a subsurface evacuation model for submarine slope failure

Asubsurface evacuationmodel for submarine slopefailureSuzanne Bulln, Joe Cartwrightn and Mads Huusewn3DLab, School of Earth Ocean and Planetary Sciences, Cardiff University, Cardiff, UKwDepartment of Geology and PetroleumGeology, College of Physical Sciences, Kings College, Aberdeen, UK

ABSTRACT

Analysis of three-dimensional (3D) seismic re£ection data from the Norwegian continental marginprovides an insight into an unusual, buried submarine slope failure, which occurred adjacent to thelater Holocene-age Storegga Slide.The identi¢ed failure, informally named the ‘SouthV�ring Slide’(SVS), occurs in ¢ne-grained hemipelagic and contourite sediments on a slope of 0.51, and ischaracterised by a deformed seismic facies unit consisting of closely spaced pyramidal blocks andridges bound by small normal faults striking perpendicular to the slope.TheSVS contrastswith otherpreviously described submarine slope failures in that it cannot be explained by a retrogressive model.The de¢ning characteristic is the high relative volume loss.The area a¡ected by sliding has thinned bysome 40%, seen in combinationwith very modest extension in the translation direction, with linelength balancing yielding an extension value of only 4.5%.The volume loss is explained by themobilisation of an approximately 40m thick interval at the lower part of the unit and its removal frombeneath a thin overburden, which subsequently underwent extensional fragmentation. Evidence forthe mobilisation of a thick ¢ne-grained interval in the development of a submarine slope failure froma continental margin setting may have implications for the origins of other large-scale slope failureson the Norwegian margin and other glacially in£uenced margins worldwide.

INTRODUCTION

Submarine slope failures are a common occurrence oncontinental margins (Canals et al., 2004), where they playa signi¢cant role in their evolution, in£uencing both mor-phology and stratigraphy (Pratson, 2001). Once failureinitiates, the slope failure may progress by means of anumber of mass movement processes, from translationalsliding to £uidal £owage (see Martinsen, 1994, and refer-ences therein). Althoughvarious subdivisions and classi¢ -cation schemes for these processes exist (Martinsen, 1994;Mulder & Cochonant, 1996), each process represents partof a continuum, whereby one type may evolve into or trig-ger another (Martinsen, 1994). As a result, submarineslope failure events can be highly complex and are likelyto have involved a number of processes, possibly active atdi¡erent stages of failure. Because of this complexity, sev-eral important aspects of submarine slope failure develop-ment and occurrence remain poorly understood. Forexample: (1) the physical processes involved in the transi-tion from failure to post-failure stages of development(Locat & Lee, 2002); (2) the mechanisms responsible forgenerating exceptional mobility and long run out dis-tances (Locat & Lee, 2002); and (3) how to better predict

the timing and location of future submarine slope failureevents (Pratson, 2001).

In recentyears, 3D seismic data have proven to be usefulin the continued study of submarine slope failures, andhave been used to increase our knowledge of the detailedmorphology and 3D architecture of their deposits(Huvenne et al., 2002; Frey Martinez et al., 2005, 2006;Moscardelli et al., 2006). The aim of this paper is to use3D seismic re£ection data to describe a submarine slopefailure unit from the Norwegian continental slope thatcontrasts markedly with previously described examples inboth gross morphology and process of origin. Classicalmodels suggest the formation of slope failure events de-pends critically on the necessary failure conditions beingexceeded on a discrete basal shear surface (Bjerrum, 1967;Martel, 2004; Petley et al., 2005).This basal surface evolvesmorphologically during failure to become the base of theslope failure, over which material is translated downslope.On seismic data the basal surface is invariably representedas a sharplyde¢ned seismic facies boundary separating thefailed and translated mass from the underlying, unde-formed slope sediments (e.g. Frey Martinez et al., 2005).The distribution and volume of the translated materialoverlying the basal shear surface is largely a function of itslocationwithin the slope failure, as it is generally expectedthat net depletion in the upslope realm of the slope failureoccurs due to mobilisation and translation of the failedmass downslope. This is balanced at some point down-slope in a zone of general accumulation due to arrest and

Correspondence: Suzanne Bull, 3D Lab, School of Earth Oceanand Planetary Sciences, Cardi¡ University, Main Building, ParkPlace, Cardi¡ CF10 3YE, UK. E-mail: [email protected]

BasinResearch (2009) 21, 433–443, doi: 10.1111/j.1365-2117.2008.00390.x

r 2009 The AuthorsJournal Compilationr Blackwell Publishing Ltd, European Association of Geoscientists & Engineers and International Association of Sedimentologists 433

mailto:[email protected]

mailto:[email protected]

deposition of the mass (sensuVarnes,1978; FreyMartinez etal., 2006).These volumetric changes from the pre-failureslope template, based on the original thickness and mor-phology of the slope before failure, are re£ected in the styleof deformational features present within submarine slopefailure deposits, with extension typically dominating theupslope region, and compression more prevalent down-slope (Lewis, 1971; Varnes, 1978; Farrell, 1984; Martinsen,1994; FreyMartinez et al., 2006).

The submarine slope failure described in this paper dif-fers from the classical models in a number of critical fea-tures. Firstly, it is characterised by extension throughoutthe body with no evidence of compressional deformationtowards the downslope limits. Secondly, a signi¢cant frac-tion of the initial pre-failure slope sediments have been al-most completely removed from the succession, resultingin signi¢cant volume loss that is not balanced by net accu-

mulation elsewhere. Finally, the base of the failure is notde¢ned by a discrete surface that developed due to pro-gressive shear failure. Further sections within this articlewill ¢rstly characterise and describe the failure unit using3D seismic data and propose a mechanism for its develop-ment.

