Stability, mobility and failure mechanism for landslides at the upper continental slope off Vesterålen, Norway

�������� ����� ��

Stability, mobility and failure mechanism for landslides at the Upper Conti-nental Slope off Vesteralen, Norway

J.S. L’Heureux, M. Vanneste, L. Rise, J. Brendryen, C.F. Forsberg, F.Nadim, O. Longva, S. Chand, T.J. Kvalstad, H. Haflidason

PII: S0025-3227(13)00202-8DOI: doi: 10.1016/j.margeo.2013.09.009Reference: MARGO 4987

To appear in: Marine Geology

Received date: 4 October 2012Revised date: 22 September 2013Accepted date: 23 September 2013

Please cite this article as: L’Heureux, J.S., Vanneste, M., Rise, L., Brendryen, J., Fors-berg, C.F., Nadim, F., Longva, O., Chand, S., Kvalstad, T.J., Haflidason, H., Stabil-ity, mobility and failure mechanism for landslides at the Upper Continental Slope offVesteralen, Norway, Marine Geology (2013), doi: 10.1016/j.margeo.2013.09.009

This is a PDF file of an unedited manuscript that has been accepted for publication.As a service to our customers we are providing this early version of the manuscript.The manuscript will undergo copyediting, typesetting, and review of the resulting proofbefore it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers thatapply to the journal pertain.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

1

Stability, mobility and failure mechanism for landslides at the

Upper Continental Slope off Vesterålen, Norway

J.S L’Heureux1, M. Vanneste

2, L. Rise

3, J. Brendryen

4, C.F. Forsberg

2, F. Nadim

2, O. Longva

3, S.

Chand3, T.J. Kvalstad

2, H. Haflidason

4

1- Norwegian Geotechnical Institute (NGI), Trondheim, Norway

2- Norwegian Geotechnical Institute (NGI), Oslo, Norway

3- Geological Survey of Norway (NGU), Trondheim, Norway

4- Department of Earth Science, University of Bergen (UIB), Bergen, Norway

Abstract

Several relatively small and spatially-isolated landslides with low mobility characterise the

geomorphology of the upper continental slope off the Vesterålen islands. Here, we present results

from a multidisciplinary study that integrates swath bathymetry data, high-resolution seismic

reflection profiles and a multitude of geological and geotechnical laboratory tests from a 12 m

long piston core in order to investigate the origin and hazard potential of these shallow landslides.

Four of the landslides have their upper headwall around the 500 m isobath, whereas the main

escarpments of another four landslides lie around 700 to 800 m. The slip planes of the

translational landslides lie within laminated glacial marine clays, overlying a well-defined seismic

horizon. These clays have a higher plasticity and water content compared to the surrounding soils

(sandy clays), and they exhibit a modest strain-softening behaviour in triaxial tests. The

interdisciplinary data set is used as input to various numerical analyses in order to assess the

failure and triggering mechanisms for these landslides, as well as their hazard potential. Stability

analyses, dynamic analyses and post-earthquake pore pressure dissipation modelling suggest that

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

2

the margin is essentially stable and that it would require a large magnitude earthquake to trigger

landslides. The resulting deformation and excess pore pressure generation occur primarily within

the top 10 m of the soil, and they become more pronounced towards the surface. Finally, the run-

out distance of these landslides is limited and strongly depends on the volume of displaced

material and the slope angle. Mobility analyses reveal that the acceleration phase lasts about 1

minute and that peak velocities may have reached up to 17 m/s. Hence, the consequences of such

flows during a time of active seabed exploitation or the impact with seabed infrastructure could

be devastating.

Keywords: Lofoten and Vesterålen margin; submarine landslides; slope stability; dynamic

analysis; landslide mobility; pre-conditioning and triggering factors; hazard potential.

1. Introduction

Submarine landslides represent a major threat for both offshore infrastructures and coastal

communities. Assessing the hazard posed by such landslides is a challenging task in offshore

engineering projects. It requires information about the timing and frequency-magnitude of the

events, the soil conditions, the potential triggering factors and the post-failure behaviour of the

landslide. One should also take into account the potential for reactivation in areas where slope

failures have taken place previously.

The causes of submarine mass movements on high-latitude continental margins are often linked

to factors such as high sedimentation rates, unfavourable soil layering, climate variability

affecting sedimentation processes, erosion/oversteepening, fluid flow factors (e.g. gas, hydrates,

diapirism), earthquake, or a combination of these (Locat and Lee, 2002; Masson et al., 2006).

However, for many submarine mass movements, the actual trigger is unknown and difficult to

pinpoint as multiple factors can play a role.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

3

In 2010, the Norwegian Petroleum Directorate (NPD) estimated that 200 million standard cubic

metres of oil equivalent could be found on the continental shelf and slope off Lofoten and

Vesterålen (NPD, 2010). Hence, the Lofoten and Vesterålen margin (LVM) now receives

increased attention as it may be one of the next, albeit environmentally debated, areas opened for

hydrocarbon exploration off Norway. This pristine region of the Norwegian margin is

characterised by several deep canyon systems and numerous landslides have been mapped

(Laberg et al., 2005a; 2007; Thorsnes et al., 2009; Rise et al., 2012). Swath bathymetry data also

reveals several smaller-scale landslides (Rise et al., 2012; Vanneste et al., 2012).

The goal of this study is to increase our knowledge regarding the origin, the causes and the

mobility of the observed landslides on the upper slope off the LVM. The potential failure

mechanisms and mobility are investigated using an integrated data set containing geological,

geophysical and geotechnical data combined with slope stability back-analyses and run-out

analyses. The results are of particular importance for a future hazard assessment of this area and

may have direct implication for the understanding of similar small-scale submarine landslides on

high-latitude continental margins.

2. Regional setting

The study area lies west of Andøya on the Lofoten and Vesterålen passive margin (LVM) (Fig.

1). The continental shelf off LVM is narrow compared to the much wider shelves off mid-

Norway (Rise et al., 2005) and in the Barents Sea (Andreassen et al., 2008). The shelf is only

about 25 km wide west of Andøya. The LVM is part of the north Atlantic margin, which has

experienced multiple rifting episodes since the collapse of the Caledonides in the Devonian (e.g.

Løseth and Tveten, 1996). Following the final opening of the proto-Atlantic Ocean in the Eocene,

the LVM has remained a passive margin. Since the Plio-Pleistocene, glacial processes became

prominent, with periods of shelf-edge glaciations alternated with marine open-ocean conditions

(e.g. Vorren et al., 1998). The seabed geomorphology and stratigraphy in the area distinctly

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

4

reflects the glacial history. Following the Last Glacial Maximum (LGM), glacio-isostatic rebound

took place.

Cross-shelf troughs separate shallow banks on the shelf area. The banks are typically

overconsolidated due to ice loading during peak glacial periods, whereas the throughs were

eroded by fast-flowing ice streams originating in the high mountain areas to the east (Ottesen et

al., 2005). Fifteen canyon systems occur on the LVM, several eroded to 500-1100 m below their

adjacent shoulders (Rise et al., 2013; Fig. 1). The slopes of the canyon walls generally exceed

30°.

Tills, generally less than 20 m thick, dominate the Quaternary succession on the shelf (Sættem

and Rokoengen, 1983). The sediment thickness increases to around 50 m along the outermost

shelf, but also locally in cross-shelf troughs (Rise et al., 1988). Fine-grained glaciomarine

sediments and Holocene sands drape the till in the troughs (Bøe et al., 2009). Along most of the

continental slope, the stratified sediments are mainly of contouritic and hemipelagic

(glaciomarine) origin (Laberg et al., 1999, 2005b; Rise et al., 2012).

Seismic monitoring stations are operating since 1987 in this region (Byrkjeland et al., 2000). The

seismicity of the LVM is relatively high and amongst the highest along the Norwegian

continental margin. Several earthquakes with a magnitude up to 5.7 have been recorded macro-

seismically since 1880 (Bungum et al., 1991; Havskov et al., 1992). The maximum peak ground

accelerations (PGA) with 90% probability of no exceedance for 475- and 10,000-year return

period offshore Vesterålen are shown in Table 1.

3. Data and methods

3.1 Geophysical data

The MAREANO seafloor mapping programme (www.mareano.no) released the multibeam

bathymetry data of the 175 km2

large study area . The data were collected using Simrad EM710

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

5

and EM1002 echosounder systems. The bathymetry grids have cell size of 5 m by 5 m beyond the

12 nm range from the coastline and 50 m by 50 m grid cell size within. In 2010, a total of ~280

km of high-resolution reflection data was acquired using a hull-mounted parametric Topas Sub-

bottom Profiler System (Kongsberg Topas PS18) onboard R/V G.O. Sars. The system

compensates for roll, heave and pitch. It operates with primary frequencies between 15 and 21

kHz, which generate a broadband difference energy between 0.5 and 6 kHz.

