Modeling Internal Deformation of Salt Structures Targeted for Radioactive Waste Disposal

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 551

Modeling internal deformation of salt structures targeted for radioactive waste disposal

ZURAB CHEMIA

ACTA UNIVERSITATIS UPSALIENSIS ISSN 1651-6214 UPPSALA ISBN 978-91-554-7281-8 2008 urn:nbn:se:uu:diva-9279 ! " # $%%& %'%% ( ) ( ( *) )+ ,) - . )+

/) 0+ $%%&+ 1 ( ( ( - + + 22 + 3& + 4"56 78&97 922398$& 9&+

,) ) ( ) -)) ( - ) ) ( : - ) + ,) ) ) ) ) ) ;6! < -)) - ( ( ( ))9 - + ,) ( - ( ) ) ) ( ) - ) + = ( ) ) - ) ) ( ) - ) ( ' ( - ) ) ) ( ) + ,) (( ( ) ( -)) ) ) ( ; > ( ) ) ) - ) < ) ) ( + " - 9 - ) ( )) - - ) ) ) ; <+ - -) ) ) > : - ) ) + 4 : ( ) ) ) ) ( ) -)) - ) ) ) + ) 9 ( ) (( ( ( - ) )) ( % &9 % 7 * ) ( ) : ) ( - ) ) + - ) ( ( ) ) : - ) (( ( + ,) ) ) ) )) : ) -) > : ( -) ) : : - ) ) + ? ( ) : ( ( (( - ) ) + @ ) ) ) - ) ) ( ( ) ) : (( (( ( ) + ,) ) ( ) ) : - ) ) ) ( - - ) +

) ( ) :

! "

A 0 /) $%%&

4""6 2 9 $ 3 4"56 78&97 922398$& 9& ' ''' 97$87 ;) '>> +:+> B C ' ''' 97$87< "...Though we are justices and doctors, and churchmen, Master Page, we have some salt of our youth in us..." W. Shakespeare

Dedicated to: salt of the earth...

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Chemia, Z., Koyi, H., and Schmeling, H. (2008) Numerical modelling of rise and fall of a dense layer in salt diapirs. Geophysical Journal International, 172(2):798–816 II Chemia, Z. and Koyi, H. (2008) The control of salt supply on entrainment of an anhydrite layer within a salt diapir. Journal of Structural Geology, 30(9):1192–1200 III Chemia, Z., Schmeling, H., and Koyi, H. (2008) The effect of the salt viscosity on future evolution of the Gorleben salt diapir. Tectonophysics, submitted

Reprints were made with permission from the publishers.

Contents

1 Introduction ...... 11 1.1 Objectives...... 11 1.2 Rock Salt...... 12 1.3 Processes of diapir growth...... 12 1.4 Dense inclusions...... 15 1.5 The evolution history of the Gorleben diapir...... 15 1.6 Long-term safety...... 16 2 Numerical modeling...... 17 2.1 Fundamental principles...... 17 2.1.1 Conservation of mass...... 17 2.1.2 Conservation of momentum...... 18 2.1.3 Conservation of composition...... 19 2.2 Symmetric boundary condition...... 20 3 Rheology ...... 21 3.1 Rocksalt...... 21 3.2 Anhydrite...... 22 3.3 Overburden...... 22 4 Sedimentation technique ...... 23 5 Summary of the Papers ...... 25 5.1 PaperI: Rise and fall of a dense layer...... 25 5.1.1 Modeling concept...... 25 5.1.2 Summary...... 26 5.1.3 Conclusions...... 28 5.2 PaperII: The parameters that inﬂuence salt supply...... 30 5.2.1 Modeling concept...... 30 5.2.2 Overview...... 31 5.2.3 Conclusions...... 31 5.3 Paper III: Evolution of the Gorleben salt diapir...... 34 5.3.1 Modeling concept...... 34 5.3.2 Summary...... 34 5.3.3 Conclusions...... 37 6 Concluding Remarks ...... 39 7 Summary in Swedish ...... 41 Acknowledgments ...... 43 References...... 45

List of Figures

1.1 Global distribution of basins containing salt structures ...... 13 1.2 Modes of diapir piercement ...... 14 1.3 Line drawing of the Gorleben salt diapir ...... 16

4.1 Sediment aggradation ...... 24

5.1 The initial geometric condition ...... 26 5.2 Regions of the diapir piercement ...... 27 5.3 Snapshots of models ...... 29 5.4 Illustrated areas of the diapir and entrained anhydrite ...... 31 5.5 Shapes of passive diapirs ...... 32 5.6 Salt supply and entrainment of the anhydrite layer ...... 33 5.7 The initial geometry of the model for Gorleben diapir...... 35 5.8 Evolution of the models for different salt rheologies ...... 36 5.9 Internal displacement ﬁeld of the diapir (PF-Model) ...... 37

1. Introduction

1.1 Objectives For the last 40 years, scientists and engineers have been searching for geologically suitable repositions for radioactive waste. Salt layers and structures have been targets as repositories for hazardous waste [e.g. Gorleben and Morsleben salt diapirs in Germany; WIPP site in USA and Anloo, Gasselte (Drenthe) and Winschoten (Groningen) in Netherlands]. Currently only one repository in salt is in use, the Waste Isolation Pilot Plant (WIPP) where long-lived radioactive waste is buried in deep salt beds (40 kilometres east Carlsbad, New Mexico). However, many other salt structures are targets for waste storage (e.g. Gor- leben diapir in Germany). Tectonic stability of a salt structure is a significant factor in evaluating its suitability as a repository (e.g. hydrocarbon storage, waste disposal, etc.). To evaluate safety of a repository many different scenarios have been considered such as: suberosion and diapirism, brine migration, thermo-mechanical fractures, flooding of a disposal mine, large brine inclusions, diapirism to the biosphere, solution mining etc (e.g. onshore disposal committee, 1989). Many of these scenarios have been applied to Gorleben salt diapir, which is currently used as an intermediate storage facility and was targeted as a future final repository for high-grade radioactive waste. However, the influence of dense inclusions, anhydrite layer, which is present within the Gorleben salt diapir, was not considered as a possible disturbance of the repository. Based on analogue and numerical models Koyi[2001] suggested that Gorleben diapir might be active internally due to presence of a dense anhydrite layer (blocks) within it. According to Koyi’s [2001] models, the denser blocks entrained by the diapir, sink within the externally inactive Gorleben diapir. Indications for movement of the anhydrite blocks comes from acoustic emission measurement that have recorded displacement on the boundary between rock salt and the anhydrite blocks [Spies and Eisenblätter, 2001]. Understanding the structural evolution of an initially tabular salt layer with intercalated anhydrite layer of high density and high viscosity and investigation of the parameters that influence its development are required for evalua- tion of salt structures as a "safe repository". Salt tectonics is therefore critical for radioactive waste storage in rock salt in many parts of the globe. First, a short introduction to the modern understanding of the salt tectonics is presented below, followed by geology of the Gorleben diapir, which has been used as a general guideline in this work. Second, the applied modeling

11 procedures, together with the theoretical background for the modeling are explained. Finally a brief summary of each paper including conclusions is given.

1.2 Rock Salt Halite, a colorless or white mineral sometimes tinted by impurities that is commonly known as "rock salt" is found in beds as evaporites. The term "rock salt" is used to include all rock bodies composed primarily of halite [Jackson, 1997a,b]. Salt naturally occurs in immense deposits, and occasionally in surface deposits in arid areas as the mineral halite. Salt deposits are widespread. There are major salt basins in the different parts of the world (Fig. 1.1). A number of these salt deposits are mined for halite. Salt has mechanical properties different from those of most clastic and carbonate rocks. Under unusually high strain rates, salt fractures like most other rocks. However, under typical geologic strain rates, salt flows like a fluid in the subsurface and at surface [Weijermars et al., 1993]. Salt is also relatively incompressible so it is less dense than most carbonates and all moderately to fully compacted siliciclastic rocks. Salt rheology and incompressibility make it inherently unstable under a wide range of geologic conditions. Diapirs are a major type of salt structures resulting from tectonic deformation. Bedded salt deposits are nearly horizontal, although some contain fault zones and other anomalies. Salt beds range in thickness from a few tens of meters to several thousands of meters. The differential compaction of the sediments that cover salt may produce instabilities and flows within the salt layer that may result in salt domes or diapirs. Furthermore, salt is also an effective conductor of heat, elevating the ther- mal maturity of rocks above salt structures and cooling rocks that lie below or adjacent to salt bodies. The modern investigation of the thermo-mechanical behavior of salt started in the mid-1930’s and has acquired considerable technical depth and sophistication. This has been largely due to anticipated use of domal or bedded salt deposits as ideal for storage of petroleum products, sites for special radioactive wastes repositories, and to expanded need for hydrocarbon storage caverns. In addition, interest in salt tectonics comes from the oil industry because many of the world’s great hydrocarbon provinces lie in salt basins (e.g., Gulf of Mexico, Persian Gulf, North Sea, Lower Congo Basin, Campos Basin, and Pricaspian Basin). Moreover, evaporites provide economic resources of potash salts, sodium salts, gypsum, sulfur, borates, ni- trates, and zeolites [Warren, 1999].