GEOLOGICAL SETTING

This study focuses on a deformed unit situated on thenorthern £ank of the giant Storegga Slide on the Norwe-gian continental margin (Fig. 1a). The Storegga Slide,which occurred 8200 years ago (Bryn et al., 2005a) extendsfrom the M�re and V�ring Basins and involved some3500 km3 of material, a¡ecting an area of ca. 95 000 km2

(Bryn et al., 2005a; Fig. 1a). The basins on this sectorof the Norwegian margin developed due to a prolonged

Fig.1. (a) Location map showing the study area on the Norwegian continental margin and data used. SSC, Storegga Slide Complex;SVS, SouthV�ring Slide. (b) Two dimensional seismic line showing a regional transect across the Storegga area and main stratigraphicsubdivisions of theMiocene^Pleistocene succession. Location shown in A.TWT, two-way travel time.

r 2009 The AuthorsJournal Compilationr Blackwell Publishing Ltd, European Association of Geoscientists & Engineers and International Association of Sedimentologists434

S. Bullet al.

period of extension beginning in the Carboniferous, culmi-nating in the Early Eocene with sea£oor spreading in theNorway-Greenland Sea, and related volcanism (Brekke,2000). Thermal subsidence in the Cretaceous led to thedevelopment of over 10km of sedimentary ¢ll in the V�ringBasin, which remained tectonically active throughout a largepart of the Cenozoic, until a series of compressional domeand arch structures evolved during the Tertiary (Brekke,2000). The M�re Basin was tectonically less active duringthe Cretaceous and Cenozoic, undergoing continuous,thermally driven subsidence (Brekke, 2000).

The ¢ll of the basins is made up of ¢ne-grained hemi-pelagic oozes of the Palaeogene age Brygge and Miocene-earliest Pliocene age Kai Formations, and contourites,hemipelagites and glacigenic sediments of the Plio-Pleis-tocene ageNaust Formation (Evans etal., 2002).Within thestudy area the Naust Formation is subdivided into eightunits (A^H; Evans et al., 2002) to re£ect glacial episodesthat a¡ected the margin from the Late Pliocene onwards,with ice-sheets ¢rst reaching the shelf break at ca. 1.1Ma(Berg etal., 2005). Large-scale sliding in the region, culmi-nating in the formation of theStoreggaSlide at 8.2 kaBP, isthought to be related to climatic variability following onsetof regular shelf glaciations at 0.5Ma (Bryn et al., 2005a).During peak glacial conditions, glacially derived debris£ows were deposited on the continental slope (Berg et al.,2005). Throughout comparatively longer periods of re-duced ice cover, normal marine to distal glaciomarine se-dimentation prevailed with contourites developing on theslope (Berg et al., 2005). Contourites in the study area formlaterally extensive, sheet-like deposits up to 150m thick(Bryn etal., 2005b), and consist of clays with silty and sandylaminae (Berg et al., 2005). Contourite sediments are char-acterised byhighwater contents and the tendency to devel-op excess pore pressures when rapidly loaded by glacialdebris £owdeposits (Bryn et al., 2005b). Laterally continu-ous layers of increased sensitivity within contourite driftsare thought to have served as the main glide planes forslope failures in the area, including the Storegga Slide(Bryn et al., 2005a).

DATASETANDMETHODOLOGY

The primary source of data for this study is a 3D seismicsurvey (Fig.1a), which images an area measuring 2670 km2

on the northern margin of the Storegga Slide (Fig. 1a).These data are 3D time migrated with a bin size (inlineand crossline spacing) of 25 � 25m, and are processed tonear zero phase characteristics.They are displayed accord-ing to the SEG normal polarity meaning that an increasein acoustic impedance across an interface produces a posi-tive amplitude excursion on the seismic trace, which is co-loured red in the ¢nal display.The dominant frequency atthe level of interest is 55Hz, giving a minimumvertical re-solution (l/4) of ca. 9m. Selected 2D seismic lines from aregional grid, and data from one geotechnical borehole(6404/5-GB1; labelled on Fig. 1a) complete the data set.The main focus of the paper is based on the detailed map-ping and analysis of a sub-area of a deformed unit where itis imaged by the 3D survey. An integrated 3D seismicinterpretation approach has been applied, combiningconstruction of time structure and extracted amplitudemaps, time slice interpretation and 3D visualisation(Cartwright, 2007; Bull et al., in press).

3D SEISMIC INTERPRETATION

The deformed unit onwhich this study focuses (labelled inFig. 2) occurs within Naust subdivision B on the southernouter slope of the V�ring Plateau (Figs 1 and 2). Naust B,deposited between 330 and 200 ka, is composed of thin,well-strati¢ed interglacial hemipelagites and contouriteson the outer slope (Fig. 2; Berg et al., 2005).The sedimentsare characterised by low amplitude, continuous seismic fa-cies developed in parallel re£ection con¢gurations (Fig. 2),which de¢ne downslope-thinning packages when mappedin 3D. Upslope on the plateau itself, Naust B is comprisedof seismically massive, lens-shaped bodies interpreted asglacial debris £ow deposits (GDFs of Hjelstuen et al.,2004), which inter¢nger with the hemipelagites and con-

Fig. 2. Two dimensional seismic line showing a SW^NE transect across the study area, illustrating the character and geometry of theNaustA andNaustB sediments.Note the seismically massive and transparent character of theNaustB sediments on theV�ring plateau,where they are comprised of glacial debris £owdeposits, and the transition to laterally continuous, parallel andwell-bedded contouritesheet deposits on the outer plateau slope. Also note the presence of a bottom-simulating re£ector (BSR, labelled) underlying Naust B.Location shown in Fig.1a.


Subsurface evacuationmodel for submarine slope failure

tourites on the outer slope (Fig. 2). The predominantlithology ofNaust B is silt-rich clay, with a highwater con-tent, and decreasing silt content towards the top of the unit(Berg etal., 2005;Hjelstuen etal., 2005).The deformed unitis buried at a depth of approximately 200m below theseabed (based on an assumed seismic velocity of2000ms�1 for the shallow succession) by Naust subdivi-sion A, which represents the last 250 ka of sedimentation(Fig. 3a; Hjelstuen et al., 2004). A laterally continuous bot-tom simulating re£ector (BSR; labelled in Figs 2 and 3)underlies a large portion of the northern margin of theStoregga Slide (Bˇnz et al., 2005), and is present belowthe upslope region of the deformed unit (Fig. 2). It hasbeen shown by previous studies that the BSR is due tothe presence of gas hydrates (Berndt et al., 2004), whichmay have played a minor role in the more recent large-scale sliding on the margin (Bryn et al., 2005a).

