ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2015

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1259

Analysis of Seismic Data Acquiredat the Forsmark Site for Storageof Spent Nuclear Fuel, CentralSweden

FATEMEH SHARIFI BROJERDI

ISSN 1651-6214ISBN 978-91-554-9259-5urn:nbn:se:uu:diva-251621

Dissertation presented at Uppsala University to be publicly examined in Hamberg,Geocentrum, Villav. 16, 752 36 Uppsala, Uppsala, Friday, 12 June 2015 at 10:00 for thedegree of Doctor of Philosophy. The examination will be conducted in English. Facultyexaminer: Professor Ramon Carbonell (CSIC-Inst. Earth Sciences Jaume Almera).

AbstractSharifi Brojerdi, F. 2015. Analysis of Seismic Data Acquired at the Forsmark Site for Storageof Spent Nuclear Fuel, Central Sweden. Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology 1259. 48 pp. Uppsala: ActaUniversitatis Upsaliensis. ISBN 978-91-554-9259-5.

The Forsmark area, the main study area in this thesis, is located about 140 km northof Stockholm, central Sweden. It belongs to the Paleoproterozoic Svecokarelian orogenand contains several major ductile and brittle deformation zones including the Forsmark,Eckarfjärden and Singö zones. The bedrock between these zones, in general is less deformed andconsidered suitable for a nuclear waste repository. While several site investigations have alreadybeen carried out in the area, this thesis focuses primarily on (i) re-processing some of the existingreflection seismic lines to improve imaging of deeper structures, (ii) acquiring and processinghigh-resolution reflection and refraction data for better characterization of the near surfacegeology for the planning of a new access ramp, (iii) studying possible seismic anisotropy fromactive sources recorded onto sparse three-component receivers and multi-offset-azimuth verticalseismic profiling data (VSP). Reflection seismic surveys are an important component of theseinvestigations. The re-processing helped in improving the deeper parts (1-5 km) of the seismicimages and allowing three major deeper reflections to be better characterized, one of which issub-horizontal while the other two are dipping moderately. These reflections were attributed tooriginate from either dolerite sills or brittle fault systems. First break traveltime tomographyallowed delineating an undulating bedrock-surface topography, which is typical in the Forsmarkarea. Shallow reflections imaged in 3D, thanks to the acquisition design were compared withexisting borehole data and explained by fractured or weak zones in the bedrock. The analysisof seismic anisotropy indicates the presence of shear-wave splitting due to transverse isotropywith a vertical symmetry axis in the uppermost hundreds of meters of crust. Open fractures andjoints were interpreted to be responsible for the large delays observed between the transverseand radial components of the shear-wave arrivals, both on surface and VSP data.

Keywords: Forsmark site, anisotropy, deformation zone, reflection seismic, shear-wavesplitting, fractures and traveltime tomography

Fatemeh Sharifi Brojerdi, Department of Earth Sciences, Geophysics, Villav. 16, UppsalaUniversity, SE-75236 Uppsala, Sweden.

© Fatemeh Sharifi Brojerdi 2015

ISSN 1651-6214ISBN 978-91-554-9259-5urn:nbn:se:uu:diva-251621 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-251621)

“There is always something you can do, and succeed at.

While there’s life, there is hope.”

Stephen Hawking

Dedicated toMy Mom, Dad

& My love Reza

List of Papers

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

I Fatemeh Sharifi Brojerdi, Christopher Juhlin, Alireza

Malehmir, Michael B. Stephens, (2013) Reflection seismic im-aging of the deeper structures at the Forsmark spent nuclear fuel repository site, central Sweden. Journal of Applied Geophysics, 89:21–34

II Fatemeh Sharifi Brojerdi, Fengjiao Zhang, Christopher Juhlin, Alireza Malehmir, Tomas Lehtimäki, Håkan Mattsson, Philip Curtis, (2014) High resolution seismic imaging at the planned tunnel entrance to the Forsmark repository for spent nuclear fuel, central Sweden. Near Surface Geophysics, 12:709-719

III Fatemeh Sharifi Brojerdi, Christopher Juhlin, Alireza

Malehmir, Bjarne S.G. Almqvist, (2015) Analysis of seismic anisotropy derived from sparse three-component receivers at the Forsmark spent nuclear fuel repository site, central Sweden. Submitted.

IV Fatemeh Sharifi Brojerdi, Christopher Juhlin, Alireza

Malehmir, Bjarne S.G. Almqvist, (2015) Analysis of P- and S-waves in vertical seismic profile data acquired at the Swedish site for storage of spent nuclear fuel, Forsmark. Manuscript.

Reprints were made with permission from the respective publishers.

Contents

1. Introduction ............................................................................................... 11 

2. Geological background ............................................................................. 13 

3. Data acquisition ........................................................................................ 16 3.1. Data acquisition for the reflection and three-component seismic profiles-stage 1 ......................................................................................... 16 3.2. Data acquisition for the reflection seismic profiles-stage 2 .............. 17 3.3. Data acquisition for the reflection and refraction seismic lines ........ 18 3.4. Data acquisition for the vertical seismic profile (VSP) ..................... 19 

4. An introduction to anisotropy ................................................................... 21 

5. Summary of papers ................................................................................... 24 5.1. Paper I: Reflection seismic imaging of the deeper structures at the Forsmark spent nuclear fuel repository site, central Sweden ................... 24 

5.1.1. Summary .................................................................................... 24 5.1.2. Conclusions ............................................................................... 29 

5.2. Paper II: High resolution seismic imaging at the planned tunnel entrance to the Forsmark repository for spent nuclear fuel, central Sweden ..................................................................................................... 29 

5.2.1. Summary .................................................................................... 29 5.2.2. Conclusions ............................................................................... 31 

5.3. Paper III: Analysis of seismic anisotropy derived from sparse three-component receivers at the Forsmark spent nuclear fuel repository site, central Sweden ................................................................. 32 

5.3.1. Summary .................................................................................... 32 5.3.2. Conclusions ............................................................................... 35 

5.4. Paper IV: Analysis of P- and S-waves in vertical seismic profile data acquired at the Swedish site for storage of spent nuclear fuel, Forsmark .................................................................................................. 35 

5.4.1. Summary .................................................................................... 35 5.4.2. Conclusions ............................................................................... 38 

6. Conclusions and recommendations ........................................................... 39 

7. Summary in Swedish ................................................................................ 42 

Acknowledgments......................................................................................... 44 

References ..................................................................................................... 46 

Abbreviations

1D One-dimensional 3C Three-component 3D Three-dimensional CDP Common depth point CMP Common midpoint cm Centimeter DMO Dip moveout g Gram Ga A billion years (Gigayears) ago GPS Global positioning system Hz Hertz km Kilometer m Meter mm Millimeter MPa Megapascal ms Millisecond m/s Meter per second NMO Normal moveout s Second VSP Vertical Seismic profiling

11

1. Introduction

The Forsmark site, which is located about 140 km north of Stockholm in the region of Uppland, central Sweden (Figure 2.1), has recently been selected by the Swedish Nuclear Fuel and Waste Management Company (SKB) as the repository site for Sweden’s high-level spent nuclear fuel (SKB, 2011). The goal is to store nuclear and radioactive waste in a manner that meets maximum safety for human beings and the environment for a long period of time at about 500 m depth. The proposed plan includes up to 60 kilometers of tunnels in an underground system in order to host 6000 copper canisters of spent fuel (SKB, 2009). The importance of this plan requires geological and geophysical studies to understand the tectonic evolution of the area. These are a component of the site investigations in order to ensure high safe-ty for a long time (i.e. on the order of 100,000 years) storage. Characteriza-tion of the Forsmark site to date has consisted of numerous site-specific in-vestigations (Leijon, 2005). The bedrock in the Forsmark area formed during the late Paleoproterozoic, Svecokarelian orogeny inside the Fennoscandian Shield. It consists of crystalline bedrock, which formed during the orogeny and which has since been deformed both in the ductile and brittle regimes. Three major sub-vertical (at the surface), composite ductile and brittle de-formation zones with strike in a WNW or NW direction were delineated in the area. In between these zones, the bedrock is less deformed and bounded by high-strain deformation zones making it suitable for a repository (Ste-phens et al., 2007, 2008, 2009). Extensive site investigations were performed in the area prior to selecting Forsmark for the repository (SKB, 2008). The principle objective of this thesis is to apply seismic imaging methods over geologically interesting areas at the Forsmark site. Extensive seismic sur-veys, together with cored borehole data and hydraulic tests down to 1000 m depth (Andersson et al., 2013) provided data on the uppermost kilometer of crust. Various seismic experiments were conducted (e.g., Juhlin and Ste-phens 2006; Brojerdi et al., 2013) with the main focus given to the upper-most kilometer. However, the PhD thesis presented here shows that more information could be extracted from the existing data and some of the newly acquired data.