1.3 Processes of diapir growth Modern interpretations of salt tectonics consider differential loading as the dominant force driving salt ﬂow. For most of the past 70 years, the prevailing view of salt tectonics was that its mechanics were dominated by salt buoy-

12 Figure 1.1: Equal-area Mollweide projections showing global distribution of basins containing salt structures (black areas). Basins containing only undeformed salt are omitted. (a) Global distribution of basins. (b) Detail of European basins. Symbols: AD South Adriatic; ZS Zechstein; PY Pripyat; DD Dnepr-Donetz; PC Pricaspian; EA East Alpine; TV Transylvanian; CR Carpathian; MN Moesian; CK Cankiri; HP Haymana- Polatli; JU Jura; RH Rhodanian; LG Ligurian; BL Balearic; BE Betic; AL Atlas; LV Levantine; MS Messinian; AG Agadir; SF Saﬁ; BR Berrechid; ER Essaouira; LU Lusitanian; GQ Guadalquivir; MA Maestrat; EB Ebro; CT Cantabrian-West Pyrenees; AQ Aquitaine; TY Tyrrhenian; SC Sicilian; IO Ionian; SZ Suez; DS Dead Sea; PA Palmyra; CL Cicilia-Latakia; AM Amadeus [Hudec and Jackson, 2007].

13 ancy [Jackson, 1995, 1997a,b]. The overburden layers were envisioned as a dense fluid having negligible yield strength and diapirism was explained by Rayleigh-Taylor instabilities. However, the concept of a fluid overburden fell out of support after workers began to recognize the importance of roof strength as a control on diapir growth. Although, one of the principal resisting forces opposing slat flow, the strength of the overburden, or its critical thickness was reported in early works [e.g., Biot, 1965] it became favor in the late 1980s.

(a) Reactive piercement (b) Active piercement

(c) Erosional piercement (d) Ductile piercement Figure 1.2: Modes of diapir piercement, shown in schematic cross sections. The overburden is brittle except in (d) [Hudec and Jackson, 2007].

Diapirs are no longer pictured like giant lava lamps rising through a yielding overburden. Instead, the position and shape of viscous salt bodies are seen to depend on how the brittle, mechanically competent overburden deforms. Salt structures are defined as tectonically active when they rise due to differential loading created by the overburden layers and/or by regional tectonics [Jack- son and Vendeville, 1994; Koyi, 1998; Hudec and Jackson, 2007]. A diapir may become inactive (and stops rising) when the driving forces (differential loading, regional tectonics) cease, salt layer is depleted or a strong lid covers the diapir. However, buried salt can remain static in the subsurface for tens of millions of years. Buried salt may rise as a diapir if: (i) the regional extension weaken and thin the overburden, which makes room for a reactive diapir to rise (Fig. 1.2a), (ii) The flaps of thin overburden lift, rotate, and shoulder aside, as the diapir forcibly breaks through them by active diapirism (Fig. 1.2b), (iii) Erosion may remove the roof of the diapir (Fig. 1.2c). All these processes may occur at various times during the growth of a single salt structure. Di- apirs having fluid overburdens may rise by ductile thinning of the diapir roof [Nettleton, 1934; Talbot et al., 1991, Fig. 1.2d].

14 1.4 Dense inclusions Inclusions of denser rocks are common in salt diapirs. The size and lithology of such entrained inclusions varies from place to place. Sedimentary, volcanic and even some plutonic inclusions (several kilometers in diameter) character- ize many of the diapirs in the Zagros fold-thrust belt [Kent, 1979]. Many of the inclusions are related to the original salt deposits or are older [Gansser, 1992]. On Hormuz Island, inclusions of white rhyolites and trachytes frequently occur as irregular masses up to one kilometer long [Gansser, 1992]. Many salt diapirs contain varying amounts of other evaporite rocks, especially anhydrite or its hydrated form, gypsum, and/or non-evaporite rocks. Most non-halite inclusions in salt are originally interbedded with salt (e.g. Zechstein formation). During the diapiric processes, these inclusions can be transported by salt ﬂow. The fact that the salt can carry denser blocks close to the surface was an enigma until late 1990s. Whereas the external dynamics of the diapir was intensively studied, few researchers have focused on internal dynamics of the salt induced by intercalated layers. The ability of salt diapirs to lift large inclusions of dense rocks generally depends on rising velocity of salt and negative buoyancy of the inclusions. Denser inclusions can be lifted if salt in the diapir rises faster than the inclusions sink. Sinking velocity of the inclusions is controlled by the effective viscosity of the salt, which is a function of the rheological properties, strain rate and temperature of the salt [Weinberg, 1993]. An example of a diapir which has entrained dense blocks is the Gorleben diapir in Germany [Bornemann, 1991]. Geophysical and subsurface data conﬁrm that large blocks of denser anhydrite are present at relatively shallow depth within the Gorleben salt diapir [Richter-Bernburg, 1980]. These anhydrite blocks, deformed within the diapir, form a key horizon in the Zechstein salt sequence [Bornemann, 1982; Zirngast, 1991].

1.5 The evolution history of the Gorleben diapir The Gorleben salt diapir, which roughly trends NE-SW, is about 14km long and up to 4km wide. It contains different cycles of the Zechstein salt formation (z2, z3 and z4, Fig. 1.3). The base of the Zechstein lies at about 3.1 to 3.3km below the surface. The stratigraphy and the evolution history of the Gorleben diapir is well explored by four exploratory wells and two shafts that are drilled into the Gorleben diapir, in addition to more than forty shallow wells that explored overlying beds and the salt dissolution zone at the base of the caprock [Bornemann, 1982, 1991]. It is suggested that the salt structures from the Gorleben diapir evolved above an elevation of the basement and the total primary thickness of the Zechstein salt in that area is estimated to be about 1.1 to 1.4km [Zirngast, 1996]. Halokinesis was initiated in the Early Triassic to Middle Triassic, and a thick salt pillow formed during Keuper (Late Triassic) and Jurassic time

15 NW SE q

t t

o kr kr0

u kr kru sm + j0 - Wd su k so + m s u + sm

1000m Anhydrite Z3 salt Borehole Z2 salt Z4 salt Level of plant survey

Figure 1.3: Line drawing of a northwest-southeast proﬁle of the Gorleben salt diapir in northwest Germany showing entrained segments of dense anhydrite within a diapir. Symbols: q-Quaternary, t-Tertiary, kro-Upper Cretaceous, kru-Lower Creta- ceous, so+m-Upper Bunter, su+sm-Buntsandstein. Modiﬁed after Bornemann[1991].

[Zirngast, 1991; Jaritz, 1993]. The diapiric stage was reached in Late Jurassic and Early Cretaceous and the diapir continued into the Tertiary. Maximum rise rates are calculated as 0.08 mm a−1 for the Late Cretaceous and less than 0.02 mm a−1 in Miocene to recent times [Zirngast, 1996]. The internal structure of the Gorleben diapir is very complex, including vertical or steeply inclined and overturned fold axes. A very remarkable feature is the main anhydrite, which forms a key horizon and is folded within the diapir. The main anhydrite is an important layer with respect to nuclear waste storage because of its size and density. It is about 80m thick and denser (ca. 3000 kg m−3) than the salt (ca. 2200 kg m−3).

1.6 Long-term safety Final disposal of radioactive waste is planned in deep geological formations (Fig. 1.3). Radioactivity form the waste decreases in time due to the radioactive decay. Nevertheless, in case of long-living nuclides the radiation after 100 000 years will still requires the waste to be isolated from the biosphere. Therefore, long-term periods up to one million years and even more have to be considered. Long-term safety analyses are performed to determine the ra- diological effects of the waste on the biosphere for the next one million years [Federal Institute for Geosciences and Natural Resources (BGR)].

16 2. Numerical modeling

All the mathematical sciences are founded on relations between physical laws and laws of numbers, so that the aim of exact science is to reduce the problems of nature to the determination of quantities by operations with numbers. James Clerk Maxwell

2.1 Fundamental principles All of Computational Fluid Dynamics, in one form or another, is based on the fundamental governing equations of fluid dynamics, continuity, momentum, and energy equations. They are the mathematical statements of three fundamental physical principles upon which all of fluid dynamics is based. We model the dynamics of salt tectonics driven by differential loading and compositionally induced density variations using a two dimensional finite difference code (FDCON modified after H. Schmeling). This code solves equations for an initially horizontally layered multi-composition system. Differ- ent chemical compositions (i.e. layers) may be assigned different densities and rheologies. The dynamics of such flow can be described by the equations of conservation of mass, momentum and composition [e.g. Weinberg and Schmeling, 1992].

2.1.1 Conservation of mass The law of conservation of mass states that the mass of a closed system of substances will remain constant, regardless of the processes acting inside the system [Antoine Lavoisier, 1785]. According to conservation of mass, reac- tions and interactions, which change the properties of substances, leave un- changed their total mass. The time rate of change of density of the given ﬂuid element as it moves through space is given as: ∂ρ + ∇(ρu) = 0 (2.1) ∂t where ρ is density, t is time, and u is ﬂuid velocity.

