Geomorphology of buried glacigenic horizons in the Barents ... et al 2002.pdf · Geomorphology of...

The Barents Sea is a large epicontinental sea,where bathymetry is characterized by aneast–west orientated channel, Bjørnøyrenna,which reaches a depth of 500 m, with shallowerbanks in the north and south (Fig. 1a). An upperregional unconformity (URU) separatesdipping well-bedded pre-glacial sedimentaryrocks of Tertiary age and older (Solheim &Kristoffersen 1984) from the upper, sub-hori-zontal layers of glacigenic sediments (Fig. 1b).The upper regional unconformity is a time-transgressive (diachronous) surface (Lebesbye2000) which represents mainly the incipientformation of ice sheets in the Barents Sea(Richardsen et al. 1991). Although its age isdisputed, it has tentatively been estimated ataround 0.8 million years old (Table 1; Vorren etal. 1988, 1991; Richardsen et al. 1991; Knutsen etal. 1992). Biostratigraphic evidence on thetiming together with the magnitude of erosion(Eidvin et al. 1993, 1998) indicates that thepresent upper regional unconformity landscapeformed during the late Pliocene–Pleistocene,since about 2.5 Ma. The intensity of erosionprobably varied considerably, both in time and

geographically, thus explaining the diachronousnature of the surface (Lebesbye 2000). Accord-ing to Vorren et al. (1991), Bjørnøyrenna wasformed in the mid-Oligocene as a fluvial surface,prior to the onset of late Cenozoic glaciations.At least seven glacial episodes have occurred onthe Barents Shelf since the onset of the glacia-tions (Solheim & Kristoffersen 1984) andregional ice sheets have reached the shelf edgeat least five times during the last 0.8 Ma (Vorrenet al. 1991). The stratigraphy of the glacigenicsequence in the Barents Sea, known as theBarents Sea Synthem (Vorren et al. 1989), hasbeen described and classified by several authors(Solheim & Kristoffersen 1984; Vorren et al.1990; Sættem et al. 1992a). A large proportion ofthe sediments above the upper regional uncon-formity have been interpreted either as sedi-ment deposited directly from ice sheets as till oras proximal glacimarine sediment (Solheim &Kristoffersen 1984; Vorren et al. 1990).

Previous seismic investigations of theglacigenic sequence of the southwesternBarents Sea have been based on 2-D data(Solheim & Kristoffersen 1984; Vorren &

Geomorphology of buried glacigenic horizons in the Barents Seafrom three-dimensional seismic data

B. RAFAELSEN1, K. ANDREASSEN1, L. W. KUILMAN2, E. LEBESBYE1, K.HOGSTAD3 & M. MIDTBØ4

1Department of Geology, University of Tromsø, N-9037 Tromsø, Norway(e-mail: [email protected])

2Norsk Hydro ASA, 0246 Oslo, Norway3Norsk Hydro ASA, 9480 Harstad, Norway

4Norsk Hydro ASA, Pb. 7190, 5020 Bergen, Norway

Abstract: The glacigenic sequence of the southwestern Barents Sea shelf has for the firsttime been studied using 3-D seismic data. The close spacing of 3-D lines and powerfulcomputer workstation interpretation techniques have allowed detailed mapping of theobserved features. Several generations of subglacial lineations observed on four differentpalaeo-surfaces are interpreted to reflect the flow patterns of palaeo-ice sheets. To ourknowledge, this is the first time that multiple levels of subglacial lineations have beenobserved. The lineations are 2.5–8 m in relief, 50 to 180 m wide and 0.5 to 20 km long. Allfour surfaces show a main lineation pattern comprising lineations with a N–S trend,suggesting that the dominant ice flow was directed northwards across the Barents Sea shelfat least four times during the last 0.8 Ma. Two of the surfaces display semi-circular tooblong depressions trending mainly in the same direction as the sub-parallel lineations.These depressions are 9–53 m in relief, 1.25–3.2 km wide and 1.9–9 km long. In contrast tothe buried surfaces, the sea floor is dominated by 2.5–25 m deep cross-cutting icebergplough-marks from the deglaciation phase of the last Barents Sea ice sheet. The 3-Dseismic data are conventional industry data. Despite relatively low seismic frequencies and,hence, limited vertical resolution of seismic profiles, time slices and sub-horizontal timemaps are of high spatial resolution, providing detailed images of different stages of buriedQuaternary glacial geomorphology.

From: DOWDESWELL, J. A. & Ó COFAIGH, C. (eds) 2002. Glacier-Influenced Sedimentation on High-LatitudeContinental Margins. Geological Society, London, Special Publications, 203, 259–276. 0305-8719/02/$15.00© The Geological Society of London 2002.

TN14 Rafaelsen (JG/d) 5/2/03 10:38 am Page 259

Kristoffersen 1986; Vorren et al. 1986, 1988,1989, 1990; Sættem 1992b, 1994; Sættem et al.1994). This paper presents results from the firstuse of 3-D seismic data to study the glacigenicsequence of the Barents Sea. The study area,covering 2870 km2, is located immediately south

of Bjørnøyrenna at water depths ranging from250 to 450 m (Fig. 1). Powerful computer work-station interpretation techniques have alloweddetailed 3-D mapping and visualization of areasand features of particular interest (Rafaelsen etal. 2000). Detailed interpretation of the 3-D

260 B. RAFAELSEN ET AL.

Bjørnøyrenna

I II

b1b1b2

b2?

b2?b4

b5

Pre-glacial strataupper

Upper middle

23 E°50 km

Well 7222/09-U-01(4km off line)

100

(m)

300

400

500

200

60021 E°

Weichselian deposits

SG9810

SG9804

BNW SE

URU

to

Barents Sea

NORWAY

Svalbard

Bjørnøyrenna

70°N

74°N

100

100

200300

400

400

300

200 300

300

200500

1000

1500

2000

I

II

30°E20°E

A

Well 7222/09-U-01

400

300

200300

I

II

SG9803

SG9804SG

9810

Fig. 1. (a) Location and bathymetry of the southwestern Barents Sea. The areas with studied 3-D seismic dataare indicated by black rectangles. (b) Generalized geo-seismic section, location indicated in (a), outlining thestratigraphy of glacigenic sediments in the study area. Rectangles show the approximate location of two of the3-D surveys. Modified from Sættem et al. (1992b).

TN14 Rafaelsen (JG/d) 5/2/03 10:38 am 60

seismic dataset suggests the potential contri-bution of industry 3-D seismic data to studies ofthe shallow sediments in the Barents Sea.

The objectives of this paper are, first, toemphasize the value of conventional 3-Dseismic data in order to derive detailed morpho-logical information of good horizontal resolu-tion from shallow buried horizons; and,secondly, to focus on how the observed morpho-logical features can help us to understandpalaeo-conditions better, in this case the direc-tion of palaeo-ice flow and palaeo-ice sheetextent.

Data and methods

This study is mainly based on 2870 km2 ofindustry 3-D migrated seismic data acquired in1998 (Lillevik & Lyngnes 1998), made availableto the University of Tromsø by the former SagaPetroleum ASA (now Norsk Hydro ASA). 2-Dseismic data, industry well logs and shallowborehole logs have also been used with thepurpose of stratigraphic correlation to previousseismic studies in the southwestern Barents Sea.Constant velocities of 1500 m s–1 and 1750 m s–1

have been applied for the water and theglacigenic sediment layer above the URU whenconverting time to depth (Sættem et al. 1992b).This gives a depth range for the zone of interestof 290–540 m below sea level (bsl).