This thesis is divided into two sections and consists of a summary and a collection of papers that comprises re-processing of existing reflection seis-mic data and studies of seismic anisotropy in existing surface and borehole seismic data. In addition, high resolution seismic data were acquired in con-

12

junction with the planning of an access ramp to the repository and results from reflection seismic processing and traveltime tomography are presented here. The summary is divided into seven chapters, starting with this intro-duction (Chapter 1). In Chapter 2, the geological background is provided. In Chapter 3, various types of data acquisition geometries used in this thesis are presented. Chapter 4 provides an introduction to anisotropy. Chapter 5 pre-sents a summary of the papers that are included in the second section. Con-clusions and recommendations are summarized in Chapter 6. Finally, a summary in Swedish is provided in Chapter 7.

13

2. Geological background

Forsmark is located in south-eastern Sweden in the Fennoscandian Shield and part of the Paleoproterozoic Svecokarelian orogen, which formed c. 1.9 to 1.8 Ga ago. Dominant rock types in the area include 1.91-1.89 Ga metavolcanic rocks, both older and younger metasedimentary rocks, 1.90-1.84 Ga calc-alkaline meta-intrusive rocks, and 1.87-1.84 Ga and 1.80 Ga alkali-calcic meta-intrusive rocks (Stephens et al., 2009). These rocks were subjected to at least two separate pulses of ductile transpressive deformation under low pressure conditions, with peak metamorphism under amphibolite or, locally, granulite facies conditions (Stephens et al., 2009). This ductile strain is expressed by a major belt of deformation zones and intervening lower strain tectonic lenses in the coastal area close to the Forsmark site. The belt strikes WNW-ESE or NW-SE, and extends across strike for several tens of kilometers and includes gneissic rocks and mylonite zones in what is re-ferred to as the Singö shear zone (Talbot and Sokoutis, 1988, 1995). The Forsmark deformation zone (FDZ), the Eckarfjärden deformation zone (EDZ), and the Singö deformation zone (SDZ) (Figure 2.1) are the three main deformation zones in the Forsmark area, with thicknesses up to 200 m, as constrained by geophysical and geological data. These three zones are observed in the total field magnetic anomaly map for the area and in the topographic relief (Figure 2.1) (e.g., Malehmir et al., 2011; Stephens et al., 2007, 2009). The magnetic anomalies tend to follow the strike of the shear zones. The WNW striking FDZ and SDZ are considered as the most region-ally important structures, while the NW striking EDZ has been interpreted as a Riedel splay off the FDZ (Saintot et al., 2011; Stephens et al., 2007).

Results from drilling through these zones, in combination with tunnel data along the SDZ and deterministic geological modeling work, show that the zones are sub-vertical in at least the uppermost kilometer of crust (Stephens et al., 2007) and separate the less deformed tectonic lenses from one another. The Forsmark tectonic lens, which is considered as the repository volume for the spent nuclear fuel, is situated in the north-westernmost part of one of these lenses. It is located close to profiles 1 and 4 (P1 and P4 in Figure 2.1) and bounded in essence by the EDZ and SDZ. In this volume gneissic gran-ite formed at 1.87 Ga (Hermansson et al., 2008) and, along with pegmatites and amphibolites, comprise the dominant rock types. In addition, two main sets of fracture zones, displaying only brittle deformation, are present in the area and are considered as second order structures with respect to the sub-

14

vertical WNW-ESE or NW-SE striking ductile deformation zones. These fracture zones have trace lengths at the ground surface of generally less than a few kilometers. According to Stephens et al., (2007) and Sandström et al., (2009), the style of deformation and the common occurrence of adularia and laumontite along the fractures within these zones imply that their develop-ment occurred under lower amphibolite- to greenschist- facies metamorphic conditions (Sandström et al., 2009; Stephens et al., 2007). One fracture set consists of sub-vertical zones and strikes NNE-ENE and is spatially con-strained within the higher-temperature ductile shear belts. The second set of brittle fracture zones shows a similar strike, but dips gently in a southerly to easterly direction. There are also reported sheet joints, extending for more than one kilometer and mostly horizontal, following the undulations of the ground surface, revealing that many of the joints are present in the upper-most part of the bedrock (Stephens et al., 2015). The bedrock could have started to deform in the brittle regime some time between 1.8 and 1.7 Ga during the latest part of the Svecokarelian orogeny (Söderlund et al., 2009; Stephens et al., 2008). It has been shown that the maximum principal stress is horizontal at the site, trending NW-SE (Martin, 2007; Glamheden et al., 2007). The stress model at the proposed repository depth is estimated to have a maximum σ1 σH and minimum horizontal stress of ~45 MPa and 20-25 MPa, respectively, and a vertical stress of 10-15 MPa (Juhlin and Stephens 2006; Stephens et al., 2015; Brojerdi et al., 2015). It is clear that both horizontal stresses are significantly larger than the vertical stress at repository depth. Consequently, a horizontal stress oriented NW-SE has the main influence on the bedrock at the proposed repository depth. Based on a variety of geological and geophysical data, including reflection seismic data, geological models of different scales were developed for rock domains and deformation zones and, inside the local model volume, for dif-ferent fracture domains in the bedrock between the zones (Stephens et al., 2007). These models served as the basis for site evaluation and planning, and are relatively well constrained in the uppermost 1000 m in the crust immedi-ately southeast of the nuclear power plants, due to the extensive drilling to this depth in this area.

15

Fig

ure

2.1.

(a)

Det

aile

d be

droc

k ge

olog

ical

map

of

the

Fors

mar

k ar

ea (

afte

r S

teph

ens

et a

l., 2

008)

with

the

thre

e m

ain

defo

rmat

ion

zone

s,

incl

udin

g: th

e F

orsm

ark

defo

rmat

ion

zone

(F

DZ

), th

e E

ckar

fjär

den

defo

rmat

ion

zone

(E

DZ

), a

nd th

e Si

ngö

defo

rmat

ion

(SD

Z).

The

ref

lec-

tion

seis

mic

pro

file

s in

the

area

wer

e ac

quir

ed in

sta

ges

1 an

d 2

(see

Juh

lin a

nd S

teph

ens,

200

6; B

roje

rdi e

t al.,

201

3). (

b) T

otal

mag

netic

fi

eld

anom

aly

map

. Coo

rdin

ates

are

in th

e Sw

edis

h R

T90

sys

tem

. Mag

netic

dat

a ar

e ki

ndly

pro

vide

d by

the

Geo

logi

cal S

urve

y of

Sw

eden

.

16

3. Data acquisition

The data sets used in this thesis were acquired during four different data acquisition stages. The main seismic data were acquired in two stages 1 and 2. Paper I and III are case studies based on extensive seismic reflection sur-veying performed in stages 1 and 2. The seismic data in paper II were ac-quired along only two lines with the primary purpose to determine depth to bedrock. Paper IV is about VSP surveys which were acquired in boreholes KFM01A and KFM02A and focuses solely on the former, KFM01A.

3.1. Data acquisition for the reflection and three-component seismic profiles-stage 1 Reflection seismic data were acquired in the Spring of 2002 (March 10 to May 5) using an explosive source (Juhlin et al., 2002). About 16 km of high resolution seismic data were acquired along five separate profiles (P1, P2, P3, P4 and P5) ranging in length from 2 to 5 km. The acquisition parameters were optimized to provide high resolution images of the uppermost kilome-ter of crust, using 10 m source and receiver spacing with 100 active chan-nels. The shot points and geophones were located on the bedrock as much as possible. Shot holes were drilled at the closest suitable location to a stacked point where bedrock was present, but not farther than 30 cm parallel and 1 m perpendicular to the profile of the staked point. If no bedrock was found within this area, even after removing 50 cm of soil, the shot hole was drilled at the staked point. In bedrock, dynamite with a charge size of 15 g was placed in shot holes that were drilled with a diameter of 12 mm to 90 cm depth. In areas covered with soil, 75 g of dynamite charge was placed in shot holes with a diameter of 32 mm drilled to a depth of 150 cm. All shot holes and geophone positions were located using a high precision GPS instrument in combination with a total station. In addition, most of the acquisition was carried out with fourteen 3C receivers. All the 3C stations were positioned on bedrock outcrops to enhance data quality and to avoid complexities intro-duced by near surface conditions. Every station consisted of one portable Orion seismograph and a data recorder, one 3C seismometer, one GPS an-tenna, and one car battery powering them (Juhlin et al., 2002). Twenty se-cond data records were retrieved every day for every shot, and extracted

17

from hard drives. At the end, only twelve 3C stations recorded high enough quality data to be used in this study. Acquisition details for stage 1 (cabled-receiver reflection survey and 3C stations) are briefly summarized in Table 3.1.

Table 3.1. Data acquisition parameters for the cabled-receiver reflection survey (Juhlin et al., 2002) and 3C stations.