3.2 Core data and geotechnical testing

Using a Kullenberg piston core, an 11.9 m long sediment core was recovered in 2010 from an

undisturbed part of the slope between landslide scars (for location, see Fig. 2). This core provided

material for detailed sediment physical logging and geotechnical characterization. X-ray imagery

was performed in addition to multi-sensor core logging (MSCL) for P-wave velocity, bulk

density, and magnetic susceptibility using either a loop sensor on whole cores or point sensor on

split cores. Geotechnical tests included grain size analyses using the Sedigraph method, water

content (i.e., the ratio of mass of water and mass of solids in the soil) and density measurements,

undrained shear strength (fall cone and pocket vane) of undisturbed and remoulded clays, as well

as Atterberg limits (liquid and plastic limits). In addition, four oedometer tests, four

anisotropically-consolidated triaxial tests under compression (CAUC) and two direct simple shear

tests (DSS) were conducted on specific samples from the different lithologies identified on the X-

ray data. All samples were consolidated to in situ effective stresses prior to shearing in the CAUC

and/or the DSS tests. The advanced geotechnical tests followed state-of-the-art practise at the

Norwegian Geotechnical Institute (NGI) (e.g. Sandbækken et al., 1986; Berre, 1998).

Six AMS 14

C dates from marine calcareous fossils allowed establishing an age-depth model. The

14C ages were converted to calendar ages using the Calib 6.0 software (Stuiver et al., 2005) and

the Intcal Marine09 calibration curve (Reimer et al., 2009). A reservoir age of 250±250 was used

in the calibration.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

6

3.3 Stability analyses

To shed light on the failure and triggering mechanism of the landslides off Vesterålen, a series of

deterministic slope stability back-analyses were performed. The Slope/W software was used with

limit equilibrium methods in 2D. The outcome of deterministic analyses comes in the form of

factors of safety (FoS). A slope is stable when the FoS exceeds unity, and unstable when less than

unity. In deterministic analysis, failure takes place when the mobilized shear strength exceeds the

maximum shear strength available. The simplest case occurs when a slab of soil slips on a plane

parallel to the ground surface (infinite slope analysis). This approach is well suited for modelling

translational landslides moving along specific slip planes, e.g. the Vesterålen case study. In

infinite slope stability analyses, the FoS is defined as the shear strength (s) divided by the

mobilized shear stress () (e.g., Nash, 1987):

(1)

where su is the undrained shear strength, γ´ the buoyant unit weight (kN/m3), z is the depth below

seafloor (m) and β the slope angle. By including a uniform horizontal seismic acceleration, Eq.

(1) may be written as:

(2)

where γ is the total unit weight (kN/m3) and Kx is the maximum horizontal component of seismic

acceleration in g.

A pseudo-static FoS < 1 during the peak earthquake-induced load does not necessarily indicate

slope failure with large movement of soil masses. The maximum earthquake load lasts only for a

fraction of a second. As long as the soil does not experience a significant reduction in its shear

strength, the consequence of full mobilization of soil strength along a slip surface will be limited

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

7

permanent deformations. If a slope is subject to an earthquake with peak ground acceleration

greater than the acceleration giving a factor of safety of unity, permanent deformations along the

slip surface will occur. One-dimensional (1D) non-linear earthquake response analyses were

performed in this study using the AMPLE2000 code (Pestana and Nadim, 2000). In this code, the

base rock acceleration time history is scaled to the relevant peak ground acceleration (PGA)

design criteria and the response of a 1D soil column from base rock to seabed is calculated.

AMPLE2000 has several options for the constitutive soil models. In this study, the SIMPLE DSS

model (Pestana et al., 2000) was chosen for simulating the stress-strain-strength relationships for

normally to lightly overconsolidated clays in cyclic simple shear, as this model is best suited for

slope stability assessment (Biscontin et al., 2004). This soil model allows simulating different

shear strain and stress reversal histories as well as providing a realistic description of the

accumulation of plastic shear strains and excess pore pressures during successive loading cycles.

Since AMPLE2000 was specifically designed to describe the response of submerged slopes to

dynamic loading, the inclination of soil layers is an input value describing the geometry and is

automatically accounted for by the program as a state variable. This allows the SIMPLE DSS

model to predict soil response based on the initial shear stresses acting in the sediments.

3.4 Post-failure analysis

The 1D flow dynamic model (BING) presented by Imran et al. (2001a), was used to analyse the

mobility of the failed sediment mass. BING was developed for the study of debris flows and can

be used with various rheological models (i.e. Bingham, Herschel-Bulkley and Bilinear). As no

rheology data are available from the area, both the Herschel-Bulkley and Bilinear rheologies were

included in the analysis. The visco-plastic rheology governed by the Herschel-Bulkley model is

given by the following relationship:

(3)

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

8

where τ is the flow resistance (kPa), τy is the critical yield strength (kPa), γ is the shear rate (-)

and n is the Herschel-Bulkley exponent (-). The parameter K in Eq. (3) is defined as:

(4)

where γr is the strain rate defined as

(5)

and where τya is the yield strength and μdh is the plastic viscosity. Locat (1997) proposed the

Bilinear flow model in order to represent the rheology of clayey silt mixtures which often show

pseudo-plastic flow behaviour. This model assumes that the initial phase of the flow is Newtonian

and evolves – after reaching a threshold shear rate value – into a Bingham type flow. The

constitutive equation for the Bilinear model is given by (Locat 1997):

(6)

where γ0 is the shear rate at the transition from a Newtonian to a Bingham behaviour. In BING,

the constitutive equation for the Bilinear flow is expressed as follows (Imran et al., 2001b):

(7)

where r is the ratio of strain rates expressed as:

(8)

We refer to the section 7 for further details for the rheological parameters used here.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

9

4. Geomorphology and landslide architecture

On the continental margin immediately west of Andøya, the continental shelf break is located at

about 150 m water depth (see SB; Fig. 2). The shelf itself is characterized by several moraine

ridges sub-parallel to the shelf edge (Fig. 2). From the shelf break to about 400 m water depth, the

seafloor gently dips (c. 1–2°) and shows many slope-parallel furrows interpreted as iceberg

scours (see also Laberg et al., 2007). A steeper slope segment (c. 4–5°) between the 400 m to 475

m isobaths marks a transition between the iceberg-grooved seafloor above and the smoother

seafloor morphology below. The bathymetry data from this area reveals the presence of several

smaller-scale but well defined, isolated, landslides with their upper headwall at/or just above the

500 m isobaths (SL1–SL4; Fig. 2). The landslides do not directly connect to either the glacially-

influenced area upslope or the larger canyon systems further downslope.

Landslides SL1 to SL4 have horseshoe escarpments. They are, up to 10 m high and 5 to 10°

steep, with peaks up to 20° (Figs. 2-3). The undisturbed seafloor adjacent to these headwalls has a

maximum gradient of 3–4°. Both the evacuation zone surrounded by the headwalls and

accumulation zone of the landslides are clear on the morphology. The depositional lobes are well

defined with rougher seafloor topography with respect to the surroundings. In the depositional

area, the slope dips at 1–3°. The mobility of these landslides is limited and their total length (i.e.

from the headscarp to the toe of the deposits) ranges from 1500 to 3300 m (Table 2). Another

landslide (SLW) with its headwall at about 700 m water depth has a 15 m high escarpment (Fig.

2). However, it is not possible to identify the deposits for this landslide, as the remobilized mass

was most probably dispersed as a thin veneer of sediment further downslope. The slope angle is

typically less than 1–2° between SLW and SL1-SL4. The bathymetry data also reveal a subtle,

linear depression on the seafloor between these landslides (Fig. 2). The depression is 300 m wide,

6 km long and less than 5 m deep. Several short curvi-linear depressions, 20-25 m wide, 1-2 m

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

10

deep interpreted as seafloor cracks are also detected to the northeast of the lineament in water

depths between 650 and 800 m (Fig. 2).

The upper part of the canyon walls and their shoulders on either side of the study area are

sculpted by landslide scars of various dimensions (Figs. 1-3). The largest landslide (SLA) is part

of a complex extending upslope of the Andøya Canyon system (Fig. 2). Also this landslide has its

upper escarpment along the same steeper slope segment as landslides SL1-SL4. Bathymetric data

shows that the flow of material originating from SLA eroded the seabed downstream towards the

western branch of the Andøya Canyon.

Figure 3 presents characteristic sub-bottom profiles across and along the landslide complexes 3.

The landslide deposits consist of transparent seismic facies, up to 15 m thick (assuming a velocity

of 1660 ms-1

for the sediment) that thin downslope where it eventually overrides the intact

sediment deposits. Despite the short mobilization distance, the failed mass appears entirely

remoulded and deformed resulting in the loss of all internal stratification. However, some chaotic

and discontinuously stratified facies are observed in the proximal parts of the landslides masses

(i.e. within the landslide scar). The stratification outside the landslide areas is mainly

characterized by continuous, parallel to sub-parallel reflections, typical for glacio-marine to

marine sedimentary processes. The horizons labelled a, b and c are distinct reflections and

continuous across the whole area. Horizons b and c seem to have acted as failure planes or lie

close to the slip planes for most of the landslides (Fig. 3). The clean surfaces exposed on the

bathymetric data just below the landslide scars also suggest a control of the bedding on the

location of the failure plane (e.g. SLA and SLW; Fig. 3).

Accurately determining the dimensions of the landslides is difficult due to the limited seismic

data coverage. However, an assessment based on the swath bathymetry data complemented with

the seismic lines gives a first-order volume estimate in the range 6−70·106 m

3 for these landslides

(Table 2).