17 Using the Boussinesq approximation density differences are sufﬁciently small to be neglected, except where they appear in terms multiplied by the acceleration due to gravity. Thus for a Boussinesq ﬂuid (ρ = constant) equation 2.1 can be written as ∇u = 0 (2.2)

2.1.2 Conservation of momentum The conservation of momentum is a fundamental concept of physics along with the conservation of mass. The conservation of momentum states that, within some problem domain, the amount of momentum remains constant; momentum is neither created nor destroyed, but only changed through the action of forces as described by Newton’s laws of motion [Landau and Lifsic, 1987]. Momentum is conserved in all three physical directions at the same time. Consequently the force balance for an incompressible fluid undergoing flow in two dimensions can be defined by adding the volume, surface and inertial forces together and equating their sum to zero. We neglect the inertial force associated with the acceleration of a fluid. This assumption is appropriate for the slow motion of very viscous or high Prandtl number fluids. Salt behaves as a highly viscous fluid on geological time scale [Weijermars et al., 1993]. The viscosity of salt is about 1017 − 1019 Pa s [Hunsche and Hampel, 1999; Spiers et al., 1990; Urai et al., 1986]. With assumption of incompressibility the equation of motion is given by

∂τ ji 0 = −∇P + − ρg (2.3) ∂x j where τ ji is the deviatoric stress tensor and can be expressed in terms of ve- locities using the constitutive law

τi j = 2ηkε˙i j (2.4) where ηk is the viscosity of the k-th component and ε˙i j is the strain-rate tensor deﬁned as ∂εi j 1 ∂ui ∂u j ε˙i j ≡ = + (2.5) ∂t 2 ∂x j ∂xi

To specify the density in the buoyancy term of the momentum equation we take the mean of the densities of the different materials weighted by their concentrations ρ = ∑ckρk (2.6) k

th where ck and ρk is the concentration and density of the k chemical component.

18 Introducing the stream-function ψ as ∂ψ ∂x u = − ∂z ∂ψ and using constitutive law with assumption of two dimensionality for the equation of motion we obtain 2 2 2 2 2 2 ∂ η ∂ ψ ∂ η ∂ η ∂ ψ ∂ ψ ∂ck 4 + 2 − 2 2 − 2 = −g∑ρk (2.7) ∂x∂z ∂x∂z ∂z ∂x ∂z ∂x k ∂x where x and z are horizontal and vertical coordinates respectively. Given the chemical concentrations ck(x,z,t), the biharmonic equation (2.7) is solved numerically. In equation (2.7), η is the effective viscosity defined at boundaries between the different materials, where mixture can occur. In mixtures of different compositions, an effective viscosity may be defined by assuming that the total strain rate is given as the weighted sum of the strain rates in each component, or assuming that the total stress is given as the weighted sum of the stresses in each component. Accordingly, the effective viscosity would be given as the harmonic or the arithmetic mean of the viscosities of the components. As it is not possible to specify which of these means is to be preferred in a complex flow field [see Schmeling et al., 2008, for a comparison of different averaging schemes], the effective viscosity of a mixture is defined here as the geometric mean as: logη = ∑ck logηk (2.8) k

2.1.3 Conservation of composition The conservation of composition is conveniently deﬁned as ∂c k + u∇c = 0 (2.9) ∂t k where ck is the concentration of the k-th chemical component. At any location, the concentrations of the different components add up to 1, i.e.

nmat ∑ ck = 1 (2.10) k=1 where nmat is the number of components of the problem. This definition allows consideration of the chemical mixtures of different materials as well as immissible fluids. If the fluid is immissible, ck is equal to 1 only at the positions which are occupied by the material k, otherwise it is 0. The conservation of composition is solved using a marker approach. The advantage of the marker approach is that it introduces no numerical diffusion. Following any fluid particle on its flow path, equation (2.9) may be written in

19 the Lagrangian reference system ∂c k = 0 (2.11) ∂t particle

The equation (2.11) implies that we have to solve for the flow path equation ∂x = u (2.12) ∂t particle and have to ensure that the ckvalues remain constant at x(t)particle. To solve (2.12), model domain is filled with closely spaced marker points initially distributed on a rectangular grid with small random displacement. These fluctuations, which have maximum amplitude of a half marker grid distance, are also responsible for the initial perturbations of the horizontal in- terfaces between materials of different density. Each marker is assigned its particular chemical composition, i.e. ck(xm), where xm(t) is the position of a marker at the time t. Equation (2.12) is then solved for all markers by a com- bination of fourth-order Runge-Kutta integration with a predictor-corrector step.

2.2 Symmetric boundary condition In simulations, we assume that symmetric state exists on the boundary. This treatment of the boundary condition corresponds to the physical assumption that, on the two sides of boundary, the same physical processes exist. The variable values at the same distance from the boundary at the two sides are the same. The function of such a boundary is that of a mirror that reﬂects all the ﬂuctuations generated by the simulation region. Instead of simulating the whole structure, we set the appropriate boundary conditions and reduce the problem size.

20 3. Rheology

In geology and geophysics, the term rheology refers to the study of mechanical properties and their role in the deformation and the ﬂow of the materials that form the Earth. We can describe the Earth’s rheology mathematically in many ways, but it is always a function of intrinsic, i.e., material parameters, and extrinsic, i.e., environmental parameters [Park, 1989]. The rheology of each material used in numerical calculations is described bellow.

3.1 Rocksalt Many studies have focused on experimental deformations of the steady state ﬂow properties and processes of natural rocksalt [e.g., Hunsche and Hampel, 1999; Carter et al., 1993; Urai et al., 1986]. The sets of the steady-state data ﬁt well a power law creep relation: ε˙ = A∗ · σ n · exp(−H∗/RT) (3.1) where ε˙ is the steady-strain creep rate, σ = (σ1 − σ3) is the stress difference of the maximum and minimum principal stresses, T is absolute temperature, A∗ is a material constant, n power law exponent, and H∗ = E∗ + PV is the activation enthalpy. E∗ is the activation energy, V is the activation volume, R is the universal gas constant (8.32 J mol−1), and P is the (lithostatic) pressure. Numerical models presented in this study are based on an approach where power-law ductile creep describes deformation behavior of the rock. The viscosity depends non-linearly on deviatoric stress, which is given as: H∗ η = A · exp · τ1−n (3.2) RT II

−1 n+1 ∗ where A = 3 2 · A is a pre-exponential constant, and τII is the 2nd in- variant of the stress tensor. Although salt compositions vary from one formation to another, the rheological behavior is qualitatively similar. The domal salt at Gorleben, for example, has significantly low water content (ca 0.1–0.2%) and creeps slowly [Fairhurst, 2002]. Yet, flow laws essentially similar to that in equation (3.2) can be used to study internal dynamics of the Gorleben diapir. In some of the simulations, we can assume that the rocksalt behaves as Newtonian fluid [e.g. wet salt, Weijermars et al., 1993]. The viscosity of wet

21 salt is virtually independent of stress or strain rate. Wet salt therefore has no yield strength. Data from ﬁeld observations and rock mechanics both suggest that natural faulting in wet rock salt is rare. Thus, disregarding temperature de- pendency of the viscosity, the deformation behaviour of rock can be described by equation (3.2) when H∗ = 0 and n = 1.

3.2 Anhydrite The deformation experiments on anhydrite rocks showed that the strength of the ﬁne-grained anhydrite is strongly temperature and strain-rate dependent above 300◦C whereas coarse grained anhydrite rock remains relatively insen- sitive to strain-rate and temperature even at 450◦C. Although anhydrite has a high strength at room temperature, ﬁne grained aggregates weaken rapidly above temperatures of 100◦C to 200◦C, depending on grain size and strain rate [Muller et al., 1981]. We model anhydrite with a power-law rheology independent of temperature and a low value for the exponent, i.e. n = 2 [Muller et al., 1981]. 1−n η = A · τII (3.3)

−1 n+1 ∗ ∗ − −n − where A = 3 2 · A and A = 3.08 × 10 27Pa s 1. Since the modeling approach is based on the mechanics of continua rather than fracture mechanics, the brittle rheology of anhydrite cannot be achieved in numerical models. We model anhydrite with an effective viscosity ranging between 1019 − 1021 Pa s.

3.3 Overburden The sedimentary overburden rocks simulated in all the models are assigned high viscosity (effective viscosity ranges between 1023 and 1025 Pa s), because most sedimentary rocks (clastic or carbonate) behave as non-Newtonian materials or as brittle solids, rather than ductile ﬂuids, depending on factors such as their depth of burial, etc. [Vendeville and Jackson, 1992]. Our assumption of a very high viscosity keeps ductile deformation in the overburden extremely small [Rönnlund, 1989; Hughes and Davison, 1993]. We also use high viscosity overburden to simulate postdiapirc state of the Gorleben diapir. In analogue models often the postdiapiric state is simulated by a critical thickness of the overburden [Biot, 1965; Koyi et al., 2001]. Alter- natively, sufﬁciently stiff overburden can be used to simulate inactive diapirs [Biot, 1965]. Since the thickness of the overburden for Gorleben diapir is well known and it is reported that the diapir is externally inactive we used stiff overburden approach. However, lower viscosity contrast between the overburden and the salt, which causes viscous drag along their boundaries with the salt is considered as well.