Parameter Cabled-reflection line 3C stations

Spread type End-on, shoot through 14 stationsNumber of channels 100 3 (N-S, W-E, Z)Near offset 20 m NAGeophone spacing 10 m NAGeophone type 28 Hz single 0.2 HzShot spacing 10 m 10 mCharge size 15/75 gram 15/75 gramNominal charge depth 0.9/1.5 m 0.9/1.5 mNominal CMP fold 50 1

Recording instrument SERCEL 348 Orion

Sample rate 1 ms 2 ms

Field low cut 8 Hz Out

Field high cut 250 Hz 125 Hz

Record length 4 s 20 s

Profile length / shots P1: 2950 m / 260 P2: 2740 m / 217 P3: 2050 m / 143 P4: 2410 m / 196 P5: 5280 m / 507

P1: 2950 m / 260 P2: 2740 m / 217 P3: 2050 m / 143 P4: 2410 m / 196 P5: 5280 m / 507

3.2. Data acquisition for the reflection seismic profiles-stage 2 Reflection seismic data were acquired in the Autumn of 2004 (9 October to 17 December), using both an explosive and mechanical hammer source (Juhlin and Palm, 2005). The explosive source used in this study was the same as the one described in stage 1 (Juhlin et al., 2002). A major difference compared to previous site investigations (stage 1) was the use of a VIBSIST source. The mechanical hammer source, or VIBSIST system (Park et al., 1996; Cosma and Enescu, 2001; Juhlin et al., 2002; Yordkayhun et al., 2009), was used for the first time for a surface seismic survey in Sweden. Since 2004 this source has been used extensively and has consistently pro-vided penetration to depths of 4-5 km (e.g., Juhlin et al., 2010; Lundberg and Juhlin, 2011). Reflection seismic data were acquired with the SERCEL

18

408UL system along ten different profiles (P2b, P4b, P5b, P6, P7, P8, P10, P11, P12 and P13), providing about 25 km of high resolution seismic data. For the VIBSIST source, 3 to 5 sweeps were recorded at each source point and were then used as input into the decoding software. Geophones were placed in drilled bedrock holes wherever possible. Otherwise, they were placed directly in the soil cover. High precision GPS instruments in combi-nation with a total station were used to survey all source points and geo-phone positions. Horizontal and vertical precisions greater than 10 cm were obtained using this combination. Acquisition details for stage 2 are summa-rized in Table 3.2.

Table 3.2. Data acquisition parameters for the reflection seismic profiles.

Parameter Value

Spread type Asymmetric split (40-120)Number of channels Minimum 160 / Maximum 276Near offset 0 mGeophone spacing 10 m / 5 m on profile 12Geophone type 28 Hz singleShot spacing 10 m / 5 m Profile 12Charge size 15 / 75 gramNominal charge depth 0.9 / 1.5 mNominal CMP fold 80

Recording instrument SERCEL 408

Sample rate 0.5 ms explosive / 1 ms VIBSIST

Field low cut Out

Field high cut 500 Hz explosive / 250 Hz VIBSIST

Record length 3 s / 30 s VIBSIST

Profile length / shots P2b: 2780 m / 191 P4b: 740 m / 79 P5b: 3360 m / 305 P6: 2900 m /255 P7: 4240 m / 287 P8: 2860 m / 196 P10: 4300 m / 235 P11: 1040 m / 68 P12: 1010 m / 150 P13: 1390 m / 99

3.3. Data acquisition for the reflection and refraction seismic lines The seismic data were acquired along two lines during the Summer of 2011 (August 15 to August 19) using an explosive source. Lines 2 and 4 are paral-lel to each other and have a length of about 300 m. Geophone spacing was 2

19

m while the source was located between every third geophone, giving a source spacing of 6 m. Charges of 66 g were used in shot holes 30 cm deep that were made by hand using a large steel rod. Given the simple source, the recorded seismic data are of good quality with clear first arrivals observed. When the source was active on Line 2, seismic recording was carried out on Line 2 and Line 4 and while the source was active on Line 4, the recording was done along Line 4 and Line 2. This arrangement allowed semi-3D cov-erage between the lines. In addition, data were also recorded on twelve 3-component MEMS (Micro Electro-Mechanical Systems) sensors that were placed between Line 2 and Line 4 and 12 wireless units along each line to extend them to the southeast with 10 m spacing between the geophones. The central line recorded data from sources both on Line 2 and Line 4. In this work, we did not consider the data from the wireless sensors. Instead, we produced two separate 2D traveltime inversions for Line 2 and Line 4. De-tails on acquisition parameters can be found in Table 3.3.

Table 3.3. Acquisition parameters for the reflection and refraction seismic along Line 2 and Line 4.

Parameter Value

Spread type FixedNumber of channels Line 2: 154, Line 4: 157Near offset 1 mGeophone spacing 2 mGeophone type Vertical, 28 Hz singleSource spacing 6 mCharge size 66 gRecording instrument SERCEL 428 ULSample rate 0.5 ms

Record length 3 s

Nominal fold 25

Field low cut Out

Field high cut 800 Hz

Profile length 2 ×300 m

Number of source points 80

3.4. Data acquisition for the vertical seismic profile (VSP) The data for the VSP survey were acquired in Summer of 2004 (August 2 to August 22) using the VIBSIST-1000 source, which is operated based on the Swept Impact Seismic Technique (SIST) (Park et al., 1996; Cosma and Enescu, 2001). The VSP surveys were conducted in boreholes KFM01A and

20

KFM02A, to a maximum depth of 775 m in each borehole. We focused on the borehole KFM01A with ten source points (1-10) located at different azi-muths and offsets around the borehole KFM01A. The distance from the top of the wellhead to the near offset source point 3 is about 38 m and for the far offset source point 7 is about 1168 m. The dipping angle of the borehole is 84.7° from the horizontal towards the west. The VIBSIST-1000 source uses a tractor/excavator-mounted hydraulic rock-breaking hammer, powered through a computer-controlled flow regulator, and in this survey generated about 100 impacts per sweep. The VSP survey was performed with two sources; one hammer was mounted on a 12 ton Huddig 960 tractor and an-other one an 18 ton Atlas 1504 excavator. The hammers delivered 1,500 and 2,500 Joules/impact, respectively, at 400-800 impacts/minute. The VIBSIST source was activated for a period of 15-20 seconds with different impact frequency ranges. The sweep was repeated five to ten times to obtain a high signal-to-noise ratio. For each day, the raw data files were backed up and decoded from long sweeps to primary raw data files. The 3C geophone chain Vibrometric R8-XYZ-C was used for recording the data. The X component was placed along the axis of the hole and the Y and Z components are ori-ented perpendicular to the X component. Each 3C module consisted of three 28 Hz geophones. Eight modules were used with 5 m distance and in total 136 levels were used along the borehole. For the purpose of a consistent directional response of 3C receivers, they remained fixed in the borehole at the same positions while the source moved to different source points. When two sources were used, the source points were assigned to two groups, each group containing at least four source points. The receivers were unclamped and moved only when the data had been recorded for all source points in both groups. The acquisition parameters for the borehole KFM01A are shown in Table 3.4.

Table 3.4. Acquisition parameters for the borehole KFM01A VSP survey.

Parameters Value

Number of source points 10 Number of receivers 136Near / far offset 38 m/ 1168 mGeophone spacing 5 mGeophone type 3C-28 HzRecording instrument PC-based acquisition systemSample rate 0.2 msRecord length 20 sField low cut 40 Hz

Filed high cut 1000 Hz

Borehole length 775 m

21

4. An introduction to anisotropy

Anisotropy implies that the properties of a medium vary depending upon which direction through the medium that the property is measured. Seismic anisotropy refers to the directional dependence of the velocity for both P- and S-waves and particle motion polarizations in the medium. At certain propagation directions the splitting of shear-waves will occur, as for example caused by the elastic properties of a fractured rock (Figure 4.1). Seismic anisotropy can be caused by many sources, such as crystal and mineral grain alignments from ductile deformation and metamorphism (e.g., Babuska and Cara, 1991; Malehmir et al., 2011, 2013), crack, and pore space alignments due to brittle deformation (e.g., Schild et al., 2001; Crampin and Peacock, 2005), grain-size layering (e.g., Crampin, 1993; Rowland et al., 1993), and induced anisotropy due to preferential stress, and aligned cracks caused by, for example, mining and unloading of materials (Crampin, 1987). Anisotro-py and shear-wave splitting analysis (e.g., Crampin, 1985; Crampin, 1987; Malehmir et al., 2013) may provide important information on the rock prop-erties in order to improve seismic data interpretation. The simplest aniso-tropic case of broad geophysical applications are layered rocks (Thomsen, 1986). Layered materials are considered transversely isotropic. The trans-verse isotropy (TI) is divided into two sub-groups with different alignments (symmetries). These two sub-groups are: (1) vertical axis of symmetry, which is called vertically transverse isotropic (VTI) and (2) horizontal axis of symmetry, which is called horizontally transverse isotropic (HTI) (Figure 4.2). In a VTI medium with horizontally layered rocks, elastic properties are uniform in the horizontal direction within the layers, but may vary vertically from one layer to another. In this case, compressional seismic waves gener-ally propagate faster horizontally than vertically. A VTI medium does not generate azimuthal anisotropy in the horizontal plane even though the medi-um is anisotropic. The presence of such an anisotropy system is challenging to detect from solely surface vertical component data, shear-wave splitting is expected to occur for horizontally propagating waves. A shear-wave (fast shear) will generally travel faster when polarized parallel to the layering than if polarized perpendicular to the layering (slow shear). Therefore, 3C data are ideal for understanding seismic anisotropy from this system. Identifying and quantifying this type of anisotropy is important for correlation purposes, such as comparing sonic logs in vertical and deviated wells and for compar-ing surface and borehole seismic data (Melaku, 2007).