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

11

5. Sedimentological and geotechnical properties

5.1 Sedimentology

Core GS-10-163-05PC was retrieved from the undisturbed area in between landslides SL2 and

SL3 (Fig. 2). Visual inspection of X-radiographic imagery and split core surface shows that the

core stratigraphy consists of several sedimentary units that can be distinguished by colour,

structure and texture (Figs. 4-5). Three main facies are identified:

Facies 1 is characterized by a structure-less appearance and a variable, but generally high, content

of coarse sand to pebble-sized clasts interpreted as ice rafted debris (IRD) (Fig. 5A). The IRD are

dispersed in a dark grey to brownish silty clay matrix, sometimes concentrated in diffuse strata

(Figs. 4 and 5A). The sand content in this facies often exceeds 20-40 % (weight), causing peaks

in the MSCL logs of density, P-wave velocity and magnetic susceptibility. The sandy clay

sediment is interpreted to be hemipelagically deposited with a large component of IRD (glacial-

marine clay).

Facies 2 is laminated and has much lower IRD content (Fig. 5B). This sedimentary facies occurs

in seven well-defined intervals which can be over one meter thick. IRD, when present, is

concentrated in distinct laminae as well as in some diamictons. This laminated facies have higher

clay and fine silt content compared to the sandy clays of facies 1. The MSCL-values of P-wave

velocity, density and magnetic susceptibility are typically low, but larger values occur over IRD

rich lamina and strata (Fig. 4). Facies 2 is as a silty clay.

Facies 3 is a specific dense and well-graded sandy sediment, only found at a depth of 2.3 to 3.1 m

in core GS-10-163-05PC (Fig. 4). X-ray radiographs reveal that this unit has an erosive base.

Furthermore, the sediments appear remoulded suggesting that they were rapidly deposited, e.g.,

by sediment gravity flows (e.g. a debrite). This facies is absent in nearby cores. Therefore, it is a

local feature and not representative for the general stratigraphy in the area.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

12

Six AMS 14

C dates on planktonic foraminifers and a bivalve provide the chronological framework

for the GS10-163-05PC stratigraphy (Table 3). Except for the shell, the ages fall in chronological

order. The bottom of the core is dated to >30 000 cal. yrs BP (Fig. 4). The debrite between 2.3

and 3.1 mbsf has a maximum age of ~19 000 cal. yrs BP, based on a date obtained from

immediately underneath the debrite.

5.2 Geotechnical properties

From a geotechnical point of view, the sediments fall into two geotechnical soil types: 1) the

sandy clay with a plasticity index of 15-32%; and 2) the silty clay with plasticity index around

25%. The undrained shear strength on the sediments retrieved was thoroughly evaluated using the

fall cone test, torvane tests, triaxial compression tests and direct simple shear tests (Fig. 4). The

ratios of undrained shear strength and effective vertical stress range from 0.27 to 0.52. In general,

the silty clays have lower undrained shear strength, a higher sensitivity (up to 3.5; from fall cone)

and a greater plasticity than the sandy clays with IRD. Oedometer tests on three silty clay samples

from 6.7, 8.7, and 11.6 m depth indicate that the sediments are normally-consolidated. Similarly,

these tests allowed determining a coefficient of consolidation (cv) ranging from 3.0−7.0 x 10-8

m2/s for the silty clays.

During the triaxial and DSS shear tests, shear strains reached typically 15 to 20% (Fig. 6). Modest

strain softening, (i.e., loss of strength after peak strength is reached with ongoing strain)in a silty

clay sample collected at a depth of 8.6 m. The silty clays exhibit a strength reduction of about

22% of peak strength when sheared to 20% (Fig. 6). The sandy clay deposits with high content

IRD showed little to no strain softening. Strength anisotropy in the silty clays, determined from

comparing results from samples subjected to similar consolidation stresses in compression

(CAUC) and in direct simple shear (DSS), yield a ratio of 0.75.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

13

Results from Triaxial and DSS tests were also use to obtain the drained strength parameters of the

soils. The friction angle (φ’) for the silty clays is close to 28° with a cohesion intercept (c’) of 5

kPa. The friction angle estimated from laboratory measurements agrees well with the morphology

of the eroded slopes in the study areas (i.e. slope of escarpments and canyons at c. 30°).

Although X-ray and CT-scans showed little signs of sample disturbance, the oedometer and the

CAUC tests indicate relatively poor sample quality. The sample quality was assessed based on the

volumetric change during re-consolidation to in situ effective stresses and the approach proposed

by Lunne et al. (2006). Sample disturbance can affect the geotechnical measurements, as it may

reduce the peak undrained shear strength. Also the strain-softening behaviour might become less

pronounced compared to perfect conditions and the soil's strength might be too high at large

strains. One should keep this in mind when interpreting the laboratory results in stability analysis.

Correlation of the core and seismic reflection data is hampered by the fact that we lack the P-

wave velocity log from the upper 4.5 m of the core. In addition, the along-slope seismic line ends

200 m from the core location. From 4.5 to 12 m in the core, the mean P-wave velocity from the

MSCL logs is 1660 m/s. The core does not penetrate horizon c at 19–20 ms two way time

thickness (ms twt) (16–17 m). Using a constant velocity of 1660 m/s, horizon a lies at a depth of

8.0 m below the seafloor (9.5 ms twt). Similarly, the uppermost slip plane (i.e. horizon b at 12 ms

twt) corresponds to a depth of c. 10 m below the seafloor. The laminated fine-grained layer with

few IRD-clasts at 9.8–10.2 m may therefore host the failure plane at the depth of horizon b.

6. Slope stability analysis

6.1 Static slope stability analyses

In order to perform 2D stability analyses with the Slope/W software, one requires a morpho-

stratigraphic model with information about the topography prior to failure. One also needs the

soil’s geotechnical properties. From the sub-bottom profiling data across the undisturbed part of

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

14

the Vesterålen margin, the pre-landslide surface topography can safely be reconstructed across the

landslides (Fig. 7). The two soil materials presented in Table 4 define the geotechnical

stratigraphy. Based on the integration of the core and seismic data, the depths of the high

amplitude seismic reflections such as horizon a, b and c are assumed to represent laminated, silty

clay layers. However, and for simplicity, the thinner silty clay layers were not included in the

model. This has very limited influence on the stability analyses.

The silty clays exhibits strength anisotropy. Therefore, the results of the DSS tests are used for

the undrained shear strength (suDSS) along the slip planes. The undrained shear strength is then a

function of the effective overburden stress:

(9)

where σ’v is the vertical effective stress (kPa), αDSS is the normalised undrained DSS strength, and

∆u the excess pore pressure (kPa). αDSS equals 0.27 for the silty clays (Table 3). The inclined

portions of the critical slip surfaces, as shown in Figure 7, pass through the upper layer of the

sandy clay. In such case, the normalised undrained compressive strength, αC is more

representative. In both cases, the undrained shear strength anisotropy is kept constant at 0.75 (see

Table 3). For drained analysis, the Mohr-Coulomb failure criteria is used with estimates of φ’ and

c’ as presented above.

Results from slope stability analyses show that neither the undrained shear strength nor the

drained strength parameters are sufficiently low to produce the observed slope failures off

Vesterålen. In the drained case, the stability analysis yields a FoS of 7.5 and the critical slip

surface does not match the observations, as it tend to go too deep. The calculated FoS is 3.6 in the

undrained case. Here the most critical slip surface matches the depth and partly the extent of the

observed landslide (Fig. 7).

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

15

Slope stability scenarios were further evaluated by including excess pore pressure in the

simulations for static slope conditions, using Slope/W. As a result, critical values were obtained

for the excess pore pressure ratio (Ru) needed for failure to occur. The excess pore pressure ratio

is defined as the ratio between the excess pore pressure (i.e., in situ pressure minus hydrostatic

pressure) and the difference between lithostatic and hydrostatic pressure. For the drained case,

failure requires a Ru value close to 0.9. Again, the most critical slip surface lies deep below the

seafloor, and was discarded as it does not fit with the observed landslide morphology. The excess

pore pressure ratio also affects the soil’s undrained shear strength as shown in Eq. 7. In such case

a Ru coefficient of 0.7 leads to slope failure.

The results above imply that an external load (e.g. earthquake) or a very high excess pore

pressure is necessary to set off the landslides off Vesterålen. Since both the shear strength of the

sediment and the failure surface are fixed in the analyses, the remaining instability factors for

modelling slope failure at Vesterålen are linked to 1) effect of external forces (i.e. earthquake, 2)

generation of excess pore pressure and/or 3) strength degradation. These effects will be evaluated

below.

6.2 Pseudo-static stability analyses

Infinite slope stability analyses (i.e. 1D) show that a peak ground acceleration (PGA) close to

0.18-0.20g is required to trigger a landslide at Vesterålen for 3 to 4° slope angle and undrained

shear strength from DSS tests (Fig. 8). The relative poor sample quality is accounted for in the

analyses by increasing the normalized undrained shear strength from 0.27 to 0.35. Results show

that a 10 000-yr earthquake is necessary to trigger sliding or permanent deformation (Fig. 8).

However, note that sample disturbance does not always cause a reduction in undrained shear

strength. The estimates using the lower normalized shear strength value of 0.27 are more

conservative. In summary, both results from 1D and 2D pseudo-static analysis show that the FoS

drops below one when a large-magnitude earthquake strikes the LVM margin.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

16

6.3 1D dynamic earthquake analyses

For submarine slopes in clays, the pseudo-static factor of safety presents only a rough estimate of

the seismic response of the slope during an earthquake. Earthquakes generate vibrations and mass

inertia forces, which at times cause large shear stresses in the down-slope direction. The duration

of the loads is, however, typically short and in most situations, the main effects are accumulation

of down-slope displacements accompanied by a moderate cyclic degradation of strength. To study

this aspect, 1D non-linear earthquake response analyses were performed in AMPLE2000 using

the SIMPLE DSS constitutive soil model (Pestana and Nadim, 2000). This soil model includes

pore pressure generation and degradation in shear modulus during cyclic loading.