22 4. Sedimentation technique

Albert Einstein is reputed to have warned his son Hans Albert against becoming a river engineer, saying that the physics of sediment transport was too complex . . . Philip A. Allen, 1997

Sedimentation is a process of deposition of a solid material from a state of suspension or solution in a ﬂuid (usually air or water). It is one of the important parameters that drive salt tectonics. The importance of sedimentation rate and its impact on the diapir evolution has been studied previously and it is shown that sedimentation rate can control the rate of rise of a diapir and its geometry [Vendeville et al., 1993; Vendeville and Jackson, 1992; Koyi, 1998]. We introduce here a technique of sedimentation which makes it possible to gradually down-build salt diapir and investigate the effect of sedimentation on the evolution of the diapir. Two different modes of sediment aggradation (constant and variable) are used in the simulations. Sediments accumulate on the top of the model with a prescribed rate,s ˙e. The initial position of the sediment surface is given by the top of the pre-kinematic overburden, zs0 (Fig. 4.1a). Above the initial model surface, background markers are assumed with a low viscosity and a low density. During sedimentation, the position of the surface of sediment layer, zs, rises with the prescribed rate, zs = zs0 + s˙e, thereby thickening the overburden (Fig. 4.1b). If the salt surface lies below zs, then the salt is also covered with sediments (Fig. 4.1c). During the simulation, sediments are deposited at each time step between the previous model surface at time, t − dt, zs(t − dt), and the new sedimentation surface, zs(t) [but only where zs(t − dt) is below zs(t)]. This is done by assuming all background markers, whose positions are between zs(t − dt) and zs(t) with sedimentary rock properties. In the variable sedimentation mode, the sedimentation rate varies according to the evolution of the crest of the diapir. During the simulation, the velocity at which the crest of the diapir rises (vc) is determined. A quarter of this velocity is assigned to the sedimentation rate, so the surface of the sediment layer grows depending on the evolution of the crest of the diapir. Z t 1 zs = zs0 + vcdt (4.1) 0 4

23 The choice of the factor 1/4 in the variable mode of the sedimentation is manually calibrated to form a columnar diapir. The quarter of the rising velocity of the crest of the diapir seems to be a good approximation for the formation of the diapir whose crest remains uncovered and possess a columnar shape.

Low viscosity backgraund material

zs0

pre-kinematic overburden layers

Initial Salt Layer

(a) Initial conﬁguration

Diapiric rise > Sedimentation rate zs

Aggraded sediments

zs0 pre-kinematic overburden layers

InitialSalt Salt Layer Layer

(b) After sediment deposition

Diapiric rise < Sedimentation rate

zs0 Aggraded sediments pre-kinematic overburden layers

InitialSalt Salt Layer Layer

24 5. Summary of the Papers

This chapter presents a brief summary of the papers included in this thesis. Each paper is summarized, describing the main objectives, the system of methods used, the main results and conclusions.

5.1 PaperI: Rise and fall of a dense layer PaperI addresses the issue of presence of dense inclusions within salt diapirs. In this paper, we used the results of 2-D numerical calculations to quantita- tively analyze the parameters (sedimentation rate, salt viscosity, width of the perturbation, and stratigraphic position of dense layers within a salt layer) that govern the rise and fall of dense blocks (anhydrite) within salt diapirs. Models presented in this study are not scaled to any particular salt diapir. Neverthe- less, the Gorleben salt diapir (with an intercalated dense anhydrite layer) has been used as a general guideline for the modeling.

5.1.1 Modeling concept Many investigators numerically modeled different aspects of salt diapirism [e.g., Woidt, 1978; Schmeling, 1987; Romer and Neugebauer, 1991]. How- ever, few studies have addressed the problem of down-building and multi- compositional salt structures. A two-dimensional box (4 × 4km) with a grid resolution 161 × 161, representing the model domain, consisted of a salt layer (1040m thick) simulating the Zechstein salt formation, an anhydrite layer (80m thick) embedded within the salt layer, a pre-kinematic overburden layer (160m thick), and an arbitrary material simulating water or air (Fig. 5.1). The dense anhydrite −3 layer (ρa = 2900 kg m ) is initially embedded as a horizontal layer within a viscous salt. The anhydrite layer is assigned non-Newtonian rheology [Kirby and Kronenberg, 1987]. During the experiments effective viscosity of the anhydrite ranges between 1019 and 1021Pa s. The overburden, representing a pre-kinematic layer, is placed on the salt layer apart from the perturbation to initiate salt ﬂow (Fig. 5.1). A pre-kinematic perturbation (400 or 800m wide) is initiated at the left-hand side of the upper boundary of the salt layer as a trigger for diapirism. The perturbation is down-built by aggradation of non- Newtonian sediments (n = 4, constant temperature) at a rate that is varied systematically.

25 Several parameters (sedimentation rate, salt viscosity, perturbation width and stratigraphic position of the anhydrite layer) are varied systematically to understand their inﬂuence on entrainment of the anhydrite layer whereas oth- ers are kept constant.

t=0.00 Ma 0 Low viscous background material −0.5 A B −1 C D −1.5

−2

−2.5Depth (km) Pre-kinematic overburden layer −3 Anhydite layer −3.5 Salt Layer −4 0 0.5 1 1.5 2 2.5 3 3.5 4 Length (km) Figure 5.1: Line drawing showing the initial geometric condition and the points (A, B, C and D), which are monitored throughout model evolution. The box has no-slip at the bottom, free-slip at the top and reﬂective side boundaries. The box can be considered as showing the right hand half of a symmetric structure.

5.1.2 Summary The evolution of the modeled diapirs depends on the parameters that vary during the simulations. The rate of rise of the diapir increases as the overburden aggrades and increases the driving pressure (differential pressure) on the un- derlying salt layer. An increase in differential loading accelerates salt ﬂow and yields to entrainment of the anhydrite layer. However, high sedimentation rate during the early stages of the diapir evolution bury the initial perturbation and prevents formation of a diapir. As a result, the anhydrite layer sinks within the buried salt layer. Sedimentation rate and viscosity of salt are the main parameters controlling the burial of the diapir. Diapirs of low-viscosity salt (1016 Pa s) survive for high sedimentation rates (e.g. 15 mm a−1) whereas increasing the salt viscosity requires a decrease of the sedimentation rate for the diapir to rise without being buried (Fig. 5.2). Increase of salt viscosity by an order of magnitude and decrease of sedimentation rate by the same order often result in diapirs of similar geometry.

26 Diapirs down-built by slow sedimentation rate are characterized by a slow vertical growth, because the salt extrudes and spreads laterally faster than it ascends vertically. In regions of slow vertical salt movement (e.g. within the overhang), entrained blocks may sink. In contrast, where vertical ﬂow dominates, the anhydrite blocks are entrained and dragged up to higher levels. When the overhang widens faster than the overburden thickens, the pre- kinematic overburden is bent and sinks. The entrained anhydrite segments are carried sideways into the wide overhang. Additional sediments cover the crest of the diapir forming mini-basins that segment the diapir in its later stages. The mini-basins push the anhydrite blocks down as they sink into and segment the diapir. Numerical models with salt viscosity 1017 Pa s predict that diapirs can po- tentially grow with an embedded anhydrite layer, and deplete their source layer in a very short time ranging from three to as little as 0.7 Myr. That the natural diapirs typically grow more slowly and for much longer time than the above predictions, clearly suggesting that salt does not ﬂow freely and con- tinuously throughout the growth history of the diapir. However, as it is shown here, many parameters such as sedimentation rate, salt viscosity, perturbation width and the stratigraphic location of the anhydrite layer can retard or speed up the evolution of the diapir.

1 10 Zone of no diapirism

0 10 Seddimentation rate (mm/a)

Zone of diapirism

−1 10 17 18 19 10 10 10 Viscosity (Pa s) Figure 5.2: Logarithmic plot showing regions where the diapir is piercing for sedimentation range 0.1 − 5 mm a−1 and viscosity range 1016 − 1019 Pa s. Circles show the models where piercement takes place and the squares models where the diapir is buried. Symbols are data points and the contour deﬁnes region of piercement

27 5.1.3 Conclusions Model results show that the combined effect of four parameters (sedimentation rate, salt viscosity, perturbation width, and stratigraphic location of a dense layer) shape a diapir and the mode of entrainment rather than the net value of each parameter individually.

I Sedimentation rate: Fast sedimentation suppresses the rate of diapiric rise. Dense embedded layers are likely to sink within the salt layer left beneath the thick overburden before forming a diapir. However, sedimentation rate slower than the rate of diapiric rise, can down-build a diapir that entrains and deforms the anhydrite layer (Fig. 5.3).

II Salt viscosity: For the same sedimentation rate, increasing viscosity of the salt layer decreases the rise rate of the diapir and reduces the amount (volume) of the anhydrite layer transported into the diapir. An anhydrite layer sinks easily into less viscous salt, whereas salt that is more viscous can prevent sinking of the anhydrite layer for longer time. Fur- thermore, a more viscous salt provides a larger viscous drag, and hence, assists in entrainment and carrying upward of the anhydrite layer. Model results show that viscous salt (> 1018Pa s) is capable of carrying sepa- rate blocks of the anhydrite layer to relatively higher stratigraphic levels (Fig. 5.3).

III Perturbation width: Wide perturbation allows signiﬁcantly larger vol- umes of salt supply to the stem of the diapir and can more easily entrain an embedded anhydrite layer than narrower diapirs. (Fig. 5.3).

IV Stratigraphic location: The presence and stratigraphic location of an anhydrite layer signiﬁcantly alters the time evolution of the diapir (de- celerates the diapiric growth rate by 17 percent). In addition, an anhydrite layer embedded in the middle of a salt layer is entrained more effectively than a layer embedded within the upper or lower half of the salt layer (Fig. 5.3).