22

In an HTI medium that is vertically fractured or with foliation oriented vertically (e.g., folded) the compressional wave velocity will generally be faster in the vertical direction parallel to these structures (bedding or fractur-ing) and slower in the direction perpendicular to them (horizontally) and across them. Therefore, compressional seismic waves travelling in the frac-ture direction generally travel faster than waves crossing the fractures. This variation depends on the degree of anisotropy and sometime a clear azimuth-al anisotropy can be observed (e.g., Bamford and Nunn, 1979); in this case P- and S-wave velocities vary as a function of azimuth. Observed azimuthal anisotropy may yield information about the rock stress and fracture density and orientation. Variations in shear-wave splitting can also be expected as a function of azimuth in this case and can provide information about the fast (e.g., fracture orientations) and slow (perpendicular) directions.

A combination of VTI and HTI systems (e.g., horizontally layered rocks and vertical fracture systems) results in a more complex symmetry called orthorhombic (Figure 4.2). This case can also be caused by either two or three mutually orthogonal fracture systems. Hence, orthorhombic anisotropy may be one of the simplest realistic symmetry for many geophysical prob-lems (Tsvankin, 1997). In this context, it is possible to imagine that at Forsmark steeply dipping shear zones (generating vertical shear fabrics or foliation; HTI model) and gently dipping fractures and joints (i.e. VTI mod-el) will form this system of anisotropy. However, the system's sensitivity to seismic waves needs to be studied, which is one of the objectives of this thesis.

Figure 4.1. Schematic figure showing shear-wave splitting occurring in the case of horizontally oriented layering or fracturing. The original S-wave which is polarized in a single direction enters the anisotropic medium. If the polarization is at an angle to the fractures or foliation then it will split into two components; fast and slow shear-waves. The fast shear-wave will polarize parallel to the fractures or foliation direction and slow component will polarize perpendicular to them. This can conse-quently result in the complete separation of these two components after sufficient time (delayed shear-wave arrivals).

23

Figure 4.2. (a) VTI anisotropy characterizes horizontal layering with a vertical axis of rotational symmetry that produces no azimuthal anisotropy on surface seismic data. (b) HTI anisotropy can represent vertical layering with a horizontal axis of symmetry that produces azimuthal anisotropy on surface seismic data. (c) A combi-nation of VTI and HTI symmetries leads to an orthorhombic system of anisotropy. The figure is modified from Kendall et al. (2014).

24

5. Summary of papers

This chapter presents a brief summary of the four papers that constitute the main part of this thesis. Each paper is summarized with its objectives, meth-ods, results and, finally, conclusions. A statement of my own particular con-tributions to each study is given here:

Paper I: I carried out the main processing of the seismic data, finalized the results and prepared the figures. My co-authors also contributed in this con-siderably given that this was my first processing experience. Discussions with the co-authors guided the processing and eventually produced improved results. I wrote the manuscript, however my co-authors greatly helped to improve the quality and clarity of the paper.

Paper II: I participated in the seismic data acquisition and preparation of the seismic data. In this work, I assisted with the first arrival traveltime to-mography work using the new seismic data. I wrote the manuscript with help from my co-authors. The comments from co-authors improved the quality of the paper and the interpretation of the results.

Paper III: After recovering the data in a format suitable for data analysis, I carried out the analysis, picked the arrivals and prepared most of the figures and after discussion with my co-authors, wrote the manuscript. My co-authors helped to improve the text, interpretations and presentation of the results with their comments and contributions in the writing.

Paper IV: After preparing the data and their geometry, I performed the main analysis of the VSP data including the picking, velocity calculations and rotations. I provided ideas on how to interpret the results. I wrote the manuscript with comments and suggestions for improvements provided by my co-authors.

5.1. Paper I: Reflection seismic imaging of the deeper structures at the Forsmark spent nuclear fuel repository site, central Sweden 5.1.1. Summary The main objective of this study was to reveal the deeper structures ranging from 1 to 5 km by passing lower frequencies through the processing flow at

25

the cost of poorer resolution for the near-surface structures. Extensive seis-mic reflection surveying was achieved in two stages at the Forsmark site. In stage 1, five profiles were acquired in the Spring of 2002 using an explosive source (Juhlin et al., 2002). In stage 2, ten additional profiles were acquired in the Autumn of 2004 using both explosive and mechanical hammer sources (Juhlin and Palm, 2005). This paper presents re-processed reflection seismic data from seven profiles 1, 2, 2b, 5, 5b, 7 and 10 that were acquired during stages 1 and 2 of the seismic investigations. The key re-processing steps were (1) slalom CDP stacking instead of straight CDP lines, (2) retaining lower frequencies in the spectral whitening and filtering steps and (3) dip moveout (DMO) corrections which were not attempted earlier. In addition to these, refraction static corrections were recalculated and partly contributed to the improved images of the deeper structures. Normal moveout (NMO) ve-locity had a significant effect on the stacked images at early times, but less so at later times. In order to avoid stacking in refracted waves, a front mute was applied to about 10 ms after the first arrivals. Two to three iterations of velocity analysis and residual statics were done for each profile. Two exam-ples of source gathers before and after some key processing steps are shown in Figure 5.1. Hints of reflections are already seen in the raw source gathers. These reflections become much more apparent after some processing steps, especially frequency filtering.

In order to provide an overall picture of major structures, results from profiles 2b, 2, 5 and 5b (Figure 5.2), profiles 7, 10, 2b and 2 (Figure 5.3) and profiles 7 and 1 were merged. This also helped to check the consistency be-tween reflections when lines were crossing each other. For example, the merged profiles 2b, 2, 5 and 5b cross all the three major deformation zones (SDZ, EDZ and FDZ) and merged profiles 7, 10, 2b and 2 cross the EDZ deformation zone.

The sub-horizontal to moderately dipping structures are more extensive in depth in the new re-processing work than those presented earlier (Juhlin and Stephens, 2006). Three main deeper reflective zones (A1, J (J1-J2) and C2) were better characterized, two of which dip moderately to the southwest and one that is sub-horizontal. The moderately dipping structure (A1 reflective) may exist across most of the area and could potentially cut all the three sub-vertical deformation zones at depth. Reflection sets J (J1-J2) originate from a moderately southwest dipping dolerite sill or possibly from a brittle fault system. The sub-horizontal C2 reflection may represent a 1.27-1.26 Ga dol-erite sill at about 3 km of depth. In this study, alternative geometric solutions for the relationship between the A1 structure and the steeper (at least at the surface) FDZ, EDZ and SDZ (Figure 5.4) were proposed. Although the new images and interpretation do not indicate the necessity for re-evaluating the Forsmark site for storage of spent fuel, they do influence the interpretation of deeper structures and, hence, the interpretation of the tectonic evolution of the area.

26

Figure 5.1. Example source gathers before and after some processing steps. (a) Raw gather using an explosive source from profile 5, (b) same gather after applying spec-tral whitening and bandpass filtering, (c) raw decoded gather from the VIBSIST source on profile 10, and (d) same gather after applying spectral whitening and bandpass filtering. Note the quality of the reflections marked by the arrows after the processing steps were applied to these data.

27

Figure 5.2. Merged stacked sections of profiles 2b, 2, 5 and 5b. The data have been combined where profiles cross one another (Figure 2.1) to check the consistency between reflections on the different profiles. Reflections A1, C2, F1 and J1 are dis-cussed in Brojerdi et al. (2013).

Figure 5.3. Stacked section of profile 10 merged with profiles 2b and 7. The data have been combined where profiles cross one another (Figure 2.1) to check the con-sistency between reflections on the different profiles. Reflections A1, J1 and J2 are discussed in Brojerdi et al. (2013).

28

Figure 5.4. Three conceptual models explaining the deeper structures of the Forsmark site. (a) The steeply dipping zones, FDZ, EDZ and SDZ with a dextral strike-slip component of movement (Saintot et al., 2011) extend to great depth with a near-vertical orientation and partition the A1 reflective zone between the EDZ and SDZ. This corresponds to the model outlined in Stephens et al. (2007). Reflectors C2 and J developed later and cut the SDZ (C2) and the EDZ (J). (b) Reflective zone A1 may be present below the entire area. Data quality and signal penetration limit the imaging of this structure. Reflector C2 does not appear to extend further to the southwest than that indicated here. The steeply dipping zones, FDZ, EDZ and SDZ are inferred to sole into more sub-horizontal structures at 3-5 km depth. (c) Similar to (b), but reflective zone A1 becomes more vertical with depth and is truncated by the FDZ. In this second more extensive model, the sub-vertical, regionally signifi-cant transpressive FDZ is the master zone, implying that the A1 structure forms part of a positive flower structure in this tectonic regime.

