases with ageing of the ice. During the ageing process, grain-boundary salts and microscopic salt inclusions in ice tend to migrate and form coherent brine bodies, while the salinity of solid ice decreases. Intrapermafrost cryopegs vary in size from small isolated brine inclusions to extensive brine layers in ice.

4.2.4 Taliks

Cryopegs are saline water or brine bodies formed due to fluid-phase segregation and freezing point depression during freezing, while taliks are thermally induced unfrozen bedrock domains in the areas of otherwise continuous permafrost. The most typical cases are the unfrozen bedrock below sea and below major lakes or rivers. For example, distribution of permafrost in Siberia can be seen to be distinctly controlled by the major rivers Yenisey and Lena (Figure 2.4), which have never been underlaid by permafrost (Yershov 1998).

Taliks penetrating through whole frozen bedrock sequence are open taliks, while those only reaching limited depth of the whole frozen section are closed taliks ("pseudotaliks"). Intrapermafrost taliks may develop between two freezing fronts of different ages, e.g., between a deeper relict permafrost and upper (younger) permafrost layer.

The sources of heat generating the talik may be classified as exogenous when coming from the earth's surface (solar heat) or endogenous when coming from the geological formation. Heat from the Earth's interior may be due to volcanic - hydrothermal processes or due to natural radioactivity.

Underwater taliks develop beneath all lake and river systems, which do not freeze throughout during cold periods. Depending on the size of the water system and on the temperature of water, the talik may be either open or closed. Rivers may transfer continuously large amounts of heat, thus even a narrow river may be able to maintain unfrozen conditions. Warming effect of lakes during cold periods is more based on the large heat capacity of water. Another important factor is that the coldest water in the lake tends not to be at the lake bottom, because water is at its densest a + 4° C.

In otherwise-frozen bedrock, open taliks serve as the most important flow path for groundwater, if the bedrock is permeable. Vertical flow systems may develop in a network of open, water conductive taliks, which are hydraulically connected below the permafrost. Driving force of the flow is mainly caused by the topographic differences on the surface. Possible hydraulic head differences caused by the expansion of freezing water may also seep through the taliks. Contribution of this process to volumetric flux is, however, low. 35

4.3 Frost cracking and other periglacial processes

Two different phenomena may be called 'frost cracking':

1) Cracking of soil/rock due to ice formation

2) Contraction cracking of ice

Frost cracks are developed in the seasonally thawing/freezing near-surface layer of the permafrost. Melt water of the warm season infiltrates down, and during following cold period frost cracks may form in the soil layer, because the volume of water increases in the formation of ice. The open space formed during following melting is again filled with water or other material. During repeated thawing/freezing cycles, ice wedges or sand wedges may form. Frost cracking also causes the formation of various patterns on the ground surface (e.g. polygons). Thermal expansion of water during freezing may also cause cracking of crystalline rocks if water seeped into the rock fissures freezes rapidly.

Another type of frost cracking, commonly seen in ice, is the thermal contraction cracking. Due to the decreasing specific volume of ice with decreasing temperature, tensile stress causes cracking of ice. However, this phenomenon can only take place in the surface conditions, because ice-temperature fluctuations of tens of degrees are required.

Both of these cracking phenomena are clearly connected to the near-surface active layer of the permafrost. Contraction cracking can only take place in ice-dominated systems (mainly in pure ice) in contact with climate fluctuating between very cold and mild. Cracking of geological material due to freezing of water commonly takes place in the active layer of the permafrost. The maximum depth of frost cracking in permafrost has been estimated to be 3 - 4 meters (Yershov 1998).

Heaving of ground due to freezing is a widely developed process leading to the formation of different kinds and sizes of formations, like mounds (palsas) and hillocks. Frost heaving formations are classified to two types according to the mechanism of formation, segregation heaving and intrusion heaving mounds. Segregation heaving is due to the downward migration and accumulation of moisture to the freezing front. Palsas in mire are typically formed in this process, and the difference in heat conductivity of frozen and unfrozen peat plays an important role in the process. Segregation heaving is typical to soils and loose formations, but it may also take place in porous rocks like welded tuffs (Akagava and Fukuda 1991). Intrusion heaving of ground is due to increase of underground pressure in a closed system. Freezing of a sublacustrine closed talik due to shallowing or emptying of a lake may lead to this type of intrusion heaving formations, sometimes called as pingo's. 36

5 DISCUSSION AND CONCLUSIONS

Two mutually dependent questions may be posed with respect to the item permafrost- nuclear waste disposal:

• How important or relevant issue the permafrost is in the context of the future climatic scenarios of Fennoscandia?