The slope angle in the dynamic analyses was set at 4°. The modelled soil profile is uniform and

normally-consolidated. Rise et al. (2013) shows that the sediment thickness is about 25 m on the

continental shelf and that it increases westwards of the shelf break. In addition, the Topas seismic

data more than 50 m of sediment below the seabed in most of the study area. To evaluate the

influence of the depth to bedrock, simulations were carried out for sediment thicknesses of 25, 50

and 100 m.

No shear wave velocity or shear stiffness at small strains (i.e. Gmax) measurements are available

from this area. Therefore, we rely on published data to estimate Gmax as a function of depth, z, for

normally consolidated clays (e.g. Biscontin et al., 2004):

(10)

The same relationship presented in Eq. (10) was also used in offshore studies on the Møre and

Vøring plateau (NGI, 1998) and off the Helland Hansen area (Leynaud et al., 2004). This

relationship gives a reasonable estimate of Gmax and shear wave velocity when compared to shear

wave velocity from Norwegian marine clays (e.g. Long and Donohue, 2007), and is in agreement

with NGI's rock physical model for low effective stress environments. The undrained strength for

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

17

the inclined soil is anisotropic with normalized SHANSEP properties (Ladd and Foot, 1974). The

viscous damping ratio in the soil column was set constant at 1.5%, which is typical for clay. The

other input parameters into the simple DSS soil model follow the work by Pestana and Nadim

(2000) and Biscontin et al. (2004) (i.e. failure ratio, β=0.35; slenderness parameter, m=0.5; large

strain obliquity angle, ψ=28; backbone curve parameters, Gp=10; θ=25; λ=30). The last two

parameters (i.e. θ and λ) control the effective stress path for loading and the shear stiffness during

cyclic loading, respectively.

Unfortunately, no earthquake data from this area are available. Hence, we selected two well-

known reference earthquake motions from the literature in order to evaluate the effect of different

frequency content on the soil’s response. The Friuli Tarcento event (May 6, 1976; Ms= 6.5,

duration=34.24s) contains mainly high frequencies (periods lower < 0.5 seconds). The second

event is the Loma Prieta earthquake (October 17, 1989; Ms=7.1, duration = 39.95 s) which

contains low and high frequency events. Both records are scaled to a 10 000-year return period

(i.e. 0.224 g) and are specified as rock outcrop motion (cf. Pestana and Nadim 2000). The scaled

acceleration time history and the spectral acceleration for both earthquake records are presented

in Figure 9.

Figure 10 presents the predicted maximum shear strains and displacements as a function of

depth. The results show the accumulation of deformation with increasing importance near the

surface for sediment thicknesses of 25, 50 and 100 m. The largest strains and displacements occur

within the top 10 m of the soil. Similarly, and during both earthquake events, the dynamic FoS

falls below unity for the layers in the uppermost 11.5 m. This depth range fits the depth of the

failure planes observed on the seismic data reasonably well. As seen from Figure 10, the

accumulated strains and displacements are higher for earthquakes with lower frequency

components (e.g., Loma Prieta event) compared to earthquakes with higher frequency

components (e.g., Friuli Tarcento event). The frequency component of the earthquake also has a

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

18

larger influence on the results than the sediment thickness. Figure 11 shows the impact of PGA on

the predicted maximum strains and displacements, considering the Loma Prieta event and a

sediment thickness of 100 m. The modelling returns maximum shear strains of 1% and 0.5% at

depths of 5 m and 10 m, respectively.

Figure 10 also illustrates the predictions of excess pore pressure ratio generated by the Friuli

Tarcento and Loma Prieta events (scaled to 0.224 g). The simulations indicate that the highest

excess pore pressure ratios would occur within the top 15-20 m of the soil. At larger depths, the

excess pore pressure ratios tend to be nearly constant with depth. The results furthermore reveal

large differences in generation of pore pressure with respect to the frequency components of the

earthquake records. Excess pore pressure generation is higher for a seismic event containing low-

frequency components (e.g., Loma Prieta). In this case, the model predicts an excess pore

pressure ratio around 13% at 10-15 m below seafloor at the end of the earthquake (Fig. 10).

6.4 Post-earthquake pore pressure dissipation and slope stability

Excess pore pressure generated in a normally-consolidated deposit dissipates over time following

a seismic event. The time to reach steady-state conditions may be very long for fine-grained soils

(years to decades). During the re-equilibration process, an upward fluid flow towards the seabed

may exist in the soil and this can affect the post-earthquake stability of the slope. Core results as

well as seismic reflection data show that the soil profile off Vesterålen is well stratified. Under

such conditions, upwards migrating pore water may become trapped underneath a soil unit having

lower permeability, ultimately leading to the build-up of excess pore pressure. Consequently,

effective stresses will drop and this process may well undermine the stability of the slope.

Laboratory tests by Kokusho and Kojima (2002) using stratified sandy and silty soils confirmed

this type of delayed failure.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

19

The possibility and implications of this mechanism was investigated using AMPLE2000, by

performing a consolidation analysis and considering seepage normal to the slope. The excess pore

pressure profile predicted at the end of a Loma Prieta type earthquake (100 m sediment thickness)

was considered as an input (Fig. 9). The results are sensitive to the coefficient of consolidation

(cv), which must be representative and realistic, which is challenging. Scatter may be up to

several order of magnitude depending whether the consolidation coefficient is estimated from

laboratory results or in situ (e.g. Leroueil et al., 1995; Leroueil and Hight, 2003). This parameter

also depends on the compressibility of the material, its hydraulic conductivity and the stress level.

The analyses presented here should therefore only be considered as a first approximation, but they

are nevertheless based on the best possible data from the area available. Two different scenarios

were analysed. In the first scenario (scenario I), the coefficient of consolidation does not vary

with depth and equals the average values defined from laboratory results (i.e. 5 x 10-8

m2/s). In the

second scenario (scenario II), the coefficient of consolidation for a layer at 8 to 12 m depth (silty

clay soil) uses the same average value. The other sediment units (i.e. sandy clay soil) are given a

consolidation coefficient that is two orders of magnitude higher. A higher cv below and above this

depth interval is realistic considering the higher sand content in the deposits and thus higher

hydraulic conductivities.

For scenario I, the excess pore pressure ratio decreases very slowly over time (Fig. 12). In

contrast, for scenario II, the silty clay unit with lower permeability traps the upwards flowing

water causing the build-up of excess pore pressures at a depth of 12.5 and 15 m. Excess pore

pressure reaches its peak approximately 4,750 days (c. 13 years) after the earthquake. The

increase in excess pore pressure ratio is substantial for this scenario (i.e., from 11% to 24%; Fig.

12). This leads to a reduction in effective stresses over time within specific intervals, which may

subsequently drive the slope to unstable conditions. Results from pseudo-static analyses

following the build-up of pore pressure suggest that a small earthquake with a PGA of 0.05 (i.e.,

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

20

475 years return period) may trigger slope instability when the excess pore pressure in the slope is

25% (Fig. 8). As such, repeated earthquake activity could be a potential scenario to explain the

landslides on the LVM.

7. Mobility analysis

In order to carry out a mobility analysis of the landslide off Vesterålen with BING, one needs

input on the geometry of the landslide at the onset of failure, a flow path and information about

the rheological parameters of the failed material. This is accomplished in 2D.

7.1 Geometrical parameters

For the mobility analyses, the length, the width and the depth of the failing masses come directly

from the multibeam data and the seismic data (Table 2). The run-out simulation considers that the

failure takes place instantaneously, i.e., the failure release all material at one given time. The

geophysical data support this assumption, as neither the seismic nor the bathymetry data reveal

evidence of multiphase landslide development. It is furthermore in agreement with the slope

stability simulations presented above (e.g. Fig. 7). The shape of the initial landslide masses are

half ellipses in the BING simulation. In reality, the shapes of the landslides are closer to a

rectangular shape but, according to Imran et al. (2001a; b), this has little influence on the final

results. The flow path was derived directly from the multibeam bathymetry.

7.2 Rheological parameters

Rheological data are not available for the landslides off Vesterålen. Therefore, one can has to use

an indirect approach to estimate the yield strength, i.e., based on the geomorphology in the

depositional area. For mudflows and debris flows, the minimum thickness of the remobilized

material (Hc, in metres) for a flow to take place can be approximated using the following

relationship (Hampton, 1972):

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

21

(11)

where γ’ (kN/m3) is the submerged unit weight of the sediment and β is the slope angle.

Constrained by the field data, one obtains values of yield strength, c (kPa), in the range 1.5-3.0

kPa when using an average submerged unit weight of 8 kN/m3 (slope angle in the depositional

zone of 2.0° and Hc of 5-10 m). Similarly, laboratory experiments reported by Locat (1997)

showed that the yield strength of cohesive mixtures depends on the liquidity index (IL):

(12)

This relationship was successfully used to e.g. model the mobility of the giant Storegga landslide

(Gauer et al., 2005). Eq. (12) with an IL value of 0.98 (as measured in the core) gives a yield

strength value around 2.5 kPa. The estimated values of yield strength obtained from Eqs. (11-12)

are slightly lower than the measured remoulded shear strength in the core (Fig. 4).