V The anhydrite layer is entrained as long as rise rate of the diapir exceeds the descent rate of the denser anhydrite layer. However, entrained blocks inevitably sink back if the rise rate of the diapir is less than the rate of descent of the anhydrite layer or the diapir is permanently covered by a stiff overburden in case of high sedimentation rates.

VI Our model results support earlier studies that diapirs containing denser blocks can be active internally even though they might be considered inactive externally. The Gorleben diapir in Germany, which has been considered as a repository for radioactive waste maybe such an example.

28 sedimentation rate

t = 7.8 Ma t = 1.26 Ma t = 1.26 Ma 0 0 0 1 mm/a −0.5 4 mm/a −0.5 2 mm/a −0.5

−1 −1 −1

−1.5 −1.5 −1.5

−2 −2 −2

−2.5 −2.5 −2.5

Depth (km) −3 −3 −3

−3.5 −3.5 −3.5

−4 −4 −4 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4

salt viscosity t = 10.1 Ma t = 10.1 Ma t = 10.1 Ma 0 0 0 −0.5 1.e17 Pa s −0.5 1.e18 Pa s −0.5 5.e18 Pa s −1 −1 −1

−1.5 −1.5 −1.5

−2 −2 −2

−2.5 −2.5 −2.5

−3 −3 −3 Depth (km) −3.5 −3.5 −3.5

−4 −4 −4 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4

perturbation width t = 1.26 Ma t = 1.26 Ma t = 1.27 Ma 0 0 0 600 m 800 m −0.5 400 m −0.5 −0.5

−1 −1 −1

−1.5 −1.5 −1.5

−2 −2 −2

−2.5 −2.5 −2.5 Depth (km) −3 −3 −3

−3.5 −3.5 −3.5

−4 −4 −4 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4 stratigraphic location t = 6.07 Ma t = 6.07 Ma t = 6.07 Ma 0 0 0

−0.5 800 m −0.5 480 m −0.5 200 m

−1 −1 −1

−1.5 −1.5 −1.5

−2 −2 −2

−2.5 −2.5 −2.5 Depth (km) −3 −3 −3

−3.5 −3.5 −3.5

−4 −4 −4 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4 Length (km) Length (km) Length (km)

Figure 5.3: Snapshots of models where the sedimentation rate, salt viscosity, perturbation width and the stratigraphic location of the anhydrite layer is varied.

29 5.2 PaperII: The parameters that influence salt supply Analogue and numerical models confirm that dense block can be entrained by diapiric flow if the rate of diapiric rise is greater than the rate of decent of the dense entrained blocks [Weinberg, 1993; Koyi, 2001; Chemia et al., 2008]. The influence of four parameters (sedimentation rate, viscosity of salt, stratigraphic location of the anhydrite layer within the salt layer, and the perturbation width) described in previous section strongly influence the style and the amount of entrainment of dense inclusions into a diapir [PaperI]. To investigate the entrainment of a denser layer, which depends on multiple parameters, we combine these parameters into a single parameter, which is the rate of salt supply (volume/area of the salt that is supplied to the diapir with time, i.e. cumulative flow of salt from the salt layer to the diapir). In this study, we investigate in detail the influence of these four parameters on rate and the amount of salt supply during the development of a diapir down-built from a salt layer with an initially intercalated dense anhydrite layer. Using two-dimensional numerical calculations, a whole range of values for these four parameters are explored.

5.2.1 Modeling concept The results presented in this paper are based on models, which are described in PaperI. In this article, we extend the range of the parameters and recast results in terms of salt supply. At different stages of the diapir growth, the area of the diapir and the area of the entrained anhydrite layer are calculated. The contour comprising the chemical composition of the salt layer defines the area of the diapir. The contour closes at the level of the perturbation, which might sink or rise during the model evolution and the area is calculated above this dynamic level (Fig. 5.4). Similarly, we define that the anhydrite is entrained if it is carried upward into the diapir above the perturbation level. The closed contour comprising the chemical composition of the anhydrite layer above the perturbation level defines the amount of the entrained anhydrite layer (Fig. 5.4). If the anhydrite layer is disrupted then the cumulative area of the entrained anhydrite is considered sum of the areas of the closed contours above the perturbation level (Fig. 5.4). In two dimensional numerical experiments, the parameter cumulative percentage entrainment (E) is conveniently defined as E = (Aat /Aa0 ) × 100, where Aat is the area of anhydrite layer which has moved above a reference line (set at the perturbation level) at any given time t, and Aa0 is the original area of the anhydrite layer at t = 0 (Fig. 5.4). Similarly, cumulative percentage salt supply S is defined as S = (Ast /As0 ) × 100, where Ast is the area of salt layer which has moved above a reference line (area of the diapir,

Fig. 5.4) at any given time t, and As0 is the original area of the salt layer beneath the surface at t = 0, (Fig. 5.4).

30 0 Area of the diapir -0.5 Area of the entrained anhydrite layer

-1

-1.5

-2 Depth (km) -2.5

-3

-3.5

-4 0 0.5 1 1.5 2 2.5 3 3.5 4 Length (km) Figure 5.4: Illustrated areas of the diapir and the entrained anhydrite layer used for calculated salt supply and entrainment

5.2.2 Overview The geometry of an active diapir is molded by six parameters: rate of salt supply, dissolution, sediment accumulation, erosion, extension, and shortening [Koyi, 1998]. The rate of salt supply is the rate at which salt ﬂows from the source layer into the diapir and contributes to its growth. In a system where lateral movement (extension and shortening) is insigniﬁcant, salt supply and sedimentation rate govern the evolution of a down-built diapir [Vendeville and Jackson, 1993; Koyi, 1998]. If the rate of salt supply is less than the rate of sediment accumulation, upward-narrowing diapirs form (Fig. 5.5a). In contrast, upward-widening diapirs form when the salt supply is greater than the rate of sediment accumulation and columnar diapirs form when the rate of salt supply is equal to the rate of sediment accumulation [Fig. 5.5, Jackson and Talbot, 1991; Vendeville and Jackson, 1993; Talbot, 1995; Koyi, 1998].

5.2.3 Conclusions We conclude that salt supply controls the amount of entrainment, distribution of initially interbedded dense blocks/layers within a diapir, and dictates the internal structure of the diapir. With the range of parameters used in our models, entrainment of the anhydrite layer, which has been inevitable, was directly governed by the rate of salt supply (Fig. 5.6). On the other hand, the entrained anhydrite segments sink within the diapir when salt supply decreases dramat- ically or ceases entirely. The rate of sinking at this stage depends strongly on the viscosity of the salt; anhydrite segments sink slower within more viscous salt.

31 (a) Diapir rise > Aggradation rate (b) Diapir rise = Aggradation rate

(c) Diapir rise < Aggradation rate Figure 5.5: The cross-sectional shapes of passive diapirs are tied to the relative rates of net diapir rise (salt rise minus erosion and dissolution) and sediment aggradation. Modiﬁed from Giles and Lawton (2002).

Although the viscosity of salt signiﬁcantly alters salt supply, down-built diapirs grow in height at essentially similar rates. The variation in salt supply results in the formation of overhangs of different width. However, the effect of the viscosity variation on rate of rise of the crest of the diapir is negligible. Salt supply is governed by parameters (sedimentation rate, viscosity of salt perturbation width, and stratigraphic location of the anhydrite layer), which control the internal and external evolution of the diapir.

32 Salt Supply Cumulative Entrainment 100 100 2 mm/a 1 mm/a 80 80 0.5 mm/a 1 mm/a 0.25 mm/a 2 mm/a 0.1 mm/a 0.5 mm/a 0.05 mm/a 60 0.25 mm/a 60 0.1 mm/a 0.05 mm/a 40 40

20 20 Cumulative salt supply, S (%) Cumulative entrainment, E (%)

0 0 0 1 2 3 4 5 6 7 0 0.5 1 1.5 2 2.5 Time (Myr) Time (Myr) (a) Effect of the sedimentation rate

100 100

5×1017 Pa s 1017 Pa s 1018 Pa s 80 80 5×1017 Pa s 1017 Pa s 1018 Pa s 5×1018 Pa s 5×1018 Pa s 1019 Pa s 60 60

40 40

20 20 Cumulative salt supply, S (%) Cumulative entrainment, E (%)

0 0 0 5 10 15 20 25 30 35 0 5 10 15 20 25 Time (Myr) Time (Myr) (b) Effect of the salt viscosity

100 100

80 80 P =1600 m w P =1200 m w 60 60 P =400 m P =800 m w w P =500 m P =500 m w w P =600 m P =400 m 40 w 40 w P =800 m w P =1200 m w P =1600 m 20 w 20 Cumulative salt supply, S (%) Cumulative entrainment, E (%)

0 0 0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 Time (Myr) Time (Myr) (c) Effect of the perturbation width

100 100

NO Anhy. L = 800 m L = 800 m a a 80 80 L = 480 m L = 200 m a a L = 200 m L = 480 m a a 60 60

40 40

20 20 Cumulative salt supply, S (%) Cumulative entrainment, E (%)

0 0 0 2 4 6 8 10 0 2 4 6 8 10 Time (Myr) Time (Myr) (d) Effect of the stratigraphy

Figure 5.6: Effect of the sedimentation rate, salt viscosity, perturbation width (Pw) and stratigraphic location of the anhydrite layer (La) on cumulative salt supply and entrainment of the anhydrite layer.

