multidisciplinary near-surface investigation of a quick-clay...

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1775

Multidisciplinary Near-SurfaceInvestigation of a Quick-ClayLandslide Prone Area in SouthwestSweden

SILVIA SALAS ROMERO

ISSN 1651-6214ISBN 978-91-513-0575-2urn:nbn:se:uu:diva-375568

Dissertation presented at Uppsala University to be publicly examined in Hambergsalen,Geocentrum, Villavägen 16, Uppsala, Friday, 29 March 2019 at 10:00 for the degree ofDoctor of Philosophy. The examination will be conducted in English. Faculty examiner:Associate Professor Laura Valentina Socco (Department of Environment, Land andInfrastructure Engineering, Politecnico di Torino, Torino, Italy).

AbstractSalas Romero, S. 2019. Multidisciplinary Near-Surface Investigation of a Quick-ClayLandslide Prone Area in Southwest Sweden. Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology 1775. 70 pp. Uppsala: ActaUniversitatis Upsaliensis. ISBN 978-91-513-0575-2.

Quick-clay landslides are considered one of the most important geohazards in Sweden, Norway,and Canada. The deposits involved are glacial and postglacial clays and silty clays, which arevery sensitive to increased stress that may collapse their structure and cause liquefaction.

A multidisciplinary approach was adopted in this study of quick clays in an area ofsouthwest Sweden that is prone to landslides. It was mainly based on geophysical methods,but was complemented and validated with geotechnical, geological, and hydrological data.Downhole geophysics, land and river reflection seismics, radio magnetotellurics, P-waverefraction tomography, magnetics, and multichannel analysis of surface waves comprised themain geophysical methods used in this research. Laboratory measurements of core samples, suchas grain size analysis, mineral magnetic properties, fossil content, X-ray fluorescence, cationexchange capacity, X-ray powder diffraction, electrical conductivity, and pH were also taken.Hydrological modelling was used to obtain information on the properties of groundwater withina coarse-grained layer, given the importance of this information in the formation of quick clays.

The evaluation of the physico-chemical properties of the coarse-grained layer revealedhigh values of magnetic susceptibility, probably as a result of fluvial sorting, which tends toaccumulate denser minerals such as magnetite. Potential quick clays were visually observedabove this layer, and their presence was also confirmed by geotechnical data acquired inprevious studies. Marine fossils identified within the coarse-grained layer confirmed theglaciomarine origin of the clays. Geophysical results revealed the presence of large-scalestructures, an undulating fractured bedrock and a coarse-grained layer sandwiched between claydeposits, with leached sediments on top and unleached sediments below. This layer, importantfor the development of quick clays in the area, was 3D modelled in a regional context. Magneticdata also revealed that the coarse-grained layer together with quick clays, have the potential toact as a sliding prone layer. Multichannel analysis of surface waves helped to geotechnicallycharacterize the area. Although the results could not reach the deeper layers, it did yieldinformation about the shallower layers of clay, silt and sand. The VS30 values indicated thepresence of soft soils, as well as soft clays/silts with high plasticity index and high water content.

This PhD is expected to improve the current knowledge of quick-clay landslides and howthey are preconditioned. Climate change will probably affect the study area in the near future,most likely increasing landslide risk, therefore, research should continue and advance to newlevels. Application of other geophysical methods such as borehole or geophysical monitoring,and induced polarization, could provide more information about the formation of quick claysand associated landslides.

Keywords: Quick-clay, reflection seismic, landslide, southwest Sweden, borehole logging,physico-chemical properties, hydrogeological modelling, multichannel analysis of surfacewaves, magnetics

Silvia Salas Romero, Department of Earth Sciences, Geophysics, Villav. 16, UppsalaUniversity, SE-75236 Uppsala, Sweden.

© Silvia Salas Romero 2019

ISSN 1651-6214ISBN 978-91-513-0575-2urn:nbn:se:uu:diva-375568 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-375568)

Dedicated to my family, especially to Tristan

Supervisor Professor Alireza Malehmir Department of Earth Sciences – Geophysics, Uppsala University, Uppsala, Sweden Assistant Supervisors Professor Ian Snowball Department of Earth Sciences – Natural Resources and Sustainable Develop-ment, Uppsala University, Uppsala, Sweden Professor Christopher Juhlin Department of Earth Sciences – Geophysics, Uppsala University, Uppsala, Sweden Faculty Opponent Associate Professor Laura Valentina Socco Department of Environment, Land, and Infrastructure Engineering Politecnico di Torino, Torino, Italy Examination Committee Dr. Gunnel Göransson Swedish Geotechnical Institute, Linköping, Sweden Professor Hemin Koyi Department of Earth Sciences – Mineralogy, Petrology and Tectonics, Uppsala University, Uppsala, Sweden Associate Professor Sara Bazin Department of Geosciences – Physical Geography and Hydrology, University of Oslo, Oslo, Norway

List of Papers

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

I Salas-Romero, S., Malehmir, A., Snowball, I., Lougheed, B. C., and Hellqvist, M. (2016) Identifying landslide preconditions in Swedish quick clays–insights from integration of surface geo-physical, core sample- and downhole property measurements. Landslides, 13(5), 905–923. https://doi.org/10.1007/s10346-015-0633-y

II Salas-Romero, S., Malehmir, A., Snowball, I., and Dessirier, B. (2019) Subsurface structures of a quick-clay sliding prone area revealed using land-river reflection seismic data and hydrogeo-logical modelling. Solid Earth Discuss. https://doi.org/10.5194/se-2019-22 (in review)

III Salas-Romero, S., Malehmir, A., Snowball, I., and Brodic, B. (2019) Geotechnical site characterization using multichannel analysis of surface waves–a case study of an area prone to quick-clay landslides in southwest Sweden (manuscript)

Reprints were made with permission from the respective publishers. Addition-ally, other publications written during my PhD studies, but not included in the thesis are:

• Salas-Romero, S., Malehmir, A., and Snowball, I. (2013) Analysis

of seismic data and correlation with downhole geophysical meas-urements in the assessment of a Swedish area prone to quick-clay landslides. SEG Technical Program Expanded Abstracts 2013, 1939–1943. https://doi.org/10.1190/segam2013-1000.1

• Malehmir, A., Salas-Romero, S., Shan, C., Lundberg, E., Juhlin, C., Bastani, M., Persson, L., Krawczyk, C., Polom, U., Adamczyk, A., Malinowski, M., Gurk, M., and Ismail, N. (2013) A multidisci-plinary geophysical and geotechnical investigation of quick-clay

landslides in Sweden. SEG Technical Program Expanded Abstracts 2013, 1254–1258. https://doi.org/10.1190/segam2013-0211.1

• Salas-Romero, S., Malehmir, A., and Snowball, I. (2016) Com-bined land and river high-resolution reflection seismic imaging of an area prone to quick-clay landslides in Sweden. In 22nd European Meeting of Environmental and Engineering Geophysics (Near Sur-face Geoscience 2016), Extended Abstract. https://doi.org/10.3997/2214-4609.201601912

Contributions

The papers included in this thesis are the result of close collaborations with several colleagues. My individual contributions in each paper are summarized below:

I I participated in the fieldwork campaign of 2013, during which

core samples, seismic and downhole logging data, were col-lected. With the help of B. Brodic, I collected all the core sub-samples that were subsequently analysed in the lab. I performed part of the laboratory measurements, along with B. C. Lougheed and M. Hellqvist. I processed and modelled all the laboratory re-sults and downhole geophysical data. With the help of the co-authors, I interpreted all the integrated data (including other types of data provided by SGI and C. Shan). I prepared all the figures, wrote the first draft of the manuscript, and worked alongside the co-authors of this paper to make significant changes to the text and improve its clarity.

II I processed all but one of the seismic lines collected during the fieldwork campaign of 2013, plus the two river lines collected by SGU in this area with the help of one of the co-authors, A. Malehmir. I modelled the 3D elevation surfaces for the top and bottom of the coarse-grained layer and the top of bedrock, ob-tained by interpolating the seismic horizons picked along all the seismic lines, resistivity, borehole and LiDAR data. I prepared the data as required for the hydrological modelling, which was performed by one of the co-authors, B. Dessirier. I performed the analysis, correction and processing of the magnetic data. I inter-preted all the integrated data (including other types of data pro-vided by SGU, SGI, and S. Wang) with the help of my co-au-thors. I prepared all the figures, wrote the first draft of the manu-script, and worked on it together with the rest of authors to make significant changes to the text and improve its clarity.

III I performed multichannel analysis of surface waves in all the

seismic lines collected during the fieldwork campaign of 2013. I

processed and interpreted the results, and obtained new data from them. I prepared all the figures and wrote the first draft of the manuscript. All the authors, including myself, worked on this draft to make significant changes to the text and improve its clar-ity.

Contents

1 Introduction ............................................................................................131.1 Motivation and objectives ................................................................131.2 Outline of the thesis .........................................................................14

2 Study area...............................................................................................152.1 Geology and geomorphology ...........................................................152.2 Previous studies...............................................................................202.3 Recent studies .................................................................................20

3 Quick clays .............................................................................................223.1 What are quick clays? ......................................................................223.2 Physico-chemical properties ............................................................223.3 Areas susceptible to contain quick clays ..........................................23

4 Methodology ..........................................................................................244.1 Geophysical measurements ..............................................................24

4.1.1 Borehole logging......................................................................244.1.2 Reflection seismics ..................................................................244.1.3 Magnetics ................................................................................254.1.4 Multichannel analysis of surface waves ....................................26

4.2 Laboratory measurements ................................................................264.2.1 Electrical conductivity, pH and salinity ....................................264.2.2 Mineral magnetic properties .....................................................274.2.3 Grain size analysis ...................................................................274.2.4 Macro- and microfossil identification .......................................274.2.5 X-ray fluorescence ...................................................................284.2.6 X-ray powder diffraction ..........................................................284.2.7 Cation exchange capacity .........................................................28

4.3 Other type of data used in the interpretation .....................................284.3.1 RMT ........................................................................................284.3.2 P-wave refraction tomography and full-waveform inversion .....284.3.3 CPT and CPTU-R ....................................................................29

4.4 Hydrogeological modelling .............................................................29

5 Summary of papers .................................................................................305.1 Paper I: Identifying landslide preconditions in Swedish quick clays–insights from integration of surface geophysical, core sample- and downhole property measurements ...................................................30

5.1.1 Summary .................................................................................305.1.2 Conclusions .............................................................................36

5.2 Paper II: Subsurface structures of a quick-clay sliding prone area revealed using land-river reflection seismic data and hydrogeological modelling ..............................................................................................37


5.3 Paper III: Geotechnical site characterization using multichannel analysis of surface waves–a case study of an area prone to quick-clay landslides in southwest Sweden .............................................................45


6 Outlook: Amplitude Versus Offset ..........................................................526.1 Motivation and objectives ................................................................526.2 A simplified model ..........................................................................526.3 AVO modelling ...............................................................................536.4 AVO analysis ..................................................................................56

6.4.1 Synthetics ................................................................................566.4.2 Reflection seismic data.............................................................58

6.5 Conclusions and further investigations .............................................58

7 Conclusions ............................................................................................60

8 Summary in Swedish ..............................................................................62

Acknowledgements ...................................................................................64

References .................................................................................................66

Abbreviations

1C 3C 2D 3D Al AGC API ATEM AVO BGA Ca CDP CMP CPT CPTU-R CSTMT E EM ERT Fe g/cm3

g/l G GPR GPS GWB Hz K kg km kN kPa LIAG LiDAR m MASW

Single-component Three-component Two-dimensional Three-dimensional Aluminium Automatic gain control American Petroleum Institute Airborne transient electromagnetic Amplitude versus offset Branschens Geotekniska Arkiv Calcium Common depth point Common midpoint Cone penetration test Cone penetration test - resistivity measurements Controlled-source tensor Young modulus Electromagnetics Electric resistivity tomography Iron Gram per cubic centimetre Gram per litre Shear modulus Ground-penetrating radar Global positioning system Geoscientists Without Borders Hertz Potassium Kilogram Kilometre Kilonewton Kilopascal Leibniz Institute for Applied Geophysics Light detection and ranging Metre Multichannel analysis of surface waves

MEMs Mg MPa mT nT m/s Na NGI NMO NORSAR ppm rms RMT s SEG SGI SGU St

VP VS XRD XRF

Micro-electro mechanical system Magnesium Megapascal Militesla Nanotesla Metre per second Sodium Norwegian Geotechnical Institute Normal moveout Norwegian Seismic Array Parts per million Root mean square Radio magnetotellurics Second Society of Exploration Geophysicists Swedish Geotechnical Institute Geological Survey of Sweden Sensitivity Compressional-wave velocity Shear-wave velocity X-ray powder diffraction X-ray fluorescence

13

1 Introduction

1.1 Motivation and objectives Near-surface geophysical methods have proved to be useful for investigating ultrashallow and near-surface materials (Büker t al., 2005; Cardimona et al., 1998; Donohue et al., 2012; Schmelzbach et al., 2005). Furthermore, the com-bination of several methods, such as reflection seismic and resistivity, has been shown to result in a more complete interpretation for environmental ap-plications (Malehmir et al., 2016). A multidisciplinary approach using near-surface geophysical, geotechnical and geological investigations has success-fully been applied to study quick clays (Bastani et al., 2017; Dahlin et al., 2013; Malehmir et al., 2016; Lundström et al., 2009; Rankka et al., 2004; Sol-berg et al., 2012; Sauvin et al., 2013 and 2014; Wang et al., 2016).

Quick clays, which include sensitive glacial and postglacial sediments, can be found in Sweden, Norway, Finland, Canada, United States (Alaska), Russia or Japan; in some of these countries quick-clay landslides are considered one of the most important geohazards, being responsible for many deaths and sig-nificant economic losses. Catastrophes such as Rissa (Gregersen, 1981), Tuve (Larsson and Jansson, 1982), or more recent ones like Saint-Jude (Locat et al., 2017) or Munkedal (Nadim et al., 2008) show the importance of studying this type of landslides.