Values related to plastic viscosity, i.e. the strain rate, γr, and the ratio of strain rates, r, are selected

to match the geomorphology of the landslides and in particular the observed run-out distances. An

r value of 1000 gives adequate results, and this value is comparable to rheological testing results

from other clays (e.g. Locat 1997, Locat and Lee 2005).

7.3 Run-out distances and flow dynamics

The Bilinear flow model was used in BING to compute the relationship between the initial

volume of failed sediments and the run-out distances using the soil parameters given above (Fig.

13). For landslides SL1-SL4, SLW and SLA, the modelled run-out varies between 1800 and 3700

m. In order to reach the run-out distances observed from the geophysical data, each landslide

must have occurred rapidly as one single event (complete volume mobilization, which confirms

the above). Therefore, the various features observed in the failure zone were most likely created

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

22

during the main event(s). The modelled run-out distance for landslide SLA is much larger than

the other landslides, about 6000 m, suggesting that the failed mass must have reached the western

branch of the Andøya canyon. Once entering the canyon, the mass was fed into the main channel

and transported towards the lower part of the continental slope. This is supported by the scouring

and erosive seafloor morphology at the base of SLA (Fig. 3).

BING also allows estimating the velocity of the failing mass, which is an essential parameter with

respect to the potential hazards imposed by these small landslides (i.e. tsunami and/or impact on

seafloor structures). Figure 14 presents the velocity of the debris front during the flow of SL3,

i.e., as a function of the distance travelled from the headscarp. The simulations yield peak flow

velocities of 16 m/s and 17 m/s for the Bingham and Bilinear flow models, respectively. The

velocity increases rapidly over the first 100 m travelled, after which it stabilises near its peak

velocity for nearly 800 to 1000 m before a rapid deceleration. Results from both rheological

models are very similar, in both the behaviour over time and the peak velocities values. The time

between release and final deposition for these landslides is in the order of 10 to 20 minutes. The

acceleration phase alone is completed after only 1.1 minute. The maximum acceleration for

landslide SL3 is in the order of 0.20 m/s2.

8. Discussion

8.1 Failure and triggering mechanisms

The landslides in the study area are interpreted as translational landslides resulting from the

development of a shear surface along well-defined, parallel strata interpreted as glacial marine

sediments. As most of the landslides have common slip planes (i.e. horizons b and c) and

environmental controls, it is likely that the individual landslides have the same preconditioning

factors and triggering mechanism(s). The headwall for landslides SL1-SL4 and SLA all stopped

at a similar water depth at/or just above the 500 m isobaths. This may be explained by 1) a slight

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

23

change in slope angle (c. 0.5 to 1° change), and 2) a thinning of the sedimentary cover above

horizons b and c upslope (see Fig. 3B). Both factors lead to a reduction in the potential energy

available for remoulding the sediments and to an increase in safety factor (e.g. Tavenas et al.,

1983).

Unfortunately, core GS10-163-05PC did not penetrate horizon c. However, samples at the depth

corresponding to horizons a and b show laminated silty clay with higher plasticity than the

surrounding core material (i.e. sandy clay). They also have higher water content, greater fine

content and a contractive behaviour when subjected to shear. The softer and weaker properties of

these horizons are, however, not sufficient to explain how the failure can occur on slope angles

less than 3-4°.

Excess pore pressure is often postulated for triggering submarine landslides since the movement

of large sediment slabs on slope angles less than 3-4° requires very low shear resistance (e.g.

Kvalstad et al., 2005). In the marine environment, excess pore-pressure can results from one, or a

combination, of the following processes: erosion, high sedimentation rate, groundwater seepage

forces resulting from a coastal aquifer, gas hydrate dissociation, earthquakes, mud

diapirism/volcanism, diagenetic processes, and fluid accumulation underneath a permeability

barrier.

At this stage, the Vesterålen area lacks direct and indirect evidence of fluid flow phenomena (free

gas, pockmarks, mud diapirs, hydrate) from geophysical data or core samples. As such, we

exclude these factors in the discussion in order to avoid speculation.

Erosion in clay material causes a reduction in mean stress and an increase in shear stresses which

together lead to the development of negative excess pore pressure (Bishop and Bjerrum 1960). As

the clay deposits swells, the negative excess pore pressure will dissipate until steady-state

conditions are reached. During this process, the factor of safety decreases towards its long term

conditions. In the study area, however, erosion at the base of the continental slope is likely to be a

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

24

slow process which would be difficult to link to excess pore pressure generation required for

destabilizing the slope.

Excess pore-pressure generation due to high sedimentation rates during peak glacial time is an

important factor contributing to submarine slope failure on high-latitude continental margins, e.g.,

the Norwegian-Barents-Svalbard margin (e.g. Dymakis et al., 2000; Bryn et al., 2005; Kvalstad et

al., 2005; Leynaud et al., 2007) and the Canadian east coast (Piper and McCall, 2003; Piper et al.,

2003; Mosher et al., 2004). The rate of overpressure generation depends on the sedimentation

rate, sedimentation duration, sediment compressibility and its permeability. In addition, slope

geometry and flow focusing play a role for overpressure originating from sediment loading (e.g.

Flemings et al., 2002; Stigall and Dugan, 2010; Urgeles et al., 2010; Dugan 2012). Offshore the

Vesterålen margin, however, the sedimentation rate is relatively low (i.e. nearly no sedimentation

during the last 10 000 years and less than 1 m/ka over the last 30 000 years; see Fig. 4). The

reason for these low sedimentation rates is that the high mountain ranges of the Lofoten and

Vesterålen islands formed a natural barrier, such that the mainland ice masses did not connect

directly to the LVM. These mountains diverted the large ice streams into the deep Andfjorden and

Vestfjorden troughs (Fig. 1). Consequently, only small amounts of glacial debris was transported

and distributed on the continental slope (Rise et al., 2013). Excess pore pressures originating from

high sedimentation rate, sediment consolidation and/or coastal aquifers can therefore most

probably be ruled out to explain the landslides on the upper continental slope off Vesterålen.

The required excess pore pressure for failure to take place may have been of limited duration.

This is in agreement with the steep canyon slopes (generally exceeding 30°) and the steep

landslide escarpments. Indeed, and assuming a Mohr–Coulomb failure criterion, the slope angle

in the canyons off Vesterålen and the different landslide scarps would present much lower

gradients (i.e. only a few degrees) if elevated excess pore pressure would still remain in the

slopes.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

25

Earthquakes can easily trigger large submarine instabilities as they typically increase the loads on

a slope and the pore pressures (Locat and Lee 2002). Our analyses show that, during an

earthquake, the largest strains and displacements occur within the top 10 m of the soil which

corresponds well to the landslide thickness (Figs. 2-3). Furthermore, the calculated earthquake-

induced permanent displacement, in the order of 10-20 cm (Figs. 10-11), could explain the

development of the small seafloor cracks and the longer linear depression observed (Fig. 2).

Nadim and Kalsnes (1997) and Nadim et al. (2007) showed that creep strains and reduction of

static shear strength becomes significant in marine clays when the earthquake-induced cyclic

shear strain exceeds c. 1-2%. Since the soils off Vesterålen exhibit a modest contractive

behaviour and have sensitivity up to 3.4, a large earthquake and/or repeated seismicity along the

LVM may have induced a reduction in static shear strength and hence stability. In summary, the

results from pseudo-static calculations, dynamic slope stability analyses and post-earthquake pore

pressure dissipation support the idea of earthquake loading as the main triggering mechanism

(Table 5). However, it is unclear whether or not the landslides occurred because of one single,

large earthquake or repeated earthquake activity. Only dating of the landslide could elucidate this,

however. Unfortunately, accurate dating of the landslides is difficult with the currently available

samples. A larger data set of cores and in situ geotechnical tests would also be beneficial to

understand the failure mechanism following an earthquake in the region as this is governed by

local soil conditions and stratigraphy. Therefore, even if the landslides mapped off Vesterålen

were governed by the same pre-conditioning and triggering mechanisms, they might have

occurred at different times.

8.2 Hazard potential

The seismic activity of the LVM is relatively high and amongst the highest along the Norwegian

continental margin (Byrkjeland et al., 2000). There is macro-seismic evidence of several strong

earthquakes since 1880 (Bungum et al., 1991; Havskov et al. 1992). Globally, however, the

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

26

seismicity is modest, and there are other areas with significantly higher seismicity (e.g., off Japan,

Eastern Mediterranean, Marmara Sea, Chili). Seismicity related to glacio-isostatic rebound

following deglaciation after the LGM was much higher than today. One of high-magnitude

earthquake due to glacio-isostatic rebound was the likely trigger for the giant Storegga landslide

(Kvalstad et al., 2005). As shown from the stability analyses, only a large earthquake or repeated

seismic activity, can also set off landslides in the Vesterålen area. It must be stressed, however,

that repeated seismicity may also result in re-consolidation and excess pore pressure dissipation,

and therefore, strengthening of the soils rather than weakening them. Due to the lack of high

quality soil samples and in situ measurements, the present analyses remain preliminary, and

future research should take into account uncertainty and regional variability in soil parameters.

As seen from the mobility analyses, the run-out distances of the different landslides are limited.

The run-out strongly depends on the volume of displaced material and on the slope angle. The

subtle change in dip downslope of the initial landslide may also be in part responsible for the

rapid deceleration. The acceleration phase is completed within 1 minute for most of the

landslides, and they reach peak velocities up to 17 m/s. These velocities fall in the same range as

those recorded during historical events (e.g. Grand banks landslide, Kuenen (1952);

Orkdalsfjorden landslide, Andersen and Bjerrum (1967) and the Rissa landslide; L’Heureux et al.,

(2012). The socio-economic consequences of such debris flows during a time of active

exploitation or with seabed infrastructure could be devastating.