The three conventional 3-D seismic surveys

SG9803, SG9804 and SG9810 (Fig. 1a) wereacquired with a depth of source and receiver inthe range of 5–8 m. The acquisitions consist ofdual sources (flip-flop shooting) with 75 mseparation and 6 streamers with 100 m separ-ation (Lillevik & Lyngnes 1998).

With the 3000 m long streamers consisting of240 channels and a shot point interval of 25 m,the recorded fold is 30 and the recorded bin sizeis 12.5 m by 37.5 m (inline by cross-line, inlinesare orientated parallel to the survey vessel whilecross-lines are orientated perpendicular to theinlines). As the applied Hale-McClellan migra-tion algorithm carried out post stack requiresinput bins to be quadratic, the data were inter-polated to 12.5 by 12.5 m. The bin size wasfinally converted to 25 by 25 m.

The final fold of the data is 20, after spatialdecimation from 1440 to 720 channels (30 fold),by common receiver interpolation (60 fold),production of super common mid-point (CMP)input to radon transform demultiple (120 fold)and application of partial stack for the dip-moveout (DMO) and pre stack time migrationprocessors (20 fold). With the applied mute, theprocessing-generated fold in the glacigenic sedi-ments varies from 1 to maximum 3.

The 3-D seismic surveys SG9803 and SG9804differ from SG9810 by having a shot pointinterval of 18.75 m, thus giving a recorded foldof 40. However, the end product is the same; afinal fold of 20 and a 25 by 25 m bin size.

GEOMORPHOLOGY OF GLACIGENIC HORIZONS 261

Table 1. Relationship between the present seismo-stratigraphic units defined from 3-D data and earlier workbased on 2-D data

Horizons / units Corresponding units and their agesUnits in Sættemet al. (1992a)

D

E

C

B

A

b4

b1

b2?

b5 5E

4E

b4

I

D2

22–18 ka (Vorren et al. 1989, 1990)

Solheim & Kristoffersen (1984)

200–130 ka (Sættem et al. 1992b)

Younger than URU (< c. 800 ka)

c. 28 ka (Vorren et al. 1989, 1990)

< 27 320 14C years BP (Hald et al. 1990)

bE

bD

bC

bB

bA


Early in the processing, the number ofsamples of the data was reduced by resamplingfrom 2 to 4 ms using a bandpass filter in frontwith a low cut-off of 6 Hz (24 dB/octave) and ahigh cut-off of 90 Hz (72 dB/octave) to attenuatethe swell noise and to avoid anti-aliasing, respec-tively. From filter-tests, the highest frequencywith geological information in the shallow part islimited up to 80–85 Hz, resulting in a maximumvertical resolution (the ability to distinguish indi-vidual reflecting surfaces) of approximately 5 m(=1/4 wavelength). For a dominating frequencyof around 40 Hz the corresponding resolution isapproximately 10 m.

Velocity analyses were run for every 14 3-Dinlines and for every 42 3-D cross-lines to give asquare velocity grid of 525 by 525 m. Based onan early phase interpretation of fast-track data,the processing contractor was required to pickvelocities on selected target and regionalhorizons, including the shallow part.

The horizontal resolution (the ability to distin-guish separate features on a horizon, given by thepost-migrated Fresnel zone; Fig. 2; Brown 1999;Sheriff 1999) is approximately the same as thevertical resolution. Including the relatively high-density grid in the data and the application of the

3-D migration, the total horizontal resolution isgood (Fig. 2). By comparison, conventional 2-Dmigrated seismic data would have an ellipticFresnel zone (Fig. 2) with improved horizontalresolution only along the line.

In 3-D seismic interpretation, the density ofdata-points combined with a snap interpretationtechnique makes it possible to detect sub-sample features. The snap interpretation tech-nique (when searching on maximum amplitude)searches until a true maximum value is foundand calculates a sub-sample time value by fittinga quadratic polynomial to the nearest threesamples. The peak value of this function is usedas a maximum, giving an accuracy of 0.1 msinstead of the sample rate (which is 4 ms). Bymeans of the snap interpretation technique theprogram can detect the depth of features belowthe sample interval, explaining why ploughmarks less than 3 ms (two-way traveltime,TWT) of depth are detectable (Fig. 3). Further,this accuracy is far below the vertical resolution.

The studied 3-D surveys are 16-bit datasets,giving an amplitude range of –32768 to +32767,compared with –127 to +126 in 8-bit datasets.Thus 16-bit datasets give the Automatic SeismicArea Picker (ASAP) more sensitivity in ampli-tude variation, making the automatic interpre-tation easier and subtle amplitude variationsdetectable.

The GeoQuest software GeoFrame 3.8Charisma IMain, running on a Unix work-station, has been used to interpret the seismicdata. Mapping of horizons was achieved mainlyby tracing seismic reflections with the ASAP.The ASAP tool provides a more objective wayof mapping reflections than hand-drawninterpretations and is able to detect and accu-rately visualize details on the interpretedhorizons. Subtle changes in reflection characterthat may look like minor disturbances or noisein the seismic section are generally easily recog-nized on time slices and horizon maps, and mayturn out to be interesting geological featuresthat can be traced over significant areas. A goodreference is plough marks and lineations (Figs 3& 4); features which, in the worst case, wouldnot be detected by manual mapping or, at best,would take several days to interpret in detail.

Comparison of maps generated fromdifferent specific amplitude extremes, such asmaximum amplitude (peak), minimum ampli-tude (trough) or upper/lower zero crossing(where the amplitude changes from positive tonegative, or the opposite), shows that upper andlower zero crossings give the most detailedmaps of the shallow palaeo-surfaces. In order topreserve the details from the sub-sampled snap


Post-migrationFresnelzone

= λ = V

T = 0.5 sF = 40 HzV= 1750 m s-1

λ = Wavelength

Pre-migration Fresnel zone

= V(T/F)0.5 = 195.66 m

10.94 m

Fig. 2. Effect of 2-D and 3-D migration on Fresnelzone size and shape. The Fresnel zone beforemigration (large circle), after 2-D migration (ellipticshape) and after 3-D migration (small black circle).Modified from Brown (1999).



1 km

D N

391.44 565.28

Mound

P

P`

B

500 m500 mP P`

50 mN

TWT

550 -

450 -

391.44 565.28(ms)

A

5 km391.44 565.28

NC

ED

B

NSG9804

C

Sea floor

Fig. 3. (a) Seismic profile showing a cross-section of an asymmetrical curved furrow at the seabed with a ridgeof ploughed sediments on each side. (b) An illuminated time–structure map showing the furrow in (a). (c) Anilluminated time–structure map with curved furrows in a chaotic pattern. (d) An illuminated time–structuremap showing twinned furrows. (e) The approximate location of the illuminated time–structure maps (b), (c)and (d) in 3-D survey SG9804. In all illuminated time–structure maps the light source is located east of thehorizons. Colours show depth in ms (two-way traveltime) below sea level.


interpretation, no smoothing has been appliedto the final maps.

In addition to application of amplitude infor-mation, different attribute maps like instan-taneous frequency, reflection strength, cosine of

phase, apparent polarity, instantaneous phase,dip and azimuth have been generated in anattempt to extract more information from thedataset regarding specific features.