According to the model predictions for climate change over the next 100 years, Sweden is likely to become warmer and wetter (Swedish Government Official Reports, 2007). It is probable that precipitation will increase in the west of the country, resulting in more runoff and higher river discharge. This change in river conditions may create unstable slopes, and hence, increase the likelihood of landslides in both new areas and areas already prone to slope failure (Swedish Government Official Reports, 2007). Furthermore, a range of studies have investigated the effects of climate change on landslides in other parts of the world (e.g. Jakob & Lambert, 2009; Schildgen et al., 2016). Hence, the potential effects of climate change on landslides is important globally, as well as within Sweden. Therefore, an important area for research is to better understanding the processes of landslide initiation and how climate change could worsen the landslide risk.

This thesis focuses on studying quick clays in an area located in southwest Sweden, and should be considered within a Swedish context. Quick clays may have different properties in other countries, e.g. Canada or Norway.

14

Nevertheless, methodologies and interpretations may be extrapolated to other places in the world where similar postglacial landscapes can be found. The main objectives of this thesis are (i) to study the response of quick clays in a range of different geophysical methods, (ii) to study properties of quick clays and their relationship with associated materials, such as coarse-grained mate-rials, (iii) to make an integrated interpretation of the subsurface structures us-ing multidisciplinary data, and (iv) to provide a geotechnical characterization of the subsurface materials in a pilot study area in southwest Sweden.

1.2 Outline of the thesis The thesis is composed of two parts, first, eight chapters summarizing the most important aspects of the research conducted during my PhD, and second, three papers that include most of the work. Chapter 2 describes the geology and geomorphology of the study area, and also the previous and more recent stud-ies performed there. Chapter 3 deals with the fundamentals of quick clays es-sential to these investigations. The methodologies applied during the study (e.g. acquisition and processing) are explained in Chapter 4. Chapter 5 in-cludes the summaries of each paper; all three papers are focused on multidis-ciplinary investigations, mostly geophysical (e.g. borehole logging, reflection seismic and multichannel analysis of surface waves–MASW), in an area prone to quick-clay landslides in southwest Sweden. The outlook focused on AVO (amplitude versus offset) analysis is described in Chapter 6, and the final two chapters cover the conclusions and summary of the thesis in Swedish.

Note that the research on AVO analysis is not included in any of the papers, and should be considered as an experiment to better understand quick clays and their associated materials in the study area.

15

2 Study area

2.1 Geology and geomorphology The research area, Fråstad is located to the north of the municipality of Lilla Edet, in southwest Sweden. The Göta River, which crosses this area, flows from Lake Vänern (the largest lake in Sweden) to and through Gothenburg on the west coast of Sweden. The river is around 93 km long, and acts, as the source of drinking water for approximately 700,000 people, as a transport route, and the main outflow of Lake Vänern (Swedish Geotechnical Institute–SGI, 2012a). Many landslide scars can be found along the Göta River valley, and most of them are thought to be quick-clay landslides (SGI, 2012a). Apart from the evident risk for human lives and economic losses, a large landslide mass could dam the river, affect transportation, or even contaminate the water (for example if the landmass contains contaminated soil, Göransson et al., 2009).

Figure 2.1 shows several photos from a landslide that affected a sulfite fac-tory located next to the Göta River (in the southern part of Lilla Edet) in June 1957. The retrogressive landslide, which caused three deaths and structural damage, transported large amounts of materials into the river, produced a wave of 5 to 8 m height and covered approximately 32 hectares (Húlten, 2006; Odenstad, 1958). The infiltration of chemicals into the ground (such as sul-fite), could have reduced the clay shear strength, which, in addition to the ero-sion at the riverbank and the presence of quick clay, may have triggered the landslide (Odenstad, 1958).

16

Figure 2.1. Different perspectives of the landslide that affected a sulfite factory lo-cated next to the Göta River in Lilla Edet (June 1957). Photos: Edet Group’s ar-chive.

17

The sediments within the study area are composed mostly of glacial clays, with patches of postglacial and fluvial sediments containing clay, silt, and sand, and sandy-silty till in a few places. In a number of places the bedrock is exposed, where outcrops of granite and granodiorite are visible (© SGU, Fig-ure 2.2a). Many landslide scars can be observed at the river shorelines (Figure 2.2a–b). The 2011 3D survey (Lundberg et al., 2014; Shan et al., 2016) was done at the position of a bottleneck landslide scar (Figure 2.2a–b). The Göta River valley at the study area shows a number of morphological or geological lineaments, which are typically fracture zones (Figure 2.2a). A bedrock fault zone follows the river channel (SGI, 2012b).

Based on LiDAR elevation data (© Lantmäteriet, Figure 2.2b) the topog-raphy reaches almost 100 m above sea level at the exposed bedrock outcrops. However, most of the landscape has an elevation of between 20 and 28 m. Next to the river and gullies the elevation is below 20 m (lines 6 and 7 are located in these low-elevated areas in the western side of the study area, where a lateral spread landslide scar lies, Demers et al., 2017).

The landscape of the Göta River valley is the product of erosional and land-slide processes (SGI, 2012a). Figure 2.3 provides three examples of how the riverbed and riverbanks appear using side-scan sonar (© SGU) and bathymet-ric (© SGI) data. Figure 2.3a–e belongs to the area next to the bottleneck land-slide scar, and Figure 2.3f–j and Figure 2.3k–m are located slightly south of the study area (see Figure 2.2c). Figure 2.3 shows steep slopes at the riverbanks, subaquatic landslide scars and deposits (from landslides and/or tributaries or gullies) on the riverbanks or at the centre of the riverbed.

18

Figure 2.2. Quick-clay study location. (a) Geological map of the study area (© SGU). The legend on the right-hand side shows the geological materials. The legend below the map shows the symbols for different geological structures and features, and the investigations carried out by Uppsala University, SGU and SGI. (b) LiDAR elevation map of the study area (© Lantmäteriet). SGI boreholes are represented as black and white circles, and Uppsala University boreholes as blue circles. (c) Sketch of the Göta River that includes the position of the study area (Figure 2.2b), and the position of Figure 2.3. The red arrows indicate the beginning and end of the river seismic lines.

19

Figure 2.3. Side-scan sonar data (© SGU) overlaid on the bathymetric data (© SGI) on three segments of the Göta River. See position in Figure 2.2c. (a) Cross section next to the study area, showing the position of line 5–5b. Scale 1H:8V. (b) Elevation along profile AA’. (c) Slope (absolute values) along profile AA’. (d) Elevation along profile BB’. (e) Slope (absolute values) along profile BB’. (f) Cross section of the river. Scale 1H:3V. (g) Elevation along profile AA’. (h) Slope (absolute values) along profile AA’. (i) Elevation along profile BB’. (j) Slope (absolute values) along profile BB’. (k) Cross section of the river. Scale 1H:4V. (l) Elevation along profile AA’. (m) Slope (absolute values) along profile AA’. Areas delineated with a dashed red line show accumulated material at the bottom of the river, and areas delineated with a dashed blue line show probable subaquatic landslide scars.

20

2.2 Previous studies The study area was initially investigated by SGI, who acquired geotechnical, geophysical and hydrogeological data (Dahlin et al., 2013; Löfroth et al., 2011). These investigations were conducted as part of a project to map areas next to the Göta River that were considered to contain quick clays. The main conclusions of these studies were (i) that the clay was more leached in the eastern part than the western part of the study area, (ii) that potential quick clays can be found above a coarse-grained layer or in the upper section of the soil in the western part, (iii) that the presence of quick clay was only confirmed in several points, which may be due to the high amount of magnesium ions in the pore water. The increased thickness of the quick clays was correlated with increased thickness of the coarse-grained layer, (iv) that saline clays were found above bedrock, and (v) that there is a downward groundwater move-ment.

In year 2011, a joint geophysical project led by Uppsala University, spon-sored by the Geoscientists Without Borders (GWB) program of Society of Ex-ploration Geophysicists (SEG) was initiated. This project had the main objec-tive of studying quick-clay landslides using geophysical methods and to make an integrated interpretation together with other types of data to characterize the subsurface down to bedrock. Among the collaborators of this initial project were SGU, Leibniz Institute for Applied Geophysics (LIAG), University of Cologne, Syiah Kuala University, Polish Academy of Sciences, Norwegian Seismic Array (NORSAR), Norwegian Geotechnical Institute (NGI), Institute for Geosciences at the University of Oslo, Geotechnical Group at the Norwe-gian University of Science and Technology and Geological Survey of Nor-way. 2D and 3D P- and S-wave reflection and refraction seismic, electric re-sistivity tomography (ERT), radio magnetotellurics (RMT), controlled-source tensor magnetotellurics (CSTMT), ground-penetrating radar (GPR), ground magnetics and gravity measurements were surveyed. Figure 2.2 shows the po-sition of the seismic, RMT and CSTMT measurements. These surveys (Adamczyk et al., 2013 and 2014; Comina et al, 2017; Malehmir et al. 2013a and 2013b; Lundberg et al., 2014; Polom et al., 2013; Shan et al., 2014 and 2016) helped to image the subsurface in good resolution and suggested coarse-grained layers underlying potential quick clays or quick clays in some places.

2.3 Recent studies In year 2013, a new fieldwork campaign within the scope of the same project and with extra funding from the Trust2.2-GeoInfra project was carried out, extending the initial study area to the north and south. Several geophysical measurements (2D P-wave reflection seismic, RMT, electromagnetics–EM and borehole logging) were performed. In order to ground truth previous

21

geophysical interpretations, three boreholes (BH1 to BH3, Figure 2.2b) were drilled down to bedrock using the SONIC technique (Ewing, 2009). Core sam-ples were recovered in each borehole with minimum disturbed conditions, however, samples were not always retrieved down to the depth of the bedrock. In BH1 and BH3 the core length was less than 40 m, and the drill length about 36 and 48 m respectively. The core and drill length in BH2 were 62 and 78 m, respectively. The boreholes were typically located on a seismic line, or at the intersection between seismic lines; BH1 is close to lines 1 and 2, BH2 on line 4 and BH3 on line5. BH2 and BH3 were closer to the river and to the bottle-neck landslide scar located in the area. Figure 2.4a shows part of the landslide scar and Figure 2.4b a quick-clay core sample. PVC casing was installed in each borehole, including a perforated section of 2 m at the suspected depth of the coarse-grained materials for two reasons. First, it was necessary to have direct downhole logging access to this formation, and second, to allow the groundwater to infiltrate the cased hole to allow its properties to be measured.

In addition to downhole geophysical measurements, core samples were used for laboratory measurements, such as grain size analysis, mineral mag-netic properties, fossil identification and X-ray powder diffraction (XRD). The measurements were carried out at Lund University, the University of Trier and Uppsala University. Processing and analysis of the data were performed in successive years, and are the central subjects of Papers I, II and III presented in this thesis.

Figure 2.4. (a) Landslide scar at the study area. (b) Quick-clay sample recovered in BH1. Photos: Silvia Salas-Romero.

22

3 Quick clays

3.1 What are quick clays? Quick clays are sensitive glacial and postglacial clayey sediments, deposited in a shallow marine or fresh water environment (Osterman, 1963; Torrance, 2012). In Scandinavia, quick clays are mainly glaciomarine sediments depos-ited during the deglaciation of the Fennoscandian ice sheet, between ca. 14,000 and 10,000 years ago (De Geer, 1940). During the ablation of the ice sheet, dispersed silt and clay particles flocculated in the saline marine envi-ronment. The flocculates formed clay-rich deposits with high water content, and after being isostatically uplifted above the sea level, their marine salts were gradually leached out by infiltrated meteoric waters, and their soil prop-erties were altered (Rosenqvist, 1953). When quick clays are under increased stress or their shear strength is reduced, they may liquefy and form a landslide if a minimum slope angle is exceeded and there is a place to flow to. Heavy rainfall, high-low water river flow, river erosion, changes in groundwater level, or even constructions or loading/unloading may also trigger a landslide (Thakur et al., 2014).

3.2 Physico-chemical properties To define a clay sample as quick clay, undisturbed sampling and laboratory geotechnical measurements are necessary. Using these methods, the sensitiv-ity (St) can be estimated, which is the ratio of undrained undisturbed to re-moulded shear strength (Skempton, 1953). In Sweden quick clays are defined as clays with St > 50 and remoulded shear strength < 0.4 kPa.

The preconditions to the formation of Swedish quick clays (Paper I), which may help identifying them are:

• Flocculated structures (Torrance, 2012), • The presence of low-activity, non-swelling secondary clay miner-

als, such as illite and chlorite (Quigley, 1980; Torrance, 2012), • Primary minerals such as quartz or feldspar (Quigley, 1980; Tor-

rance, 2012), • Low concentration of pore-water cations (Andersson-Sköld et al.,

2005; Rosenqvist, 1953; Torrance, 2012),

23

• Higher water content than their liquid limit (Torrance, 2012), • Basic environment (Bjerrum, 1967; Torrance and Pirnat, 1984), • Salinity values below 2 g/l (Torrance 1974 and 1979), • The presence of organic dispersants (Söderblom, 1966), • The presence of iron oxides (Torrance, 2012).

3.3 Areas susceptible to contain quick clays Rankka et al. (2014) provided an extensive review of quick clays in Sweden, describing the geological and geohydrological requirements for their for-mation (some of these conditions are discussed in Paper II). These are:

• Areas with glaciomarine sediments, • Thin clay deposits (leaching of salts is slower in thicker clay depos-

its), • Underlying coarse-grained materials (these permeable layers are

conduits for fresh water), • Undulating bedrock that can concentrate groundwater (leaching in-

creases at these positions), • Artesian groundwater pressure (higher pressure if permeable layers

underlie clay deposits), • Permeable layers intercalated with clay deposits, • Height above sea level (elevated sediments may have been above

sea level for a longer time and thus have been leached for a longer time),

• Organic soils, • A large catchment area, • Multidirectional infiltration.

The multidisciplinary approach that is used in this thesis for studying an area prone to quick-clay landslides is a good example of how to combine different near-surface data in a local or regional context. The results show a more com-prehensive description of the subsurface, improving our knowledge about quick clays and the landslide hazard assessment.

24

4 Methodology

A brief description of each method or type of data set used during my research is given in this chapter.