29

5.1.2. Conclusions Three main reflective zones (A1, J (J1-J2) and C2) were identified that are correlated over a number of the seismic profiles. Different conceptual mod-els for their relationships and subsurface geology were presented. In the first scenario, the A1 reflective zone is limited in its lateral extent and does not cross the main sub-vertical deformation zones at depth. In the second scenar-io, this zone is more extensive and may cross these zones at depth either partially or completely. The more extensive model implies that the sub-vertical zones sole into the A1 reflective zone and that the A1 structure rep-resents a major tectonic element in the Forsmark area. In addition, the more extensive model regards the sub-vertical regionally significant transpressive FDZ as the master zone while indicating that the A1 structure forms part of a positive flower structure in this tectonic regime. Since reflections with simi-lar apparent dip were observed on the northern end of the Dannemora P1 line (Malehmir et al., 2011) a further 4-5 km to the southwest (Figure 2.1), the possibility that this structure extends towards southwest across the FDZ is not excluded.

5.2. Paper II: High resolution seismic imaging at the planned tunnel entrance to the Forsmark repository for spent nuclear fuel, central Sweden 5.2.1. Summary The proposed repository for the storage of high-level radioactive spent nu-clear fuel at the Forsmark site is located in the northern end of a relatively undeformed tectonic lens at a depth of about 470 m. The facilities will in-clude a new access ramp, shafts, rock caverns, transport and deposition tun-nels, and deposition holes (SKB, 2009). To provide information at shallow depths for the preparation of the access ramp, a new high-resolution seismic survey was conducted during August 2011 using two reflection and refrac-tion seismic lines parallel to each other and about 300 m long each and spaced about 35 m apart. Geophone and source spacing were 2 and 6 m, respectively. An explosive source was used with data recording on all geo-phone stations simultaneously. Traveltime tomography was applied to the first break picks using a modified version of the code PStomo_eq (Benz at al., 1996; Tryggvason et al., 2002, 2009) that can also handle active source data (e.g., Bergman et al., 2004; Yordkayhun et al., 2009). Results from the first break traveltime tomography indicate a depth to bedrock that is greater than that found from geotechnical observations along the two lines (Figure 5.5). The reason for this discrepancy can be explained by the uppermost

30

bedrock being highly fractured and having a velocity significantly below that expected from the intact bedrock deeper down.

Given the tight spacing between the receivers and large number of source stations, reflection data processing was also conducted as part of this study. The main challenge in the reflection data processing was the severe effects of the near surface statics. Figure 5.6 shows the improvement in the coher-ency of the data in a selection of shot gathers after various processing steps (particularly the first arrival statics), indicating a reflection at about 20 ms (or about 60 m).

Figure 5.5. Traveltime tomography models projected onto the reflection seismic sections (unmigrated but time-to-depth converted) of (a) Line 2 and (b) Line 4. Note the sub-horizontal reflection (G8) at about 60 m. Dashed lines indicate depth of bedrock from statics while the total sounding bedrock depths are shown by red cir-cles. The closest cored boreholes KFM21 and KFM14 are projected into seismic Lines 2 and 4, respectively (Brojerdi et al., 2014).

31

Figure 5.6. (a) Raw source gather from Line 2 recorded on receivers active on Line 2 spread. (b) Same gather as (a) after static corrections. (c) Same gather as (b) after refraction statics and spectral equalization. (d) Raw source gather from Line 4 rec-orded on receivers active on Line 2 spread. (e) Same gather as (d) after static correc-tion. (f) Same gather as (e) after refraction statics and spectral equalization. Reflec-tions from c. 60 m depth are marked with arrows in both lines.

Although none of the cored boreholes actually penetrate the reflector where it is imaged by the seismic data, projections of core mapping onto the seis-mic profiles indicate the reflection at c. 20 ms is from a thin fracture zone. This fracture zone may be part of a larger fracture zone mapped by core drilling farther to the east. The newly mapped reflector may be crossed by the ramp when excavation begins. The tomographic results indicate veloci-ties of about 5500 m/s at their deepest level of penetration of about 25 m, showing that the G8 reflection (Figure 5.5) originates from well within the bedrock.

5.2.2. Conclusions Results from the first break tomography indicate an undulating bedrock sur-face, which is typical for the Forsmark area. If the 5000 m/s contour is con-sidered as the top of intact bedrock, then depths to bedrock are overestimated by up to 5-20 m compared to geotechnical observations. The discrepancy may be explained by the uppermost bedrock being highly fractured in the

32

survey area. When a lower velocity (3000 m/s) is set as a criterion for the presence of bedrock, the traveltime tomography results and the total sound-ing penetration tests (a method similar to cone penetration tests; see Nilsson and Forssman (2004)) are in better agreement. Lower velocity zones in the bedrock can be associated with depressions in the bedrock, suggesting these zones can be deemed as fracture zones. It may be that one pair of low veloci-ty zones can be correlated between Lines 2 and 4. The results do not show any features that necessitate changing the location of the proposed ramp entrance.

5.3. Paper III: Analysis of seismic anisotropy derived from sparse three-component receivers at the Forsmark spent nuclear fuel repository site, central Sweden 5.3.1. Summary In this study, the analysis of seismic anisotropy of large offset 3C seismic data was performed. Active explosive sources fired on five separate profiles (Juhlin et al., 2002), stage 1, were retrieved from fourteen passively operat-ed, sparse 3C receivers placed within the profiles. Prior to performing the data analysis, the horizontal components were rotated in the direction of the source and receiver azimuths (Figure 5.7). Clear P- and S-wave arrivals were observed at offsets up to 4 km. Large time delays in the shear-wave arrival times between the transverse and radial components, with the transverse generally having earlier arrival times were observed (Figure 5.8). No evi-dence of offset dependence in the time delays nor azimuthal variations were observed. The generally earlier arrival of the shear wave on the transverse component is best explained by sub-horizontally oriented highly fractured bedrock in the top 100 m, implying transverse isotropy with a vertical sym-metry axis.

33

Figure 5.7. Example receiver gathers as recorded on the (a) transverse component (H1), (b) radial component (H2) and (c) vertical component with their corresponding traveltime picks (red squares). Note the quality of the arrivals that were used in the anisotropy analysis.

Figure 5.8. Comparison of reduced traveltime versus azimuth for (a) horizontal components (transverse with the green color and radial with the red color), (b) time delay between the shear-wave arrivals as a function of both azimuth and offset, and (c) reduced traveltime for the vertical component data. Color shading ranges from light to dark showing offsets ranging from 500-1500 m, 1500-2500 m and 2500-5500 m, respectively (a and c).

34

A 1D P-wave velocity model based on the first arrival picks suggested that the velocity increases rapidly in the uppermost 100 m. Therefore, most of the observed shear-wave splitting was suggested to be due to a high density of sub-horizontal fractures in this depth range. To complement this analysis, conventional differential effective medium (DEM) theory to predict seismic velocities and anisotropy considering several different fracture aspect ratios

(McLaughlin, 1977; Mainprice, 1997; Mavko et al., 2009) (Figure 5.9) was tested. Shear-wave anisotropy was calculated using: 200 / . To obtain the observed time delays between the shear-wave components, a large aspect ratio was concluded to be required for reasonable porosities.

Figure 5.9. Shear-wave velocity as a function of variation in fracture aspect ratio

and volume of fractures, obtained by using the differential effective medium (DEM) theory. A Cartesian coordinate system , , is used to define direc-tional dependence of seismic velocities, where - is oriented in the plane normal to the VTI symmetry axis (i.e., horizontal plane) and is oriented parallel to the axis of symmetry in the VTI medium. (a) P-wave velocity along the -axis and in the - symmetry plane. (b) S-wave velocity along the -axis and in the - symmetry plane. Maximum P- and S-wave anisotropy are shown in (c) and (d), respectively.

35

5.3.2. Conclusions Shear-wave traveltime delays were presented as evidence for seismic anisot-ropy in the sparse 3C data. The anisotropy was attributed due to the upper-most 100 m of bedrock being highly jointed and fractured, with these fea-tures having a preferred sub-horizontal orientation. No evidence of azimuth-al anisotropy was observed, although acquisition of a more extensive data set could reveal some azimuthal dependence. Combined with the effective me-dium modeling using available data from the site, it can be argued that trans-verse isotropy (TI anisotropy system) is present, but not due to fabrics asso-ciated with the deformation zones rather due to stress control on the upper-most fractured interval in the study area. Based on the modeled delay times, it can be deduced that the uppermost region contains c. 1-2 % water-filled fractures (or effective porosity) and that in the deeper levels of the bedrock (>100 m) the volume of these open fractures is likely smaller.

5.4. Paper IV: Analysis of P- and S-waves in vertical seismic profile data acquired at the Swedish site for storage of spent nuclear fuel, Forsmark 5.4.1. Summary The main research aim of this study was to check if a multi-offset-azimuth vertical seismic profiling (VSP) survey acquired down to a depth of 775 m in borehole KFM01A (Figure 2.1) could validate or refine the anisotropy model obtained in Paper 3. Ten-source points were activated around the borehole and 3C geophones were used for the measurements. Seismic velocities and traveltime delays, after correcting for the orientation of the receivers, were calculated. The data were rotated into a coordinate system aligned with the borehole for better picking of arrival times of the P-wave on the axial com-ponent and the shear-wave on the transverse and radial components. All source points show strong P-wave energy in the axial component and S-waves primarily on the radial and partially on the transverse components. Figure 5.10 shows the vertical and horizontal components before and after rotation for azimuth and incident angle for one of the ten source points.