• What are the effects of frozen bedrock on the performance of a deep nuclear waste disposal system, and which are the most critical phenomena?

5.1 Future climatic conditions

According to the classical astronomical theory, Quaternary climatic fluctuations were assumed to be controlled mainly by the variations in the inclination of the Earth's spinning axis. The models derived from the original Milankovich theory (i.e., cyclic fluctuations of the northern 65° midsummer radiation being the climatic forcing factor) forecast that the future fluctuations will probably continue in a similar manner. According to other theories, long-term climatic fluctuations may be due to the internal dynamics of the major climate affecting entities (e.g., sea currents, glaciers etc.), in which the apparently random processes may in certain time scales be organised to cyclic patterns. There is also geological evidence that major climatic fluctuations may have taken place within very short periods of time, as for example the 'Heinrich Events' in the northern hemisphere during the end of last glaciation. Some of these abrupt changes are clearly associated with the dynamics of the Gulf Stream, and may have drastic consequences for the climate of Fennoscandia.

As a strong heat (infrared) absorber, atmospheric carbon dioxide has always in the Earth's geological history been a climate-controlling factor by maintaining the conditions suitable for life by means of the so-called 'greenhouse' effect. Carbon dioxide concentration of the Earth's atmosphere has probably been stable for millions of years, but only concentration and variations during about the last half million years are known in details. Current trend in the evolution of atmospheric CO2 concentration indicates a very drastic increase in the nearest future, if compared with the geological time scale (Figure 5.1). New climate variation models taking account the CO2 -climatic forcing indicate that low atmospheric CO2 is a prerequisite for the cooling of the climate. Anthropogenic effects on the long-term evolution of the future climate cannot be ruled out.

It may be concluded that very large uncertainties are involved in forecasting the future climatic development. However, existence of permafrost conditions in Fennoscandia in the future cannot be ruled out. 37

400 a. 300 O o 200

•400 •300 -200 -100 0+10 Age ka

Figure 5.1. Variation of atmospheric CO2 during last 400 000 years (solid line) and current/future trend of increase (dotted line). Past data (10 000 BP and earlier) from Shackleton 2000. Future trend for the atmospheric CO2 evolution corresponds to time span of next 40 to 100 years according to different scenarios of IPCC (Intergovernmental Panel on Climatic Change, http://www.ipcc.ch/) (Note that thickness of the line corresponds to about 1000 years).

5.2 Effects of permafrost on a deep nuclear waste repository

Effects of permafrost on the deep nuclear waste disposal system may be classified to different categories:

• Effects on the integrity and performance of an intact repository

• Effects on the radionuclide mobilisation and transport in case of canister failure

• Effects on the biospheric behaviour of radionuclides

If the depth of permafrost is of the same order as the disposal depth, mechanical effects due to freezing of rock around the repository have to be taken into considerations. At the moment, there is very little detailed information on the mechanical behaviour of crystalline rock during and after freezing. Preliminary observations from deep mines in the permafrost domain do not indicate discernible breaking due to freezing of water in the rock fissures. Evidently the worst case in this respect would be a situation in which a freezing front would be oscillating within the depth range of the repository.

Temperature of the bedrock decreases substantially in the permafrost conditions. If the permafrost front reaches the depth of the repository, temperature gradient around the heat-generating repository may increase, depending on the heat generation (i.e., time elapsed since the burial) and on the thermal conductivity and moisture content of the bedrock. 38

A basal cryopeg zone (zone of cold brine) may develop beneath the freezing front. If this zone reaches the depth of the repository, performance of the technical barriers in the saline environment is the key question.

In case of a broken canister (broken canisters), radionuclide mobilisation from the spent fuel and migration through the litosphere are the key questions. Primary step on the mobilisation of the main components of the spent fuel is the radiolytic disintegration of the UC>2-surface. This process is not affected by the ambient temperature conditions. Development of deep permafrost ascertains the permanence of reducing conditions within and beneath the freezing front, which in turn results in the immobility of most of the important radionuclides.

Radionuclide transport in permafrost conditions is limited by the lack of liquid water phase, decreased hydraulic conductivity due to freezing of water in the fractures and due to the low hydraulic gradient. Most important possible transport pathways in the permafrost conditions are the open taliks. Transport pathways are thus more 'channeled' and the role of dispersion decreases. Ground-surface topography is the main driving force of flow through a network of taliks. 39

6 REFERENCES

Adams, J., Maslin, M., Thomas, E. 2001. Sudden climate transitions during the Quaternary. Progress in Physical Geography (in press) (also available in: http://www.esd.ornl.gov/projects/qen/transit.html).