Relationships between landslide parameters and maximum surface elevation can be used to

provide a first approximation of the tsunami hazard from potential landsliding in the Lofoten-

Vesterålen area. A parametric study by Løvholt et al. (2005) for similar water depths on the

Storegga slide escarpment showed that a landslide with a volume 20 times the ones observed off

Vesterålen (e.g. SL3), would only produce surface elevations in the order of decimetres in coastal

areas (landslide F6 in Løvholt et al., 2005). In that particular analysis, the landslide had a run-out

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

27

distance three times the one observed off Vesterålen and maximum velocities in the range 10-20

m/s. Similar results are obtained if the acceleration of the landslides is considered. It is thus

reasonable to assume that waves from a potential landslide in the Lofoten-Vesterålen area will be

felt locally but will attenuate rapidly.

9. Conclusions

In this study, a combination of geophysical, geological and geotechnical data with slope stability

and mobility analyses was used to understand the occurrence of smaller-scale isolated landslides

off the Lofoten and Vesterålen margin. The main conclusions of the work are:

The slip planes of the isolated landslide features occur in well-defined seismic facies,

which correlate, to laminated glacial marine clays, above a distinct and continuous

reflection. These soils exhibit modest strain-softening. As the landslides have a common

slip planes and the same environmental controls (sedimentation history), it is likely that

these individual landslides have the same preconditioning factors and triggering

mechanism(s).

Compared to many other high-latitude continental margins, the continental slope off

Vesterålen has lower sedimentation rates (nearly no sedimentation during the last 10 000

years and less than 1 m/ka over the last 30 000 years). Therefore, excess pressure due to

sediment loading is highly unlikely to develop. The data currently available do not

witness evidence of fluid flow. Neither the samples nor the seismic data indicate shallow

gas.

The margin per se is stable (Factor of Safety well beyond 1). Results from pseudo-static

analysis, dynamic analysis and post-earthquake pore-pressure dissipation show that

severe earthquake loading with a very low probability of occurrence or repeated

earthquake activity was necessary for these landslides to occur. Large and frequent

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

28

earthquakes may have occurred following the last glaciations due to glacio-isostatic uplift

of the area.

The mobility of these landslides is in general limited. Run-out strongly depends on the

volume of displaced material as well as the slope angle. For most of the landslides, the

acceleration phase only last about 1 minute and the peak velocities may reach up to 17

m/s. Hence, the consequences of such flows during a time of active exploitation or with

seabed infrastructure could be devastating.

The simulations depend strongly on the input parameters, which are determined from a

limited number of samples, geotechnical tests and geophysical data. Therefore, the results

must be seen as preliminary. Additional sediment cores and in situ geotechnical tests data

are planned to gain further insights into the variety of slope processes in this pristine

region of the Norwegian margin.

Acknowledgements

The authors acknowledge funding from the SEABED-Project through the Norwegian DeepWater

Programme (www.ndwp.org) and the MAREANO programme (www.mareano.no) for providing

the multibeam bathymetric data. We are also thankful to Nabil Sultan, Roger Urgeles and David

Piper for their constructive comments, which helped to improve the paper. This is contribution #

XXX from the International Centre for Geohazards.

References

Andreassen, K., Laberg, J.S., Vorren, T.O., 2008. Seafloor geomorphology of the SW Barents

Sea and its glaci-dynamic implications. Geomorphology 97, 157-177.

Andresen, A., Bjerrum, L., 1967. Slides in subaqueous slopes in loose sand and silt. Richards,

A.F. (Ed.), Marine Géotechnique. University of Illinois Press, 221-239.

Berre, T., 1998. Triaxial testing at the Norwegian Geotechnical Institute. Geotechnical Testing

Journal 5: 3-17.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

29

Biscontin, G., Pestana, J.M., Nadim, F., 2004. Seismic triggering of submarine slides in soft

cohesive soil deposits. Marine Geology, 203: 341-354.

Bishop, A.W., Bjerrum, L. 1960. The relevance of the triaxial test to the solution of stability

problems. In Proceedings of the ASCE Research Conference on Shear Strength of Cohesive Soils,

Boulder, Colo., June 1960. ASCE, New York. pp. 437–501.

Bungum H, Alsaker, A., Kvamme, L.B., Hansen, R.A., 1991. Seismicity and seismotectonics of

Norway and surrounding continental shelf areas. Journal of Geophysical Research 96, 2249-2265.

Bryn, P., Berg, K., Forsberg, C.F., Solheim, A., Kvalstad, T.J., 2005. Explaining the Storegga

slide. Marine and Petroleum Geology 22: 11-19.

Byrkjeland, U., Bungum, H., Eldholm, O., 2000. Seismotectonics of the Norwegian continental

margin. Journal of Geophysical Research 105 (B3), 6221-6236.

Bøe, R., Bellec, V.K., Dolan, M.F.J., Buhl-Mortensen, P.B., Buhl-Mortensen, L., Rise, L., 2009.

Giant sand waves in the Hola glacial trough off Vesterålen, North Norway. Marine Geology 267,

36-54.

Dugan, B., 2012. A Review of Overpressure, Flow Focusing, and Slope Failure. In Yamada et al.

(eds) Submarine Mass Movements and Their Consequences (pp. 267-276). Springer Netherlands.

Dimakis, P., Elverhøi, A., Høeg, K., Solheim, A., Harbitz, C., Laberg, J.S.,Marr, J., 2000.

Submarine slope stability on high-latitude glaciated Svalbard–Barents Sea margin. Marine

Geology 162(2), 303-316.

Flemings, P.B., Stump, B.B., Finkbeiner, T., Zoback, M., 2002. Flow focusing in overpressured

sandstones: Theory, observations, and applications. American Journal of Science, 302(10), 827-

855.

Gauer, P., T. J. Kvalstad, C. F. Forsberg, P. Bryn, K. Berg (2005), The last phase of the Storegga

Slide: Simulation of retrogressive slide dynamics and comparison with slide-scar morphology,

Marine and Petroleum Geology 22, 171– 178.

Havskov, J., Kvamme, L.B., Hansen, R.A., Bungum, H., Lindholm, C.D., 1992. The northern

Norway seismic network: Design, operation, and results. Bulletin of the Seismological Society of

America, 82(1), 481-496.

Hampton, M.A., 1972. The role of subaqueous debris flow in generating turbidity currents.

Journal of Sedimentary Research, 42(4), 775-793.

Imran, J., Harff, P., Parker, G., 2001a. A numerical model of submarine debris flow with

graphical user interface. Computers and Geosciences 27, 717–729.

Imran, J., Parker, G., Locat, J., Lee, H., 2001b. A 1-D numerical model of muddy subaqueous and

subaerial debris flows. Journal of hydraulic Engineering 127, 959–968.

Kokusho, T., Kojima, T., 2002. Mechanism for post-liquefaction water film generation in layered

sand. Journal of Geotechnical and Geoenvironmental Engineering 128: 129-137.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

30

Kuenen, P.H., 1952. Estimated size of the Grand Banks turbidity current. American Journal of

Science. 250, 874– 884.

Kvalstad, T.J., Andresen, L., Forsberg, C.F., Berg, K., Bryn, P., Wangen, M. 2005. The Storegga

slide: evaluation of triggering sources and slide mechanics. Marine and Petroleum Geology

22:245-256.

Laberg, J.S., Vorren, T.O., Kenyon, N.H., Ivanov, M., Andersen, E.S., 2005a. A modern canyon-

fed sandy turbidite system of the Norwegian Continental Margin. Norwegian Journal of Geology

85, 267-277.

Laberg, J.S., Stoker, M.S., Dahlgren, K.I.T., de Haas, H., Haflidason, H., Hjelstuen, B.O.,

Nielsen, T., Shannon, P.M., Vorren, T.O., van Weering, T.C.E., Ceramicola, S., 2005b. Cenozoic

along-slope processes and sedimentation on the NW European Atlantic margin. Marine and

Petroleum Geology 22, 1069-1088.

Laberg, J.S., Guidard, S., Mienert, J., Vorren, T.O., Haflidason, H., Nygård,A., 2007.

Morphology and morphogenesis of a high-latitude canyon area; the Andøya Canyon, Norwegian

Sea. Marine Geology 246, 68–85.

Ladd, C.C., Foott, R., 1974. New design procedure for stability of soft clays, Proceedings of

ASCE, GT7 1974 pp. 763–786.

Leroueil, S., Demers, D., La Rochelle, P., Martel, G., Virely, D., 1995. Practical applications of

the piezocone in Champlain sea clays. Proc. Int. Symp. on Cone Penetration Testing, CPT-95,

Linköping, pp. 515–522.

Leroueil, S., Hight, D.W., 2003. Behaviour and properties of natural soils and soft rocks. In T. S.

Tan et al., eds. Characterisation and Engineering Properties of Natural soils. Swets & Zeitlinger,

pp. 29-254.

Leynaud, D., Mienert, J., Nadim, F., 2004. Slope stability assessment of the Helland Hansen area

offshore the mid-Norwegian margin. Marine Geology 213, 457-480.