The top and base of the deformed unit are both de¢nedby high amplitude positive (red) re£ections (labelled ‘Hor-izonX’and ‘HorizonY’, respectively onFig. 3b).HorizonXis distinctive in that it exhibits a highlydisrupted,‘crinkled’character (Fig. 3b). Detailed study of this horizon revealsthat its appearance is due to the presence of numerous, lat-erally equivalent inclined segments, separated by abruptterminations.The inclined segments dip at up to 301, andform positive relief of up to 12m. Inclined segments oftenshare a common sense and magnitude of dip as individual‘rafts’of re£ectionswithin the interior of theUnit (Fig. 3b).Time structure mapping and extraction of the dip attri-bute for Horizon X has allowed examination of the topUnit surface morphology in detail (Fig. 4a). The inclinedsegments and intervening terminations form an unusual

pattern of interconnected ‘troughs’ separating ‘peaks’ andsometimes more laterally continuous, narrow structuralhighs to give an unusual ribbed or ‘¢ngerprint’ morphol-ogy. The base Unit re£ection, Horizon Y (Fig. 3b), alsoshows evidence of deformation, exhibiting an undulatorycharacter similar to that seen on Horizon X (Fig. 3b).Theplanform morphology of Horizon Y is therefore remark-ably similar to that of Horizon X, although Horizon Yis not a¡ected by faults and the deformation it exhibitsappears closer to ductile folding than brittle fracture.

Correlation of the top and basal surfaces reveals that theUnit forms a downslope tapering wedge across a slopeof 0.51 inclination, characterised internally by inclined,discontinuous ‘rafts’ of re£ections (Fig. 3a). Regional 2Dseismic lineswere used to establish the position of the edgesof the Unit and show it to exhibit an almost elliptical plan-form geometry covering an area measuring approximately850km2 (Fig. 1a).The Unit has a maximum thickness in itsupslope region of 65m, and thins to approximately 35mdownslope (Fig. 3a). At its upslope margin, deformed slopesediments are juxtaposed against updip undeformed slopesediments (Fig. 3a and b), and a key observation is that it ispossible to correlate Horizon Xupdip into the undeformedslope section (Fig. 3b). At this point a marked thicknesschange occurs, with Horizon X being observed to rampup through the interval at an angle of ca.181 in the landwarddirection, by ca. 40m. Landward of this feature, Horizon Xis continuous and parallel to the dip of the slope (Fig. 3b). Asharp topographic break with onlap is observed directlyabove this position inNaustA (right handmargin ofFig. 3b).

Downslope, the Unit terminates above the updipmargin of a further seismic facies unit characterized by

Fig. 3. (a) Dip line showing the up- and downslope margins of the deformed unit. Note that both margins are in¢lled byNaust Asediments. Location shown in Fig. 2. (b) Dip line through upslope region of the deformed unit showing upslope margin and high-amplitude, deformed character of the top and basal re£ections, labelled ‘Horizon X’and ‘HorizonY’, respectively. Note the thicknesschange across the upslope margin, and howHorizon X is readily correlated from the upslope deformed region updip into theundeformed slope sediments.


S. Bullet al.

contorted and disrupted re£ections, with discrete blocksof intact and rotated, but laterally discontinuous re£ec-tions (labelled ‘Unit D’ in Fig. 5). The base of Unit Doccurs at a stratigraphically lower level than the base ofthe deformed unit, by some 200m (Fig. 5).The deformedunit is overlain and in¢lled by youngerNaustB sediments,which are in turn succeeded by those of Naust A (Fig. 3a).Downslope correlation of these packages shows that UnitD is also overlain and in¢lled by lateral and thereforeage-equivalent Naust B sediments (Fig. 3a).

INTERPRETATION

Using established recognition criteria (Moore et al., 1976;Embley & Jacobi, 1977; Woodcock, 1979; Trincardi & Nor-mark, 1989; O’Leary, 1993; Martinsen, 1994; Hamptonet al., 1996; Frey Martinez et al., 2005), the deformed unit isinterpreted as a submarine slope failure, which we infor-mally name the ‘SouthV�ring Slide’ (SVS) from its locationon the outer slope of theV�ringPlateau (Fig.1a). It should benoted, however, that use of the term ‘slide’ is for ease of re-ference and should not be taken to imply a dominant slopefailure mechanism.This interpretation is based on: (1) thecorrelation of distinctive top and basal surfaces (HorizonsX andY) bounding a chaotic seismic facies unitwhich con-trasts signi¢cantly with the more continuous, undisruptedre£ections characteristic of the sedimentary units aboveand below; (2) the abrupt transition from undeformed sedi-

ments updip, to downdip deformed sediments; (3) theabrupt thickness change at the steep ramp which is inter-preted as a headscarp, and upslope of which is preservedthe intact, pre-failure stratigraphy (undeformed slope tem-plate, labelled on Fig. 6a); (4) the alignment of rotated faultsblocks orthogonal to the direction of slope; (5) its planformgeometry, and (6) its context and location in close vicinity toother slope failures (Fig.1; Evans et al., 1996).This interpre-tation is summarised in Fig. 6.

Fig.4. Dip-attribute map of Horizon X showing the remarkableribbed morphology, which characterises the top of the deformedunit.The pattern is comprised by laterally continuous ‘troughs’,shown on the dip map by areas of lighter shading, whichseparated less continuous ‘highs’, shown on the dip map by thedarker areas (both labelled). Location shown in Fig.1a.

Fig. 5. (a) Seismic section showing the downslope margin of thedeformed unit, where it terminates above an abrupt change fromundeformed slope sediments to a further deformed seismicfacies unit (labelled ‘Unit D’), characterised by contorted anddisrupted re£ections with discrete blocks of intact and rotated,but laterally discontinuous re£ections. Note the positions ofHorizons X and Y. Location shown in Fig.1a. (b) Zoomed insection showing in more detail the relationship between thedeformed unit andUnit D. It is possible to correlate Horizons Xand Y (labelled) into the Unit D area.