33 5.3 Paper III: Evolution of the Gorleben salt diapir The Gorleben diapir, which has been targeted for radioactive waste disposal, contains large blocks of anhydrite. Numerical models that depict the geometri- cal conﬁguration of the Gorleben diapir are used to understand internal structure of diapir caused by movement of the anhydrite blocks for various salt rheologies. The Gorleben salt is modeled with rheological parameters chosen from laboratory experiments on different salt formations to test which rheology is able to activate previously inactive anhydrite blocks and if the mobility of anhydrite blocks can disturb any repository within the diapir.

5.3.1 Modeling concept In this study, we use documented configuration of the internal structure of the Gorleben diapir as the initial setup of the models (Fig. 1.3). The complex internal stratigraphy of the Gorleben diapir is simplified so that only the Main Anhydrite layer is used in the modeling (Fig. 5.7). The Main Anhydrite layer is modeled as an average 80m thick, (density = 2900 kg m−3) embedded within the Zechstein salt formation. In all models an effective viscosity of anhydrite is between 1019 and 1021 Pa s. Overburden layers are unified and given corresponding depth-independent density (2600 kg m−3). The sedimentary overburden rocks simulated in all the models are assigned high viscosity (effective viscosity ranges between 1023 and 1025Pa s). However, few models are deployed with lower viscosity contrast between the overburden and the salt to elucidate the importance of the overburden layers. The different cycles of the Zechstein salt z2, z3 and z4 represent the salt layer above the basement (Figs 1.3 and 5.7). The Zechstein salt formation is modeled as a power law salt using the currently available material parameters for different salt formation [Table 5.1, Kirby and Kronenberg, 1987]. A temperature distribution within the model linearly varies with depth (at the surface 0◦C and 140◦C at the bottom). A mean value of the density for rock salt is assumed to be 2200 kg m−3 [Landolt-Boernstein, 1982]. For comparison, we model Newtonian salt with an average viscosity ranging between 1017 and 1018Pa s [Urai et al., 1986; Spiers et al., 1990; Carter et al., 1993; Van Keken et al., 1993].

5.3.2 Summary If the rheology of the salt from the Gorleben diapir is to be approximated to Newtonian salt with viscosity of 1017Pa s, then the anhydrite layer within the diapir will sink with a maximum average velocity of ca 5 mm a−1. This observation suggests that if the average effective viscosity of salt ranges from 1017 to 1018 Pa s [e.g., Urai et al., 1986; Spiers et al., 1990; Hunsche and Hampel, 1999] diapirs with intercalated denser layer must be internally active even though the surface of the diapir does not rise. The fact that the internal

34 t = 0 Myr 0

−500 a b c −1000

−1500

−2000 Depth (m) −2500 Anhydrite Layer Overburden −3000 Salt −3500

−4000 0 1000 2000 3000 4000 5000 6000 Lenght (m) Figure 5.7: The initial geometry of the model. The temperature distribution within the box linearly varies with depth at the surface 0◦C and 140◦C at the bottom. Points a, b and c deﬁne passive markers used to monitor deformation of the plant survey.

movement of the diapir is due to the presence of the anhydrite layer is verified in the model where the anhydrite layer was removed from the initial setup, but all the other parameters were kept the same. In this model salt remained inactive (maximum vertical velocity 9.6 × 10−5 cm a−1) both externally and internally for more than 5 million years. Based on the movement of passive markers, it is suggested that for the salt modeled with rheological parameters of the Paradox Formation (weakest), the western margin of the potential repository (adjacent to larger anhydrite blocks, point a) will sink with an average rate of 3.2 mm a−1 during a subsequent time period of 0.1 Ma. The eastern margin of the potential repository (point c) will sink with a rate of 0.2 mm a−1 within the same time. Unlike marginal points of the potential repository, the mid-point (point b) rises with a rate of 0.73 mm a−1. The deformation rate of the potential repository is sufficiently reduced if the effective viscosity of salt is high (models: AI, SD, RD, and PB, Fig. 5.8). This observation suggests that if the average effective viscosity of salt ranges from 1019 to 1020 Pa s the diapirs with intercalated denser layer must deform with a rate that is not significant. It should be noted that the predicted fast deformation rates (weak PF-rheology) lie in the range to be observed by geodetic measurements. Depending on size and orientation of the anhydrite layer, deformation pattern is significantly different within the diapir (Fig. 5.9). Horizontal blocks sink much slower than vertical blocks. Furthermore, the rate of descent differs on different sides of the diapir. Descent rate, mode and pattern depend on vertical and horizontal location and orientation of the anhydrite blocks within the

35 PF: t = 0.010 Myr PF: t = 0.047 Myr PF: t = 0.208 Myr 0 −0.5 −1 −1.5 −2 −2.5 Depth (km) −3 −3.5 −4

(a) (b) (c)

AI: t = 1.038 Myr AI: t = 2.597 Myr AI: t = 4.805 Myr 0 −0.5 −1 −1.5 −2 −2.5 Depth (km) −3 −3.5 −4

(d) (e) (f)

VD: t = 0.011 Myr VD: t = 0.104 Myr VD: t = 1.003 Myr 0 −0.5 −1 −1.5 −2 −2.5 Depth (km) −3 −3.5 −4

(g) (h) (i)

N17: t = 0.010 Myr N17: t = 0.104 Myr N17: t = 0.274 Myr 0 −0.5 −1 −1.5 −2 −2.5 Depth (km) −3 −3.5 −4 0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6 Length (km) Length (km) Length (km) (j) (k) (l) Figure 5.8: Evolution of the models for different salt rheology. (a-c) Paradox For- mation (PF), (d-f) Avery Island (AI), (g-i) Vacherie Dome (VD), (j-l) Newtonian salt with viscosity 1017 Pa s.

36 0

−0.5

−1

−1.5

−2 deformed undeformed displacement

Depth (km) −2.5

−3

−3.5

−4 0 1 2 3 4 5 6 Length (km) Figure 5.9: Internal deformation of the diapir (PF-Model) shown after 85 kyr. Arrows deﬁne velocity vectors at t = 85 kyr. diapir. In addition, in models with variable salt viscosity descent rate depends on the rheology of the salt, which is in direct contact with the anhydrite layer.

5.3.3 Conclusions Numerical models show that salt structures with intercalated dense layers are internally active. We have shown that the mobility of anhydrite blocks depends on the effective viscosity of salt which has to be lower than a threshold value of around 1018 − 1019 Pa s. During the post-depositional stage, the effective viscosity may fall below this threshold (e.g. due to changing temperature or migrating/diffusing water). The internal deformation of the salt diapir by the descending blocks increases with decrease in effective viscosity of salt. Un- like salt viscosities beyond the values 1020 Pa s, for the common range of the effective viscosity of salt (1017 − 1020 Pa s) relatively high rate of descend is observed for the intercalated dense layers. The mobility of these dense layers directly inﬂuences any repository within the diapir. However, the rate of deformation that the repository undergoes is strongly dependent on salt rheology.

37 38 ∗ ∗ Table 5.1: Model Parameters: A is a pre-exponential constant, H -activation enthalpy, n-power-law exponent, Tt − Tb is temperature at top and bottom, respectively. ηmin, ηmax and η are minimum, maximum and arithmetic mean of the salt viscosity occurring in models, respectively. ∗ ∗ Model Natural Rocksalt A n H Tt − Tb Effective viscosity of salt [Pa s] Comments −n −1 −1 ◦ [Pa s ] [kJ mol ] [ C] ηmin ηmax η PF Paradox Formation 6.9183E-16 1.39 28.80 0-140 2.29E16 2.29E18 2.6E18 PF-Cold Paradox Formation 6.9183E-16 1.39 28.80 0 8.6E17 1.9E19 1.09E19 PF-NA Paradox Formation 6.9183E-16 1.39 28.80 0-140 7.14E16 3.2E19 9.7E18 No anhydrite AI Avery Island 1.4454E-32 4.10 33.6 0-140 5.9E18 1.7E23 1.29E20 AI-OV21 Avery Island 1.4454E-32 4.10 33.6 0-140 1.7E18 2.7E22 5.2E19 Overburden viscosity is 1021 Pa s AI-OV22 Avery Island 1.4454E-32 4.10 33.6 0-140 6.2E18 1.7E22 8.9E19 Overburden viscosity is 1022 Pa s VD Vacherie Dome 4.1687E-16 2.22 62.90 0-140 5.7E16 1.0E21 1.05E19 RD Richton Dome 2.2387E-32 5.01 82.30 0-140 2.9E19 1.0E25 4.3E20 SF Salado Formation 1.5488E-35 4.90 50.20 0-140 4.0E19 1.0E25 4.2E20 PB Permian Basin 4.6774E-30 4.50 72.00 0-140 2.9E19 9.8E24 4.0E20 N17 Newtonian 1017 1 - 0-140 - - - N18 Newtonian 1018 1 - 0-140 - - - 6. Concluding Remarks