4.1 Geophysical measurements

4.1.1 Borehole logging Borehole measurements were performed using ‘temperature-conductivity’ and ‘full-waveform triple sonic’ probes in each borehole and included natural gamma radiation, fluid temperature, fluid conductivity and interval transit time, which serves for estimating slowness (the reciprocal of the velocity).

Natural gamma values are higher in fine-grained clay or shale deposits, where radioactive elements tend to accumulate, than in sands or gravels (RG WinLogger, 2007). Soil textures were classified on the basis of the natural gamma radiation data and visual observations of the core samples.

The PVC casing did not allow the correct measurement of sonic velocities. The casing had higher velocity than the surrounding materials, and hence, no information from the sediments was retrieved except at the perforated section in the casing. At this position, the sonic measurements provided a mixed ve-locity between the casing and the surrounding materials, the latter having more influence.

4.1.2 Reflection seismics Reflection seismic surveys allow to obtain information about the Earth’s sub-surface contained in recorded reflected wavefield. A controlled seismic source is used for transmitting the energy into the ground, and receivers on land or water (which are able to detect ground motion or pressure disturbance) record the reflected energy over a specific duration.

Paper II shows reflection seismic data collected on land and along the Göta River. Land seismic data were collected in year 2013, whereas river seismic data were acquired from the survey vessel (SV) Ocean Surveyor operated by SGU in year 2000.

Four land reflection seismic lines were acquired, lines 2b, 5b, 6 and 7, to-talling a length of around 3.8 km. The data extended the previous study area

25

(the 2011 survey) along the N-S direction. Cabled geophones of 28 Hz (4 m receiver spacing) were used in all the lines, except in the northern part of line 5b, where wireless stations were deployed every 10 m. These stations con-sisted of alternating single-component (1C) receivers of 10 Hz and three-com-ponent (3C) broadband digital MEMs (micro-electro mechanical systems) sensors. Dynamite was used as the energy source in line 5b every 20 m. In line 2b a 45-kg accelerated weight drop was used, and in lines 6 and 7 a 5-kg sledgehammer was used. More details on data acquisition can be found in Pa-per II.

Two types of river reflection seismic data were recorded, single-channel (3.5 kHz echo sounder) and six-channel data (Figure 2.2a–c). Both lines had similar lengths and ran parallel. Shot and receiver positions were the same in the single-channel data, measuring every 3 m. In the six-channel data the max-imum offset is 21 m, and the receiver spacing 3 m. A 10 in3 sleeve gun was used as the seismic source. More details on data acquisition can be found in Paper II.

Reflection seismic data processing for both types of data followed standard procedures (see details in Paper II). Land seismic data for lines 2b and 5b were merged with those of their respective lines collected in 2011. Lines 2–2b, 6 and 7 were processed in a similar way, while line 5–5b required further pro-cessing. The wireless data were processed separately as the distance between the stations was longer than between the cabled geophones, and hence, a dif-ferent geometry was needed.

The processing of the river reflection seismic data was simpler, and con-sisted mainly on the removal of multiple reflections.

4.1.3 Magnetics During a magnetic survey the strength of the Earth’s magnetic field is meas-ured and mapped. The aim of this survey was to acquire information about the magnetic properties of the crustal rocks and reveal, e.g., bedrock topography.

In Paper II, ground magnetic results are shown for an area initially covered during the fieldwork campaign of 2011 (Malehmir et al., 2013a). Total-field magnetic and vertical gradient were measured using a walking mode GPS-mounted magnetometer over the course of five days. A base station was in-stalled every day to correct for diurnal variations and instrumental drift (a background, International Geomagnetic Reference Field–IGRF, value of 50600 nT, corresponding to Earth’s magnetic field at the site was subtracted from these measurements to produce to residual magnetic anomaly data).

Nevertheless, an initial exploratory analysis of the data highlighted the need to apply more specific corrections for removing several inconsistencies, such as multiple values from repeated measurements at the same position, lev-elling issues between measurement days, and elongated features parallel to the measuring paths. These inconsistencies could have been due to following

26

different directions during the measurements, and/or a metallic objects carried inadvertently by surveyors during data acquisition. The corrections involved the addition or subtraction of a constant value to the measurements, and changing the polarity. The elongated features were filtered following the mi-cro-levelling procedure (Minty, 1991), which resulted in a smoother and more homogeneous magnetic map than before, with some residual errors remaining that seemed to coincide with the sampling directions.

4.1.4 Multichannel analysis of surface waves This geophysical method allows to obtain a VS (shear-wave velocity) model of the subsurface, which can provide an indication of the stiffness of the ground. Combined with other physical properties, such as VP (compressional-wave velocity) and density, a geotechnical characterization of the survey site is feasible.

The reflection seismic data used in this research were not collected for the purpose of MASW, and hence, the MASW results may reach shallower depths compared to other studies.

The reflection seismic data required some previous preparations before MASW could be applied. For example, the record length had to be extended through recovering the original data, near offsets were removed according to the horizontal travelling plane-wave criteria, body waves and noise were muted, and a division of the seismic data in smaller sections of 100 m each was performed.

MASW was carried out using RadExProTM software. The procedure in-volves picking the dispersion curves where possible, and then inverting these to obtain 1D VS profiles, which are combined to produce a pseudo-2D velocity profile of the subsurface. More details can be found in Paper III.

4.2 Laboratory measurements Detailed descriptions on how the sediments were subsampled and prepared, may be consulted in Paper I. Samples from BH2 were analysed for all the laboratory measurements, since the longest core was retrieved from this bore-hole and was the closest to the river. Nevertheless, certain measurements such as electrical conductivity, pH, or mineral magnetic properties were taken for the three boreholes.

4.2.1 Electrical conductivity, pH and salinity Electrical conductivity measures/quantifies the amount of electric current that a material can transmit. Salinity measures the amount of dissolved salts in the water, and is related to electrical conductivity (which is also influenced by the

27

amount of dissolved ions in the water). Conductivity measurements (Paper I) were performed using a Thermon OrionTM 115AplusTM device.

pH provides information on the hydrogen ion content in a solution. A pH100 model (VWR InternationalTM) was used for taking measurements.

4.2.2 Mineral magnetic properties Mineral magnetic properties can provide information about past environmen-tal conditions (Thompson & Oldfield, 1986). The studied parameters included magnetic susceptibility (which indicates the degree of magnetization induced in a material in response to an external magnetic field), anhysteretic remanent magnetization (ARM) which is the remanent magnetization due to the summed effects of an intense alternating field and a weaker direct current field), saturation isothermal remanent magnetization (SIRM) which is the as-sumed maximum remanent magnetization induced by a strong magnetic field, and S-ratio (which indicates the relative amounts of ferrimagnetic and antifer-romagnetic minerals). The equipment used and procedures followed are de-scribed in detail in Paper I.

4.2.3 Grain size analysis A Micromeritics SediGraph III 5120 particle size analyser was used for the grain size analysis in the <63 µm fraction. Samples belonging to the >63 µm fraction were sieved out prior to the grain size analysis and weighted. Grain size intervals were defined according to the Wentworth grain size scale.

Given that grain size analysis data were only available for BH2, approxi-mated grain size distributions for BH1 and BH3 were derived from other prop-erties using linear regression (see more details in Paper I). Considering that all boreholes were drilled in similar material, the estimated linear relation be-tween grain size and natural gamma radiation in BH2 was used to determine the percentages of clay in the other two boreholes from their respective natural gamma radiation values. Similarly, the >63 µm fraction in BH1 and BH3 was determined using the linear relationship between grain size and magnetic sus-ceptibility in BH2. The silt content in BH1 and BH3 was obtained by subtract-ing the sum of the clay and >63 µm fraction percentages from 100 %.

4.2.4 Macro- and microfossil identification Biological, fossil, content of the samples was identified using a light micro-scope, literature references and compared to reference materials, such as mod-ern barnacles.

28

4.2.5 X-ray fluorescence X-ray fluorescence (XRF) serves to determine the relative abundances of the most common elements found in sediment samples, such as Al, Si, Mg or K. XRF spectra were obtained using a Thermo Fisher Scientific Niton XL3t handheld XRF analyser.

4.2.6 X-ray powder diffraction XRD provides qualitative and quantitative information on crystalline materi-als (e.g. clay minerals). All measurements were performed with a Siemens XRD Diffractometer D500.

4.2.7 Cation exchange capacity Cation exchange capacity (CEC) measures the amount of cations contained in soil particle surfaces. The amount of Na, K, Ca, and Mg was defined using flame atomic absorption spectrometry.

4.3 Other type of data used in the interpretation 4.3.1 RMT RMT is a passive-source EM method, which uses radio transmitters (14 kHz-250 kHz) as sources at enough distance to meet the plane EM wave criteria (Bastani, 2001; Bastani & Pedersen, 2001). Bastani (2001) developed the sys-tem used in this research for collecting RMT data in the field, named Enviro-MT. During the data acquisition, three components of the magnetic field (Hx, Hy, and Hz) and two horizontal components of the electric field (Ex and Ey) are measured simultaneously. Electrical resistivity values are estimated from these components through an inversion process.

Deploying and measuring with this RMT system is relatively fast, plus it is easy to transport. RMT resistivity data complement reflection seismic data especially at shallower depths. The 2D RMT resistivity models obtained by Shan et al. (2014) and Wang et al. (2016) were used for correlation with the reflection seismic data presented in Papers I, II and III.

4.3.2 P-wave refraction tomography and full-waveform inversion P-wave refraction tomography provides P-wave velocity models of the sub-surface. In comparison to refraction tomography, full-waveform inversion al-lows to obtain much higher resolution velocity models following an iterative least-square optimization technique fitting both phase and amplitude of the recorded wavefield (Adamczyk et al., 2014). The 2D P-wave velocity models

29

obtained by Adamczyk et al. (2014) and Wang et al. (2016) were compared with the downhole geophysical and reflection seismic data in Papers I and II.

4.3.3 CPT and CPTU-R Cone penetrations tests (CPT) and cone penetration tests with resistivity meas-urements (CPTU-R) were performed by SGI in the study area. Rankka et al. (2004) described a method for estimating potential quick clays using CPTU and total penetration force measurements (this method is widely accepted in Sweden). These measurements are also able to detect coarse-grained materi-als. Geotechnical data (Löfroth et. al, 2011; Branschens Geotekniska Arkiv–BGA, 2018) were used for calibrating the geophysical measurements in Pa-pers I and II.

4.4 Hydrogeological modelling Hydrogeological modelling was carried out in three steps. First, the 3D geo-logical modelling of the subsurface materials was performed using the hori-zons picked in the interpreted seismic sections (coarse-grained layer and bed-rock), total sounding data from boreholes (BGA, 2018) and LiDAR data (© Lantmäteriet), although it was necessary to fix the elevation of bedrock out-crops. Elevation surfaces for the top of the coarse-grained layer and bedrock were estimated with a natural neighbour interpolation after a Delaunay trian-gulation of the scattered sample points was generated. In the second step, the elevation surface of the bottom of the coarse-grained layer was estimated us-ing the RMT resistivity (Lindgren, 2014; Shan et al., 2014; Wang et al., 2016) and the total sounding data where possible. An interpolated thickness of the coarse-grained layer was then obtained, and subtracted from the elevation sur-face of the top of the coarse-grained layer, resulting in the elevation of the contact between the bottom of the coarse-grained layer and the underlying sediments.

The aforementioned two steps provided a plan of the thickness and eleva-tion of the coarse-grained layer, which was simulated in a single-layer 2D groundwater model. See Paper II for more details on hydrological modelling.

30

5 Summary of papers

5.1 Paper I: Identifying landslide preconditions in Swedish quick clays–insights from integration of surface geophysical, core sample- and downhole property measurements 5.1.1 Summary Downhole geophysical and laboratory physico-chemical property measurements carried out in year 2013 in an area prone to quick-clay landslides in southwest Sweden, are correlated with existing surface geophysical data (Malehmir et al., 2013a and 2013b; Polom et al., 2013; Shan et al., 2014) for (i) improving our knowledge of subsurface structures and geological preconditions related to the formation of quick clays, (ii) geological and stratigraphical site characterization of the area, focusing on quick clays, and (iii) obtaining more information on the relationship between coarse-grained materials and quick clays.

Downhole geophysical measurements included natural gamma radiation, which was used in conjunction with visual observations of the core samples for deriving the soil texture, fluid temperature and conductivity, and sonic measure-ments, from which VP was estimated. In the laboratory, electrical conductivity, salinity, pH, mineral magnetic properties, grain size analysis, fossil content, XRF, XRD and CEC were measured.

Figure 5.1 shows the results for BH2 (which shares many similarities with those of BH1 and BH3). These findings demonstrate that clay and silty clay are the most abundant soil textures, and also reveal the presence of a layer of approx-imately 10 m of sand and sandy silt at a depth of around 30 m. This layer has low natural gamma radiation and high magnetic susceptibility values, and is consid-ered to be a preserved beach-type deposit, where wave action removed the fine-grained and low density particles, allowing the concentration of magnetite, which is relatively dense. The groundwater level coincides with a sudden increase in temperature at 11 m of elevation. The electrical conductivity is constant down to the sandy silt layer, where it gradually increases and reaches a maximum value at the bottom of this layer, before decreasing again. The results for pH indicate acid conditions above the groundwater level and basic conditions a few metres below it. Quick clays are more likely to be found in basic conditions (Torrance & Pirnat, 1984).

31

The reduced VP record shows very low velocities, below 500 m/s, at the top of the sandy silt layer, which may be explained by the presence of a flammable gas (most likely methane) that emerged from the borehole when the sandy silt layer was penetrated.

Figure 5.1. Downhole geophysical and physical property measurements for samples from BH2. From left to right: the soil texture obtained from natural gamma radiation data and our visual observations, natural gamma radiation, fluid temperature, mag-netic susceptibility, electrical conductivity, pH, VP, near-receiver waveform, mid-receiver waveform, seismic data with the identified reflections S1 and S2 (Malehmir et al., 2013b), and the waveform tomography velocity profile (Adamczyk et al., 2014).