Fig. 6. (a) Unintepreted seismic section showing the upsloperegion and headscarp of the SouthV�ring Slide (SVS). (b)Summary interpretation diagram for the SVS. Note theundeformed slope template landward of the headscarp, thedegree of thinning observable in the deformed region, and thetransparent interval removed during development of the SVS.



Within the SVS, the abrupt re£ection terminations ando¡sets bounding each of the inclined segments, whichmake up the remarkable ‘crenulate’ character of HorizonX are interpreted as normal faults (Figs. 3a, 7a and b).These tip out at the base of the unit, and delimit inclinedhorizon segments that de¢ne individual fault blocks(Fig. 5a). Examination of seismic sections from both theupslope and downslope limits of the SVS (Fig. 7a and b)reveal that the entire SVS is a¡ected by these extensionalfaults, with fault density only slightly decreasing towardsthe downslope termination. Detailed strain analysis of thefaults was based on line balancing of four representativedip lines taken across the SVS, along with measurementsof strike, dip and throw.The predominant strike of the faultsis NW^SE, normal to the local slope direction (Fig. 7d).Maximum fault throws and dips were used to calculate anextension value of 4.5%, and a net extension of1.25km in thedowndip cumulative slip direction measured from averageheave vectors (assuming dip slip on individual fault planes).

To quantify the degree of thinning observed in theextended region of the SVS, comparisons were madebetween the SVS and the ca. 90m thick undeformed slopesediments preserved updip of the headscarp (labelled‘undeformed slope template’ in Fig. 6b). From measure-ments made on isopach maps of the deformed interval,and by comparing the thickness change observed acrossthe headscarp into the undeformed upslope template withthe thickness of the SVS, it is estimated that a mean thin-ning of 40% has occurred. A total volume loss of 25 km3

is estimated for the SVS when compared with the likelypre-failure con¢guration. Uncertainties in the pre-failure

con¢guration are estimated as o10%, because the 3Dseismic data allows for excellent correlation along the strikeof the slope into a stratigraphically equivalent position.

A critical observation based on stratal correlationupdip into the undeformed slope sediments is that whileHorizon X and Y can be tracked directly updip into theundeformed slope sediments landward of the headscarp(Figs. 3b and 6a), an approximately 40m thick seismicallytransparent and laterally homogeneous unit present in theundeformed slope template is almost completely missingwithin the deformed region of the SVS (Fig. 6).This criti-cal missing interval is labelled ‘transparent interval’onFig.6b, and can be seen from the 3D seismic datato consist almost entirely of this low-amplitude seismicfacies unit. Importantly, we consider the most likely causeof this volume loss to be its remobilisation and removalduring failure. This interpretation of the wholesaleevacuation or depletion of a speci¢c stratigraphic intervalis central to the genetic model presented below. Fromcorrelation with a geotechnical borehole 2 km north of theheadscarp (Fig. 1a), the transparent unit is most probablycomposed of a high water content, clay-rich intervaldeposited as part of a regionally extensive contouritesystem (Berg et al., 2005; Forsberg & Locat, 2005).

An interesting observation concerning the basal re£ec-tion, Horizon Y, is the distinct lack of shear observable(Fig. 3b). Recent studies using 3D seismic data have facili-tated the imaging of large tracts of the basal surfacesof submarine landslides and led to the identi¢cation ofvarious associated features (Frey Martinez et al., 2005;Gee etal., 2005;Moscardelli etal., 2006; Bull etal., in press).

Fig.7. (a) Zoomed in seismic section from near the updip margin of the SouthV�ring Slide (SVS), showing closely spaced normal faultsand fault-bound blocks.Note howHorizonsX andYexhibit parallelism,with‘highs’ inHorizonXunderlain by a high inHorizonY, andvice versa. Location shown on Fig.1a. (b) Zoomed in seismic section from near the downdip margin of the SVS, showing that normalfaults are also present in the downslope region. Location shown in Fig.1a. (c) 3D visualisation illustrating the remarkable ribbed andextended morphology of the SVS due to the fault graben and fault-bound ridges and blocks. Location shown in Fig. 4. (d) Rose plotshowing the dominant orientation of the faults.The arrow indicates the local slope direction.


S. Bullet al.

Basal surfaces are dominated by shear deformation andrecord information related to the dynamic emplacement ofmass wasting events (Gee etal., 2005), such as the passage ofslide blocks which disintegrate and erode the substrate (i.e.the basal surface), forming erosive features such as striationsor grooves which are typically orientated downslope andrecord directly the movement of failed material downslope(Posamentier & Kolla, 2003; Gee et al., 2005). Horizon Ydoes not show evidence of the translation of material down-slope, and is insteaddominated byductile fold-type features(Fig. 3a), which when mapped in 3D, appear to elongateperpendicular to the slope direction, forming a system ofridges, which closely resembles the pattern of ridges andtroughs exhibited byHorizon X.

The SVS is associated with a further downslope failureunit, Unit D, above which it terminates abruptly (Fig. 5aand b). Unit D is also interpreted as a submarine slopefailure (hereinafter referred to as ‘Slide D’), and exhibitsfeatures typical of those associatedwith large-scale, retro-gressive sliding in the region (e.g. Evans et al., 1996; King etal., 1996). Such features include well de¢ned, steep head-scarps with rotated and translated blocks of failed materialseparated by normal faults in the close vicinity, andprogressive disaggregation of material with increasingdistance downslope from the headscarp (Evans etal.,1996).

The close proximity of SlideD to the SVS o¡ers the op-portunity to compare the morphological characteristics ofthe SVS with those more commonly observed in submar-ine slope failures on the Norwegian margin. In someplaces at the downslope margin of the SVS, it is possibleto correlateHorizonsX andYacross the headscarp of SlideD where they overlie Slide D material (Fig. 5b).This rela-tionship is taken as evidence for Slide D occurring beforethe SVS, which then partially in¢lled the Slide D scarp.This relationship is not visible everywhere, however, dueto the resolution limitations of the 3D data. Both the SVSand Slide D are overlain and in¢lled by the same seismicfacies unit, interpreted as Naust A sediments (Fig. 3a).Therefore, it is probable that the two slope failuresoccurred in close temporal succession,withSlideD occur-ring in a downslope position before the later SVS.Deposi-tion of Naust A sediments, which in¢ll the scarps of theSVS and Slide D began at 250 ka (Hjelstuen et al., 2004),providing a minimum age constraint for the slides.