In order to assess the potential importance of sinking or mobility of entrained blocks within a diapir, it is important to understand how an anhydrite layer was initially entrained into a diapir. Two time-scales have to be considered: the time scale of diapiric growth t1 (e.g. by down-building) and the time scale of sinking of the blocks, t2. Keeping all material parameters the same, entrainment of dense blocks requires that t1 < t2, i.e. during diapiric rise, the blocks do not sink faster than the diapir grows. The time scale of the diapiric growth (t1) is governed by four main parameters (sedimentation rate, salt viscosity, perturbation width and stratigraphic location of dense layer), which are systematically studied in this thesis. In a system where lateral movement (extension and shortening) is insignif- icant, keeping all material parameters constant, the condition for entrainment of dense blocks by a down-built diapir is that the sedimentation rate should be less than a maximum values ˙max, so the diapir remains uncovered and it has to be greater than a minimum sedimentation rates ˙min, otherwise no entrainment will take place. Therefore, sedimentation rate within the range where the conditions ˙max < s˙e < s˙min is fulﬁlled is required for the entrainment. Depending on the evolution history of diapirs modeled in this thesis, we distinguished three ranges of sedimentation rate: I High sedimentation rates (> 4 mm a−1), which bury the initial perturbation and results in an immediate sinking of the embedded anhydrite layer within the salt layer. In this case, salt supply to the perturbation is zero and the anhydrite layer is not entrained. II Intermediate sedimentation rates (3–0.5 mm a−1) that are initially less than the rate at which the crest of the diapir is growing so that the crest of the diapir remains uncovered and provides highest salt supply. Mod- els with intermediate sedimentation rate are characterized by high percentage entrainment (80-90 percent). III Slow sedimentation rates (0.25–0.05 mm a−1) that segments the diapir due to formation of mini-basins and reduces salt supply. Entrainment of the anhydrite layer in these models is limited (less than 30 percent). Down-building and formation of a diapir is a slow process (c.f. the time scale of 10–100 Myr) and requires relatively stiff salt rheology (effective viscosity > 1019 Pa s) for efﬁcient entrainment of dense blocks at geologically reasonable rates. However, during the post-depositional stage, when the salt supply to the diapir ceases and the diapir becomes inactive, entrained segments inevitably sink back into the diapir. As the time scale of sinking of dense blocks (t2), controls the mobility of the blocks, the condition t2 > t1

39 has implicitly been taken as a guarantee for the safety of the stability of any repository built within a diapir where dense blocks exist. However, as it is shown in this thesis if the effective viscosity of salt falls below the threshold value of around 1018 − 1019 Pa s, the mobility of anhydrite blocks might in- ﬂuence any repository within a diapir. Moreover, redistribution of stresses at boundary between salt rocks with distinctly different creep properties (e.g., if the salt layers within Zechstein Formation (z2, z3 and z4) from the Gorleben diapir differ in rheology) can control the direction and the rate of sinking of the dense blocks. An accurate monitoring of the current deformation within the Gorleben diapir will help us understand advanced diapirism of evaporites with a layered rheology. The structural evolution of a formerly tabular salt layer in which an anhydrite layer of high density and high viscosity is entrained is a key for understanding the internal deformation of any repository built in a diapiric structure. The development of such structure must reliably be forecasted within the next one million years.

40 7. Summary in Swedish

I den här avhandlingen används resultat från numeriska modeller för att visa att externt inaktiva saltstrukturer, vilka utgör potentiella förvarsplatser för radioaktivt avfall, skulle kunna vara internt aktiva på grund av närvaro av inre tunga lager eller block. I de tre uppsatser som avhandlingen bygger på används saltdiapiren i Gor- leben, vilken diskuterats i samband med den slutgiltiga förvaringen av högak- tivt radioaktivt avfall, som en allmän referens. I de första två uppsatserna presenteras systematiska studier över de parametrar som kontrollerar utvecklingen av en saltdiapir och vad som händer om den innehåller ett tungt anhydritlager. Reslutat från de numeriska modellerna visar att inkorporering av ett tungt anhydritlager i en saltdiapir beror av fyra parametrar: sedimentationshastighet, saltets viskositet, störningens bredd, samt den stratigraﬁska lokaliseringen av det tunga lagret. Den kombinerade effekten av dessa fyra parametrar, vilka har direkt betydelse för salttillförselns hastighet (volym/yta förhållandet av det tillförda saltet som funktion av tiden), kommer att bestämma diapirens form och sättet på vilket inkorporeringen av det tunga lagret sker. Saltdiapirer, "nedåtbyggda" med högviskösa sedimentära enheter har potentialen att växa med ett inbäddat anhydritlager och utarma källan (salttillförseln upphör). När saltkällan snabbt minskar eller upphör helt och hållet kommer emellertid det inbäddade anhydritlagret/segmentet att börja sjunka inom diapiren. I inaktiva diapirer är detta ofrånkomligt och starkt beroende av reologin hos det salt som är i direktkontakt med anhydritlagret. Under det post-depositionella stadiet, och om saltets effektiva viskositet är under tröskelvärdet 1018 −1019 Pa s, skulle anhydritblockets mobilitet kunna tänkas påverka ett slutförvar inne i diapiren. Daipirens interna deformation som funktion av de nedåtgående anhydritblocken kommer emellertid att minskas när saltets effektiva viskositet ökar. Resultaten i avhandlingen visar att saltstrukturer som innehåller tunga och viskösa lager/block med stor sannolikhet kommer att undergå interna defor- mationsprocesser när de tunga lagren börjar sjunka inne i diapiren. Blockens storlek och orientering bestämmer deformationsmönstret. Applicering av resultaten från modelleringarna på Gorlebendiapiren visar att sjunkhastigheten hos det interna anhyritlagret skiljer sig på olika sidor av diapiren vilket an- tyder icke symmetrisk deformation. Av detta kan man dra slutsatsen att en sjunkning av anhydritblocken inne i Gorlebendiapiren kommer att initiera ett inre, icke symmetriskt ﬂöde.

Acknowledgments

I have crossed paths with several colleagues since the beginning of my studies and would not have been able to carry out all of this work alone. I would like to express my deepest gratitude to all of those who have been involved and supported this work. Especially, I would like to mention the supervisor of this work, Professor Hemin Koyi, who introduced me to this ﬁeld of geology and offered me the possibility to do research. He shaped my learning with his valuable geological insight and I would sincerely like to express my deep feelings towards him. I would also like to acknowledge Professor Harro Schmeling, whose useful suggestions have helped to give this project its current shape, as he was my co- supervisor and the author of the FDCON numerical code used in this research. Without his valuable contribution, this work would not have been possible. Many thanks also go to Professor Christopher Talbot for raising my interest in Global Tectonics and for patiently answering all of my questions related to salt tectonics. I would like to thank the people who shared their valuable knowledge and inspired me in the very beginning of my PhD research. I would like to express my gratitude to Jacek Gutowski who kindly introduced me to basics of geology, Stanislaw Burliga for sharing his knowledge about salt tectonics. I would like to extend deep gratitude to my former colleague in Uppsala and dear friend Faramarz Nilfouroushan, who in the early years of my postgradu- ate studies helped me greatly to really get started. I thank Faramarz for proof reading this thesis. Thank you to Örjan Amcoff, who kindly translated the summary of the thesis into Swedish, and to Taher Mazloomian, Anna Caller- holm and Leif Nyberg who helped me to print the thesis and who provided technical assistance critical for completing the project. I am thankful to Murad Odisharia and Guram Odisharia, who paved my way into science. They were instrumental in my coming to Uppsala University as a Master student. I studied under the supervision of Sergej Zilintikevich and the co-supervision of Igor Ezau. I would like to express my profound gratitude to them for inspiring me and helping to deﬁne my research topic. I feel everlasting gratitude for lecturers and professors of the Tbilisi State University, where I enrolled with encouragement from my physics teacher Jimsher Qajaia and my mathematics teacher Tengo Chemia. They instilled in me from an early age the desire and skills necessary to obtain a University degree. I greatly enjoyed collegiate life at Tbilisi University, where I made lifelong friends that I have always missed during my stay in Uppsala.

43 Uppsala is a place that gives great opportunity to meet people from all over the world. I have to extend my thanks to those people that I met during my stay in Uppsala: Albert Mihranyan , Giorgi Metrevely, Davit Khoshtaria, Juan Diego Martín, Jose Luis Perez, Petros Mouyiasis, Stela Tagliati, Tomas Björk- lund and Caroline Burberry. I would like to convey thanks to all of you for your valued support and friendship. Next, I wish to thank the complete collegium: Sadoon Morad, Hans An- nersten, Hans Harryson, Peter Lazor, Håkan Sjöström, Ulf Andersson, Howri Mansurbeg, Khalid Al-Ramadan, Erik Ogenhall, Soﬁa Winell, Zuzana Konop- kova, Kristina Zarins, Osama Helal, Lijam Hagos Zemichael, Tuna Eken, Majid Beiki, David Gee , Christoph Hieronymus, Juliane Hübert, Nazli Is- mail, Nina Ivanova, Thomas Kalscheuer, Artem Kashubin, Hesam Kazemeini, Alireza Malehmir, Peter Schmidt, Hossein Shomali, Arnaud Pharasyn and Frances Mary Deegan. All personnel in the department of Earth Sciences, who, during these past several years, in innumerable ways helped me to carry out this work. I also thank VR, The Swedish Research Council for providing the necessary support for this research. Finally, special and dearest thanks to my family and my girlfriend who light up my life and who have given me something else to think about during these past four years.