Figure 5.2 shows three interpreted reflection seismic sections for lines 1Sa (Figure 5.2a), 4 (Figure 5.2b) and 5 (Figure 5.2c) from 2011 (see Malehmir et al., 2013a and 2013b for more details about the interpretation). Two reflec-tions from a coarse-grained layer (S1 and S2) and one from the bedrock (B1) were identified. The borehole results for natural gamma radiation, magnetic susceptibility and electrical conductivity are shown superimposed on each section. The interface between clayey sediments and the shallower coarse-grained layer (S1) coincides with the strongest variations in natural gamma radiation and magnetic susceptibility data. Quick clay was recognised upon visual inspection and sampling of the cores. Electrical conductivity generally increases with depth but starts to decrease at the bedrock level. Figures 5.2b–c also contain geotechnical data (CPTU-R and CPT data, Löfroth et al., 2011) and 2D RMT resistivity models (Shan et al., 2014). The resistivity values in line 4 only delineate well the coarse-grained layer in the eastern part, where the values are approximately 100 Ωm. In contrast, in line 5 the coarse-grained layer is also well delineated by the resistivity model with values around 100

32

Ωm. Several areas in both profiles above and below the coarse-grained layer showed resistivity values between 10 and 80 Ωm, suggesting the presence of potential quick clay (leached clay, Solberg et al., 2012). Unleached marine

Figure 5.2. Three seismic profiles collected in 2011 (Malehmir et al., 2013a and 2013b; Polom et al., 2013). (a) SH-wave seismic data for line 1Sa, (b) P-wave seis-mic data for line 4, and (c) P-wave seismic data for line 5. The interpreted reflec-tions (S1, S2 and B1), bedrock position estimated from the drilling, potential quick clays and coarse-grained layer obtained from SGI data (Löfroth et al., 2011) are in-dicated (see legend on the right-hand side of the figure). Borehole data (BH1-line 1Sa, BH2-line 4 and BH3-line 5), natural gamma radiation, magnetic susceptibility and electrical conductivity, are superimposed in each section. RMT resistivity re-sults (Shan et al., 2014) for lines 4 and 5 are also superimposed in the seismic sec-tions. Colour bar showing the range of values can be found at the bottom of the fig-ure

33

clays with resistivity values between 1 and 10 Ωm mostly exist below the coarse-grained layer. The correlation with the geotechnical data is good, showing strong variations at the coarse-grained layer, and slight decreases at the suspected position of quick clays.

Figure 5.3. The variation of mineral magnetic properties with elevation for samples from BH1 (a)–(c), BH2 (d)–(f), and BH3 (g)–(i). On the left-hand side of the figure log χ versus log SIRM, on the centre log SIRM versus log ARM, and on the right-hand side of the figure χ versus S-ratio 100 mT.

Figure 5.3 shows the cross-plots of the mineral magnetic properties (χ–mag-netic susceptibility, SIRM, ARM, S-ratio 100 mT) for the samples from the three boreholes. Results for BH1 in Figure 5.3a–b show linear and curvilinear trends (with some outliers) that indicate the presence of a single magnetic remanence-carrying mineral (a mineral with intermediate to high magnetic concentrations and low coercivity, probably titanomagnetite).

34

Titanomagnetite can be concentrated in sands due to wave-assisted fluvial sorting. The data in Figure 5.3c indicate the presence of one or more high coercivity magnetic minerals at low magnetic concentrations (titanohematite). BH2 (Figure 5.3d–f) and BH3 (Figure 5.3g–i) show similar data for all the cases. Figures 5.3d and Figure 5.3g show two linear trends that indicate the presence of at least two magnetic minerals, one with low SIRM/χ ratio and the other with high SIRM/χ ratio (these samples also have low XARM/SIRM ratio and belong to the clays located below the coarse-grained layer). Hence, a third magnetic remanence-carrying mineral seems to be present in BH2 and BH3, with intermediate coercivity, high SIRM/χ ratio and low XARM/SIRM ratio (most probably greigite). This ferromagnetic iron sulphide is associated with glaciomarine sediments in southwest Sweden, organic matter and deeper wa-ter conditions. BH1 is located further from the river than the other two bore-holes, but it is still quite similar to them.

Grain size distribution (on the basis of the Wentworth classification scale) and fossil content for samples from BH2 are shown in Figure 5.4. The material distribution correlates well with the soil texture data (Figure 5.1). Fine clays are the most abundant materials on top and below the coarse-grained layer (high-energy environment) between -9 and -19 m of elevation. Silt is more abundant in the shallower levels. Fossils have been found in almost all the sediments. Clay and silt deposits at deeper levels are indicative of a low-en-ergy sedimentary environment, such as a marine basin. Fossils of barnacles, which live in a shallow marine environment, were found below -10 m of ele-vation. Mussel remains below 5 m of elevation suggest a marine or brackish origin; the latter agrees with the presence of terrestrial insect remains. BH1 and BH3 also show grain size distributions similar to the soil textures derived from the analysis of natural gamma radiation data and visual observations.

Figure 5.5 presents XRF results for samples from BH2 and the estimated salinity values. The most important results are (i) at the coarse-grained layer Si is high, and phyllosilicates and iron oxides are low, (ii) at clay locations Al, Fe, K, and Mg are abundant, and where there is silt these amounts are lower and Ca is high (maybe due to the presence of carbonaceous shells), (iii) at the elevation where quick clay was observed, Fe, Al, and K concentrations are high (indicating phyllosilicates), and (iv) salinity is constant and below 2 g/l down to the top of the coarse-grained layer, from where it increases gradually towards the bottom of the layer up to 4 g/l, then decreases again, staying above 2 g/l.

The salinity values are evidence of a downward movement of groundwater. The coarse-grained layer acts as a conduit for infiltrated fresh water from out-crops or fractures, helping to leach salts and drive the water towards the Göta River. Torrance (2012) established a salinity limit of 2 g/l for the formation of quick clays. All the three boreholes show salinity values below 2 g/l above the coarse-grained layer, and, in the case of BH3, also below it.

35

Figure 5.4. (a) Grain size distribution in percentages, scale 1H:2V, and (b) biologi-cal content for samples from BH2 (legend below the figure).

XRD and CEC results for samples from BH2 indicate that the clay sediments consist mainly of illite or muscovite and chlorite. The samples contained quartz and feldspars, and low amounts of Ca and Mg at the quick clay sus-pected position. Cation exchange capacity increased with depth in all the bore-holes.

36

Figure 5.5. From left to right: the soil texture obtained from natural gamma radiation data and our visual observations, the amounts of Si, Al, K, Ca, Mg, Fe, Cl, and the estimated salinity for samples from BH2.

5.1.2 Conclusions A combination of surface geophysical, core sample- and downhole property measurements was used to characterize an area prone to quick-clay landslides in southwest Sweden. It provides new information, in a Swedish geological context, about the properties and origin of quick clays and their associated materials, and their preconditions to slope failure.

The distribution of grain sizes depicts a coarse-grained layer sandwiched between silty-clayey deposits. As seen in all boreholes, natural gamma radia-tion generally decreases and magnetic susceptibility increases at the coarse-grained layer. Salinity values within this layer increase progressively with depth, indicating a downward groundwater flow towards the river. Infiltration of meteoric water through outcrops or fractures into the coarse-grained layer, up to 10 m thick in some parts, allows the leaching of salts from the overlying clays, thus accelerating the formation of quick clays. Integrated interpretation with geotechnical data demonstrates that the thicker the coarse-grained layer, the thicker the quick clays above. Therefore, although this layer is not a nec-essary requirement for the formation of quick clays, without it the leaching of salts would be slower and the quick clays would be shallower.

37

The combination of complementary methods, such as reflection seismic and resistivity, significantly improves the quality of the geological and strati-graphical information obtained. Seismic data make it possible to delineate at a higher resolution the transition between clays and coarse-grained materials. Resistivity data, on the other hand, are sensitive to salinity content, and hence, the distinction between leached and unleached clays may be possible.

5.2 Paper II: Subsurface structures of a quick-clay sliding prone area revealed using land-river reflection seismic data and hydrogeological modelling 5.2.1 Summary A multidisciplinary approach, involving the use of numerous techniques and data sets, such as land and river reflection seismics, resistivity, P-wave refrac-tion tomography, magnetics, borehole logging, side-scan sonar, and bathym-etry was followed in order to study an area prone to quick-clay landslides in southwest Sweden. The objectives of this research were (i) to construct a 3D geological/geophysical model of the large-scale subsurface structures above and within bedrock (e.g. fracture zones), (ii) to gain a better understanding of the role of a coarse-grained layer and its spatial relationship with the bedrock surface in the formation of quick clays, (iii) to predict the movement of groundwater within the coarse-grained layer using hydrological models, and (iv) to study the morphology of the Göta River valley, its riverbanks, bed, and mass-movement deposits.

Land and river (© SGU) high-resolution reflection seismic data were cor-related with resistivity (Bastani et al, 2017; Wang et al., 2016), P-wave refrac-tion tomography (Wang et al., 2016), downhole geophysical (Paper I) and ge-otechnical (BGA, 2018) data where available.

Figure 5.6 shows an example of these correlations, and how it was used for interpreting the subsurface structures along the line 5–5b. Interpreted horizons for the coarse-grained layer (S1) and the top of bedrock (B1), previously iden-tified in Paper I and in Malehmir et al. (2013a and 2013b), are delineated along almost the entire length of the line. S1 is discontinuous between common depth points (CDPs) 400 and 480 likely due to either or some of the following: 1) a lower fold (shown in Figure 5.6a), 2) bedrock movement, 3) fractures in the bedrock, and 4) deposits disturbed by human activity. Diffraction signa-tures in the bedrock at around CDP 440 (Figure 5.6b) are interpreted as a frac-ture indicating the presence of a strong bedrock curvature or an edge.

Variations in borehole data in Figure 5.6b coincide with the position of the coarse-grained layer. B1 is undulating, reaching the surface at CDP 500, and

38

Figure 5.6. Merged land seismic line 5–5b. The top part of the figure shows the po-sitions of line 5 collected in 2011, the seismic lines that intersect this line, the land-slide scar and the Göta River. (a) Processed seismic data. (b) Interpreted seismic data, including natural gamma radiation (Paper I) from BH3 in green (located at 0.23 m from the seismic line, values go from 59 to 200 API), magnetic susceptibility (Pa-per I) from BH3 in purple (values go from 0.07·10-6 to 2.7·10-6 m3/kg), and total sounding (BGA, 2018) from borehole 7062 in blue (located at 90.3 m from the seis-mic line, values go from 0 to 12 kN). At the bottom of the figure, close-ups for each borehole are shown. S1= coarse-grained materials, B1= bedrock, F= fractured or dis-turbed materials. (c) P-wave refraction tomography model (Wang et al., 2016) su-perimposed on the interpreted seismic data. (d) RMT resistivity results (Wang et al., 2016) superimposed on the interpreted seismic data. The resistivity values are con-sidered reliable above the thin dashed white line. (e) ATEM resistivity results (Bas-tani et al., 2017) superimposed on the interpreted seismic data. Colour bars on the right side of the figure indicate range of values for each parameter.

39

thereafter descends progressively towards the river. Displacement and oblique translation of some reflections below the landslide scar suggest that the sedi-ments slid towards the river. P-wave refraction tomography results shown in Figure 5.6c delineate reflector B1 reasonably well in the southern portion of the line, showing velocity values above 3000 m/s. By contrast, only in the last 300 m of the northern portion reflector B1 is clearly delineated, whilst the section between the river and this last stretch reveals high-velocity anomalies above B1. The deposits that overlay the bedrock have lower velocities, be-tween 300 and 1500 m/s. In the northern portion, S1 has higher velocities compared to the clayey deposits below and above; however, such differences are not observed in the southern part, which may indicate that the tomography was not effective at resolving the interface within this stretch.

Figure 5.6d–e shows the resistivity models of RMT and ATEM, respec-tively. Bedrock shows high resistivity values, >100 Ωm in both cases. At the S1 interface, values reach around 100 Ωm, providing a clearer delineation of the section between CDPs 500 and 700. Above S1, resistivity values drop (10-80 Ωm), indicating the presence of leached sediments (i.e. possible potential quick clays according to Solberg et al., 2012). Malehmir et al. (2016) inter-preted a fault (low resistivity zone) between CDPs 730 and 750 at a depth of 30 m below sea level.

The analysis of the six-channel river seismic data (© SGU), shown in Fig-ure 5.7, allows the discrimination of several geological features down to a depth of 100 m. Filled channels (feature C) are delineated in the shallower parts. The interface of the channels matches the strong variations observed in the borehole data (BGA, 2018). The bedrock interface (B1) shows an undu-lated structure and fractures at several positions along the line, e.g. between CDPs 800 and 1000, or 2350 and 2500. These fracture zones coincide with fracture zones observed in the geological data shown in Figure 2.2a (© SGU). The amplitude of the bedrock reflection decreases after CDP 3000 until the end of the line, likely due to being overlapping surface reflections. Other frac-tured or disturbed materials are identified at shallower depths.

40

Figure 5.7. Final results for the six-channel river seismic data (© SGU). (a) Pro-cessed seismic data. (b) Interpreted seismic data that also includes total sounding data (BGA, 2018) from boreholes 11014 (located at 45 m from the seismic line, val-ues go from 0 to 7.2 kN), 11034 (located at 10 m from the seismic line, values go from 0 to 9.2 kN) and 11094 (located at 25.7 m from the seismic line, values go from 0 to 11.2 kN), all in blue. At the bottom of the figure, close-ups for each bore-hole are shown. C= filled channels, B1= bedrock, F= fractured or disturbed materi-als (no reflection continuity). The land seismic lines that intersect this seismic lines are shown on top of Figure 5.7b. Scale 1H:30V.

The fault interpreted in Figure 5.6e coincides with one of the fracture zones interpreted in Figure 5.7b. However, no such fault could be inferred from the land reflection seismic data in Figure 5.6b (line 5–5b). In this respect, it is worth noting that a difference in time of about 1 s between cabled geophones and wireless seismic sections was detected during the processing of data, which could correspond to the presence of a fracture in the river. The combi-nation of a coarse-grained layer, together with the undulating bedrock mor-phology and the existence of bedrock fractures can promote the formation of quick clays, and are likely preconditioning factors for landslides (see e.g. L’Heureux et al., 2017).