DISCUSSION: A SUBSURFACEEVACUATION MECHANISM FORSUBMARINE SLOPE FAILURE

The de¢ning characteristic of the SVS, which sets it apartfrom other previously described slope failures, is theanomalous volumetric depletion of the lower part of thepre-failure slope unitwith respect to the modest concomi-tant extension of the deformed body in the downslope di-rection.The ca. 40% thinning of the SVS combined withthe observation of the missing transparent interval is con-sistent with a failure mechanism that involves the whole-

sale depletion of a substantial volume of the lower part ofthe original slope template, combinedwith only mild exten-sion and limited downslope translation of material (maxi-mum of 1.25km). The top SVS re£ection, Horizon X, canbe correlated directly across the SVS headscarp (Figs 3band 6a) where it shows a distinct downward shift in strati-graphic level in the downslope direction consistent withthe degree of thinning caused by the removal of the trans-parent interval. Based on these main characteristics, it isclear that the SVS is strikingly di¡erent from the classicalmodels of slope failures described from continental marginsettings (cf.Varnes, 1978; FreyMartinez et al., 2006).

What process then, can account for the detailed mor-phology, internal structure, and critically, the loss of vo-lume from the lower part of the original undeformedslope template? In answer to this question, we suggest ananalogy can be drawn with a process, which is well knownto occur onshore inNorway, and in other glaciated regions.Landslides onshore in glaciated regions can occur as a re-sult of liquefaction remobilisation of deposits of quick clay.Quick clay is a highly sensitive, high water content clayfound in glaciated and uplifted regions such as Canadaand Scandinavia. Such ‘quick clay’ slides are known fortheir rapid and destructive nature and occur when quickclay deposits, which commonly occur shallowly buried be-neath a relatively thin overburden, readily liquefy followingmechanical disturbance (Rosenqvist, 1966; Bjerrum, 1967;Carson, 1977). Critically, the lique¢ed layer then £ows outfrombeneath the immobile overburden,with enlargementby retrogression through expansion of the liquefactionfront (Carson, 1977; Fig. 8). Evacuation of the mobilised

Fig. 8. Diagram to illustrate the progression of a quick claylandslide. Note the remobilisation of the sensitive (quick) clayand its extrusion from beneath the more competent overburden,which then breaks up, with individual blocks undergoingsubsidence or even being rafted by the mobile clay.Modi¢edfrom Abbott (1996).



clay layer results in subsidence, collapse, and break-up ofthe immobile crust, often leaving ridges and rafts strandedon the residuum (Odenstad, 1951; Mollard & Hughes,1973). The base of this landslide is not a discrete shearplane, but rather a contact between the residuum of themobilised layer and the substrate.This sequence of eventsresults in a ¢nal geometry that strongly resembles the3D seismic interpretation of the SVS (e.g. Carson, 1977),in that the SVS also exhibits unambiguous evidence formobilisation and evacuation of almost a complete layer,with an extended immobile unit left juxtaposed on the un-deformed substrate (Fig.7c). Based on this striking geome-trical similarity, subsurface evacuation is suggested as thelikeliest process to result in the development of the SVS.

Using the sequence of deformation reported for quickclay failures, we suggest a model for the developmentof the SVS based on subsurface sediment mobilisationand evacuation (Fig. 9).The model is summarised below:(1) a highwater content, ¢ne-grained unit (the transparentinterval) is deposited and rapidly buried (Fig. 9a); (2)Down£ank undermining (Slide D) releases lateral con¢n-ing pressure at the toe of the transparent interval (Fig. 9b)and initiates liquefaction; (3) the mobilised material issqueezed out from beneath the relatively competent over-burden and extruded into the water column. The rigidoverburden begins to extend, fracture and subside due tothe volumetric depletion of the underlying unit (Fig. 9c),which continues until all of the mobilised material isextruded, forming the collapsed, extended crust observedtoday (Fig.9d). It is quite possible that the evacuated mate-rial, behaving in a highly mobile manner, travelled for asigni¢cant distance downslope and was eventually depos-ited some distance away from the study area and thereforeoutside of the area covered by our data.

The depletion of the transparent interval from beneath amore competent overburden is the key diagnostic evidencearguing for the near complete subterranean evacuation ofthe remobilized sediments. The sediments, once remobi-lised, subsequently escaped downslope to be extrudedinto the water column. Such a model requires a plausiblemechanism for the remobilisation of the material withinthe transparent interval (Fig. 6b). From correlation with anearby borehole (Fig. 1a), the transparent interval is inter-preted to constitute part of the widespread sheet-like con-tourite drift system widely present on the southernV�ringPlateau (Berg et al., 2005; Bryn et al., 2005b; Hjelstuen et al.,2005), which is known to have a high clay content andwatercontent. Contourites are known to form substantially thick,high water content sediment bodies, which readily developoverpressure and sensitivity due to their ¢ne-grainedlithology and rapid accumulation, and are thought to havehosted glide planes of slope failures which occurred in theadjacent Storegga Slide area (Bryn et al., 2005b).