44 References

Biot, M. A. (1965). Theory of viscous buckling and gravity instability of multilayers with large deformation. Bulletin of Geological Society of America 76(3), 371–378.

Bornemann, O. (1982). Stratigraphy and tectonics of the Zechstein in Gor- leben salt dome using drilling results. Zeitschrift der Deutschen Geologischen Gesellschaft 133(2), 119–134.

Bornemann, O. (1991). Zur geologie des salzstocks Gorleben nach den Bohrergeb- nissen. Hannover, Germany, Bundesamt fur Strahlenschutz Schriften, Salzgit- ter 4(91), 67.

Carter, N. L., S. T. Horseman, J. E. Russell, and J. Handin (1993). Rheology of rocksalt. Journal of Structural Geology 15(9-10), 1257–1271.

Chemia, Z., H. Koyi, and H. Schmeling (2008). Numerical modelling of rise and fall of a dense layer in salt diapirs. Geophysical Journal International 172(2), 798–816.

Fairhurst, C. (2002). Geomechanics issues related to long-term isolation of nuclear waste. Comptes Rendus Physique 3(7-8), 961–974.

Gansser, A. (1992). The enigma of the Persian salt-dome inclusions. Eclogae Geo- logicae Helvetiae 85(3), 825–846.

Hudec, M. and M. Jackson (2007). Terra inﬁrma: Understanding salt tectonics. Earth-Science Reviews 82(1-2), 1–28.

Hughes, M. and I. Davison (1993). Geometry and growth kinematics of salt pillows in the southern north sea. Tectonophysics 228, 239–254.

Hunsche, U. and A. Hampel (1999). Rock salt-the mechanical properties of the host rock material for a radioactive waste repository. Engineering Geology 52(3-4), 271–291.

Jackson, J. A. (1997a). Glossary of Geology. Alexandria, Virginia: American Geo- logical Institute, 4th edition.

Jackson, M. (1995). Retrospective salt tectonics. In M. Jackson, D. Roberts, and S. Snelson (Eds.), Salt Tectonics: A Global Perspective, Volume 65, pp. 1–28. AAPG Memoir.

Jackson, M. and C. Talbot (1991). A Glossary of Salt Tectonics. Bureau of Economic Geology, Geological Circular 91-4: The University of Texas at Austin.

Jackson, M. P. A. (1997b). Conceptual breakthroughs in salt tectonics: A Histor- ical Review, 1856-1993. University of Texas at Austin, Austin, Texas: Bureau of Economic Geology.

45 Jackson, M. P. A. and B. C. Vendeville (1994). Regional extension as a geologic trigger for diapirism; with suppl. data 9401. Geological Society of America Bul- letin 106(1), 57–73.

Jaritz, W. (1993). Die geowissenschaftliche Untershchung des Salzstocks Gorleben auf seine Eingnung für ein Endlager für radioaktive Abfalle-Stand 1993. Geologis- ches Jahrbuch 142, 295–304.

Kent, P. E. (1979). The emergent Hormuz salt plugs of southern Iran. Journal of Petroleum Geology 2, 117–144.

Kirby, S. and A. K. Kronenberg (1987). Rheology of the lithosphere: selected topics. Reviews of Geophysics 25(6), 1219–1244.

Koyi, H. (1998). The shaping of salt diapirs. Journal of Structural Geology 20(4), 321–338.

Koyi, H. (2001). Modeling the inﬂuence of sinking anhydrite blocks on salt diapirs targeted for hazardous waste disposal. Geology 29(5), 387–390.

Koyi, H. A., N. S. Mancktelow, and H. Ramberg (2001). Tectonic modeling : a volume in honor of Hans Ramberg. Memoir / Geological Society of America, 193. Boulder: Geological Society of America. edited by Hemin A. Koyi and Neil S. Mancktelow ill., diagr., tab.

Landau, L. D. and E. M. Lifsic (1987). Fluid mechanics (2. ed.). Course of theoretical physics, 6. Oxford: Pergamon Press. by L.D. Landau and E.M. Lifshitz ; translated from the Russian by J.B. Sykes and W.H. Reidill.

Landolt-Boernstein (1982). Numerical data and functional relationships in science and technology group, v: Geophysics and space research. In Physical Propertes of Rocks, Volume 1a, pp. 373. New York: Springer.

Muller, W. H., S. M. Schmid, and U. Briegel (1981). Deformation experiments on anhydrite rocks of different grain sizes: Rheology and microfabric. Tectono- physics 78(1-4), 527–543.

Nettleton, L. (1934). Fluid mechanics of salt domes. AAPG Bulletin 18, 1175–1204.

Park, R. G. (1989). Foundations of structural geology (2nd ed.). Glasgow: Blackie. R.G. Park. Previous ed.: 1983.

Richter-Bernburg, G. (1980). Salt domes in Northwest Germany. Centres de Recherches Exploration-Production, Bulletin, Elf Aquitaine 4, 373–393.

Rönnlund, P. (1989). Viscosity ratio estimates from natural rayleigh-taylor instabilities. Terra Nova 1, 344–348.

Romer, M. M. and H. J. Neugebauer (1991). The salt dome problem: a multilayered approach. Journal of Geophysical Research 96(B2), 2389–2396.

Schmeling, H. (1987). On the relation between initial conditions and late stages of Rayleigh-Taylor instabilities. Tectonophysics 133(1-2), 65–80.

46 Schmeling, H., A. Babeyko, A. Enns, C. Faccenna, F. Funiciello, T. Gerya, G. J. Go- labek, S. Grigull, B. Kaus, G. Morra, S. M. Schmalholz, and J. van Hunen (2008). A benchmark comparison of spontaneous subduction models - towards a free surface. Physics of the Earth and Planetary Interiors.

Spiers, C., P. M. T. M. Schutjens, R. H. Brzesowsky, C. J. Peach, J. L. Liezenberg, and H. J. Zwart (1990). Experimental determination of constitutive parameters governing creep of rocksalt by pressure solution. In Deformation mechanisms, rheology and tectonics, Volume 54, pp. 215–227. Geological Society, London; Special Publi- cation.

Spies, T. and J. Eisenblätter (2001). Acoustic emission investigation of microcrack generation at geological boundaries. Engineering Geology 61(2-3), 181–188.

Talbot, C., P. Rönnlund, H. Schmeling, H. Koyi, and M. Jackson (1991). Diapiric spoke patterns. Tectonophysics 188, 187–201.

Talbot, C. J. (1995). Molding of salt diapirs by stiff overburden. In Salt tectonics: A Global Perspective, Volume 65, pp. 61–75. AAPG Memoir.

Urai, J. L., C. J. Spiers, H. J. Zwart, and G. S. Lister (1986). Weakening of rock salt by water during long-term creep (nuclear waste disposal). Nature 324(6097), 554–557.

Van Keken, P. E., C. J. Spiers, A. P. Van den Berg, and E. J. Muyzert (1993). The effective viscosity of rocksalt: Implementation of steady-state creep laws in numerical models of salt diapirism. Tectonophysics 225(4), 457–476.

Vendeville, B. and M. Jackson (1992). The rise of diapirs during thin-skinned extension. Marine and Petroleum Geology 9, 331–353.

Vendeville, B., M. Jackson, and R. Weijermars (1993). Rates of salt ﬂow in passive diapirs and their source layers. In Rates of geologic processes: tectonics, sedimentation, eustasy and climate: implications for hydrocarbon exploration: Gulf Coast Section, pp. 269–276. Houston, Texas: Soc. Econ. Paleontol. Mineral. Found. Res. Conf., 14th.

Vendeville, B. C. and M. P. A. Jackson (1993). Rates of extension and deposition determine whether growth faults or salt diapirs form. In Rates of geologic processes: tectonics, sedimentation, eustasy and climate: implications for hydrocarbon exploration: Gulf Coast Section, pp. 263–268. Houston, Texas: Soc. Econ. Paleontol. Mineral. Found. Res. Conf., 14th.

Warren, J. (1999). Evaporites: Their Evolution and Economics. Oxford: Blackwell Science.

Weijermars, R., M. P. A. Jackson, and B. Vendeville (1993). Rheological and tectonic modeling of salt provinces. Tectonophysics 217(1-2), 143–174.

Weinberg, R. F. (1993). The upward transport of inclusions in Newtonian and power- law salt diapirs. Tectonophysics 228(3-4), 141–150.

Weinberg, R. F. and H. Schmeling (1992). Polydiapirs: multiwavelength gravity structures. Journal of Structural Geology 14(4), 425–436.

47 Woidt, W. D. (1978). Finite element calculations applied to salt dome analysis. Tectonophysics 50, 369–386.

Zirngast, M. (1991). Die entwicklungsgeschichte des salzstocks gorleben-ergebnis einer strukturgeologischen bearbeitung. Geologisches Jahrbuch 132, 3–31.

Zirngast, M. (1996). The development of the Gorleben salt dome (northwest Ger- many) based on quantitative analysis of peripheral sinks. In Salt tectonics, Volume 100, pp. 203–226. London: Geological Society, Special Publication.

Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 551

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science and Technology, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology”.)

ACTA UNIVERSITATIS UPSALIENSIS Distribution: publications.uu.se UPPSALA urn:nbn:se:uu:diva-9279 2008