In the search for geological features not


AP

P`

1 km

500 -

600 -

TWT50 m

C

P`

P

1 km

(ms)

B

Footprints

496.08 633.62

N

Lineations

SG9804

bC

URU

Sea

floor

Fig. 4. (a) Seismic profile showing irregularities on horizon bC in SG9804. The green line represents theinterpreted horizon bC. (b) Map with time-structure contours indicating the location of (c) within 3-D surveySG9804. (c) An illuminated time-structure map of horizon bC showing lineations interpreted to be formedsubglacially. The ‘footprints’ aligned parallel to the course of the survey vessel (inlines) are artefacts related tothe acquisition of the 3-D seismic data. Depth in ms (two-way travel-time) below sea level is shown on thecolour bar in (c), but applies for (b) as well. The light source is located east of the horizon and verticalexaggeration (z-axis) is 8�.


related to horizons, a set of volumetric attributesand horizon slices has been generated. Forvisually enhancing the subtle features, a programcalled GeoViz has been used (allowing the userto manipulate colours, select the position andaspect of the light source, the vertical exaggera-tion (z-axis) and the angle of the horizon).

So-called ‘footprints’ are artefacts occurringin the 3-D seismic data. They resemblelineations parallel to the inline direction (Fig.4). The footprints have the same spacingthroughout the dataset (with a separation equalto every 12 common mid point lines) and shouldnot be regarded as geological information, butrather as a result of the pattern of movement ofthe survey vessel.

Stratigraphy of the glacigenic sequence

In the study area, the glacigenic sequence is upto 146 m thick. We have identified five seismichorizons (plus seabed) bounding five glacigenicunits, termed A (oldest) to E (youngest) (Fig. 5).The oldest seismic unit, A, is only observed in 3-D area SG9810 and is so thin that the lineationsobserved at its basal reflection are interpreted tobe dominated by interference from the seismicresponse of the top of unit A. The other, over-lying younger seismic units, B, C, D and E areseparated by pronounced seismic reflectors, heretermed horizon bB, bC, bD and bE (bB = baseunit B, etc.). The seismic horizons (bA, bB, bC,bD, bE and the sea floor) are here interpreted torepresent unconformities formed during severalglaciations. The upper regional unconformity isa diachronous surface and is interpreted to be

represented by horizon bD in SG9803, bB inSG9804 and bA in SG9810. The seismic units (A,B, C, D & E) probably represent deposition ofglacimarine sediments or basal till (Solheim &Kristoffersen 1984). Units A–C in the 3-Dsurveys are probably of mid-Pleistocene age,whereas units D and E are of upper Pleistoceneage (Sættem et al. 1992b).

The glacigenic sequence of the southwesternBarents Sea, called the Barents Sea Synthem(Vorren et al. 1989), has been mapped anddivided into several seismostratigraphic units(Solheim & Kristoffersen 1984; Vorren et al.1990; Sættem et al. 1992b). A shallow borehole(7222/09-U-01), located close to the 3-D surveysof our study area (Fig. 1), has been correlated to2-D seismic data by Hald et al. (1990) andSættem et al. (1992a). Their work is the basis forthe correlation of our 3-D seismic data to thisshallow borehole and to 2-D seismic data fromthe same area (Fig. 5; Table 1).

Unit A corresponds with unit b1 in Sættem etal. (1992b) and is younger than the upperregional unconformity but older than unit B(Table 1). Units B and C (Fig. 5) probably corre-late with the lower and upper part of unit b2 ofSættem et al. (1992b), respectively. Unit b2 isdescribed by Sættem et al. (1992b) as a soft, darkclaystone, where the lower part contains struc-tures that may have been formed by icebergturbation or glaciotectonics. Unit b2 wasprobably deposited 200–130 ka BP, and thelower part of unit b2 has been interpreted to bedeposited in an ice-proximal glacimarineenvironment whereas the upper part has beeninterpreted to be a till (Sættem et al. 1992a).


Fig. 5. Seismo-stratigraphic units and horizons from the 3-D seismic surveys correlated with the units ofSættem et al. (1992b). Unit b2? is inferred to be present only in SG9804 while unit b1 is inferred to be presentonly in SG9810. Units b4 and b5 (Sættem et al. 1992b) correlate to units 4E and 5E (Vorren et al. 1989, 1990),respectively (Table 1). Broken lines indicate uncertain correlations. Not to scale. Bedrock is mainly of Triassic,Cretaceous and Palaeocene age.



UnitBUnit CUnit D

Unit E

5 km

ASea floor

100 m

III

500 -

400 -

TWT

(ms)

I

II

SG9804

Youngest

Oldest

C

Horizon bD

486.31 637.77Depth i ms (TWT) below sea level

5 km

N

Depth i ms (TWT) below sea level

5 km

Youngest

Oldest

Horizon bE

466.07 543.22

F

OldestYoungest

5 km

Horizon bC

496.08 633.62


Horizon bB (URU)

5 km

OldestYoungest

E

504.69 689.30


DN

NN

B

Fig. 6. (a) Stratigraphy from the 3-D survey SG9804. (b) Horizon bC from SG9804 shown with time–structurecontours, indicating the location of the seismic section in (a). (c–f) Sections of illuminated time–structure mapsfrom the interpreted horizons in 3-D survey SG9804, showing subglacial depressions and different generationsof subglacial lineations. A summary of the dominant orientations is shown in Fig. 8 and Table 2. Depth in ms(two-way travel-time) below sea level is given on the colour bars. The light source is located to the east of thehorizon and vertical exaggeration (z-axis) is 8�.


According to Vorren & Laberg (1996), mostof the southern Barents Sea was deglaciatedduring the Arnøya Interstadial (29–24 ka BP;Andreassen et al. 1985), when our unit D wasprobably deposited. A glacier advance justbefore this interstadial probably formedhorizon bD (this paper). Sediments near thebase of unit b4 of Sættem et al. (1992b), whichprobably correspond to our unit D (Fig. 5; Table1), are radiocarbon dated to be younger than27 320 ± 735 years BP (unit b4 of Hald et al.1990). This corresponds well with the timeframe of the Arnøya Interstadial. Unit b4 ofSættem et al. (1992b) is over-consolidated

(Sættem et al. 1992b) and corresponds to unit 4Efrom Vorren et al. (1989, 1990), which is inter-preted to be a glaciofluvial sandy sediment,probably deposited by meltwater (Vorren et al.1989). In addition, unit D (this paper) correlateswith sequence I in Solheim & Kristoffersen(1984).

Lineations on horizon bE (Fig. 6) indicatethat our study area was again ice-covered,probably in the beginning of LGM I (LateGlacial Maximum I, 24–22 ka BP; Vorren &Laberg 1996). After LGM I, an ice-free periodcalled the Andøya Interstadial (22–19 ka BP)took place, and unit E was probably depositedin this period. Our unit E corresponds with unit5E from Vorren et al. (1989, 1990; Table 1),which is interpreted to be deposited in a distalglacimarine environment. The sediments maybe derived from meltwater rivers in the southand deposited from suspension together withsome ice-rafted debris during the ice-free periodbetween LGM I and LGM II (Vorren & Laberg1996).

Our 3-D seismic data are acquired froman area where other authors (Sættem et al.1992b; Vorren & Laberg 1996; Lebesbye 2000)have had difficulty in correlating glacigenicseismic units between the eastern and westernBarents Sea. The correlation problem is duemainly to the fact that most of the unitsbelong either to a western or an easternprovince and wedge out in and around our studyarea (Figs 5 & 6a). Additional 2-D and 3-Dseismic data are required to resolve thisproblem.