Mobilisation in the shallow subsurface can occur as aresult of many di¡erent processes but is commonly linkedto the development of overpressure, where additional loadis borne by the pore £uid which is unable to de-wateradequately (Maltman & Bolton, 2003). Overpressure can

develop in near surface sediments as a result of initiallyhigh porosity andwater content being maintained becausethe sediment has too low a permeability to de-watere⁄ciently (Collinson, 1994). This e¡ect is particularly

Fig.9. Schematic diagram showing the development of theSouthV�ring Slide. (a)Deposition of the ¢ne-grained, highwatercontent transparent interval (as part of the regionally extensivesheet-drift contourite system), and rapid burial beneath a thincover of more competent sediment. (b) Down£ank underminingby Slide D removes the downslope con¢ning pressure su⁄ce toinitiate mobilisation of the transparent interval. (c)Mobilisedmaterial is extruded from beneath the more competentoverburden, which undergoes extension, fracture andsubsidence. (d) Extrusion of mobilised material and subsidenceof the overburden continues until mobilised material is almostcompletely evacuated. (e) Renewed interpretation summarydiagram for the SVS, showing the ‘transparent interval’ presentwithin the undeformed ‘slope template’. In the deformed region,mobilisation and evacuation of the transparent interval has led toextension, fracturing and subsidence of the overburden, whichnow forms an ‘extended crust’, which comprises the top of theSVS.


S. Bullet al.

common in clay-rich sediments, which have a high pro-portion of platy or acicular particles, leading to a looselypacked structure and a highwater content. Davies & Clark(2006) described an example of a buried submarine slopefailure which occurred on the NE Atlantic margin, whichresulted from overpressure generated by £uid expulsionalong a silica chemical reaction front (Volpi et al., 2003).Similarly, the failure involved mobilization of a lower unitabove which a rigid, coherent unit was translated andbroken up.The failure described byDavies & Clark (2006)di¡ers from the SVS in that no signi¢cant volume lossoccurred, with the exception of localised venting of mate-rial, which escaped upwards along faults in the overbur-den. The failure developed in a con¢ned manner (sensuFreyMartinez et al., 2006), with both the lower mobilizedunit and overlying coherent material buttressed againstdownslope undeformed sediments.

To account for the almost complete depletion of thelower unit in the SVS, we suggest that mobilisation musthave involved a substantial degree of liquefaction of all orpart of the lower unit. Liquefaction involves the total lossof strength of sediment, in which pore £uid pressuresreaches lithostatic values. Although clay-rich sedimentsare generally thought of as being less prone to liquefactionthan coarser-grade sediments due to their cohesivenature, the liquefaction of clayey deposits has been ob-served in number of instances (Gratchev et al., 2006), in-cluding quick clay failures, and has been implicated in thesubmarine environment by Bugge (1981).The susceptibil-ity to liquefaction depends on the content and mineralogyof clays within the sediment (Gratchev et al., 2006). Giventhe large proportional volume of material removed fromthe SVS and the e⁄ciency with which it has been evacu-ated, we favour the involvement of a true liquefactionmechanism as opposed to failure due to overpressurebuild-up and resulting hydroplastic deformation of thefailed interval.The lack of evidence for shear along the baseof the SVS (Fig. 3b) supports this, as slope failures whichoccur due to overpressure build up often result in shearfailure along a discrete layer (the basal shear surface).

Invoking a liquefaction-remobilisation process toaccount for the depletion evident in the SVS is consideredhere the most likely explanation of the observed geometry,but can we infer anything about the speci¢cs of the lique-faction process from what we know of the depositionalcontext? Liquefaction is the mechanism responsible formobilisation of clays in geometrically similar onshorequick clay slides, where it is attributed to a number ofpossible processes linked to pore £uids. These includegroundwater invasion, where the saline pore £uid ofthe clay is partially replaced by fresher water. Changes inthe pore water chemistry result in catastrophic loss ofstrength and liquefaction (Rosenqvist, 1966). Althoughgroundwater invasion is unlikely as a mechanism in theslope setting of the Norwegian margin, there are severalmechanisms that could be viable in a marine setting. Forexample, low-salinity £uids may have interacted withshallowly buried sediments through diagenetic processes.

The contourite drifts of the Naust Fm in this area areunderlain by the polygonally faulted Kai Fm (Fig. 1b),where £uid £ow indicators suggest £uid expulsion hasbeen occurring since Miocene times (Berndt et al., 2004).Another possible £uid interaction could occur as a resultof hydrate dissociation. Gas hydrates are known to bepresent in the study area (Berndt et al., 2004; Brown et al.,2006), as seen from the base hydrate re£ection (BSR;Figs 2 and 3). Dissociation of hydrates in the past wouldhave released signi¢cant quantities of freshwater (Tre¤ hu etal., 2003), and dilution of salinity might have initiatedliquefaction.

In addition to the pore £uid chemical mechanismsfor liquefaction are a group of mechanisms that can beregarded as purely mechanical.Triggering of mass wastingevents in the Storegga area has been linked to seismicitydue to isostatic rebound (Evans et al., 2002). In particular,cyclic loading due to the oscillatory transmission of seis-mic waves has been heavily implicated in submarine slopefailures due its ability to induce elevated pore £uid pres-sures, which often fail to be completely dissipated beforethe next pore £uid response (Maltman,1994).The rapidityof such loading can therefore have the a¡ect of reducingthe e¡ective stress to zero, and liquefaction can occur(Maltman, 1994). Notwithstanding the likely occurrenceof small magnitude earthquakes in the study area at thetime of the SVS, we suggest a mechanical trigger for lique-faction that develops during undermining of a layer(Carson, 1977). Undermining by failure of lateral supportis a common precursor to onshore quick clay slides, withthe removal of lateral con¢ning pressure at the toe of thesusceptible unit often proving su⁄cient to initiate remo-bilisation (Carson,1977). In the SVS area, the developmentof Slide D down£ank from what would become the SVSwould have created a steep, unsupported scarp of ca.120m, which extends along a signi¢cant proportion of theSVS distal margin.With correlation of key horizons fromwithin the SVS into the Slide D area suggesting that SlideD preceded the SVS, unloading at such a substantial andlaterally extensive headwall may well have removed su⁄ -cient lateral support such that liquefaction initiated.Therefore, we favour a triggering mechanism due toundermining as the most likely cause for the SVS.

CONCLUSIONS

1. A submarine slope failure, the SVS, identi¢ed using 3Dseismic data, has been recognised and mapped under-lying the northern margin of the Storegga Slide on theNorwegian margin.

2. In contrast to the many other submarine slope failuresin the area, the SVS exhibits a departure from typicalretrogressive models of slope failure development, andis instead better explained by a subsurface sedimentmobilisation and extrusion mechanism.