Leynaud, D., Sultan, N., Mienert, J., 2007. The role of sedimentation rate and permeability in the

slope stability of the formerly glaciated Norwegian continental margin: the Storegga slide model.

Landslides 4, 297-309.

L’Heureux, J.S., Eilertsen, R.S., Glimstad, S., Issler, D., Solberg, I.-L., Harbitz, C.B., 2012. The

1978 quick clay landslide at Rissa, mid-Norway: subaqueous morphology and tsunami

simulations. In: Y. Yamada et al. (Eds.), Submarine Mass Movements and Their Consequences,

Advances in Natural and Technological Hazards Research, 31, 507-516, Springer.

Locat, J., 1997. Normalized rheological behaviour of fine muds and their low properties in a

pseudoplastic regime. ASCE, First Conference on Debris Flows Hazards Mitigation, Mechanics,

Prediction and assessment, pp. 260–269.

Locat, J., Lee, H.J., 2002. Submarine landslides: advances and challenges. Canadian Geotechnical

Journal 39, 193-212.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

31

Locat, J., Lee, H.J., 2005. Subaqueous debris flows. Chapter 9. In: Mathias, Hungr (Eds.), Debris-

flows Hazards and Related Phenomena. Springer, pp. 203–245.

Long, M., Donohue, S., 2007. In situ shear wave velocity from multichannel analysis of sur-face

waves (MASW) tests at eight Norwegian research sites. Canadian Geotechnical Journal 44: 533-

544.

Lunne, T., Berre, T., Andersen, K.H., Strandvik, S., Sjursen, S., 2006. Effects of sample

disturbance and consolidation procedures on measured shear strength of soft marine Norwegian

clays. Canadian Geotechnical Journal, Vol. 43, 726-750.

Løseth, H., Tveten, E., 1996. Post-Caledonian structural evolution of the Lothten and Vesterfilen

offshore and onshore areas, Norwegian Journal of Geology 76, 215-229.

Løvholt, F., Harbitz, C.B., Haugen, K.B., 2005. A parametric study of tsunamis generated by

submarine slides in the Ormen Lange/Storegga area off western Norway. Marine and Petroleum

Geology 22: 219-231.

Masson, D.G., Harbitz, C.B., Wynn, R.B., Pedersen, G., Løvholt, F., 2006. Submarine landslides:

processes, triggers and hazard prediction. Philosophical Transactions of the Royal Society A:

Mathematical, Physical and Engineering Sciences, 364(1845), 2009-2039.

Mosher, D.C., Piper, D.J.W., Campbell, D.C., Jenner, K.A., 2004. Near-surface geology and

sediment-failure geohazards of the central Scotian Slope: American Association of Geologists,

Bulletin, v. 88, p. 703-723.

Nadim, F., Kalsnes, B., 1997. Evaluation of clay strength for seismic slope stability analysis.

Proc., XIV ICSMFE, Hamburg, Germany, Vol. 1, 377-379.

Nadim, F., Biscontin, G., Kaynia, A., 2007. Seismic triggering of submarine slides. In: Offshore

technology conference ′07, OTC paper 18911, Houston, 2007

Nash, D., 1987. A comparative review of limit equilibrium methods of stability analysis. In:

Anderson, M.G., Richards, K.S. (Eds.), Slope Stability. John Wiley and Sons, pp. 11– 75.

Norwegian Geotechnical Institute (NGI), 1998. Møre and Vøring soil investigation 1997, vol. I

and II. Report No. 972521-1, Oslo, Norway.

NORSAR, 1998. Development of a Seismic Zonation for Norway. Norwegian council for

Building Standardization (NBR). 162 p.

NPD. 2010. Petroleum resources in the sea areas off Lofoten, Vesterålen and Senja. 54 pp.

http://www.npd.no/en/Publications/Reports/Petroleum-resources-in-the-sea-areas-off-Lofoten-

Vesteralen-and-Senja/

Ottesen, D., Dowdeswell, J. A., Rise, L., 2005. Submarine landforms and the reconstruction of

fast-flowing ice streams within a large Quaternary ice sheet: The 2500-km-long Norwegian-

Svalbard margin (57–80 N). Geological Society of America Bulletin, 117(7-8), 1033-1050.

Pestana, J.M., G. Biscontin, F. Nadim, Andersen, K.H., 2000. Modeling cyclic behavior of lightly

overconsolidated clays in simple shear. Soil Dynamics and Earthquake Engineering 19, 501–519.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

32

Pestana, J.M., Nadim, F., 2000. Nonlinear site response analysis of submerged slopes. Tech. Rep.

UCB/GT/2000-04, Department of Civil and Environmental Engineering, University of California,

Berkeley, CA.Pestana, J.M. and Biscontin, G,. 2000. A simplified model describing the cyclic

behaviour of lightly overconsolidated clays in simple shear, Geotechnical Engng Report No.

UCB/GT/ 2000-03, Dep. of Civil and Environ. Eng., University of California, 67 pp. Berkeley,

CA.

Piper, D.J.W., McCall, C., 2003. A synthesis of the distribution of submarine mass movements on

the eastern Canadian margin, In: Locat, J. and Mienert, J., eds., Submarine Mass Movements and

Their Consequences: Advances in Natural and Technological Hazards Research, 19: Dordrecht,

Kluwer Academic Publishers, p. 291-298.

Piper, D.J.W., Mosher, D.C., Gauley, B.-J., Jenner, K., Campbell, D.C., 2003. The chronology

and recurrence of submarine mass movements on the continental slope off southeastern Canada,

In: Locat, J. and Mienert, J., eds., Submarine Mass Movements and Their Consequences:

Advances in Natural and Technological Hazards Research, 19: Dordrecht, Kluwer Academic

Publishers, p. 299-306.

Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Bronk Ramsey ,

C., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Hajdas,

I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCormac, F.G.,

Manning, S.W., Reimer, R.W., Richards, D.A., Southon, J.R., Talamo, S., Turney, C.S.M., van

der Plicht, J. and Weyhenmeyer, C.E., 2009. INTCAL 09 and MARINE09 radiocarbon age

calibration curves, 0-50,000 years Cal BP. Radiocarbon, 51 (4). pp. 1111-1150. ISSN 0033-8222

Rise, L., Bøe, R., Riis, F., Bellec, V. K., Laberg, J. S., Eidvin, T., Thorsnes, T., 2013. The

Lofoten-Vesterålen continental margin, North Norway: Canyons and mass-movement activity.

Marine and Petroleum Geology 45: 134-149.

Rise, L., Chand, S., Haflidason, H., L'Heureux, J.-S., Hjelstuen, B.O., Bellec, V., Longva, O.,

Brendryen, J., Vanneste, M., Bøe, R., 2012. Investigations of Slides at the Upper Continental

Slope off Vesterålen, North Norway. In: Yamada, Y., Kawamura, K., Ikehara, K., Ogawa, Y.,

Urgeles, R., Mosher, D.C., Chaytor, J., Strasser, M. (Eds.), Submarine Mass Movements and

Their Consequences (5th International Symposium). Springer, pp. 167-176.

Rise, L., Ottesen, D., Berg, K., Lundin, E., 2005. Large-scale development of the Mid-Norwegian

margin during the last 3 million years. Marine and Petroleum Geology 22 (1-2), 33-44.

Rise, L., Rokoengen, K., Sættem, J., Bugge, T., 1988. Thickness of Quaternary Deposits on the

Mid Norwegian Continental Shelf. M 1:1 000 000. Continental Shelf Institute (IKU). Publication

No. 119, 1988.

Sandbækken, G., Berre, T., Lacasse, S., 1986. Oedometer testing at the Norwegian Geotechnical

Institute. In Consolidation of Soils: Testing and Evaluation, RN Yong and FC Townsend (Eds),

329-353. Philadelphia: American Society for Testing and Materials

Stigall, J., Dugan, B., 2010. Overpressure and earthquake initiated slope failure in the Ursa

region, northern Gulf of Mexico. Journal of Geophysical Research: Solid Earth, 115(B4).

Stuiver, M., Reimer, P. J., Reimer, R. W., 2005. CALIB 5.0. [WWW program and

documentation]. http://calib.qub.ac.uk/calib/

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

33

Sættem, J., Rokoengen, K., 1983. Shallow bedrock geology and Quaternary thickness of northern

Helgeland, Vestfjorden and Lofoten. Continental Shelf Institute (IKU) report P-155/2/83, 44 pp.

Tavenas, F., Flon, P., Leroueil, S., Lebuis, J., 1983. Remolding energy and risk of retrogression in

sensitive clays. Symposium on slopes on soft clays, Linkoping, Swedish Geotechnical Institute

Report, No. 17: 205-262.

Thorsnes, T., Erikstad, L., Dolan, M. F., Bellec, V. K., 2009. Submarine landscapes along the

Lofoten–Vesterålen–Senja margin, northern Norway. Norwegian Journal of Geology, 89(1), 5-16.

Urgeles, R., Locat, J., Sawyer, D.E., Flemings, P.B., Dugan, B., Binh, N.T.T., 2010. History of

pore pressure build up and slope instability in mud-dominated sediments of Ursa Basin, Gulf of

Mexico continental slope. In Mosher et al. (eds) Submarine Mass Movements and Their

Consequences (pp. 179-190). Springer Netherlands.