In order to study the extension of the coarse-grained layer in a regional context, and its spatial relationship with the bedrock surface, 3D geological models were developed (Figure 5.8). The bedrock surface is undulating and outcrops at the surface in the southern part of the study area. Two elongated depressions next to the river are distinguished and interpreted as possible faults (F), coinciding with fracture zones shown in the geological data map (Figure 2.2a). The surface of the coarse-grained layer shows less undulation, and extends to the other side of the river. The filled channel interfaces inferred from the river seismic data seem to coincide with the surface of the coarse-

41

grained layer. Both elevation surfaces are shallowest at the centre of the sur-vey area; from here, the bedrock surface begins to dip gradually towards the river. This dipping causes water to accumulate in the portion of the coarse-grained layer closest to the river, in the depressions and in the thicker parts of the layer. The thickness of the coarse-grained layer is positively correlated with the thickness of the potential quick clay (Löfroth et al., 2011; Paper I), resulting in a higher risk of landslides at these depressions.

Figure 5.8. 3D geological modelling of the subsurface at the study area. (a) Aerial photo projected on the LiDAR data (© Lantmäteriet). (b) 3D view of the land and river seismic processed lines. (c) Elevation surface of bedrock. F indicates the posi-tion of interpreted faults, also shown in Figure 2.2a. (d) Elevation surfaces of the coarse-grained layer and bedrock. Colour bars indicating the range of values can be found at the bottom of the figure, including a horizontal scale.

The hydrological modelling show a gradually decreasing groundwater eleva-tion from the site of the outcrop to the river (Figure 5.9a). The measured hy-draulic head in the boreholes matches in most cases the modelled groundwater level. Transmissivity values (the product of the coarse-grained layer thickness with a calibrated hydraulic conductivity value) in Figure 5.9b are high around the land seismic lines (the coarse-grained layer is thicker at these positions).

The differences in the registered level of the groundwater table between samplings suggest that groundwater levels are subject either to large seasonal variation, or sporadic connections to a deeper groundwater system via bedrock fracture zones. The model does not predict artesian conditions during the low-water season (January-May), but conditions can change in the high-water sea-son (June-December). Calculation of the water residence time resulted in ap-proximately 20 years, assuming a porosity of 20 %. This is a relatively short

42

time period, and may therefore indicate that the salinity of groundwater is lower in the coarse-grained layer compared to the overlying clays, probably as a result of the transfer of ions from the clays to the groundwater in the coarse-grained layer, either by diffusion or slow infiltration (Torrance, 1979).

Figure 5.9. Hydrological modelling of a selected region within the study area. (a) Water table elevation. (b) Transmissivity distribution. (c) Mean groundwater veloc-ity distribution and vector field of the Darcy flow. The Göta River, outcrop area, land seismic lines (Malehmir et al. 2013a and 2013b; Paper I) and boreholes (BGA, 2018; Paper I) used in the modelling and interpretation are shown in the figure. Col-our bars on the right side indicate the range of values for each parameter.

In Figure 5.9c, the mean groundwater velocity (Darcy flow vector amplitude divided by porosity) and vector field of the Darcy flow (from the site of the outcrop to the river) are shown. The mean groundwater velocity is higher where transmissivity is lower, and vice versa. If the leaching of salts from the clays increases with groundwater velocity, the area where the velocity is higher may be more susceptible to contain quick clays.

43

Figure 5.10. Residual magnetic anomaly data on a LiDAR data contour map (© Lantmäteriet).The colour bar below shows the range of values. The land seismic lines acquired in 2011 and 2013 (Malehmir et al., 2013a and 2013b; Paper I) and boreholes (BGA, 2018: Löfroth et al. 2011; Paper I) used in the interpretation are also shown. A sketch with the measurement and base station positions can be found in the right bottom corner, and a legend just below it.

Figure 5.10 shows the residual magnetic anomaly data for the study area cov-ered during the 2011 field campaign. Overall, values in the northern sector are higher than in the southern sector. The high residual magnetic anomaly values seen at the bottleneck landslide scar may be related to the coarse-grained layer, given that this layer has high magnetic susceptibility values compared to the clayey sediments (Paper I). This suggests that the coarse-grained layer to-gether with the quick clays could have acted as the sliding prone layer. Ac-cording to a study by L’Heureux et al. (2012) the 1996 Finneidjord landslide was initiated by the presence of a ‘weak layer’ composed of softer and more sensitive clays and sands than those of the surrounding materials that acted as a slide prone layer. In our study area, changes in the river water level and/or

44

an increase in pore pressure, and/or toe erosion are possible triggering mech-anisms for the landslide.

The comparison of the residual magnetic anomaly and borehole data (in terms of the elevation of the top of the coarse-grained layer) suggests that both the depth and thickness of the coarse-grained layer are related to the high mag-netic values registered in other places of the study area. The southwestern sec-tor does not show the same results, possibly due to a thinner and/or deeper coarse-grained layer in this area.

Bedrock close to the surface generates high residual magnetic anomaly val-ues at the centre of the survey site, specifically at borehole 7067. The elon-gated anomaly running from borehole 7067 to the south of the survey area in a SW-NE direction, may be an effect of the scarp formed between bedrock and the lower elevated sediments in the west (Figure 2.2).

5.2.2 Conclusions A pioneer extensive reflection seismic investigation on land and along the Göta River, and its integration with geophysical, geotechnical, geological and topographical data, allowed the interpretation of the subsurface structures in an area prone to quick-clay landslides in southwest Sweden. Delineation of the leached and unleached clay and coarse-grained deposits, bedrock and frac-ture zones within a regional context has been possible thanks to this multidis-ciplinary data set.

Correlation of the reflection seismic, resistivity and P-wave refraction to-mography results show the presence, morphology and extension of the under-lying coarse-grained layer and bedrock, as well as the thickness and type of clay deposits. 3D geological modelling of the coarse-grained layer and bed-rock elevation surfaces illustrates the multiple potential locations, outcrops and fractures, that allow the infiltration of fresh water into the coarse-grained materials.

Hydrogeological modelling of the coarse-grained layer confirms its hydro-logical potential as a conduit for infiltrated fresh water from outcrops and frac-tures. The size of the catchment area and possible leaching processes during the low-water season (diffusion and slow infiltration) were also studied. Arte-sian groundwater conditions, which may occur during the high-water season (summer and autumn) are likely to promote the formation of quick clays.

Ground magnetic data seem sensitive to the coarse-grained layer and bed-rock. In the north of the study area the residual magnetic anomaly values are generally higher than in the south, which could be indicative of a shallower and/or thicker coarse-grained layer. At the bottleneck landslide scar, the high residual magnetic anomaly data suggest the presence of the coarse-grained layer at the surface. A possible interpretation is that this layer could have acted as sliding prone layer together with the quick clays.

45

Landslide scars and mass-movement deposits are observed along the Göta River valley and reflect the erosional processes in this landscape. The river seismic data reveal the presence of filled channels and fractured bedrock along the whole studied section.

These findings prove the potential of the study area to contain quick clays, as the geological and geohydrological requirements described in Chapter 3 are met according to this multidisciplinary investigation.

The study area, classified by SGI as a medium-high landslide risk area, will most probably be affected by the climate change, as the climate model scenar-ios suggest (SGI, 2012a). This underscores the pressing need for further stud-ies in the future, focusing on obtaining detailed information on the subsurface geology in order to improve the assessment of landslide hazard.

5.3 Paper III: Geotechnical site characterization using multichannel analysis of surface waves–a case study of an area prone to quick-clay landslides in southwest Sweden 5.3.1 Summary Multichannel analysis of surface waves was performed in active-source re-flection seismic data (Malehmir et al. 2013a and 2013b; Paper II) collected in an area prone to quick-clay landslides in southwest Sweden. The main objec-tives of this investigation were (i) to obtain VS models for every 2D seismic line collected in the area, (ii) to compare these models with the VP models obtained by Adamczyk et al. (2014) and Wang et al. (2016) and calculate VP/VS and Poisson ratio along the same seismic sections, (iii) to compare the models with the geological data obtained in nearby boreholes, and to make an integrated interpretation of the study area, and (iv) to find relationships be-tween the models that could be related to the different lithologies in the area.

MASW was not considered when the reflection seismic data were acquired, and hence, some aspects of the data acquisition had a detrimental effect on the frequency content, penetration depth, and resolution of the data. For example, the geophones used in the seismic surveys had a natural frequency of 28 Hz, thus low frequency data could be not recorded. The type of seismic source used, receiver spacing and maximum receiver-source offset also influence the frequency content of the data.

Figure 5.11 shows the VP (Adamczyk et al., 2014), VS, VP/VS, Poisson ratio and RMT resistivity (Lindgren, 2014) models for the seismic line 1 (Malehmir et al., 2013a). Previous studies identified two sub-horizontal interfaces, S1 and S2, interpreted to belong to coarse-grained materials underlying quick clays or marine clays, and the bedrock interface B1. The VS model in Figure

46

Figure 5.11. Reflection seismic section of line 1 (Malehmir et al., 2013a) with su-perimposed (a) Full-waveform tomography , (b) VS, (c) VP/VS, (d) Poisson ratio, and (e) RMT resistivity (Wang et al., 2016) models. A simplified version of the soil tex-ture results for BH1 are superimposed on all the sections (Paper I). The dashed lines and labelled reflections correspond to the interpreted interfaces in previous studies (Malehmir et al., 2013a).

47

5.11b shows good continuity in the first top 10 m, but resolution decreases at greater depth. Both VP and VS velocity models show a low velocity layer (VP is below 1500 m/s and VS is about 100 m/s) sandwiched between two high velocity layers (VP is around 2000 m/s and VS is higher than 120 m/s). VP/VS and Poisson ratio models (Figure 5.11c–d) show high values at the interfaces S1 and S2, and lower values in between these materials. The comparison with borehole data confirms that these high velocity layers are silt and sandy silt materials (Paper I). The RMT resistivity model (Figure 5.11e) has similar fea-tures, S1 and S2 are more resistive (10-100 Ωm, leached clay deposits accord-ing to Solberg et al., 2012) than the low velocity zone (1-10 Ωm, unleached cleays). .

Figure 5.12. Reflection seismic section of line 2–2b (Paper II) with superimposed (a) P-wave refraction tomography (P-wave refraction ), (b) VS, (c) VP/VS, (d) Pois-son ratio, and (e) RMT resistivity (Wang et al., 2016) models. A simplified version of the soil texture results for BH1 are superimposed on all the sections (Paper I). The dashed lines and labelled reflections correspond to the interpreted interfaces in pre-vious studies (Paper II). The ranges of values for every parameter and types of soil textures can be found in Figure 5.11.

48

Figure 5.12 shows the VP (Wang et al., 2016), VS, VP/VS, Poisson ratio and RMT resistivity (Wang et al., 2016) models for the seismic lines 2–2b. The undulating bedrock interface B1 is well delineated by all the models between CDPs 400 and 600, where the bedrock reaches closer to the surface. The in-terpreted interface S1 (the coarse-grained layer) does not coincide with the velocity models between CDPs 150 and 350. VP/VS and Poisson ratio models seem to delineate a gentle undulating layer, which is most probably the coarse-grained layer. This means that the interpreted seismic horizon is not well po-sitioned and should be adjusted due to these new findings. The RMT resistiv-ity model (Figure 5.12e, Wang et al., 2016) show a resistive layer (>100 Ωm), which appears to coincide with the S1 position between CDPs 350 and 450.

Figure 5.13. 3D views of the models for (a) VP, (b) VS, (c) VP/VS and (d) Poisson ra-tio. High amplitude anomalies can be delineated for each parameter.

3D models for VP, VS, VP/VS and Poisson ratio are shown in Figure 5.13. Sim-ilar features can be identified in all the seismic lines, and anomalies can be delineated using the highest values for each parameter. This demonstrates the good correlation between the models. The ranges of values are also similar in all the lines.

Figure 5.14 corresponds to the cross-plots of VP versus VS (Figure 5.14a), VP versus VP/VS (Figure 5.14b), and VP versus Poisson ratio (Figure 5.14c) for line 1. The data cluster in elevation intervals from 25 to 17 m, from 17 m to approximately 10 m (silt, sandy silt and quick clays were observed at these elevations), and from 10 to -2 m of elevation. VP and VS values are lower in the top 10 m (yellow-greenish colours), and higher at deeper levels (VP is around 1000-3500 m/s and VS is around 90-130 m/s).

49

Figure 5.14. Cross-plots of (a) VP versus VS, (b) vP versus VP/VS, (c) VP versus Poisson ratio, are colour-coded according to the elevation in line 1 (Malehmir et al., 2013b). In cross-plots for other seismic lines, the data did not cluster at different depth intervals as in line 1.

50

Figure 5.15. Cross-plots of different properties for BH1 (a), (b), (g), BH2 (b), (e), (h), and BH3 (c), (f), (i) (Salas-Romero et al., 2016), are colour-coded according to the soil textures identified in Salas-Romero et al. (2016). Data from BH1 and BH2 can be classified according to the type of soil textures, which is not the case for BH3.

Figure 5.15 shows cross-plots of different properties for each borehole based on the identified soil textures (Paper I). Data in BH1 (Figs. 5.15a, 5.15d, 5.15g) and BH2 (Figs. 5.15b, 5.15e and 5.15h) cluster in groups of different soil textures. In comparison with BH1, sandy silt materials show lower VP in BH2, which could be related to the presence of biogenic gas in this layer (Paper I). Silt, silty clay, and clay have lower VP values in BH1. The cross-

51

plotted data describe a continuous curve corresponding to the order of sedimentation. No data clusters are observed in BH3.

5.3.2 Conclusions MASW was performed on seven active-source seismic lines acquired specifically for reflection imaging and first-break traveltime tomography in an area prone to quick-clay landslides in southwest Sweden.

VP/VS and Poisson ratio delineate reasonably well the coarse-grained layer that underlies quick clays. Previous interpretations of the coarse-grained materials (Paper II) should be adjusted to the new findings related to the positive relationship between VS and RMT resistivity models. Cross-plots of the properties in BH1 and BH2 show data clusters depending on the type of soil texture. However, BH3 does not show the same features or they are difficult to distinguish.