A

B

Fig. 7. Sketches illustrating the difficulty indetermining whether a feature is positive or negative.(a) Closely spaced features prevent the user in givinga positive or negative determination. (b) Negativefeatures may be identified when they occur asisolated lineations.

Table 2. Size and predominant orientation of subglacial lineations from the four subsurfaces in 3-D surveysSG9804 and SG9810 (youngest generation first, oldest generation last). The dominant orientations are349º–50º/169º–230º. No lineations were found on the sea floor in the 3-D surveys. Depth calculated using1750 m s–1 as the velocity of sound in sediments (Sættem et al. 1992b)

3-D survey Horizon Length Width Relief Dominant orientation N↑

(Fig. 5)Y O Y O

SG 9804 bE 1.2–20 km 100–170 m 2.5–8 m 22º / 202º & 30º / 210º

bD 2.5–12 km 60–150 m 2.5–8 m 26º / 206º & 176º / 356º

bC 0.8–7 km 60–180 m 3.5–6 m 10º / 190º & 28º / 208º

bB 2–7 km 60–150 m 2.5–7 m 50º / 230º & 169º / 349º

SG 9810 bE 7–8 km 50–60 m 2.5–3.5 m 48º / 228º

bD 0.5–20 km 50–120 m 2.5–6 m 45º / 225º & 5º / 185º

bA 0.5–20 km 50–120 m 3.5–6 m 45º / 225º & 5º / 185º ↔↔

↔↔↔

↔↔

↔↔

↔↔

↔↔


Glacial erosion surfaces

The units defined here are separated by fourhorizons interpreted to have been formed bysubglacial erosion. Features observed on thehorizons in the 3-D seismic surveys are mainlynegative features (although it is not clearwhether the lineations are positive or negative)and have been divided into three main classes:cross-cutting furrows, straight lineations anddepressions. The depressions are furthersubdivided into type I (smooth edges), type II(indented edges) and type III (formed inbedrock) depressions.

Cross-cutting furrows are mainly observed onthe sea floor, but a few occur on one of theburied horizons (bE in SG9810). The cross-cutting furrows observed in the study area areinterpreted to be iceberg plough marks, andindicate a glacimarine environment.

Straight lineations and depressions are inter-preted as indications of palaeo-ice flow direc-tions which, in turn, may be related to palaeo-icesheet form and extent. The straight lineationsare observed on mapped horizons bB to bE andare interpreted to be subglacial bedforms. Thisimplies that warm-based, or at least periodicallywarm-based, grounded ice sheets have advancedover the study area at least four times. Oblongdepressions with similar orientation to thelineations occur on horizon bD in 3-D surveySG9804. The depressions are interpreted ashaving been formed by subglacial erosion and torecord palaeo-ice flow direction, thus supporting

the indications from the lineations of dominantpalaeo-ice flows towards the north or south.

Cross-cutting furrows (iceberg ploughmarks)

Description. Curved furrows of negative reliefthat cross-cut each other in a chaotic patternare the dominant morphological features on thesea floor (Fig. 3). Similar curved furrows alsooccur on buried horizon bE in SG9810, but noton any other buried horizon in the study area.The sea floor furrows are 30–500 m wide,1–20 km long and 2.5–25 m (3–28 ms TWT)deep. In profile, their cross-section is either U-or V-shaped (and sometimes asymmetrical),and some have a small rise on each side (Fig.3a, b). The relief of the furrows is measuredfrom the bottom of the feature to the top of theflanking rise, or to the surrounding sea floorlevel if there is no rise on either side of thefeature. The furrows usually have a constantdepth along the entire feature, and on the seafloor they show a slightly east–west dominantorientation. In some cases, the cross-cuttingfurrows occur in pairs spaced apart up to 700 m(e.g. Fig. 3d).

Interpretation. The cross-cutting curved furrowsare interpreted as representing ploughmarksformed in a submarine environment by driftingicebergs, a phenomenon well known in the litera-ture (Lien 1983; Vorren et al. 1983; Barnes et al.1984; Stoker & Long 1984; Solheim et al. 1988;Longva & Bakkejord 1990; Dowdeswell et al.1992; Vogt et al. 1994; Crane et al. 1997; Long &Praeg 1997; Polyak 1997; Solheim 1997). Theplough-marks in our study area have a maximumrelief of 25 m and are up to 500 m wide and 20 kmlong. Lien (1983) described plough-marks fromthe Norwegian continental shelf that are up to27 m deep, and Crane et al. (1997) mentionplough-marks 450–850 m bsl from the southernYermak Plateau that are up to 10 m deep, 1000 mwide and, according to Polyak (1997), 15 kmlong. The largest icebergs may extend deepunderwater and have sufficient mass andmomentum to force the keel down into the seafloor (Lien 1983). Thus, provided other factorsare equal, icebergs reaching the sea floor in deepwaters may plough to much greater depths intothe underlying sediment than those that runaground in shallower waters. The plough-markstypically have a U- or V-shaped cross-sectionwith a ridge of ploughed sediments on each side(Lien 1983).

The size and shape of the plough-marks


W E

32

1

N

S

Fig. 8. Rose diagram showing the 13 dominantorientations of subglacial lineations on the horizonsin the 3-D surveys. The main trends are349º–50º/169º–230º.


varies greatly. Buried plough-marks 246 msbelow the sea floor have been described fromthe central North Sea and the Norwegian conti-nental shelf (Long & Praeg 1997). Plough-marks have been observed on the YermakPlateau NW of Svalbard at depths of up to 850m below the present sea level (Crane et al.1997). In the Barents Sea, plough-marks havebeen found at all depths down to 450 m,although the keels of present-day icebergsrarely exceed 100 m (Solheim 1997).

Paired furrows observed in our study area(Fig. 3d) are interpreted as having been formedby icebergs with two keels, each formingseparate plough-marks, a phenomenon des-cribed previously by Barnes et al. (1984) andLongva & Bakkejord (1990). Our observationsimply that the iceberg forming the largest pairedplough marks in the 3-D surveys was at least700 m wide at the sea floor.

Plough-marks commonly occur on the seafloor in the study area, but only a few werefound on buried horizon bE in SG9810. Thereason why plough-marks are not found onother buried horizons is probably that subse-quent glaciations have removed previouslyformed plough-marks.

Straight lineations (subglacial bedforms)

Description. Sub-parallel lineations are ob-served on most of the buried horizons of theglacigenic sequence (Figs 4 & 6) and can betraced over long distances (up to 20 km). Mosthave a straight linear outline, although a few arecurvilinear. When aligned parallel to each otherwith close spacing (130–250 m) between eachlineation, it cannot be determined with certaintywhether they are positive or negative features(Fig. 7a). However, detailed studies of isolatedlineations and of lineations lying relatively farapart indicate that some of them may benegative rather than positive features (Fig. 7b).

The lineations on buried surfaces in the studyarea are mostly U-shaped in cross-section and ofconstant depth along most of the lineation,becoming shallower at both ends before theygradually disappear. These lineations may occuron all parts of some buried horizons, or only inparticular areas. They appear on both bedrockhorizons (URU) of early Cretaceous age(Sigmond 1992) and horizons within theglacigenic sequence. Usually a surface is domi-nated by lineations with one or two main orien-tations, although a few isolated lineations withother orientations may also be present (Fig. 6a).Most lineations range in size from 0.5 to 20 kmlong, 50–180 m wide and have a relief of 2.5–8 m

(Table 2). The orientations of the lineations arein the range 349º–50º and 169º–230º (Table 2 &Fig. 8). When determining the age relationshipbetween lineations of different orientations on asingle horizon in a 3-D survey, the set that clearlycross-cuts another is interpreted to be theyounger. When lineations cross-cut one another,the older must be removed or replaced by theyounger to make the age relation unequivocal.