Vanneste, M., L'Heureux, J.-S., Baeten, N., Brendryen, J., Vardy, M.E., Steiner, A., Forsberg,

C.F., Kvalstad, T.J., Laberg, J.S., Chand, S., Longva, O., Rise, L., Haflidason, H., Hjelstuen,

B.O., Forwick, M., Morgan, E., Lecomte, I., Kopf, A., Vorren, T.O., Reichel, T., 2012. Shallow

Landslides and their Dynamics in Coastal and Deepwater Environments, Norway. In: Yamada,

Y., Kawamura, K., Ikehara, K., Ogawa, Y., Urgeles, R., Mosher, D.C., Chaytor, J., Strasser, M.

(Eds.), Submarine Mass Movements and Their Consequences. Springer, pp. 29-41.

Vorren, T.O., Laberg, J.S., Blaume, F., Dowdeswell, J.A., Kenyon, N.H., Mienert, J., Rumohr, J.,

Werner, F., 1998. The Norwegian–Greenland Sea continental margins: morphology and late

Quaternary sedimentary processes and environment. In: Elverhøi, A. (Ed.), Glacial and Oceanic

History of the Polar North Atlantic Margins, Vol. 17. Quaternary Science Review, pp. 273–302.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

34

Table caption

Table 1: Maximum peak ground acceleration with 90% probability of no exceedance for 475- and

10,000-year return period offshore Vesterålen (NORSAR 1998).

Table 2: Estimated size, volume and run-out distances for landslides SL1-SL4, SLW and SLA. (See

Figures 2-3 for location). Length, thicknesses and width data are given for the failure area.

Table 3: Radiocarbon ages obtained from foraminiferas and a shell in core GS-10-163-05PC.

Table 4: Summary of soil conditions and soil parameters used for the stability analyses at Vesterålen.

Table 5: Summary of stability analyses for the Vesterålen slopes.

Figure caption:

Figure 1: Location of the Lofoten – Vesterålen continental margin. The Andøya Canyon (AC) is the

second northernmost of c. 15 deep features intersecting the continental slope. The study area is

shown with the rectangle. The yellow circles on the Norwegian map represent recorded earthquakes

with a magnitude larger than 3.5 since 1979.

Figure 2: Shaded relief bathymetric image of the study area (100 m contour interval) with named

landslides as well as cracks and a pronounced linear depression (modified from Rise et al. 2012).

Both the landslide scars and accumulation zone are shown, illustrating the low mobility of these

landslides. SB: Shelf break.

Figure 3: 3D bathymetric projection towards the southeast showing detail morphology of the study

area. Section AA’, BB’ and CC’ present sections of high resolution seismic profiles (Topas) through

the landslides.

Figure 4: Stratigraphy and geotechnical data from core GS-10-163-05PC.

Figure 5: X-ray imagery showing A) a typical sandy clay facies (depth in core 390-425 cm) and B) a

typical silty clay with laminae (depth in core 181-216 cm).

Figure 6: (left) Normalised shear stress and (right) normalised pore pressure response upon axial

strain or shear strain from CAUC and DSS strength tests for six samples. Prior to shearing in the

CAUC the samples A2, C6, C2 and D2 were consolidated at 114.7 kPa, 89.9 kPa, 79.8 kPa, and 69.9

kPa, respectively. Samples C4 and D2 were initially consolidated at 85.0 and 71.0 kPa before

shearing in DSS, respectively. Note the modest contractive response for sample C6 collected at a

depth of 8.6 m.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

35

Figure 7: Simplified morpho-stratigraphic model used in the 2D slope stability analysis with

SLOPE/W. The 10 most critical failure surfaces are shown in red while the green band represents the

most critical surface. The post-failure profile is also shown together with the depth of seismic

horizons a, b and c (see Figs. 3 and 4). (Vertical exaggeration [V.E.] is 10).

Figure 8: Factor of safety from infinite slope and 2D slope stability analyses as a function of the peak

ground acceleration (PGA) for landslide SL2. For the 2D analyses with Ru=0 and Ru=0.25 the results

are shown for normalized undrained shear strength (αDSS) ranging from 0.27 to 0.35.

Figure 9: Acceleration time history (scaled to 0.22 g) and spectral acceleration for the Ms 6.5 Friuli

Tarcento earthquake and the Ms= 7.1 Loma Prieta earthquake.

Figure 10: Maximum shear strain, maximum displacement and excess pore pressure ratio at the end

of a Friuli and Loma Prieta type earthquake scale to the 10 000 year return period off Vesterålen

(PGA= 0.22 g). The largest strains and displacements occur within the top 10 m of the soil. The

results are shown for a sediment thickness varying from 25 to 100 m below the seafloor.

Figure 11: Effect of the PGA on the maximum shear strain (left) and displacement (right) at 5 and 15

m below the seabed for a Loma Prieta type earthquake and a sediment thickness of 100 m.

Figure 12: Pore pressure dissipation following a Loma Pierta type earthquake a sediment thickness

of 100 m. Case I: cv is constant with depth at 5 x 10-8

m2/s. Case II: cv for a layer at 8 to 12 m below

seafloor is set to 5 x 10-8

m2/s whereas the overlying and underlying sediments are given a cv that is

two orders of magnitude higher.

Figure 13: Comparison of initial landslide volumes with run-out distances from multibeam

bathymetric data and with the results from BING analyses.

Figure 14: Results from mobility analysis using both Bingham and Bilinear models in BING and

showing the velocity profile of the frontal element as a function of the distance for SL3 off Vesterålen.

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

36

Figure 1

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

37

Figure 2

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

38

Figure 3

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

39

Figure 4

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

40

Figure 5

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

41

Figure 6

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

42

Figure 7

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

43

Figure 8

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

44

Figure 9

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

45

Figure 10

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

46

Figure 11

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

47

Figure 12

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

48

Figure 13

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

49

Figure 14

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

50

Table 1: Maximum peak ground acceleration with 90% probability of no exceedance for 475- and

10,000-year return period offshore Vesterålen (NORSAR 1998).

90% probability of no

exceedance during

time (years)

Return period

(years)

PGA

(g)

50 475 0.051

1000 10 000 0.224

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

51

Table 2: Estimated size, volume and run-out distances for landslides SL1-SL4, SLW and SLA. (See

Figures 2-3 for location). Length, thicknesses and width data are given for the failure area.

Landslide Length (m) Average thickness (m)

Width (m) Landslide volume

(· 106

m3)

Max run-out (m)

SL1 1200 10 1250 15.0 3000

SL2 1300 10 1230 16.0 3350

SL3 650 10 1300 8.5 2000

SL4 500 10 2135 10.6 1460

SLW 700 10 900 6.3 >2000?

SLA 1500 10 4500 >70 > 4000?

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

52

Table 3: Radiocarbon ages obtained from foraminiferas and a shell in core GS-10-163-05PC.

Lab

number

Sample depth

interval (cm) Material

Radiocarbon

age (BP)

(BP = 1950)

Error Delta R Calibrated age (cal BP)

– median probability

Top Base

Beta-

304868 316 318

Foraminifera,

Nps 16 390 60 250 18 949

Beta-

305841 639 640

Foraminifera,

Nps 20 480 80 250 23 691

Beta-

305842 705 706

Foraminifera,

Nps 21 080 80 250 24 377

Beta-

305843 1008 1009

Foraminifera,

Nps 25 190 130 250 29 342

Beta-

304869 1086 1087 Shell 25 870 120 250 30 008

Beta-

304870 1114 1116

Foraminifera,

Nps 25 590 110 250 29 834

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

53

Table 4: Summary of soil conditions and soil parameters used for the stability analyses at Vesterålen.

Material γ (kN/m3)

Clay

content

(%)

Sand

content

(%)

w (%) St αDSS αC αDSS / αC

Sandy clay 17.7 < 40 20-40 45 1.0-

3.0 - 0.52 -

Silty clay 18.4 50-65 < 10 40-47 3.4 0.27 0.46 0.75

γ: bulk unit weight, w: water content, St: sensitivity, αDSS: undrained shear strength ratio from direct simple

shear test, αC: undrained shear strength ratio from anisotropically consolidated undrained triaxial test

(CAUC).

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

54

Table 5: Summary of stability analyses for the Vesterålen slopes.

Type of analysis Factor of

safety,

FoS

Remarks

Static drained FoS > 7 Need an excess pore pressure of 90 % to reach FoS =1.

The depth and length of the critical failure surface does

not match landslide morphology.

Static undrained FoS > 3 Critical slip surface matches the depth and extent of

actual failure.

Pseudo-static analysis

su/’vcDSS

=0.27

su/’vcDSS

increased to 0.35 (to account

for sample disturbance)

FoS > 2

FoS=1

For 475-yr earthquake

For 10 000-yr earthquake

1D dynamic earthquake analyses FoS = 1

Fails in

layer at

11.5m

Accumulation of deformations in top 10 m, 10 000-yr

earthquake.

Post-earthquake slope stability analysis

coupled with pseudo-static analysis

FoS = 1 Accumulation of excess pore-pressure (Ru= 25%)

following a 10 000-yr earthquake.

Pseudostatic analyses with a 475-yr earthquake leads to

slope failure (i.e. repeated earthquake activity).

ACC

EPTE

D M

ANU

SCR

IPT

ACCEPTED MANUSCRIPT

55

Highlights

Landslides on the upper continental slope off the Vesterålen islands are investigated.

The slip planes of the landslides lie within laminated glacial marine clays .

Modest strain-softening behavior is observed in laminated glacial marine clays.

Earthquake analyses suggest that the margin is essentially stable.

Mobility analyses reveal peak velocities up to 17 m/s for these landslides.