VS measurements can delineate coarse-grained layers often associated with quick clays. Estimated VS30 values indicate the presence of soft soils and liquefiable deposits, which are susceptible to quick-clay landslides. Therefore, MASW, if combined with other types of data, can help to characterize regions at a high risk for quick-clay landslides.

52

6 Outlook: Amplitude Versus Offset

6.1 Motivation and objectives AVO is a geophysical method that is typically used for determining the pres-ence of hydrocarbons. AVO measures the variation of seismic reflection am-plitude with changing distance or offset between shot and receiver positions at a layer boundary, which separates zones of different physical properties. It provides information about the subsurface properties, such as fluid content, density, porosity, velocity or lithology (Castagna, 1993).

AVO analysis was performed in order to gain more information on fluid content and possible AVO responses at different water saturations at the sus-pected interface between quick clays and the coarse-grained layer. In this study, the modelled reservoir was the coarse-grained layer that serves as con-duit for fresh water infiltrated from outcrops or fractures. This layer might also serve as a reservoir for biogenic gas (probably methane seeped out from the top of the coarse-grained layer in BH2 as described in Paper I).

As mentioned earlier, the coarse-grained layer promotes the development of quick clays at this site, and hence, thicker coarse-grained materials correlate with thicker quick clays (Löfroth et al., 2011; Paper I). Thus, knowing the fluid content and saturation is rather important in this context. The software Hampston-Russel was used for doing AVO analysis.

6.2 A simplified model Borehole BH2 was chosen for the experiment as it was the borehole with the thickest coarse-grained layer and where biogenic gas was detected. Addition-ally, the decrease in amplitude in some parts of the reflection seismic data in line 4 (Figure 5.2, Paper I; Malehmir et al., 2013b) hinted at the presence of fluids at the depth of the coarse-grained layer.

Nevertheless, the amount of borehole data available for AVO modelling was limited, given there were no full data sets for density and VP. As explained in Paper I and in Chapter 4 of this thesis, VP (Figure 5.1) could not be recorded along the entire length of the borehole length because the PVC casing only allowed for measurements to be made at the height of the perforated section. Here, the recorded VP was the result of a mixture between the casing and the surrounding materials, with a higher influence of the latter. For this reason, a

53

simplified VP model was built using both data obtained from the perforated sections of all three boreholes (Paper I) and data extracted from the scientific literature (Albers, 2011), assigning a specific VP value depending on the li-thology derived from natural gamma radiation data. Bulk density values were calculated using mineral magnetic subsamples that were taken from core sam-ples. The high bulk density values at the coarse-grained layer were found to be related to the high magnetic susceptibility values (these materials were probably sorted by wave action when deposited, which helped concentrating titanomagnetite). As the measurements were not made in situ, it is difficult to assess how representative these values are for the geological materials. Sev-eral AVO modelling tests showed that density did not decrease when gas sat-uration increased, hence the titanomagnetite effect was assumed to be strong. In order to attain more realistic in situ bulk density values, the Gardner’s equa-tion (Gardner et al., 1974) was used for calculating new values for the coarse-grained layer.

Normal moveout (NMO) corrected reflection seismic data of line 4 (Malehmir et al., 2013b) were used to perform AVO analysis (without apply-ing automatic gain control–AGC, which could affect the amplitudes of the seismic data). A super gather, also known as a common offset stack, was gen-erated for this purpose. This type of stack enhances the signal-to-noise ratio, without changing the offset dimension.

6.3 AVO modelling AVO modelling (creation of fluid replacement models and synthetic seismo-grams based on well logs–in situ and those with modified saturation levels) correlates lithologies or fluid content with seismic data.

Since a VS log was not available for borehole BH2, one was derived using VP and bulk density well logs. Assuming the reservoir was within the coarse-grained layer, located at a depth between 31.1 and 39 m (this thickness is enough to avoid the tuning thickness effect), two different methods for calcu-lating VS were used: 1) the Greenberg-Castagna method (Greenberg & Castagna, 1992) for the reservoir layer, and 2) Castagna’s equation (Castagna et al., 1985) for the area outside of the reservoir. Correlation of the well logs (using a Ricker wavelet of 200 Hz) with the super gather data corresponding to line 4 allowed to estimate a new VP correlated log. This step was necessary to obtain a match between the synthetic seismograms and the reflection seis-mic data.

54

Table 6.1. Fluid properties in BH2, at 8 ºC and salinity equal to 1.325·10-6 ppm, used during the AVO modelling.

Water Methane Density (g/cm3) 0.999 0.00215

Bulk modulus (MPa) 2070.05 0.41

The next step was to perform fluid replacement modelling following the Biot-Gassman theory (Gregory, 1977). The variables considered included matrix content, in situ fluid content and fluid properties (see Table 6.1). The matrix reservoir was assumed to be composed of 90 % quartz and 10 % clay, based on the values on the grain size distribution chart shown in Figure 5.4. Given the lack of information on the relative concentrations of methane and water, we built a starting model assuming a 50 % water / 50 % gas content. Three types of saturation levels were tested: homogeneous mix, medium patches and big patches, and also two different output saturations, 70 % and 85 % water (i.e. 30 % and 15 % gas, respectively). Table 6.2 gives the average values of each of the physical properties analysed under each of the three patchy satu-ration scenarios, whereas Figure 6.1 shows the results of the well logs for the case of medium patches scenario only. Once that the models were created, synthetics seismograms for every model that was calculated in the previous step were created based on Zoeppritz equations (Aki & Richards, 1980; Shuey, 1985) and using a Ricker wavelet of 200 Hz.

55

Figure 6.1. Fluid modelling results for BH2, showing the modelled physical proper-ties for different water saturation models (in black the starting model, in blue 70 % water saturation and in red 85 % water saturation) at the reservoir indicated by two horizontal green lines.

Table 6.2. Averaged physical properties of the overburden and the top part of the reservoir (using data from Figure 6.1), before and after replacing methane by water, using different types of patchy saturation models.

Overburden 50 % water 70 % water 85 % water Homogeneous

VP (m/s) 852 834 VS (m/s) 737 721 Density (g/cm3)

1.65 1.67

Medium patches VP (m/s) 1312 1128 1377 1524 VS (m/s) 656 761 737 721 Density (g/cm3)

1.71 1.63 1.65 1.67

Big patches VP (m/s) 1543 1612

500 1000 1500 2000-40

-30

-20

-10

0

10

20El

evat

ion (m

)P-wave (m/s)

400 800 1200-40

-30

-20

-10

0

10

20S-wave (m/s)

1 1.5 2-40

-30

-20

-10

0

10

20Density (g/cm3)

50 % water saturation model70 % water saturation model85 % water saturation model

56

VS (m/s) 737 721 Density (g/cm3)

1.65 1.67

The results in Figure 6.1 show that VP increases with water saturation, and the same effect is observed for density. As expected, the change in VS is not sig-nificant due to its low sensitivity to pore fluid content.

Results in Table 6.2 indicate that VP is the only physical property that changes with different saturation levels. The values of VP are higher in the big patches scenario than in the homogeneous or medium patches models. The difference in VP for the homogeneous saturation at 70 % and 85 % water is small compared to the patchy saturations.

In contrast to VP, density and VS show exactly the same values with differ-ent amounts of water for each type of patchy saturation. This may indicate that this model requires the input of additional information in order to discriminate correctly between different saturation patterns.

6.4 AVO analysis Computing the AVO attributes, serves to study AVO responses in relation to a certain lithology or fluid content (e.g. Rutherford & Williams, 1989, defined a classification of AVO responses for different hydrocarbon saturation for-mations).

6.4.1 Synthetics AVO intercept and gradient were calculated for the different saturation sce-narios according to their respective models. Results in Table 6.3 demonstrate that intercept values become more positive as the percentage of water in-creases in the patchy saturations. However, the differences in the seismic re-sponse in the homogeneous scenario between different water saturations are very small. The largest differences are observed in the medium patches sce-nario. The gradient behaves in a similar way, becoming less negative with increasing water saturation.

57

Table 6.3. AVO intercept and gradient before and after replacing methane by water using different patchy saturation levels.

50 % water 70 % water 85 % water Homogeneous

Intercept -0.0379672 -0.0340158 Gradient -0.0125889 -0.0121107

Medium patches Intercept -0.0466132 0.0513386 0.0805792 Gradient 0.0103371 0.0421232 0.0573751

Big patches Intercept 0.074138 0.0922251 Gradient 0.0515016 0.0602853

Figure 6.2. AVO intercept well log model for medium patches saturation and water saturation of (a) 50 %, (b) 70 %, and (c) 85 %.

58

Figure 6.3. AVO gradient well log model for medium patches saturation and water saturation of (a) 50 %, (b) 70 %, and (c) 85 %.

6.4.2 Reflection seismic data Unfortunately, the data quality of the reflection seismic data was poor to allow to perform a proper AVO analysis and compare it with the AVO responses of the modelled well logs, due to seismic noise level and the fact that the reflector from the coarse-grained layer was not prominent enough to be used in the AVO analysis.

6.5 Conclusions and further investigations AVO analysis was performed at the boundary between quick clays and a coarse-grained layer in order to study the AVO responses to different water and gas saturations. The results of the AVO modelling show that increasing water saturation increases VP and density, and thus the synthetic models show a clear variation of amplitude versus offset. This investigation shows that AVO analysis can generate interesting results about the fluid content in the area. Time-lapse seismic measurements would allow monitoring the changes in the water saturation indices in the coarse-grained layer over time, which may be important for landslide hazard assessment.

59

This experiment posed a series of limitations which affected the perfor-mance of the AVO analysis. More specifically, (i) borehole data were obtained from a simplified model of the in situ properties, (ii) the percentage of water and gas saturations in the starting model were unknown (the presence of the fluids was confirmed but it was not clear in what proportions), and (iii) the quality of reflection seismic data was unsuitably poor.

The model uncertainties would significantly improve should additional downhole geophysical measurements be taken, even of a different type, such as neutron-porosity or vertical seismic profile. Such methods would help to overcome the problems caused by the PVC casing and obtain in situ measures for all properties (VP and porosity). Besides, reflection seismic measurements should be repeated in the same area on a regular basis.

Future investigations, not involving AVO, should include monitoring of the drilled boreholes for the properties already studied (Paper I) or new ones, e.g. to obtain new hydrological data. The drilling of a new borehole, closer to the bedrock outcrops in the study area, would be useful for obtaining new sa-linity measurements at the coarse-grained layer depth in order to improve the groundwater models.

60

7 Conclusions

A multidisciplinary near-surface investigation was carried out for studying an area prone to quick-clay landslides in southwest Sweden. Geophysical meas-urements, land and river reflection seismic, P-wave refraction tomography, resistivity, magnetics, MASW, and borehole logging data, were the foci of this investigation. Laboratory measurements of physico-chemical properties (grain size analysis, mineral magnetic properties, pH, electrical conductivity, salinity, XRF, XRD, CEC, and fossil identification) were performed on core samples from three boreholes drilled in the area. These data were compared and correlated with geotechnical, geological, and topographical data for ob-taining an integrated interpretation of the quick-clay site.

The first objective of this thesis, to study the response of quick clays using a range of geophysical methods, is dealt with in Papers I , II, and III. The presence of quick clays must be confirmed by performing laboratory geotech-nical measurements on undisturbed soil samples. Given the latter, geophysical methods are not considered suitable for this purpose. However, when geo-physical methods are combined, the position of associated materials, such as coarse-grained layers, and areas of leached and unleached clay deposits, can be defined with a high resolution (Papers I and II).

The second objective of this work was to study the properties of quick clays and their relationship with associated materials, such as the coarse-grained layer (Paper I). As previously stated, the combination of different geophysical methods allows a good delineation of the coarse-grained layer, but still some calibration is necessary to ground truth the geophysical measurements, and hence, correlation with geotechnical or laboratory measurements is necessary. The comparison with the geotechnical data provided by the SGI (Löfroth et al., 2011; BGA, 2018) yielded a good correlation with the position of the coarse-grained layer and of the potential quick clays. The results in this study area show that thicker coarse-grained layers are related to thicker quick clays. Measurements of the physico-chemical properties of core samples confirmed with more precision the thickness of the coarse-grained layer and its proper-ties, as well as the properties of potential quick clays and the clay deposits below the coarse-grained layer; grain size distributions depicted a coarse-grained layer sandwiched between silty-clayey deposits, and salinity values increased across this layer from top to bottom, indicating a downwards groundwater movement towards the river.

61

The third objective of this thesis was to make an integrated interpretation of the subsurface structures using a multidisciplinary approach (Papers II and III). Delineation of the leached and unleached clay and coarse-grained depos-its, bedrock and fracture zones in a regional context was possible thanks to this multidisciplinary approach. 3D geological modelling of the coarse-grained layer and bedrock elevation surfaces illustrated the multiple potential locations, outcrops and fractures, that allow the infiltration of fresh water into the coarse-grained materials. The development of quick clays would probably be slower without this layer. Hydrogeological modelling of the coarse-grained layer confirmed its hydrological potential as a conduit for infiltrated fresh wa-ter. Residual magnetic anomaly data at the bottleneck landslide scar suggest that this layer is potentially part of a ‘weak layer’, which could have acted as a sliding prone layer together with the quick clays.

The fourth and last objective of this work was to provide a geotechnical characterization of the subsurface materials using MASW (Paper III). VS mod-els were obtained for all the lines down to 15-20 m of depth, and compared with VP models obtained in previous studies (Adamczyk et al., 2014; Wang et al., 2016). VP/VS and Poisson ratio delineated the coarse-grained layer that underlies quick clays reasonably well; comparison between VS and RMT re-sistivity models improved the regional interpretation of previous studies (Malehmir et al, 2013a and 2013b; Papers I and II). Cross-plots for BH1 and BH2 showed data clusters at different depths, which seemed to be related to different soil textures. Estimated VS30 values indicated the presence of soft soils and liquefiable deposits, which are susceptible to quick-clay landslides.