Interpretation. The long, straight and relativelynarrow features on the buried horizons areinterpreted as subglacial bedforms indicatingthe direction of palaeo-ice flow. They closelyresemble large-scale lineations and megaflutesdescribed in the literature (Boulton & Clark1990; Clark 1993). Glacial flutes are defined asnarrow ridges of sediment aligned parallel toglacier flow (Benn 1994) and flutes on thesea floor have been described from thenorth–central Barents Sea at water depths of150–340 m (Solheim et al. 1990). They have apositive relief of less than 1 m, widths of 4–8 m,lengths typically 100–500 m, and are thus muchsmaller lineations than those observed on ourburied horizons. From the Ross Sea continentalshelf, Antarctica, lineations described to be100–200 m wide, up to several tens of kilometreslong and with a relief of the order of severalmetres are interpreted as mega-flutes (Shipp &Anderson 1997).

Some of the straight lineations in the 3-Dseismic areas occur in bedrock and mayresemble grooves. Pudsey et al. (1997) describegrooves from the Antarctic Peninsula shelf withrelief of 2–3 m offshore and 10–20 m inshore at200–700 m water depth. Bennett & Glasser(1996) interpret glacial grooves, ranging from afew metres to several kilometres long, to bestreamlined depressions formed by areal iceflow. Other authors refer to grooves that wereprobably formed by subglacial erosion withmeltwater as the active erosive agent (Kor et al.1991; Rains et al. 1993), but the features theydescribe are not as straight and persistent as thelineations in our 3-D datasets. Morphologically,glacial grooves are comparable to glacial stria-tions, except that grooves are larger and deeper(Bennett & Glasser 1996). Both glacial stria-tions and grooves indicate the direction of localglacier movement and indicate a warm-basedglacier. Based on the above discussion, wecannot exclude the possibility that some of thelineations on the upper regional unconformityare grooves.

In recent years, continental lineations com-posed of glacial sediment have been recognizedon satellite images (Punkari 1993, 1997; Boulton




5km

Dep

thin

ms

(TW

T)

bel

ow

sea

level

Q`

Q

P`

P

You

nges

t

Old

estTru

nca

ted

typ

eII

dep

ress

ion

Dat

aav

aila

ble

but

not

inte

rpre

table

due

toch

aoti

c

refl

ecti

on

pat

tern

No

dat

aT

yp

eII

dep

ress

ion

s

inan

enec

hel

on

-

patt

ern

486.3

1637.7

7

N

N

P

100

m

1km

P`

TW

T

700

-

600

-

(ms)

A

TW

T

Q1

km

Q`

600

-

500

-

100

m

(ms)

C

DB

bD

UR

U

Typ

eI

dep

ress

ion

s

Sea

floor

Fig

. 9.

(a)

Seis

mic

pro

file

show

ing

buri

ed ty

pe I

dep

ress

ions

in th

e no

rthw

este

rn p

art o

f 3-D

sur

vey

SG98

04. (

b) A

n en

larg

ed a

rea

of (

d), s

how

ing

type

I d

epre

ssio

ns.

(c)

Seis

mic

pro

file

show

ing

buri

ed ty

pe I

I de

pres

sion

s in

the

nort

heas

tern

par

t of 3

-D s

urve

y SG

9804

. (d)

Illu

min

ated

tim

e–st

ruct

ure

map

of h

oriz

on b

D. W

hite

arro

ws

indi

cate

the

mai

n or

ient

atio

ns o

f the

line

atio

ns a

nd th

eir

rela

tive

age

(26

º/20

6º =

you

nges

t, 17

6º/3

56º =

old

est)

. Dep

th in

ms

(tw

o-w

ay tr

avel

-tim

e) b

elow

sea

leve

l is

show

n on

the

colo

ur b

ar. T

he li

ght s

ourc

e is

loca

ted

east

of t

he h

oriz

on a

nd v

erti

cal e

xagg

erat

ion

(z-a

xis)

is 8

�.


et al. 2001) and classified as mega-scale glaciallineations (Bennett & Glasser 1996). Typicallengths range between 8 and 70 km, widthsbetween 200 and 1300 m and the spacingbetween lineations may vary between 300 and5000 m. On the ground, their morphology isoften difficult to detect. According to Clark(1993), mega-scale glacial lineations are formedby differential subglacial sediment deformationcaused by variations in bed characteristics, andtheir long length reflects rapid ice flow and/orlong periods of time for development. Usinglate Quaternary chronologies for the James Baylowlands, where mega-scale glacial lineationswere generated by the last Laurentide Ice Sheet,Clark (1993) concluded that these large stream-lined landforms were produced by a fast-flowingglacier at velocities typical of modern icestreams (400–1600 m a–1). On land in northernEurope and the northwestern part of Russia,0.1–10 km long erosional forms have beenobserved on satellite imagery and aerial photo-graphs (Punkari 1993). In the same region, flow-parallel sediment ridges, 10s–1000s of metres inlength, metres to 10s of metres in height and 10sto 100s of metres in width have been observed(Boulton et al. 2001).

Three-dimensional marine reflection seismictime–structure maps (Figs 4 & 6) providesimilar information about the sea floor andburied surfaces that satellite images do forcontinental surfaces. The lineations observed on3-D seismic images of buried surfaces in ourstudy area bear many similarities to subglaciallineations described from satellite images(Boulton & Clark 1990; Bennett & Glasser1996). Differential compaction at burial or thenature of the sediments in and between thelineations might have affected the observedlineations. For the interpretation of their origin,it may be significant that the large-scalelineations are observed on buried horizonswhereas the flutes and mega-scale flutes aredescribed from sea floor and continentalsurfaces. Flutes have a lower preservationpotential than negative features because theyare readily degraded by wind and water and,according to Boulton & Dent (1974), they arefar more common on modern forelands than onolder terrain. We suggest that the lineationsobserved on buried horizons of the glacigenicsection are large-scale lineations formedbeneath a warm-based ice sheet, and thusindicate dominant ice-flow directions.

Deformation of subglacial bedforms isstrongest near the margin of warm-basedglaciers where old lineations are obliterated(Bennett & Glasser 1996). Ice streams can

effectively erode the orientated elements repre-senting earlier flow phases, but in areas of weakmarginal erosion indications of older flow direc-tions of the last ice sheet may remain (Punkari1993). Several generations of lineations fromthe last glaciation have been observed innorthern Europe and northwestern Russia,where the older ones are suggested to haveformed during melting-bed phases of ice-sheetgrowth and preserved by frozen bed conditionsduring the glacial maximum (Boulton et al.2001). A change in the ice margin may explainwhy there are two dominant directions of iceflow on several horizons in our study area. Asan ice sheet retreats, the shape of the ice marginand the nature of the calving front may change,causing the ice flow to change accordingly. Thelineations trending north–south and NNE–SSWin the study area may, therefore, have formedduring deglaciation phases. If this is so, possiblelineations from earlier phases may have beenremoved as the glacier retreated and subjectedthe study area to strong deformation. Anotherpossibility, although less likely, is that the oldestdirections are preserved features from an earlierphase, such as from the ice sheet advance or atits maximum position, and that only theyoungest were formed close to the ice marginduring ice sheet retreat.