Ahonen, L. 1999. Effect of saline water on metallic copper. POSIVA Working Report 99-58, 28 p. ISSN 0359-548X.

Akagava, S. and Fukuda, M. 1991. Frost heave mechanism in welded tuff. Permafrost and Periglacial processes 2, 301 - 309.

Anderson, D.M. and Morgenstem, N.R. 1973. Physics, chemistry, and mechanics of frozen ground: A review. In: Proceedings of the 2nd International Conference on Permafrost, Yakutsk, USSR, July 13- 28, 1973, 269-273.

Berger, A. 1978. Long-term variations of daily insolation and Quaternary climatic changes. J. Atmos. Sci. 35, 2362 - 2367.

Berger, A. and Loutre, M.-F. 1997. Paleoclimate sensitivity to CO2 and insolation. Ambio 26(1), 32-37.

Biggar, K.W., and Sego, D.C. 1993. The strength and deformation behaviour of model adfreeze and grouted piles in saline frozen soils. Can. Geotechnical J. 30, 319-337.

Bily, C, and Dick, J.W.L. 1974. Natural occurring gas hydrates in the Mackenzie Delta, Northwest Territories: Bulletin of Canadian Petroleum Geology 22(3), 340-352.

Brown, J., Ferrians Jr., O.J., Heginbottom, J.A., Melnikov, E.S. 1997. Circum-Arctic map of permafrost and ground ice conditions. U.S. Geol. Survey, U.S. Department of Interior, Circum-Pasific Map Series CP-45.

Cailleux, A. 1965. Quaternary secondary chemical deposition in France. Geological Society of America, Special Paper 84, 125-138.

Carslaw, H.S. and Jaeger, J.C. 1959. Conduction of heat in solids. Oxford University Press. 510 pages. ISBN 0-19-853368-3.

Collett, T.S., 1983, Detection and evaluation of natural gas hydrates from well logs, Prudhoe Bay, Alaska, Proceedings of the Fourth International Conference on Permafrost, Fairbanks, Alaska: National Academy of Sciences, Washington D.C., p. 169-174.

Collett, T.S., Bird, K.J., Kvenvolden, K.A., and Magoon, L.B., 1988, Geologic interrelations relative to gas hydrates within the North Slope of Alaska: U.S. Geological Survey Open-File Report 88-389, 150 p.

CRC 1994. CRC Handbook of chemistry and physics. (D.R. Lide and H.P.R. Frederikse, eds.). CRC Press, Inc. 40

Dansgaard, W., Johnsen, S.J., Clausen, H.B., Dahl-Jensen, D., Gundestrup, N.S., Hammer, C.U., Hvidberg, C.S., StefFensen, J.P., Sveinbjornsdottir, A.E., Jouzel, I, Bond, G. 1993. Evidence for general instability of past climate from a 250 kyr ice-core record. Nature 3 64, 218 - 220.

Davidson, D.W., El-Defrawy, M.K., Fuglem, M.O., and Judge, A.S. 1978. Natural gas hydrates in northern Canada: Proceedings of the 3rd International Conference on Permafrost, National Research Council of Canada, v. 1, p. 938-943.

Delisle, G. 1998. Numerical simulation of permafrost growth and decay. J. Quaternary Sci. 13(4), 325-333.

Dorsey, N.E. 1940. Properties of ordinary water-substance in all its phases: water- wapor, water, and all the ices. American Chemical Society, Monograph Series. Reinhold Publishing Corporation. 673 pages.

Eronen, M. 1992. Milloin tulee seuraava jaakausi? (When will the next ice age come?) Terra 104(4), 249 - 262. (In Finnish with English abstract).

Fischer, H., Wahlen, M., Smith, J., Mastroianni, D., Deck, B. 1999. Ice core records of atmospheric CO2 around the last three glacial terminations. Science 283, 1712 - 1714.

Forsstrom, L. 1999. Future glaciation in Fennoscandia. . Report POSIVA 99-30, 31 pages. ISBN 951-652-085-5.

Freeze, R.A. and Cherry, J.A. 1979. Grounwater. Prentice-Hall, Inc. 604 pages. ISBN 0- 13-365312-9.