3. A ¢ne-grained unit of approximately 40m thicknesswhich makes up part of a regionally extensive contouritesystem has undergone mobilisation and been extrudedinto the water column, forming a heavily depleted,850km2 unit.

4. Based on in¢lling by Naust subdivision A, the SVSoccurred before 250Ka, coincident with the initiationof sliding in the adjacentStoreggaSlide complex (Evanset al., 2002).

5. The recognition of a slope failure that developed due tothe mobilisation of a signi¢cantly thick subsurface unitin a region of such proli¢c large-scale sliding raisesimplications for its role in the development of otherslope failure events, both on the Norwegian margin andother glaciated margins worldwide. If the conditions forremobilisation are met, it maybe that such units arewidespread, but are yet to be identi¢ed.

ACKNOWLEDGEMENTS

We are grateful to Statoil (Andreas Helsem) for providingdata, and to Schlumberger for providing seismic interpre-tation of software. We thank David James for discussionof topics presented in this paper, and three anonymousreviewers whose comments and suggestions greatlyimproved the manuscript.

REFERENCES

Abbott, P.L. (1996)Natural Disasters.Wm. C. Brown PublishingCo, Dubuque, IA.

Berg, K., Solheim, A. & Bryn, P. (2005) The Pleistocene torecent geological development of the Ormen Lange area.Mar.Pet. Geol., 22, 45^56.

Berndt, C., Bˇnz, S., Clayton,T.,Mienert, J. & Saunders,M. (2004) Seismic character of bottom simulating re£ectors:examples from the mid-Norwegian margin. Mar. Pet. Geol.,21, 723^733.

Bjerrum, L. (1967) Progressive failure in slopes of overconsoli-dated plastic clay and clay shales. J. Soil Mech. Foundations Div.Am. Soc. Civil Eng., 93, 1^49.

Brekke, H. (2000) The tectonic evolution of the NorwegianSea and continental margin with emphasis on the V�ring andM�re Basins. In: Dynamics of the Norwegian Margin (Ed. byA. N�ttveddt, et al.)Geol. Soc. Spec. Publ., 167, 327^378.

Brown,E.H.,Holbrook,W.S.,Hornbach,M.J.&Nealon, J.(2006) Slide structure and role of gas hydrate at the northernboundary of the storegga slide, o¡shore Norway. Mar. Geol.,229, 179^186.

Bryn,P.,Berg,K.,Forsberg,C.F.,Solheim,A.&Kvalstad,T.J.(2005a) Explaining the storegga slide.Mar. Pet. Geol., 22, 11^19.

Bryn, P.,Berg,K., Stoker,M.S.,Haflidason,H.& Solheim,A. (2005b) Contourites and their relevance for mass wastingalong theMid-Norwegian margin.Mar. Pet. Geol., 22, 85^96.

Bugge,T. (1981) Submarine slides on theNorwegian continentalmargin, with special emphasis on the Storegga area. Institutt forkontinentalsokkelunders�kelser, 110, 1^152.

Bull, S., Cartwright, J. & Huuse,M. A review of kinematicindicators from MassTransport Complexes using 3D seismicdata.Mar. Pet. Geol. (in press).

Bˇnz, S.,Mienert, J.,Vanneste,M.&Andreassen,K. (2005)Fluid £ow impact on slope failure from 3D seismic data: a casestudy in the Storegga Slide.Basin Res., 17, 109^122.

Canals, M., Lastras, G., Urgeles, R., Casamor, J.L.,Mienert, J., Cattaneo, A., de Batist, M., Haflidason,H., Imbo,Y.,Laberg, J.S., Locat, J., Long,D., Longva,O.,Masson,D.G., Sultan,H.,Trincardi, F. & Bryn, P. (2004)Slope failure dynamics and impacts from sea£oor and shallowsub-sea£oor geophysical data: case studies from the COSTAproject.Mar. Geol., 213, 9^72.

Carson, M.A. (1977) On the retrogression of landslides insensitive muddy sediments.Can. Geotech. J., 14, 582^602.

Cartwright, J. (2007) The impact of 3D seismic data on theunderstanding of compaction, £uid £ow and diagenesis insedimentary basins. J. Geol. Soc. Lond., 164, 881^893.

Collinson, J. (1994) Sedimentary deformational structures. In:TheGeologicalDeformation of Sediments (Ed. byA.Maltman), pp.127^165. Chapman &Hall, London.

Davies, R. & Clark, I. (2006) Submarine slope failure primedand triggered by silica and its diagenesis.Basin Res., 18, 339^350.

Embley,R.W. & Jacobi,R. (1977) Distribution and morphologyof large submarine sediment slides and slumps on theAtlanticcontinental margins.Mar. Geotechnol., 2, 205^208.

Evans,D.,King,E.L.,Kenton,N.H.,Brett,C.&Wallis,D.(1996) Evidence for long-term instability in the Storegga Slideregion o¡ western Norway.Mar. Geol., 130, 281^292.

Evans,D.,Mcgiveron, S.,Harrison, Z., Bryn, P. & Berg,K.

(2002) Along-slope variation in the late Neogene evolution ofthe mid-Norwegian margin in response to uplift and tecton-ism. In:Exhumation of theNorthAtlanticMargin:Timing,Mechan-isms and Implications for Petroleum Exploration (Ed. by A.G. Dore¤ ,J.A. Cartwright, S. Stoker, J.P. Turner & N.White), Geol. Soc.Spec. Publ.,Vol. 196, 139^151.

Farrell, S.G. (1984) A dislocation model applied to slumpstructures, Ainsa Basin, South Central Pyrenees. J. Struct.Geol., 6, 727^736.

Forsberg, C.F. & Locat, J. (2005) Mineralogical and micro-structural development of the sediments on the Mid-Norwe-gian margin.Mar. Pet. Geol., 22, 109^122.

Frey Martinez, J., Cartwright, J. & Hall, B. (2005) 3DSeismic interpretation of slump complexes: examples fromthe continental margin of Israel.Basin Res., 17, 83^108.