Depressions (subglacial erosion)

Description. Two kinds of incised semi-circularto oblong depressions are observed on horizonbD in 3-D survey SG9804, and a third kind isobserved on the horizon bB (URU) in thesame 3-D survey. Type I depressions (Fig. 9a& 9b) are 35–53 m (40–60 ms TWT) deep,1560–1875 m wide and 1875 to more than9000 m long. Their long axes are orientated26º/206º. These depressions have a concentricshape, are widest in the middle and becomenarrower as they shallow. Their deepest point isslightly NNE of their centre. A common featureof the depressions is that they have a smoothsurface and even edges, which are not markedby the straight lineations that are observed onthe surface surrounding the depressions.

Type II depressions (Fig. 9c & 9d) are rela-tively shallow, 3750–5625 m in length, 1250–1700 m in width and 9–32 m (10–37 ms TWT) indepth. Their longest axes are orientated 8º/188ºand 169º/349º. They are elliptical in shape,deepest in the middle and become shallowertowards the edges. The edges are rough andindented and some of the depressions seem tooccur in series, where the depressions are mostlyaligned parallel to each other, resembling an



en echelon pattern (Fig. 9d). All of the depres-sions in such a series are about the same size.

Type III depressions are observed on thehorizon bB (URU) in SG9804, in earlyCretaceous bedrock (Fig. 6e). They are semi-circular with a slight north–south trendingelongation, ranging from 3125–3200 m inwidth, 3500–3850 m in length and 9–13 m(10–15 ms TWT) deep. Their edges seem to bea combination of type I and II depressions, inthat they have relatively smooth edges eventhough the lineations on the horizon reach theiredge (Fig. 6e). This is caused by the fact that thedepth of the lineations gradually decreases asthey reach the edge of the depressions.

Interpretation. The depressions (Figs 6e & 9)are thought to have been formed by subglacialerosion. After the depressions were formed,they may have been modified by meltwater.Type I depressions (Fig. 9a & 9b) may haveformed either coeval with or after the formationof the straight lineations, since the depressionshave smooth edges. We believe that it is mostlikely that the type I depressions formed afterthe nearby lineations, mainly because theirorientations differ significantly. Depressionswith their deepest point NNE of their centremay indicate that the glacier moved towards theNNE, but ice flow in the opposite directioncannot be ruled out from the morphologicalobservations alone.

Type II depressions (Fig. 9c & 9d) may havebeen formed in both directions along theirlongest axis as their deepest point is in theircentre. The depressions were probably formedbefore the straight lineations because thestraight lineations extend all the way to the edgeof the depressions, making the edges rough andindented. If these interpretations are correct,type II depressions are probably older than thelineations and type I depressions are youngerthan both lineations and type II depressions.Another possibility is that both type I and typeII depressions formed before the lineations, butlater only the eastern part of the 3-D survey(type I depressions) was modified by meltwater.

Type III depressions (Fig. 6e) may indicate anice flow towards the north or south, but becausethey are so rounded this is speculative. Theymay have been formed in areas where the bedwas easier to erode.

Depressions formed in bedrock by subglacialerosion vary in size from a few metres to severalhundred metres in diameter (Bennett &Glasser 1996). Subglacial meltwater erosion inunconsolidated sediments is also thought tohave formed depressions of about 200 m in

diameter (Rampton 2000); such depressionsformed in bedrock are called S-forms and areabout 0.5 by 1 km (Beaney & Shaw 2000).Negative, longitudinal forms formed by melt-water erosion are called spindle flutes in theclassification of Kor et al. (1991), regardless ofsize. They are narrow, shallow negative formsthat are longer than they are wide. Theybecome slightly wider downstream and may beasymmetrical. The shape of these spindle flutesis similar to type II depressions on horizon bDin SG9804 (Fig. 9c & 9d). Examples of erodeddepressions (S-forms) formed by meltwater aredescribed from Alberta (Beaney & Shaw 2000)and from the northeastern coast of GeorgianBay, Canada (Kor et al. 1991). The depressionsobserved in this study are interpreted as havingbeen formed by erosion beneath a warm-basedice sheet, and some (especially type I depres-sions) are suspected to have been modified bymeltwater.

Discussion

The morphology of the buried surfaces of lateCenozoic age, including the upper regionalunconformity (URU) that separates the upperunconsolidated sediments from underlyingconsolidated rocks (bedrock), suggests thatextensive glacial erosion and deposition hastaken place in the study area.

Palaeo-ice flow and extent

Interpretation of 3-D seismic morphologicalfeatures of the four buried horizons of the studyarea provides important information aboutglaciations and deglaciations in the southwest-ern Barents Sea.

Glacial advances. The existence of subglaciallineations on the four main buried horizons ofour study area (horizon bB, bC, bD & bE; Fig.6; Table 2) indicates at least four glacialadvances in the southern Barents Sea. Inaddition, there may have been a glacial advanceprior to these four, as horizon bA is interpretedto represent the URU. Unfortunately, themorphological features on horizon bA are ques-tionable. On the horizons bB, bC, bD and bEthe large size and persistence over several tensof kilometres of the lineations suggests that theywere formed subglacially. Their orientationindicates that ice flow directions mainly towardsnorth and NNE (or south and SSW) werepresent in all four glacial advances. The slightlydifferent orientations of the two sets oflineations on all four buried horizons of 3-D




Fig

. 10.

(a)

Ext

ent o

f ice

she

et d

urin

g ad

vanc

e 5

(LG

M I

I; V

orre

n &

Lab

erg

1996

). (

b) H

ypot

heti

cal e

xten

t of i

ce s

heet

s du

ring

ret

reat

s 1

to 4

pro

pose

d in

this

stu

dy.


areas SG9804 and SG9810 (Table 2) may reflectdifferent ice-flow patterns.

An ice-flow direction towards the west haspreviously been suggested for the study areaduring glacial maxima when the ice sheetreached the shelf edge (Fig. 10a; Landvik et al.1998; Vorren et al. 1990; Vorren & Laberg 1996).Most of the features in our study area are orien-tated north–south, and both bathymetry, sub-aerial topography and ice-sheet modelling(Dowdeswell & Siegert 1999) suggests an ice-flow direction towards north. The Bjørnøyrennais relatively deep (up to 500 m below present sealevel) and may have acted as a calving bayduring a glacial retreat. The observed lineationsand depressions may have formed during an ice-sheet retreat (Fig. 10b), when Bjørnøyrenna wasice free, obliterating features formed during theearlier stages of glaciation. Though it is lesslikely, the possibility that the oldest generationof lineations on a horizon have formed during aglacial advance and been preserved beneath theice sheet, while the youngest generation oflineations were formed during deglaciation,cannot be excluded. A third possibility is thatthese features were formed while the ice sheetreached the shelf edge, and that ice flow turnedfrom trending north–south in our study are to amore east–west orientation when entering theBjørnøyrenna. Lineations in unlithified glaci-genic sediments beneath a warm-based ice sheetare probably degraded relatively rapidly, andwe believe that it is more likely that older gener-ations of lineations and depressions have beenremoved and that just the last couple of gener-ations descending from deglaciation phases(Fig. 10b) have been preserved.