Gallee, H., van Ypersele, J.P., Fichefet, Th., Tricot, C, Berger, A. 1992. Simulation of the last glacial cycle by a coupled sectorially averaged climate - ice sheet model. II. Response to insolation and CO2 variation. J. Geophys. Res. 97, 15713 - 15740.

Gascoyne, M. 2000a. A review of published literature on the effects of permafrost on the hydrogeochemistry of bedrock. Report POSIVA 2000-09. 53 pages. ISBN 951-652- 095-2.

Gascoyne, M. 2000b. Dissolved gases in groundwaters at Olkiluoto. Posiva working report 2000-49. 38 pages.

Geological Survey of Canada (2000). http://sts.gsc.emr.ca/permafrost/general.htm

Heginbottom, J.A., Dubreuil, M-A. and Harker, P.A. (compilers) 1995: Canada ~ Permafrost; National Atlas of Canada, 5th edition, Plate 2.1, Ottawa: Geomatics Canada, National Atlas Information Service, and Geological Survey of Canada, scale 1:7,500,000 (MCR 4177). (http://sts.gsc.nrcan.gc.ca/tsdweb/geoserv_permafrost.asp).

Hivon, E.G. and Sego, D.C. 1995. Strength of frozen saline soils. Canadian Geotechnical Journal, 32, 336-354. 41

Imbrie, J. and Imrie, J. 1980. Modeling the climatic response to orbital variations. Science 207, 943 - 953.

IPA 1995. Frozen Ground. The News Bulletin of the International Permafrost Association, number 17, June 1995.

Jouzel, J., Barkov, N.I., Barnola, J.M. Bender, M, Chappellaz, J., Genthon, C, Kotlyakov, V.M., Lipenkov, V., Lorius, C, Petit, J.R., Raynaud, D., Raisbeck, G., Ritz, C, Sowers, T., Stievenard, M., Yiou, F., Yiou, P. 1993. Extending the Vostok ice-core record of palaeoclimate to the penultimate glacial period. Nature 364, 407 - 412.

Khain, V.E. 1985. Geology of the USSR. First part: Old cratons and Paleozoic fold belts. Gebruder Borntraeger, Berlin - Stuttgart. ISBN 3-443-11017-7.

Kukkonen, I. 1986. The effects of past climatic changes on bedrock temperatures and temperature gradients in Finland. Geological Survey of Finland, Report YST-51, 70 pages. (In Finnish with English abstract). ISBN 951-690-341-X.

Kukkonen, I. T.; Safanda, J. 2001. Numerical modelling of permafrost in bedrock in northern Fennoscandia during the Holocene. In: Beltrami, H. & Harris, R. N. (eds.) Inference of climate change from geothermal data : selected papers from the symposium "Heat, stress and climate - the legacy of Arthur H. Lachenbruch", held at the AGU Fall Meeting, San Francisco, CA, December 6-10, 1998. Global and Planetary Change 29 (3-4), 259-273.

Kukla, G., Berger, A., Lotti, R., Brown, J. 1981. Orbital signatures of interglacials. Nature 290, 295 - 300.

Kvenvolden, K.A., 1993. Gas hydrates - geologic perspective and global change. Rev. Geophys. 31, 173-187.

Lahermo, P., Umasti, M., Juntunen, R., Taka, M. 1990. The geochemical atlas of Finland. Part 1. the hydrogeochemical mapping of of Finnish Groundwater. Geological Survey of Finland. 66 pages. ISBN 951-690-356-8.

Lamorey, G.W., Doyle, G.A., Alley, R.B., Grootes, P.M., Mayewski, P.A., White, J.W.C., Barlow, L.K. 1993. The 'flickering switch' of late Pleistocene climate change. Nature 361, 432-436.

Makogon, Y. F., Trebin, F. A., Trofimuk, A. A., Tsarev V. P., Cherskiy, N. V. 1972. Detection of a pool of natural gas in solid (hydrated gas) state. Doklady Academy Science USSR, Earth Science Section 196, 197-200.

McEwen, T. and de Marsily, G. 1991. The potential significance of permafrost to the behaviour of a deep radioactive waste repository. Swedish Nuclear Power Inspectorate, SKI Technical report 91:8, 69 p.

Mohamed, A.M.O., Shoospasha, I., Yong, R.N. 1997. Diffusion of metals in frozen capillary barriers. Engineering Geology 47, 1-5. 42

National Board of Survey 1988. Atlas of Finland, folio 131: Climate. National Board of Survey, Helsinki. 31 pages. ISBN 951-47-0970-5.