Additional investigations included AVO analysis of the modelled well logs at the boundary between quick clays and the coarse-grained layer. Although several limitations were encountered when analysing the well log and seismic data, the investigation suggests the potential of AVO analysis for studying the fluid content at this interface.

To conclude, this research is expected to improve the information on quick-clay landslides and the preconditions to trigger slope failure. Based on climate change predictions for this area landslide risk will likely worsen in the future. The need is therefore pressing to conduct further research and move on to new levels. Application of additional geophysical methods, such as borehole or ge-ophysical monitoring, and induced polarization, could provide more infor-mation about the formation of quick clays and associated landslides.

62

8 Summary in Swedish

Kvicklera består av glaciala och postglaciala sediment och finns avlagrade i Sverige, Norge, Finland, Kanada, USA (Alaska), Ryssland och Japan. I några av länderna är jordskred orsakade av kvicklera en av de vanligaste och allvar-ligaste orsakerna till naturkatastrofer och leder till många dödsfall och bety-dande ekonomiska förluster. När kvicklera utsätts för spänning eller dess skjuvhållfasthet reduceras kan den övergå i flytande form och orsaka ett jord-skred om topografin gör att en minsta lutningsvinkel överskrids om det sam-tidigt finns en plats för leran att flyta till. Kraftigt regn, hög eller låg vattennivå i ett vattendrag, erosion i ett vattendrag, förändringar i grundvattennivån och till och med byggnation eller lastning och lossning av gods kan också utlösa ett jordskred. Katastrofer såsom Rissa (Norge), Tuve (Sverige) eller mer nyli-gen Saint-Jude (Kanada) och Munkedal (Sverige) visar hur viktigt det är att forska på denna typen av jordskred.

Denna avhandling är inriktad på en tvärvetenskaplig undersökning av yt-nära jord som har utförts för att studera ett område i sydvästra Sverige. Om-rådet klassas som högriskområde när det gäller jordskred orsakade av kvick-lera. Undersökningen är i huvudsak baserad på landseismik och seismik utförd på älven, P-vågsrefraktionstomografi, resistivitet, magnetmätningar, MASW och borrhålsloggning. För att erhålla fysiokemiska egenskaper (kornstorleks-analys, magnetiska egenskaper hos mineralkorn, pH, elektrisk ledningsför-måga, salinitet, XRF, XRD, CEC och identifikation av fossil) utfördes mät-ningar i laboratorie. Denna data har jämförts med och korrelerats till geotek-nisk, geologisk och topografisk data för att erhålla en integrerad tolkning av området med kvicklera.

Det första syftet med avhandlingen var att studera responsen från kvicklera på olika geofysiska metoder. Detta har testats i artiklarna I, II och III. Att be-kräfta förekomsten av kvicklera kan inte göras med geofysiska mätningar. För detta krävs geotekniska laboratoriemätningar på ostörda jordprov. Men när geofysiska mätningar kombineras kan dessa med hög upplösning ge svar på var det finns associerade material som till exempel grovkorniga lager eller lager med urlakad/ej urlakad lera (artiklarna I och II).

Syfte nummer två med avhandlingen har varit att studera egenskaperna hos kvicklera och förhållandet till andra material, t.ex. grovkorniga lager. Som nämnts ovan så kan man genom att kombinera olika geofysiska metoder loka-lisera var det finns grovkorniga material. Dock krävs det en kalibrering av detta, för vilken det är nödvändigt att utföra antingen geotekniska mätningar

63

på plats eller i labb. Jämförelse med geotekniska data från SGI visade på bra träffsäkerhet hos de kombinerade geofysiska metoderna när det gällde att fast-ställa var det finns grövre lager och potentiell kvicklera. Resultaten i studien visar att mäktigare grovkorniga lager förekommer tillsammans med mäktigare lager av kvicklera. Fysiokemiska mätningar på jordkärnor visar med större exakthet mäktighet och egenskaper på både de grovkorniga lagren och poten-tiella lager med kvicklera. Kornstorleksdistributionen visar att de grövre lag-ren finns inlagrade mellan siltiga-leriga avsättningar. Salthalten ökar nedåt i dessa lager vilket indikerar ett nedåtriktat grundvattenflöde mot älven.

Det tredje syftet med avhandlingen har varit att göra en integrerad tolkning av strukturer under markytan med hjälp av ett tvärvetenskapligt tillvägagångs-sätt (artiklarna II och III). Karaktäriseringen av urlakad och ej urlakad lera, grovkorniga avlagringar, berggrunden och sprickzoner i ett regionalt samman-hang är möjligt tack vare detta tillvägagångssättet. Geologisk 3D-modellering av grovkorniga lager och berggrundstopografin visar alla möjliga platser, häl-lar och sprickor där sötvatten kan infiltreras i de grovkorniga lagren. Bildandet av kvicklera skulle troligen vara långsammare utan de grövre lagren. Hyd-rogeologisk modellering av de grovkorniga lagren bekräftar att de kan trans-portera infiltrerat sötvatten. Residualmagnetisk anomalidata där jordskredet har skurit ner i marken visar att detta lager kan vara det svaga lagret som har agerat som glidkänsligt lager tillsammans med kvickleran.

Det fjärde och sista syftet med denna avhandlingen har varit att leverera en geoteknisk karaktärisering av materialet under markytan i studieområdet (ar-tikel IV). Modellering för estimering av VS-, VP-, VP/VS- och Poissons tal ut-fördes för alla linjer ner till 15-20 m djup. VP/VS-talet visar på ett rimligt sätt de grovkorniga lagren på några linjer och förbättrar den regionala tolkningen av tidigare studier (artikel I och II). Cross-plottar för varje borrhål och för en linje visar datakluster på olika djup och kan vara relaterat till litologin. Det uppskattade VS30-värdet att jorden är mjukare I undersökningsområdet.

Undersökningar har också gjorts på gränsen mellan kvicklera och grovkor-niga lager med metoden AVO-analys på de modellerade borrhålloggarna. Även om det fanns flera begränsningar när borrhålsloggarna och den seism-iska data analyserades så visade studien att det finns potential att använda AVO-analyser för att studera fluidinnehållet i detta gränsskiktet.

Sammanfattningsvis så ämnar detta forskningsprojektet att öka kunskapen om jordskred orsakade av kvicklera och förutsättningarna för kvicklera att ut-lösa jordskred. Den förutspådda klimatförändringen för det undersökta områ-det kommer påverka och förmodligen öka risken för jordskred. Därför måste denna forskning fortsätta och utvecklas till nya nivåer. Användande av andra geofysiska metoder, t.ex. kontinuerlig borrhålsmätning inducerad polarise-ring, kan bidra med mer information om hur kvicklera bildas och orsakar jord-skred.

64

Acknowledgements

Our time here finally came to an end (hopefully happy ending...). I wish Upp-sala was brighter in winter and warmer in summer, but these are just little things with what I learned to live and complain along the time. Fred and I formed a family in Uppsala, and for this reason Sweden will be always in a special place in my heart. Being far from home was never easy, but with so many good friends we had many unforgettable moments, and I want to thank you all for your friendship and all this time.

I would like to start thanking Alireza Malehmir for giving me the oppor-tunity to do a PhD and for his support and friendship along these years. I think I have learned a lot from you and hope to be as good researcher as you are in a near future. Thank you also for offering me the chance to help you teaching several courses. It was a great experience, stressful sometimes but very en-riching, with very good students from all over the world.

Thank you Ian Snowball and Chris Juhlin. We did not work so closely, but I have always felt your support when needed and had your good advice.

I would like to mention especially Monika Ivandic, who helped me so much in the last year of my studies. Without your help AVO would not have been possible. Not only you are an excellent researcher, but a better person and friend. I still remember your warm welcome during our first day in Geocen-trum.

Thank you to my co-authors in the different papers, Bryan C. Lougheed, Magnus Hellqvist, Benoit Dessirier and Bojan Brodic. I am also very grateful with Kenneth Torrance that improved so much my first paper and from whom I learned about quick clays. Thank you Mehrdad Bastani for your help and kindness any time we cross. Also thanks to Fatima, Jeanette and Anna from administration, who were so helpful, and Leif who was always so kind and cheerful. I have also good memories from Hasse Palm, from whom I learned so much in fieldwork. Thank you wherever you are.

Thank you to all my ex-corridor mates in the ground floor, especially to my dear Shunguo, Remi, Zeynab, Claudia, Aggela, Samar, Bjarne, Giulia, Sissa and Mohsen. What a great time down there! Thank you also to all my field-work companions along the time, PhD and MSc. students, especially to Su-man, Georgiana, and Bojan, we were the fantastic team! Bojan thank you for helping me collecting my smelly mud and with MASW.

65

Daniel, you and your family were a great support for us in Uppsala, and we had so much fun together! Maria, Iwa, and Ruta thank you for your friendship and good times along these years.

Thank you to the ‘new generation’, especially to Alba, Sebastian, and Mi-chael, and to the ‘old generation’, who are not anymore in Geocentrum, Peter, Omid, Ping, Fei, Harbe, Chunling, Jochem, Azita, Sahar, Mahboubeh, Pouya, Marc, Agnes, Babis, and many more. Also thanks to other friends out of Ge-ocentrum, Alicja, Christos, Joao, Oscar, Arash, and Konstantinos.

Thank you to all my office mates for sharing space, talks, food and laugh. Thank you to my favourite neighbour Magnus, thank you for showing us Swe-den and for helping me with the summary in Swedish of my thesis. Thank you Emil for ‘letting me use your desk’, and for making me space in your shelves ;) Thank you Le for your funny and positive attitude. Last but not least, thank you Tegan (my dear Tigri). I still remember when Chris Juhlin asked me if I wanted a new office mate. Thank you Chris for bringing Tegan to my life; te quiero mucho amiga!

Special thanks to all the people that reviewed this thesis: Daniel, Monika, Lidia ☺, Fred, Vero, Sebis and of course, my supervisors!

When I left Spain 8 years ago I let many things and people behind. I want to thank all my friends in Madrid for being always so supportive whenever we came back and for sharing your time with us. Thanks to Fred’s family for the great times together in Luxembourg and Sweden, especially to Jéis and An-drea.

Thank you to my family, through all the happy but also sad moments, for your support and warm welcome every time that we visited you. Gracias mamá, papá y Gustavo, estamos lejos pero os llevo siempre en mi corazón, gracias por ayudarnos tanto en los últimos meses.

Thank you Fred, for your continuous support and love. Coming with you to Uppsala was my best decision in life. Thank you for adopting Bambi, our dog makes me extra-happy every time I come back home. Finally, thank you Tristan for your patience and for behaving so well. I hope that someday you understand all the sacrifices that we have done along this final year and forgive us ☺

Silvia Salas Romero, Uppsala, February 2016

66

References

Adamczyk, A., Malinowski, M., & Malehmir, A. (2013) Application of first arrival tomography to characterize a quick clay landslide site in southwest Sweden. Acta Geophysica, 61(5), 1057–1073. https://doi.org/10.2478/s11600-013-0136-y

Adamczyk, A., Malinowski, M., & Malehmir, A. (2014) High-resolution near-surface velocity model building using full-waveform inversion–a case study from south-west Sweden. Geophysical Journal International, 197(3), 1693–1704. https://doi.org/10.1093/gji/ggu070

Aki, K., & Richards, P. G. (1980) Quantitative seismology: theory and methods. W. H. Freeman and Co.

Albers, B. (2011) Linear elastic wave propagation in unsaturated sands, silts, loams and clays. Transport in Porous Media, 86(), 537–557. https://doi.org/10.1007/s11242-010-9638-0

Andersson-Sköld, Y., Torrance, J. K., Lind, B., Odén, K., Stevens, R. L., Rankka, K. (2005) Quick clay–a case study of chemical perspective in Southwest Sweden. Engineering Geology, 82, 107–118. https://doi.org/10.1016/j.enggeo.2005.09.014

Bastani, M. (2001) EnviroMT–a new controlled source/radio magnetotelluric system. Ph.D. thesis, ISBN 91-554-5051-2, Uppsala University, Uppsala.

Bastani, M. & Pedersen, L. B. (2001) Estimation of magnetotelluric transfer functions from radio transmitters. Geophysics, 66(4), 1038–1051. https://doi.org/10.1190/1.1487051

Bastani, M., Persson, L., Löfroth, H., Smith, C. A., & Schälin, D. (2017) Analysis of ground geophysical, airborne TEM, and geotechnical data for mapping quick clays in Sweden. In V. Thakur et al. (Eds.), Landslides in sensitive clays, Ad-vances in natural and Technological Hazards Research (Vol. 46, pp. 463–474). Springer. https://doi.org/10.1007/978-3-319-56487-6_41

Bjerrum, L. (1967) Engineering geology of Norwegian normally-consolidated marine clays as related to settlements of buildings. Geotechnique, 17(2), 83–118.

Branschens Geotekniska Arkiv (2018) Geosuite borrhål (och tillhörande Geosuite projektområden). Swedish Geotechnical Institute. Retrieved from http://bga.swedgeo.se/bga/.

Büker, F., Green, A. G., & Horstmeyer, H. (1998) Shallow seismic reflection study of a glaciated valley. Geophysics, 63(4), 1395–1407. https://doi.org/10.1190/1.1444441

Cardimona, S. J., Clement, W. P., & Kadinsky-Cade, K. (1998) Seismic reflection and ground penetrating radar imaging of a shallow aquifer. Geophysics, 63(4), 1310–1317. https://doi.org/10.1190/1.444432

Castagna, J. P., Batzle, M. L. & Eastwood, R. L. (1985) Relationships between com-pressional-wave and shear-wave velocities in clastic silicate rocks, Geophysics, 50(4), 571–581. https://doi.org/10.1190/1.1441933

67

Castagna, J. P. (1993) AVO analysis–tutorial and review. In Castagna & Backus (Eds.), Offset-dependent reflectivity–theory and practice of AVO analysis, Inves-tigations in Geophysics (Vol. 8, pp. 3–36). Society of Exploration Geophysicists. https://doi.org/10.1190/1.9781560802624

Comina, C., Krawczyk, C. M., Polom, U., & Socco, L. V. (2017) Integration of SH seismic reflection and Love-wave dispersion data for shear wave velocity deter-mination over quick clays. Geophysical Journal International, 210(3), 1922–1931. https://doi.org/10.1093/gji/ggx276

Dahlin, T., Löfroth, H., Schälin, D., & Seur, P. (2013) Mapping of quick clay using geoelectrical imaging and CPTU-resistivity. Near Surface Geophysics, 11(6), 695–670. https://doi.org/10.3997/1873-0604.2013044

De Geer, G. (1940) Geochronologia Suecica Principles. Kungl. Svenska Vetenskaps-akademiens Handlingar, 3rd series, 18(6), 1–367.