Plough-marks

The plough marks on the sea floor indicate aglacimarine environment. Their chaotic cross-cutting pattern shows a slightly east–westdominant orientation, suggesting that theprevailing wind and ocean-current direction inthe study area were towards the east or west.When the study area was ice free, but stillproximal to the ice, katabatic winds and melt-water discharge may have affected the directionof iceberg drift. Vorren et al. (1990) indicate thatBjørnøyrenna, and part of the area just south ofBjørnøyrenna, may be among the first parts ofthe southwestern Barents Sea to become icefree during ice retreats, thus forming a large ice-free bay. In this bay, the current direction islikely to have been towards the east in thesouthern part and towards west in the northernpart of the bay (Vorren et al. 1990). In the early

stages of the deglaciation, the dominant direc-tion of iceberg drift may have been east–westtrending, but as larger parts of the southwesternBarents Sea became ice-free the direction oficeberg drift became more chaotic.

Conclusions

1. The present study illustrates that con-ventional industry 3-D seismic data, acquiredfor the purpose of studying deeper targets, canprovide detailed morphological information onshallow buried glacigenic sequences.

2. The morphology of the buried horizons,together with the repeated glacigenic sequence,suggests that extensive erosion and depositionhas taken place.

3. Long straight, negative lineations ob-served on four buried horizons, are interpretedas having formed beneath a warm-based icesheet, reflecting the dominant ice-flow direc-tions. The lineations provide conclusiveevidence that grounded ice has reached thesouthern part of Bjørnøyrenna at least fourtimes since the formation of the upper regionalunconformity.

4. A dominantly north–south orientation ofglacial lineations observed on all buriedhorizons suggests that they were formed duringice sheet retreats.

5. Plough-marks occur mainly on the seafloor and indicate that icebergs have reacheddepths of 456 m below present sea level, andthat some icebergs were up to 700 m wide at thesea floor.

Norsk Hydro ASA and the European Communitiesproject TriTex (IST–1999–20500) are acknowledgedfor funding the research project. Norsk Hydro ASA,Statoil ASA, Norsk Agip A/S, Fortum Petroleum A/Sand former Saga Petroleum ASA are acknowledgedfor providing the seismic and well data. TheUniversity of Tromsø acknowledges support fromGeoQuest via computer software and help on techni-cal issues. We offer our sincere thanks to D. Praeg andA. Solheim who critically reviewed the manuscriptand to P. E. B. Armitage who corrected the Englishlanguage.

ReferencesANDREASSEN, K., VORREN, T. O. & JOHANSEN, K. B.

1985. Pre-Late Weichselian glacimarine sedi-ments at Arnøy, North Norway. GeologiskaFøreningens i Stockholm Førhandlingar, 107,63–70.

BARNES, P. W., REARIC, D. M. & REIMNITZ, E. 1984. Icegouging characteristics and processes. In:BARNES, W., SCHELL, D. M. & REIMNITZ, E. (eds).The Alaskan Beaufort Sea: Ecosystems and



Environments. Academic Press, Orlando,184–212.

BEANEY, C. L. & SHAW, J. 2000. The subglacialgeomorphology of southeast Alberta: evidencefor subglacial meltwater erosion. CanadianJournal of Earth Sciences, 37, 51–61.

BENN, D. I 1994. Fluted moraine formation and tillgenesis below a temperate valley glacier:Slettmarkbreen, Jotunheimen, southern Norway.Sedimentology, 41, 279–292.

BENNET, M. R. & GLASSER, N. F. 1996. GlacialGeology – Ice Sheets and Landforms. John Wiley& Sons Ltd., West Sussex.

BOULTON, G. S. & CLARK, C. D. 1990. A highly mobileLaurentide ice sheet revealed by satellite imagesof glacial lineations. Nature, 346, 813–817.

BOULTON, G. S. & DENT, D. L. 1974. The nature andrates of post-depositional changes in recentlydeposited till from south-east Iceland.Geografiska Annaler, 56A, 121–134.

BOULTON, G. S., DONGELMANS, P., PUNKARI, M. &BROADGATE, M. 2001. Palaeoglaciology of an icesheet through a glacial cycle: the European icesheet through the Weichselian. QuaternaryScience Reviews, 20, 591–625.

BROWN, A. R. 1999. Interpretation of Three-Dimensional Seismic Data, 5th edition. AmericanAssociation of Petroleum Geologists, Tulsa,Oklahoma, Memoir, 42.

CLARK, C. D. 1993. Mega-scale lineations and cross-cutting ice-flow landforms. Earth SurfaceProcesses and Landforms, 18, 1–29.

CRANE, K., VOGT, P. R. & SUNDVOR, E. 1997. DeepPleistocene iceberg plowmarks on the YermakPlateu. In: DAVIES, T. A., BELL, T., COOPER, A. K.,JOSENHANS, H., POLYAK, L., SOLHEIM,A., STOKER,M. S. & STRAVERS, J. A. (eds) Glaciated Conti-nental Margins: An Atlas of Acoustic Images.Chapman & Hall, London, 140–141.

DOWDESWELL, J. A. & SIEGERT, M. J. 1999. Ice-sheetnumerical modeling and marine geophysicalmeasurements of glacier-derived sedimentationon the Eurasian Arctic continental margins.Geological Society of America Bulletin, 111,1080–1097.

DOWDESWELL, J. A., WHITTINGTON, R. J. & HODGKINS,R. 1992. The sizes, frequencies and freeboards ofEast Greenland icebergs observed using shipradar and sextant. Journal of GeophysicalResearch, 97, 3515–3528.

EIDVIN, T., JANSEN, E. & RIIS, F. 1993. Chronology ofTertiary fan deposits off western Barents Sea:implications for the uplift and erosion history ofthe Barents Sea shelf. Marine Geology, 112,109–131.

EIDVIN, T., GOLL, R. M., GROGAN, P., SMELROR, M. &ULLEBERG, K. 1998. The Pleistocene to MiddleEocene stratigraphy and geological evolution ofthe western Barents Sea continental margin atwell site 7316/5-1 (Bjørnøya West area). NorskGeologisk Tiddskrift, 78, 99–123.

HALD, M., SÆTTEM, J. & NESSE, E. 1990. Middle andlate Weichselian stratigraphy in shallow drillingsfrom the southwestern Barents Sea: foraminferal,

amino acid and radiocarbon evidence. NorskGeologisk Tidsskrift, 70, 241–257.

KNUTSEN, S.-M., RICHARDSEN, G. & VORREN, T. O.1992. Late Miocene-Pleistocene sequence stratig-raphy and mass-movements on the westernBarents Sea margin. In: VORREN, T. O.,BERGSAKER, E., DAHL-STAMNES Ø. A., HOLTER,E., JOHANSEN, B., LIE, E. & LUND, T. B. (eds)Arctic Geology and Petroleum Potential. NPFSpecial Publication, Elsevier, Amsterdam, 2,573–606.

KOR, P. S. G., SHAW, J. & SHARPE, D. R. 1991. Erosionof bedrock by subglacial meltwater, GeorgianBay, Ontario: a regional view. Canadian Journalof Earth Sciences, 28, 623–642.

LANDVIK, J. Y., BONDEVIK, S., ELVERHØI, A., FJELD-SKAAR,W., MANGERUD, J., SALVIGSEN, O., SIEGERT,M. J., SVENDSEN, J. I. & VORREN, T. O. 1998. Thelast glacial maximum of Svalbard and the BarentsSea area: Ice sheet extent and configuration.Quaternary Science Reviews, 17, 43–75.