36

Figure 5.10. Example source gather data as recorded on the vertical and horizontal components before rotation (top panel) and axial, transverse and radial components after rotation (bottom panel). Red squares are the picked traveltimes. The example shows clear and strong P-wave arrivals on the axial component after the rotation. A bulk shift of -50 ms and a source static correction were applied to the rotated data prior to their display.

Velocity information was obtained from the picked data assuming a straight path from the source to the receivers. Then interval velocities using a 6-layer model as a function of depth were estimated. Source static corrections, ob-tained using one of the far offset shots, for the P-wave were applied to the rotated data to remove the effect of the near-surface in the vicinity of the source points. Source static corrections for the shear-wave were applied us-ing a value twice the corresponding P-wave source static correction (e.g., Malehmir et al., 2015). Then P- and S-wave continuous average velocities were calculated for all source points (calculated by dividing distance by traveltime corrected for statics). Both the interval and average velocities show the greatest variations above 500 m depth. The velocity-depth curves indicate the largest difference between the two S-wave components occurs at the shallowest levels. Shear-wave anisotropy was then calculated for all the source points except for one of them, because of poor quality data at that location. Similar to the results concluded in the third paper, the analysis in-dicates no clear azimuthal variations and no evidence of an offset depend-ence in shear wave arrival delays. The low degree of anisotropy between the two shear wave components in the far offset data may be due to oblique div-

37

ing shear-waves that are less sensitivite to the sub-horizontal fractures in the uppermost few hundreds of meters. To complement the velocity analysis, linear moveout (LMO) with an average transverse component velocity of

3500 m/s was applied to both the transverse and radial components to observe the behavior of the shear-waves with respect to this correction. The LMO corrected data indicate large traveltime delays for shallow receivers (or depths) that may be explained by significant sub-horizontal fracturing in the uppermost couple of 100 or so meters. Figure 5.11 displays an example LMO-corrected source gather with very clear delayed arrivals, particularly in the radial component.

Figure 5.11. Example LMO corrected source gather. Figure shows the transverse and radial components with their corresponding traveltime picks (red squares). Large delays in arrival times on the radial component for receiver locations down to 20 (200 m along the borehole) are obvious.

38

5.4.2. Conclusions The analysis of the 3C-VSP data shows clear delayed arrivals between the two horizontal components with a lower degree of anisotropy at greater depths. This work may be used to complement the third paper that also indi-cated shear-wave splitting and anisotropy in the uppermost hundred meters. Calculated velocity functions indicate a rather rapid increase in velocity for P- and S-waves down to about 500 m depth. While below this depth the ve-locities increase less rapidly. The results suggest that the closing of cracks due to increased stress are responsible for the velocity increase, with sub-horizontal cracks closing preferentially with depth.

39

6. Conclusions and recommendations

The presented articles illustrate different approaches applied to the seismic data acquired to study geophysical and geological structures in the Forsmark area. The methods included the re-processing of existing reflection seismic data, acquisition and imaging of high-resolution reflection and refraction seismic data in order to better understand the near surface and deep struc-tures of the site, and also surface and borehole seismic data analysis to ex-tract information about potential seismic anisotropy in the area. They cover various scales and depth regions; but, like studies presented in papers 3 and 4, are complementary and support each other.

In Paper I, results from the re-processing of the seismic data in the Forsmark area allowed for improvements in imaging in deeper depth inter-vals (1-5 km) by shifting the processed signal band to lower frequencies. Three main deeper reflective zones were identified, one that is sub-horizontal and two that dip moderately to the southwest. In this work, it was observed that:

The slalom CDP stacking lines maximized the amount of infor-

mation that could be obtained from the seismic data.

Careful selection of filters improves images of deeper structures. In Paper II, results from the first break traveltime tomography and reflection seismic processing of the two about 300 m long and parallel profiles near the planned ramp entrance were presented. The simultaneous data recording while shooting in one line while also recording on the other line allowed obtaining additional information between the two lines and also helped to treat the data as a semi-3D dataset. If possible, similar data acquisition ap-proaches are recommended when detailed delineation of the subsurface is one of the objectives. Thanks to the 3D nature of the data, reflection seismic processing of the data revealed a sub-horizontal reflection at about 20-25 ms (about 60 m) that could be traced under much of the surveyed area. The traveltime tomography allowed delineating depth to bedrock and to map potential fracture zones in the bedrock to ensure that no unexpected defor-mation zones would be encountered in the vicinity of the planned ramp en-trance. It was observed that:

40

The estimated depth to bedrock could reliably be identified, but care must be taken in case the bedrock is highly fractured. Bore-hole or sounding data are then important to calibrate such results.

Increasing the number of receivers, modifying the acquisition ge-ometry of refraction surveys, would allow additional information to be obtained through reflection seismic processing of the data. It is possible that a sub-horizontal fracture system or joint is running in where the access ramp will be constructed. This remains to be verified when the excavation starts.

In Paper III, results from the 3C data helped to identify possible seismic anisotropy in the area due to brittle structures. Analyzing the time delays between shear-wave arrivals recorded on the horizontal components provid-ed key evidence for sub-horizontal structures generating seismic anisotropy. These observations may be significant for the repository site since it implies that the fracture zones in the uppermost 100 m can be mapped as these are the most important structures at the seismic scale down to the repository depth. These fracture zones are kept open by horizontal stresses that are greater than the vertical stress. Two sources for seismic anisotropy in the area were postulated; vertical shear zones along with penetrative vertical tectonic foliation and sub-horizontal open and hydraulically conductive frac-ture zones and joints. The implication is that:

If a full 3D-3C azimuth-offset data acquisition is not possible, at least sparse 3C data acquisition could be attempted. This would however never replace a proper 3D-3C data acquisition. The cur-rent 3C data lack good azimuthal and offset converges and thus complimentary studies should be done to investigate if azimuthal seismic anisotropy is present in the area.

In Paper IV, results from the 3C-VSP data were presented and indicated that the differences in arrival times for the two shear-waves are best explained by a high density of sub-horizontal fractures at shallow levels. This is consistent with the results obtained using the 3C sparse receivers and also fracture modeling studies (Olofsson et al., 2007) independently conducted in the study area.

From all these studies and my experience gathered during the PhD time, I conclude that with additional efforts and time spent on processing steps that may look trivial, one would be able to improve either the imaging of shallow or deeper structures that together may provide a better explanation of the geology than one alone. While many interesting findings were observed, there are still many possibilities for future improvements, especially using a

41

combination of high-resolution reflection seismics and refraction seismics for site characterizations. This may be particularly important in the Forsmark site where the bedrock is often highly fractured and jointed. High-resolution 3D reflection seismic techniques are suggested to overcome 3D effects and better characterize the shallow subsurface; this has not yet been attempted at Forsmark in a proper scale. They can be employed for mapping diverse facies and complex geometries of the near surface. Responding to the need for cost effective 3D seismic surveying, the potential and limitations of dif-ferent source-receiver acquisition patterns by appropriately decimating a comprehensive shallow 3D seismic reflection data could also be considered.

42

7. Summary in Swedish

Forsmark är beläget ca 140 km norr om Stockholm i de centrala delarna av Sve-rige. Detta område består av paleoproterozoisksvekokarelisk bergrund tillhöran-de den del av den fennoskandiska skölden som bildades för ca 1.9 till 1.8 miljo-ner år sedan. Tre huvudsakligen brantstående, plastiska och spröda deforma-tionszoner med västnordvästlig eller nordvästlig strykning är utmärkande i om-rådet, dessa är Forsmarkzonen, Eckarfjärdenzonen och Singözonen. Berggrunden mellan dessa zoner är generellt mindre deformerad och anses lämpliga för slutförvar av kärnbränsle. Svensk Kärnbränslehantering AB (SKB) har nu valt detta område för ett framti-da slutförvar av högaktivt använt kärnbränsle i Sverige. Flera undersökningar genomfördes i området innan Forsmark valdes som plats för slutförvar. Ca 16 km högupplöst seismiskt data samlades in längs fem separata profiler under våren 2002 (etapp 1). Längs dessa profiler användes dynamit som källa. Ytter-liggare tio profiler (2 till 5 km) mättes in under hösten 2004 (etapp 2). Vid dessa mätningar användes dels dynamit, men också en mekanisk hammare som seis-misk källa. Olika seismiska experiment utfördes med fokus på den översta kilo-metern av berggrunden. Det huvudsakliga syftet med denna avhandling är att analysera reflektionsseismisk data över geologiskt intressanta områden i Fors-mark. Omfattande seismiska undersökningar, tillsammans med data från borr-hålskärnor och hydrauliska tester ned till 1000 m djup, ger data för den översta kilometern av jordskorpan. I Artikel I omprocesserades seismisk data från sju profiler insamlade under etapp 1 och 2. Ändamålet med denna omprocessering var att förbättra bilden på de djupare delarna (1-5 km). Omprocesseringen möjliggjorde en förbättrad bild av tre djupare reflektioner, varav en är flack medan de andra har måttlig lutning. Dessa reflektioners ursprung tolkas som antingen diabasgångar eller spröda förkastningssystem.