Demers, D., Robitaille, D., Lavoie, A., Paradis, S., Fortin, A., & Ouellet, D. (2017) The use of LiDAR airborne data for retrogressive landslides inventory in sensitive clays, Québec, Canada. In V. Thakur et al. (Eds.), Landslides in sensitive clays, Advances in natural and Technological Hazards Research (Vol. 46, pp. 279–288). Springer. https://doi.org/10.1007/978-3-319-56487-6_25

Donohue, S., Long, M., O’Connor, P., Helle, T. E., Pfaffhuber, A. A., & Rømoen, M. (2012) Multi-method geophysical mapping of quick clay. Near Surface Geophys-ics, 10(3), 207–219. https://doi.org/10.3997/1873-0604.2012003

Ewing, K. (2009) What is sonic drilling and how does it work? Australasian drilling. September-October 2009. www.adia.com.au

Gardner, E. H. F., Gardner, L. W., & Gregory, A. R. (1974) Formation velocity and density–the diagnostic basics for stratigraphic traps. Geophysics, 39(6), 770–780. https://doi.org/10.1190/1.1440465

Göransson, G. I., Bendz, D., & Larson, P. M. (2009) Combining landslide and con-taminant risk: a preliminary assessment. Journal of Soils and Sediments, 9, 33–45. https://doi.org/10.1007/s11368-008-0035-z

Greenberg, M. L., & Castagna, J. P. (1992) Shear-wave velocity estimation in porous rocks: theoretical formulation, preliminary verifications and applications. Geo-physical Prospecting, 40, 195–209. https://doi.org/10.1111/j.1365-2478.1992.tb00371.x

Gregersen, O. (1981) The quick clay landslide in Rissa, Norway: The sliding process and discussion of failure modes (Publication No. 135). Oslo, Norway: Norwegian Geotechnical Institute.

Gregory, A. R. (1977) Aspects of rock physics from laboratory and log data that are important to seismic interpretation. In Seismic Stratigraphy–Applications to Hy-drocarbon Exploration (Vol. 26). AAPG memoir. https://doi.org/10.1306/M26490C2

Hultén, C., Edstam, T., Arvidsson, O., & Nilsson, G. (2006) Geotekniska förutsättningar för ökad tappning från Vänern till Göta älv (Varia 565). Linkö-ping, Sweden: Statens Geotekniska Institute.

Jakob, M., & Lambert, S. (2009) Climate change effects on landslides along the south-west coast of British Columbia. Geomorphology, 107, 275–284. https://doi.org/10.1016/j.geomorph.2008.12.009

Larsson, S., & Jansson, M. (1982) The landslide at Tuve, November 30 1977 (Report 18). Linköping, Sweden: Swedish Geotechnical Institute.

68

L’Heureux, J.-S., Longva, O., Steiner, A., Hansen, L., Vardy, M. E., Vanneste, M., Haflidason, H., Brendryen, J., Kvalstad, T. J., Forsberg, C. F., Chand, S., & Kopf, A. (2012) Identification of weak layers and their role for the stability of slopes at Finneidfjord, northern Norway. In Y. Yamada et al. (Eds.), Submarine mass movements and their consequences, Advances in Natural and Technological Haz-ards Research (Vol. 31, pp. 321–330). Springer. https://doi.org/10.1007/978-94-007-2162-3_29

L’Heureux, J.-S., Nordal, S., & Austefjord, W. (2017) Revisiting the 1959 quick clay landslide at Sokkelvik, Norway. In V. Thakur et al. (Eds.), Landslides in sensitive clays, Advances in Natural and Technological Hazards Research (Vol. 46, pp. 395–405). Springer. https://doi.org/10.1007/978-3-319-56487-6_35

Lindgren, A. (2014) Combined use of ground and borehole geophysical data to model spatial variations of quick clay and surrounding sediments at a Swedish landslides site (master’s thesis). Luleå, Sweden: Department of Civil, Environmental and Natural Resources Engineering, Luleå University of Technology.

Locat, A., Locat, P., Demers, D., Leroueil, S., Robitaille, D., & Lefebvre, G. (2017) The Saint-Jude landslide of 10 May 2010, Quebec, Canada: Investigation and characterization of the landslide and its failure mechanism. Canadian Geotech-nical Journal, 54(10), 1357–1374. https://doi.org/10.1139/cgj-2017-0085

Löfroth, H., Suer, P., Dahlin, T., Leroux, V., & Schälin, D. (2011) Quick clay map-ping by resistivity–surface resistivity, CPTU-R and chemistry to complement other geotechnical sounding and sampling (GÄU-delrapport 30). Linköping, Sweden: Swedish Geotechnical Institute.

Lundberg, E., Malehmir, A., Juhlin, C., Bastani, M., & Andersson, M. (2014) High-resolution 3D reflection seismic investigation over a quick clay landslide scar in southwest Sweden. Geophysics, 79(2), B97–B107. https://doi.org/10.1190/GEO2013-0225.1

Lundström, K., Larsson, R., Dahlin, T. (2009) Mapping of quick clay formations us-ing geotechnical and geophysical methods. Landslides, 6, 1–15. https://doi.org/10.1007/s10346-009-0144-9

Malehmir, A., Bastani, M., Krawczyk, C. M., Gurk, M., Ismail, N., Polom, U. & Persson, L. (2013a) Geophysical assessment and geotechnical investigation of quick-clay landslides–a Swedish case study. Near Surface Geophysics, 11, 341–350. https://doi.org/10.3997/1873-0604.2013010

Malehmir, A., Saleem, M. U., & Bastani, M. (2013b) High-resolution reflection seis-mic investigations of quick clay and associated formations at a landslide scar in southwest Sweden. Journal of Applied Geophysics, 92, 84–102. https://doi.org/10.1016/j.jappgep.2013.02.013

Malehmir, A., Socco, L. V., Bastani, M., Krawczyk, C. M., Pfaffhuber, A. A., Miller, R. D., Maurer, H., Frauenfelder, R., Suto, K., Bazin, S., Merz, K., & Dahlin, T. (2016) Near-surface geophysical characterization of areas prone to natural haz-ards: a review of the current and perspective on the future. In L. Nielsen (Eds.), Advances in Geophysics (Vol. 57, pp. 51–146). Elsevier. https://doi.org/10.1016/bs.agph.2016.08.001

Nadim, F., Pedersen, S. A. S., Schmidt-Thomé, P., Sigmundsson, F., & Engdahl, M, (2008) Natural hazards in Nordic countries. Episodes, 31, 176–184.

Odenstad, S. (1958) Jordskredet i Göta den 7 juni 1957. Geologiska Föreningen i Stockholm Förhandlingar, 80(1), 76–86. https://doi.org/10.1080/11035895809447207

Osterman, J. (1963) Studies of the properties and formation of quick clays. In 12th National Conference on Clays and Clay Minerals, 12(1):87–108.

69

Polom, U., Bagge, M., Wadas, S., Winsemann, J., Brandes, C., Binot, F., & Kraw-czyk, C. M. (2013) Surveying near-surface depocentres by means of shear wave seismics. First Break, 31(8), 67–79.

Quigley, R. M. (1980) Geology, mineralogy and geochemistry of Canadian soft soils: a geotechnical perspective. Canadian Geotechnical Journal, 17, 261–285. https://doi.org/10.1139/t80-026

Rankka, K., Andersson-Sköld, Y., Hultén, C., Larsson, R., Leroux, V. & Dahlin, T. (2004) Quick clay in Sweden (Report 65). Linköping, Sweden: Swedish Geotech-nical Institute.

RG WinLogger (2007) Operating manual. Version 1.0. Robertson Geologging Ltd, Deganwy, Conwy.

Rosenqvist, I. T. (1953) Considerations on the sensitivity of Norwegian quick clays. Geotechnique, 3(5), 195–200.

Rutherford, S. R., & Williams, R. H. (1989) Amplitude-versus-offset variations in gas sands. Geophysics, 54(6), 680–688. https://doi.org/10.1190/1.1442696

Sauvin, G., Lecomte, I., Bazin, S., L’Heureux, J. S., Vanneste, M., Solberg, I. L., & Dalsegg, E. (2013) Towards geophysical and geotechnical integration for quick clay mapping in Norway. Near Surface Geophysics, 11(6), 613–623. https://doi.org/10.3997/1873-0604.2012064

Sauvin, G., Lecomte, I., Bazin, S., Hansen, L., Vanneste, M., & L’Heureux, J.-S. (2014) On the integrated use of geophysics for quick-clay mapping: The Hvitting-foss case study, Norway. Journal of Applied Geophysics, 106, 1–13. https://doi.org/10.1016/j.jappgeo.2014.04.001

Schildgen, T. F., Robinson, R. A. J., Savi, S., Phillips, W. M., Spences, J. Q. G., Bookhagen, B., Scherler, D., Tofelde, S., Alonso, R. N., Kubik, P. W., Binnie, S. A., & Strecker, M. R. (2016) Landscape response to late Pleistocene climate change in NW Argentina: Sediment flux modulated by basin geometry and con-nectivity. Journal of Geophysical Research: Earth Surface, 121, 392–414. https://doi.org/10.1002/2015JF003607

Schmelzbach, C., Green, A. G., & Horstmeyer, H. (2005) Ultra-shallow seismic re-flection imaging in a region characterized by high source-generated noise. Near Surface Geophysics, 3(1), 33–46. https://doi.org/10.3997/1873-0604.2004027

Shan, C., Bastani, M., Malehmir, A., Persson, L. & Engdahl, M. (2014) Integrated 2D modelling and interpretation of geophysical and geotechnical data to delineate quick clays at a landslide site in southwest Sweden. Geophysics, 79(4), EN61–EN75. https://doi.org/10.1190/GEO2013-0201.1

Shan, C., Bastani, M., Malehmir, A., Persson, L., & Lundberg, E. (2016) Integration of controlled-source and radio magnetotellurics, electric resistivity tomography, and reflection seismics to delineate 3D structures of a quick-clay landslide site in southwest Sweden. Geophysics, 81(1), B13–B29. https://doi.org/10.1190/GEO2014-0386.1

Shuey, R. T. (1985) A simplification of the Zoeppritz equations. Geophysics, 50(4), 609–614. https://doi.org/10.1190/1.1441936

Skempton, (1953) The colloidal activity of clay. In Proceedings of the 7th Interna-tional Conference on Soil Mechanics and Foundation Engineering, 1, 57–61.

Söderblom, R. (1966) Chemical aspects of quick clay formation. Engineering Geol-ogy, 1(6), 415–431. https://doi.org/10.1016/0013-7952(66)90018-4

Solberg, I., Hansen, L., Steinar Rønning, J., Haugen, E. D., Dalsegg, E., & Tønnesen, J. F. (2012) Combined geophysical and geotechnical approach to ground investi-gations and hazard zonation of a quick clay area, mid Norway. Bulleting of Engi-neering Geology and the Environment, 71, 119–133. https://doi.org/10.1007/s10064-011-0363-x

70

Swedish Geotechnical Institute (2012a) Landslide risks in the Göta River valley in a changing climate (Final report, Part 1 – Societal consequences). Linköping, Swe-den.

Swedish Geotechnical Institute (2012b) Landslide risks in the Göta River valley in a changing climate (Final report, Part 2 - Mapping). Linköping, Sweden.

Swedish Government Official Reports (2007) Sweden facing climate change–threats and opportunities (Final report from the Swedish Commission on Climate and Vulnerability, SOU 2007:60). Stockholm, Sweden.

Thakur, V., Degago, S. A., Oset, F., Aabøe, R., Dolva, B. K., Aunaas, K., Nyheim, T., Lyche, E., Jensen, O. A., Sæter, M. B., Robsrud, A., Viklund, M., Nigussie, D., & L’Heureux, J.-S. (2014) Characterization of post-failure movements of landslides in soft sensitive clays. In J.-S. L’Heureux et al. (Eds.), Landslides in sensitive clays: from geosciences to risk management, Advances in Natural and Technological Hazards Research (Vol. 36, pp. 91–103). Springer. https://doi.org/10.1007/978-94-007-7079-9_8

Thompson, R. & Oldfield, F. (1986) Environmental magnetism. Allen and Unwin, London.

Torrance, J. K. (1974) A laboratory investigation of the effect of leaching on the com-pressibility and shear strength of Norwegian marine clays. Geotechnique, 24(2), 155–173. https://doi.org/10.1680/geot.1974.24.2.155

Torrance, J. K. (1979) Post-depositional changes in the pore-water chemistry of the sensitive clays of the Ottawa area, eastern Canada. Engineering Geology, 14(2-3), 135–147. https://doi.org/10.1016/0013-7952(79)90081-4

Torrance, J. K. (2012) Landslides in quick clay. In Clague and Stead (Eds.), Land-slides: Types, Mechanisms and Modelling (Chapter 8, pp. 83–94). Cambridge University Press. https://doi.org/10.1017/CBO9780511740367.009

Torrance, J. K. & Pirnat, M. (1984) Effect of pH on the rheology of marine clay from the site of the South Nation River, Canada, landslide of 1971. Clays and Clay Minerals, 32(5), 384–390. https://doi.org/10.1346/CCMN.1984.0320506

Wang, S., Malehmir, A., & Bastani, M. (2016) Geophysical characterization of areas prone to quick-clay landslides using radio-magnetotelluric and seismic methods. Tectonophysics, 677-678, 248–260. https://doi.org/10.1016/j.tecto.2016.04.020

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1775

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-375568

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019

multidisciplinary near-surface investigation of a quick-clay...

Documents