Frey Martinez, J., Cartwright, J. & James, D. (2006) Fron-tally con¢nedversus frontally emergent submarine landslides:a 3D seismic characterisation.Mar. Pet. Geol., 23, 585^604.

Gee, M.J.R., Gawthorpe, R.L. & Friedmann, J.S. (2005)Giant striations at the base of a submarine landslide. Mar.Geol., 214, 287^294.

Gratchev, I.B., Sassa,K., Osipov,V.I. & Sokolov,V.N. (2006)The liquefaction of clayey soils under cyclic loading.Eng.Geol.,86, 70^84.

Hampton,M.A.,Lee,H.J. &Locat, J. (1996) Submarine land-slides.Rev. Geophys., 34, 33^59.

Hjelstuen, B.O., Sejrup,H.P.,Haflidason,H.,Nyg�rd, A.,Berstad, I.M. & Knorr, G. (2004) Late Quaternary seismicstratigraphy and geological development of the south V�ringMargin, Norwegian Sea.Quaternary Sci. Rev., 23, 1847^1865.

Hjelstuen, B.O., Sejrup,H.P.,Haflidason,H.,Nyg�rd, A.,Ceramicola, S. & Bryn, P. (2005) Late Cenozoic glacial his-tory and evolution of the StreoggaSlide area and adjacent slide


S. Bullet al.

£ank regions, Norwegian continental margin. Mar. Pet. Geol.,22, 57^69.

Huvenne,V.A.I., Croker-Peter, F. & Henriet, J.P. (2002) Arefreshing 3D view of an ancient sediment collapse and slopefailure.TerraNova, 14, 33^40.

King, E.L., Sejrup, H.P., Haflidason, H., Elverh�i, A. &Aarseth, I. (1996) Quaternary seismic stratigraphy of theNorth Sea fan: glacially-fed gravity £ow aprons hemipelagic se-diments, and large submarine slides.Mar. Geol., 130, 293^315.

Lewis, K.B. (1971) Slumping on a continental slope inclined at1^41. Sedimentology, 16, 97^110.

Locat, J.&Lee,H.J. (2002) Submarine landslides: advances andchallenges.Can. Geotech. J., 39, 193^212.

Maltman, A.J. (1994) Introduction and overview. In:The Geolo-gical Deformation of Sediments (Ed. by A.Maltman), pp.127^165.Chapman &Hall, London.

Maltman, A.J. & Bolton, A. (2003) How sediments becomemobilized. In: Subsurface Sediment Mobilization (Ed. by P.VanRensbergen, R.R. Hills, A.J. Maltman & C.K. Morley), Geol.Soc. Spec. Publ., 216, 9^21.

Martel, S.J. (2004) Mechanics of landslide initiation as a shearfracture phenomenon.Mar. Geol., 203, 319^339.

Martinsen, O.J. (1994) Mass movements. In: The GeologicalDeformation of Sediments (Ed. by A. Maltman), pp. 127^165.Chapman &Hall, London.

Mollard, J.D. &Hughes,G.T. (1973) Earth£ows in the Gron-dines andTroisRivie¤ res area,Que¤ bec: discussion.Can. J.EarthSci., 10, 324^328.

Moore,D.G.,Curray, J.R.&Emmel,F.J. (1976) Large submar-ine slide (olistostrome) associated with Sunda Arc subductionzone, northeast Indian Ocean.Mar. Geol., 21, 211^226.

Moscardelli,L.,Wood,L.&Mann,P. (2006)Mass-transportcomplexes and associated processes in the o¡shore area ofTrinidad and Venezuela.AAPGBull., 90, 1059^1088.

Mulder, T. & Cochonant, H.P. (1996) Classi¢cation ofo¡shore mass movements. J. Sediment. Res., 66, 43^57.

Odenstad, S. (1951) The landslide at Sk˛ttorp on the LidanRiver. Proc. R. Swed. Geotech. Inst., 4, 1^38.

O’leary, D.W. (1993) Submarine mass movement, a formativeprocess of passive continental margins: the Munsen-Nygrenlandslide complex and the southeast New England landslidecomplex. In: Submarine Landslides: Selected Studies in the USExclusive Economic Zone (Ed. by W.C. Schwab, H.J. Lee &D.C.Twichell),USGeol. Survey Bull. 2002, 23^39.

Petley, D.N., Higuchi, T., Petley, D.J., Bulmer, M.H. &Carey, J. (2005) Development of progressive landslide failurein cohesive materials.Geology, 33, 201^204.

Posamentier,H.W.&Kolla,V. (2003) Seismic geomorphologyand stratigraphy of depositional elements in deep-water set-tings. J. Sediment. Res., 73, 367^388.

Pratson,L.F. (2001) A perspective onwhat is known andwhat isnot known about sea£oor instability in the context of conti-nental margin evolution.Mar. Pet. Geol., 18, 499^501.

Rosenqvist, I.T. (1966) Norwegian research into the propertiesof quick clay ^ a review. Eng. Geol., 1, 445^450.

Tre¤ hu, A.M., Bohrmann, G., Rack, F.R.,Torres, M.E., et al.(2003)Proceedings of theOceanDrillingProgram, Initial Reports, 204.

Trincardi, F. & Normark, W.R. (1989) Pleistocene Suveroslide, Paola basin, southern Italy.Mar. Pet. Geol., 6, 324^335.

Varnes,D.J. (1978)Slope movement types and processes. In:Land-slides, Analysis and Control. Special Report (Ed. by R.L. Schuster &R.J. Krizek),Natl. Acad. Sci. [Washington] Spec. Rep. 176, 11^33.

Volpi,V., Camerlenghi, A., Hillenbrand, C.-D., Rebesco,M. & Ivaldi, R. (2003) E¡ects of biogenic silica on sedimentcompaction and slope stability on the Paci¢c margin of theAntarctic Peninsula.Bas. Res., 15, 339^363.

Woodcock,N.H. (1979) Sizes of submarine landslides and theirsigni¢cance. J. Struct. Geol., 1, 137^142.

Manuscript received 30 April 2008; Manuscript accepted 8October 2008.



a subsurface evacuation model for submarine slope failure

Documents