I Artikel II analyserades seismiskt data längs två nya profiler (profil 2 och profil 4 sommaren 2011) både med gångtidstomografi och med 3D reflektionsseis-misk processering. Resultatet från tomografin visade på en undulerande berg-grundstopografi, vilken är typisk för området i Forsmark. Ytliga reflektioner på ett djup av mindre än 100 meter kunde, tack vare mätningens design, jämföras med existerande borrhålsdata och härledas till svaghetszoner i berggrunden.

43

Artikel III och IV syftar att analysera seismisk anisotropi, dessa tyder på klyv-ning av skjuvvågor, men inte azimutiskt varierande, och relativt begränsad till de hundratals översta metrarna. Flacka öppna sprickor och “joints” tolkas vara orsak till stora fördröjningar i ankomsten av de olika komponenterna av skjuv-vågorna, både på data insamlat på ytan och i borrhål.

44

Acknowledgments

I think it's time to look back at the past when I made the decision to do something different in my life: moving to another country alone. Living here taught me that as long as we live, we are learning. Now, I have been given the opportunity to express my gratitude to everyone who helped me scientif-ically and emotionally throughout this trip.

First of all, I would like to express my deepest sense of gratitude to my supervisor Christopher Juhlin. Chris: Thanks so much for giving me the opportunity to work as a PhD student in the Geophysics group at Uppsala University. Your endless patience, priceless advices, suggestions and guid-ance, offered with extreme kindness, were always there to support me. I believe that without your help, this work would never have been completed.

I am extremely grateful to Alireza Malehmir, my co-supervisor; a huge thank you for your support, for answering all possible questions, for a warm welcome to me whenever needed, and for your friendship. Your presence in the corridor always made me happy.

I would like to thank my second co-supervisor Bjarne Almqvist. Bjarne: you entered the stage of this work when everything seemed so hard, thank you for bringing the light; Thank you!

Special thanks to Hans Palm, our fieldwork manager. Hasse: You taught me so many things about reflection seismic data acquisition during the fieldworks. I remember that it was the second day during the fieldwork when I was not feeling very confident about the work. I was alone in the ‘White House’ during the day and I asked you for some help. Your answer was “we are very busy, you can do it alone”. Yes I did it, but I didn’t believe myself. You made me brave, thank you so much!

I would like to thank Laust Pedersen for his very useful course, inversion theory.

I would like to acknowledge Ari Tryggvason, for giving me very helpful comments in using the PStomo_eq code.

I am grateful to Fengjiao Zhang and Zeinab Jeddi for all your comments related to the tomography and other useful discussions.

To the reviewers of this thesis, Chris, Alireza, Bjarne, and Omid; Thank you so much for the great help and the time given.

I would like to thank Emil and Karin for translating the summary in Swe-dish, I really appreciate it.

45

Special thanks to Azita (Mahdieh Dehghannejad). I cannot imagine better luck than getting to meet you during my very first moments at this universi-ty! I thank you very much for the great time we shared together, for showing me everything I needed to know about living here, for helping me a lot in learning seismic processing. And, of course, thank you for your friendship. :)

I would like to thank my officemate, Monika Ivandic, for being a great friend! Thanks so much for the nice moments together, for all the laughs and fun we had, our running, parties, and for being so generous in helping me with my seismic questions. I really enjoyed my time with you in the Office EK 226. :)

I would like to thank my Iranian friends and their families for their friend-ship: Faramarz Nilfouroushan, Mehrdad Bastani, and Taher Mazloomian.

I have lots of nice memories of my friends at coffee breaks, the confer-ences we attended together, meetings, fieldworks, and parties. I would like to thank all of my friends with whom I have spent some very nice moments: Aggela, Arash, Arnaud, Azita, Bojan, Chunling, Claudia, Daniel, Darina, Dibakar, Emil, Fann, Fei, Fredrick, Georgiana, Giulia, Hamzeh, Harbe, Ida, Joachim, Jochen, Juliane, Karin, Magnus, Mahboubeh, Maria, Maryeh, Milad, Monika, Nicoleta, Omid, Peter, Ping, Rémi, Samar, Shunguo, Silvia, Soroor, Suman, Sverker, Théo, Ted, View, Yang, Zhang, Zeinab and Zoya (wow, I hope I have mentioned all of you!). ;))

At Uppsala University, I would like to thank Fatima, Sive, Susanne, Taher, and other people in the administration and the support staff at the Department of Earth Sciences for their help.

The biggest thanks goes to my family; my mother (Nayereh), father (Behzad), my sister (Farnaz) and her husband (Shahryar). We cannot choose our family, but if I would have been given the possibility, I would have cho-sen you a million out of a million times. I need to truly thank you for all the moral support, for sharing your love of life, sense of humor, excitement for every experience, for giving me the support that I needed to build a dream to chase after and for spending countless hours of fun with me. I believe there is no place like home and nothing sweeter than those memories. I miss them and I’ll always keep them close to my heart.

Last but not least, I want to thank my love, Reza, from the bottom of my heart. I cannot express how much my life has changed since I met you here. Thanks for holding my hand, for showing the real meaning of life to me and for living your life with me. I can't thank you enough for your kindness. The only thing I can say is that “I love you and I will love you forever”.

Sahar

Fatemeh Sharifi Brojerdi, April, 2015

46

References

Andersson, J., Skagius K., Winberg A., Lindborg T., and Ström, A., 2013. Site-descriptive modelling for a final repository for spent nuclear fuel in Sweden. Environmental Earth Sciences, 69, 1045-1060.

Babuska, V., and Cara, M., 1991. Seismic Anisotropy in the Earth; Modern Ap-proaches in Geophysics, Vol. 10, Springer, Netherlands.

Bamford, D., and Nunn, K.R., 1979. In Situ Seismic Measurements of Crack Anisot-ropy in the Carboniferous Limestone of Northwest ENGLAND. Geophysical Prospecting, 27, 322-338.

Benz, H.M., Chouet, B.A., Dawson, P.B., Lahr, J.C., Page, R.A., and Hole, J.A., 1996. Three dimensional P- and S-wave velocity structure of Redoubt volcano, Alaska. Journal of Geophysical Research, 101, 8111–8128.

Bergman, B., Tryggvason, A., and Juhlin, C., 2004. High-resolution seismic tomog-raphy incorporating static corrections applied to a till covered bedrock environ-ment. Geophysics, 69, 1082–1090.

Brojerdi, F.S., Juhlin, C., Malehmir, A., and Stephens, M.B., 2013. Reflection seis-mic imaging of the deeper structures at the Forsmark spent nuclear fuel reposi-tory site, central Sweden. Journal of Applied Geophysics, 89, 21–34.

Brojerdi, F.S., Zhang, F., Juhlin, C., Malehmir, A., Lehtimäki, T., Mattsson, H., and Curtis, P., 2014. High resolution seismic imaging at the planned tunnel entrance to the Forsmark repository for spent nuclear fuel, central Sweden. Near Surface Geophysics, 12, 709-719.

Brojerdi, F.S., Juhlin, C., Malehmir, A., and Almqvist, B.S.G., 2015. Analysis of seismic anisotropy derived from sparse three-component receivers at the Forsmark spent nuclear fuel repository site, central Sweden. Submitted.

Cosma, C., and Enescu, N., 2001. Characterization of fractured rock in the vicinity of tunnels by the swept impact seismic technique. International Journal of Rock Mechanics and Mining Sciences, 38, 815–821.

Crampin, S., 1985. Evaluation of anisotropy by shear-wave splitting. Geophysics, 50, 142–152.

Crampin, S., 1987. Crack porosity and alignment from shear-wave VSPs, in Shear-wave Exploration: Geophysical Developments, Danbom, S.H. and Domenico, S.N. (Ed.), 227- 251. SEG Special Publications.

Crampin, S., 1993. Arguments for EDA. Canadian Journal of Exploration Geophys-ics, 29, 18–30.

Crampin, S., and Peacock, S., 2005. A review of shear-wave splitting in the compli-ant crack-critical anisotropic Earth. Wave Motion, 41, 59-77.

Glamheden, R., Fredriksson, A., Persson, L., Röshoff, K., Karlsson, J., Bohlin, H., Lindberg, U., Hakami, H., Hakami, E., and Johansson, M., 2007. Rock mechan-ics Forsmark, Site descriptive modelling Forsmark stage 2.2. SKB (Swedish Nuclear Fuel and Waste Management Co.), Report no. R-07-31.

Hermansson, T., Stephens, M.B., Corfu, F., Page, L.M., and Andersson, J., 2008. Migratory tectonic switching, western Svecofennian orogen, central Sweden:

47

constraints from U/Pb zircon and titanite geochronology. Precambrian Re-search, 161, 250–278.