Nurmi, P. 1985. Future long-term environmental changes in Finland: their influence on deep-seated groundwater. Nuclear Waste Commission of Finnish Power Companies. Report YJT-85-21. 61 pages. (In Finnish with English summary). ISSN 0359-548X.

Osterkamp, T.E. and Gosink, J.P. 1991. Variations in permafrost thickness in response to changes in paleoclimate. J. Geophysical Res. 96(B3), 4423 - 4434.

Ruotsalainen, P., Ahokas, H., Heikkinen, E., Lindh, J., Nummela, J. 2000. Groundwater salinity at the Olkiluoto site. Posiva, Working report 2000-26.

Shackleton, NJ. 2000. The 100 000-year ice-age cycle identified and found to lag temperature, carbon dioxide, and orbital eccentricity. Science 289, 1987 - 1902.

Sherwood Lollar, B., Frape, S., Fritz, P., Macko, S.A., Welhan, J.A., Blomqvist, R., Lahermo, P.W. 1993. Evidence for bacterially generated hydrogen gas in Canadian Shield and Fennoscandian Shield rocks. Geochim. Cosmochim. Acta 57, 5073 - 5085.

Sloan, E.D. 1990. Clathrate hydrates of natural gases. Marcel Bekker Inc., New York. 641 pages.

Werner, M.R. 1987. Tertiary and Upper Cretaceous heavy oil sands, Kuparuk River area, Alaskan North Slope. In Tailleur, I.L., and Weimer, Paul, eds., Alaskan North Slope geology: Bakersfield, California, Pacific Section, Society of Economic Paleontologists and Mineralogists and the Alaska Geological Society, Book 50, v. 1, 109-118.

Vogt, T. 1991. Cryogenic physico-chemical precipitations: iron, silica, calcium carbonate. Permafrost and Periglacial Processes 1, 283 - 293.

Yershov, E.D., (editor) 1991. Geocryological map of Russia and Neighbouring republics. English language edition 1999: Collaborative Map Project. Carleton University, Ottawa, Canada. ISBN 0-9685013-0-3.

Yershov, E.D. 1998. General Geocryology. University of Cambridge, Studies in Polar Research, 580 pages. ISBN 0-521-47334-9. LIST OF REPORTS 1(1)

POSIVA REPORTS 2001, situation 10/2001

POSIVA 2001-01 Geochemical modelling of groundwater evolution and residence time at the Hastholmen site Petteri Pitkdnen, Ari Luukkanen VTT Communities and Infrastructure Paula Ruotsalainen Fintact Oy Hilkka Leino-Forsman, Vila Vuorinen VTT Chemical Technology January 2001 ISBN 951-652-102-9

POSIVA 2001-02 Modelling gas migration in compacted bentonite: GAMBIT Club Phase 2 Finat report - A report produced for the members of the GAMBIT Club B.T. Swift, A.R. Hoch, W.R. Rodwell AEA Technology, United Kingdom January 2001 ISBN 951-652-103-7

POSIVA 2001-03 Modelling of the UO2 dissolution mechanisms in synthetic groundwater solutions - Dissolution experiments carried out under oxic conditions Esther Cera, Mireia Grive, Jordi Bruno EnvirosQuantiSci, Spain KaijaOllila VTT Chemical Technology February 2001 ISBN 951-652-104-5

POSIVA 2001-04 Experimental study of Ni solubility in sulphidic groundwater and cement water under anoxic conditions Torbjorn Carlsson, Vila Vuorinen, Tommi Kekki, Hannu Aalto VTT Chemical Technology June 2001 ISBN 951-652-105-3

POSIVA 2001-05 Permafrost: occurrence and physicochemical processes Lasse Ahonen Geological Survey of Finland October 2001 ISBN 951-652-106-1 P O S I V A O Y

TfiOLflNKATU 4

FIN-O010O HELSINKI

TEL. (+358-9-)228O 30

FAX ( + 358-9-) 2280 3719

K R N R O 6 4 4.30 3 TRADE REG.

M.Snellman 2.1.2002 1(1)

ERRATA

POSIVA 2001-05 REPORT: Ahonen, Lasse. Permafrost: occurrence and physicochemical processes.

P. 27: In connection with the solubility calculations for calcite there is a misprint in the text. The correct sentence after eq. 4.4 is: At a pH 8.27, T = 0 °C, equilibrium concentrations of Ca2+ and HCCV are thus about 1 mmol/1 of each...

POSIVA OY Research Juhani Vira