Juhlin, C., Bergman, B., and Palm, H., 2002. Forsmark site investigation- Reflection seismic studies in the Forsmark area - stage 1. SKB (Swedish Nuclear Fuel and Waste Management Co.) Report no. R-02-43.

Juhlin, C., and Palm, H., 2005. Reflection seismic studies in the Forsmark area, 2004: Stage 2. Forsmark site investigation. SKB (Swedish Nuclear Fuel andWaste Management Co.), Report no. R-05-42.

Juhlin, C., and Stephens, M.B., 2006. Gently dipping fracture zones in Paleoproterozoic metagranite, Sweden: evidence from reflection seismic and cored borehole data and implications for the disposal of nuclear waste. Journal of Geophysical Research, 111, B09302.

Juhlin, C., Dehghannejad, M., Lund, B., Malehmir, A., and Pratt, G., 2010. Reflec-tion seismic imaging of the end-glacial Pärvie Fault system, northern Sweden. Journal of Applied Geophysics, 70, 307–316.

Kendall, J-M., Verdon, J.P., and Bair, A.F., 2014. Evaluating fracture-induced ani-sotropy using borehole microseismic data. Canadian Society of Exploration Ge-ophysicists Recorder, 39, No. 02.

Leijon, B. (Ed.), 2005. Forsmark site investigation. Investigations of superficial fracturing and block displacements at drill site 5. SKB (Swedish Nuclear Fuel and Waste Management Co.) Report no. P-05-199.

Lundberg, E., and Juhlin, C., 2011. High resolution reflection seismic imaging of the Ullared Deformation Zone, southern Sweden. Precambrian Research, 190, 25–34.

Mainprice, D., 1997. Modeling the anisotropic seismic properties of partially molten rocks found at mid-ocean ridges, Tectophysics, 279, 161-179.

Malehmir, A., Dahlin, P., Lundberg, E., Juhlin, C., Sjöström, H., and Högdahl, K., 2011. Reflection seismic investigations in the Dannemora area, central Sweden: insights into the geometry of polyphase deformation zones and magnetite-skarn deposits. Journal of Geophysical Research, 116, B11307.

Malehmir, A., Andersson, M., Lebedev, M., Urosevic, M., and Mikhaltsevitch, V., 2013. Experimental estimation of velocities and anisotropy of a series of Swe-dish crystalline rocks and ores, Geophysical Prospecting, 61. 153-167.

Malehmir, A., Wang, S., Lamminen, J., Brodic, B., Bastani, M., Vaittinen, K., Juhlin, C., and Place, J., 2015. Delineating structures controlling sandstone-hosted base-metal deposits using high-resolution multicomponent seismic and radio-magnetotelluric methods: a case study from Northern Sweden, Geophysi-cal Prospecting, DOI: 10.1111/1365-2478.12238.

Martin, C. D., 2007. Quantifying in situ stress magnitudes and orientations for Forsmark, Forsmark stage 2.2. SKB (Swedish Nuclear Fuel and Waste Man-agement Co.), Report no. R-07-26.

Mavko, G., Mukerji, T., and Dvorkin, J., 2009. The Rock Physics Handbook, 511pp., Cambridge Univ. Press, New York.

McLauglin, R., 1977. A study of the differential scheme for composite materials. International Journal of Engineering Science, 15, 237-244.

Melaku, M. T., 2007. Velocity anisotropy of shales and sandstones from core sam-ples and well logs on the Norwegian continental shelf. MSc Thesis, Department of Geosciences, University of Oslo.

Nilsson, G., and Forssman, G., 2004. Total Sounding – Some experience in Sweden and development potential. NGM 2004 XIV Nordic Geotechnical Meeting, vol 1, s C27 – C34, NGM, Ystad.

48

Olofsson, I., Simeonov, A., Stigsson, M., Stephens, M.B., Follin, S., Nilsson, A.-C., Röshoff, K., Lindberg, U., and Lanaro, F., 2007. Site descriptive modeling Forsmark, stage 2.2 – A fracture domain concept as a basis for the statistical modeling of fractures and minor deformation zones, and interdisciplinary coor-dination, SKB R-07-15, Svensk Kärnbränslehantering AB.

Park, C.B., Miller, R.D., Steeples, D.W., and Black, R.A., 1996. Swept Impact Seismic Technique (SIST). Geophysics, 61, 1789–1803.

Rowland, H., Booth, D., and Chiu, J.M., 1993. Shear-wave splitting from micro earthquake in the New Madrid seismic zone. Canadian Journal of Exploration Geophysics, 29, 352–362.

Saintot, A., Stephens, M.B., Viola, G., and Nordgulen, Ø., 2011. Brittle tectonic evolution and paleostress field reconstruction in the southwestern part of the Fennoscandian Shield, Forsmark, Sweden. Tectonics, 30, TC4002.

Sandström, B., Tullborg, E.-L., Larson, S.Å., and Page, L., 2009. Brittle tectonothermal evolution in the Forsmark area, central Fennoscandian Shield, recorded by paragenesis, orientation and 40Ar/39Ar geochronology of fracture minerals. Tectonophysics, 478, 158–174.

Schild, M., Siegesmund, S., Vollbrecht, A., and Mazurek, M., 2001. Characteriza-tion of granite matrix porosity and pore-space geometry by in situ and laborato-ry methods. Geophysical Journal International, 146, 111-125.

Stephens, M.B., Fox, A., La Pointe, P., Simeonov, A., Isaksson, H., Hermanson, J., and Öhman, J., 2007. Site descriptive modelling, Forsmark stage 2.2. Geology. SKB (Swedish Nuclear Fuel and Waste Management Co.), Report no. R-07-45.

Stephens, M.B., Bergman, T., Isaksson, H., and Petersson, J., 2008, Bedrock geolo-gy Forsmark. Modelling stage 2.3. Description of the bedrock geological map at the ground surface. SKB (Swedish Nuclear Fuel and Waste Management Co.), Report no. R-08-128.

Stephens, M.B., Ripa, M., Lundström, I., Persson, L., Bergman, T., Ahl, M., Wahl-gren, C.-H., Persson, P.-O., and Wickström, L., 2009. Synthesis of the Bedrock Geology in the Bergslagen Region, Fennoscandian Shield, south-central Swe-den. SGU (Geological Survey of Sweden) Ba 58. Geol. Surv. of Sweden, Upp-sala, Sweden.

Stephens, M.B., Follin, S., Petersson, J., Isaksson, H., Juhlin, C., and Simeonov, A., 2015. Review of the deterministic modelling of deformation zones and fractures domains at the site proposed for a spent nuclear fuel repository, Sweden, and consequences of structural anisotropy, accepted in Tectonophysics.

Swedish Nuclear Fuel and Waste Management Co. (SKB), 2008. Site description of Forsmark at completion of the site investigation phase: SDM-Site Forsmark. Report no. TR-08-05.

Swedish Nuclear Fuel and Waste Management Co. (SKB), 2009. Underground de-sign Forsmark Layout D2. Report no. R-08-116.

Swedish Nuclear Fuel and Waste Management Co. (SKB), 2011. Site selection — siting of the final repository for spent nuclear fuel. Report no. R-11-07.

Söderlund, P., Hermansson, T., Page, L.M., and Stephens, M.B., 2009. Biotite and muscovite 40Ar-39Ar geochronological constraints on the post-Svecofennian tectonothermal evolution, Forsmark site, central Sweden. International Journal of Earth Sciences, 98, 1835−1851.

Talbot, C.J., and Sokoutis, D., 1988. The Singö shear zone near Östhammar, eastern Sweden. UUDMP Res. Rep. 55. Uppsala Univ., Uppsala. 55 pp.

Talbot, C.J., and Sokoutis, D., 1995. Strain ellipsoids from incompetent dykes: ap-plication to volume loss during mylontization in the Singö gneiss zone, central Sweden. Journal of Structural Geology, 17, 927–948.

49

Thomsen, L., 1986. Weak elastic anisotropy. Geophysics, 51, 1954–1966. Tryggvason, A., Rögnvaldsson, S.T., and Flovenz, Ó.G., 2002. Three-dimensional

imaging of P- and S-wave velocity structure and earthquake locations beneath southwest Iceland. Geophysical Journal International, 151, 848–866.

Tryggvason, A., Schmelzbach C., and Juhlin, C., 2009. Traveltime tomographic inversion with simultaneous static corrections – well worth the effort. Geophys-ics, 74, WCB74–WCB55.

Tsvankin, I., 1997. Anisotropic parameters and P-wave velocity for orthorhombic media. Geophysics, 62, 1292 –1309.

Yordkayhun, S., Ivanova, A., Giese, R., Juhlin, C., and Cosma, C., 2009. Compari-son of surface seismic sources at the CO2 SINK site, Ketzin, Germany. Geo-physical Prospecting, 57, 125–139.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1259

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science andTechnology, Uppsala University, is usually a summary of anumber of papers. A few copies of the complete dissertationare kept at major Swedish research libraries, while thesummary alone is distributed internationally throughthe series Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology.(Prior to January, 2005, the series was published under thetitle “Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-251621

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2015