LEBESBYE, E. 2000. Late Cenozoic Glacial History ofthe Southwestern Barents Sea. PhD thesis,University of Tromsø.

LIEN, R. 1983. Pløyemerker etter isfjell på norskkontinentalsokkel. PhD thesis, Institutt forkontinentalsokkelundersøkelser, 109, 147 pp.

LILLEVIK, A. & LYNGNES, B. 1998. Processing report,Saga Petroleum ASA, Loppa East/South, Seismic3D processing on SG9803, Geco-Prakla Norway.

LONG, D. & PRAEG, D. 1997. Buried ice-scours: 2D vs3D-seismic geomorphology. In: DAVIES, T. A.,BELL, T., COOPER, A. K., JOSENHANS, H., POLYAK,L., SOLHEIM, A., STOKER, M. S. & STRAVERS, J. A.(eds) Glaciated Continental Margins: An Atlas ofAcoustic Images. Chapman & Hall, London,142–143.

LONGVA, O. & BAKKEJORD, K. J. 1990. Iceberg defor-mation and erosion in soft sediments, SoutheastNorway. Marine Geology, 92, 87–104.

POLYAK, L. 1997. Overview. In: DAVIES,T. A., BELL,T.,COOPER, A. K., JOSENHANS, H., POLYAK, L.,SOLHEIM, A., STOKER, M. S. & STRAVERS, J. A.(eds) Glaciated Continental Margins: An Atlas ofAcoustic Images. Chapman & Hall, London,136–137.

PUDSEY, C. J., BARKER, P. F. & LARTER, R. D. 1997.Glacial flutes and iceberg furrows, AntarcticPeninsula. In: DAVIES, T. A., BELL, T., COOPER, A.K., JOSENHANS, H., POLYAK, L., SOLHEIM, A.,STOKER, M. S. & STRAVERS, J. A. (eds) GlaciatedContinental Margins: An Atlas of AcousticImages. Chapman & Hall, London, 58–59.

PUNKARI, M. 1993. Modelling of the dynamics of theScandinavian ice sheet using remote sensing andGIS methods. In: ABER, J. (ed.). Glaciotectonicsand Mapping Glacial Deposits, Proceedings of theINQUA Comission on Formation and Propertiesof Glacial Deposits. Canadian Plains ResearchCenter, University of Regina, 232–250.

PUNKARI, M. 1997. Subglacial processes of theScandinavian Ice Sheet in Fennoscandia inferredfrom flow-parallel features and lithostratigraphy.Sedimentary Geology, 111, 263–283.



RAFAELSEN, B., MIDTBØ, M., KULIMAN, L. W.,LEBESBYE, E., HOGSTAD, K. & ANDREASSEN, K.2000. 3-D seismic data used to investigate theLate Cenozoic sediments of the southwesternBarents Sea. Geonytt, 1, 139.

RAINS, B., SHAW, J., SKOYE, R., SJOGREN, D. & KVILL,D. 1993. Late Wisconsin subglacial megafloodpaths in Alberta. Geology, 21, 323–326.

RAMPTON, V. N. 2000. Large-scale effects of subglacialmeltwater flow in the southern Slave Province,Northwest Territories, Canada. Canadian Journalof Earth Sciences, 37, 81–93.

RICHARDSEN, G., HENRIKSEN, E. & VORREN, T. O.1991. Evolution of the Cenozoic sedimentarywedge during rifting and seafloor spreading westof the Stappen High, western Barents Sea.Marine Geology, 101, 11–30.

SHERIFF, R. E. 1999. Encyclopedic Dictionary ofExploration Geophysics, 3rd edition. Society ofExploration Geophysicists, Tulsa.

SHIPP, S. & ANDERSON, J. B. 1997. Lineations on theRoss Sea Continental Shelf, Antarctica. In:DAVIES, T. A., BELL, T., COOPER, A. K., JOSEN-HANS, H., POLYAK, L., SOLHEIM, A., STOKER, M. S.& STRAVERS, J. A. (eds) Glaciated ContinentalMargins: An Atlas of Acoustic Images. Chapman& Hall, London, 54–55.

SIGMOND, E. M. O. 1992, Berggrunnskart, Norge medhavområder. Målestokk 1: 3 millioner. NorgesGeologiske Undersøkelse. Emil Moestue A/S,Oslo.

SOLHEIM, A. 1997. Depth-dependent iceberg ploughmarks in the Barents Sea. In: DAVIES,T. A., BELL,T., COOPER, A. K., JOSENHANS, H., POLYAK, L.,SOLHEIM, A., STOKER, M. S. & STRAVERS, J. A.(eds). Glaciated Continental Margins: An Atlas ofAcoustic Images. Chapman & Hall, London,138–139.

SOLHEIM, A. & KRISTOFFERSEN, Y. 1984. Sedimentsabove the upper regional unconformity: Thick-ness, seismic stratigraphy and outline of theglacial history. Norsk Polarinstitutt Skrifter, 179B,26 pp.

SOLHEIM,A., RUSSWURM, L., ELVERHØI,A. & BERG, M.N. 1990. Glacial geomorphic features in thenorthern Barents Sea: direct evidence forgrounded ice and implications for the pattern ofdeglaciation and late glacial sedimentation. In:DOWDESWELL, J. A. & SCOURSE, J. D. (eds).Glacimarine Environments: Processes and Sedi-ments. Geological Society, London, SpecialPublication, 53, 253–268.

SOLHEIM, A., MILLIMAN, J. D. & ELVERHØI, A. 1988.Sediment distribution and sea-floor morphologyof Storbanken: implications for the glacial historyof the northern Barents Sea. Canadian Journal ofEarth Sciences, 25, 547–556.

STOKER, M. S. & LONG, D. 1984. A relict ice-scourederosion surface in the central North Sea. MarineGeology, 61, 85–93.

SÆTTEM, J. 1994. Glaciotectonic structures along theBarents shelf margin. In: WARREN,W. P. & CROOT,

D. G. (eds) Formation and Deformation of GlacialDeposits. Balkerna, Rotterdam, 95–113.

SÆTTEM, J., POOLE, D. A. R., ELLINGSEN, L. & SEJRUP,H. P. 1992a. Glacial geology of outerBjørnøyrenna, southwestern Barents Sea. MarineGeology, 103, 15–51.

SÆTTEM, J., RISE, L. & WESTGAARD, D. A. 1992b.Composition and properties of glacigenic sedi-ments in the southwestern Barents Sea. MarineGeotechnology, 10, 229–255.

SÆTTEM, J., BUGGE, T., FANAVOLL, S., GOLL, R. M.,MØRK, A., MØRK, M. B. E., SMELROR, M. &VERDENIUS, J. G. 1994. Cenozoic margin develop-ment and erosion of the Barents Sea: Coreevidence from southwest of Bjørnøya. MarineGeology, 118, 257–281.

VOGT, R. P., CRANE, K. & SUNDVOR, E. 1994.Deep Pleistocene iceberg plowmarks on theYermak Plateu: Sidescan and 3.5 kHz evidencefor thick calving ice fronts and a possible marineice sheet in the Arctic Ocean. Geology, 22,403–406.

VORREN, T. O. & KRISTOFFERSEN, Y. 1986. LateQuaternary glaciation in the south-westernBarents Sea. Boreas, 15, 51–59.

VORREN, T. O. & LABERG, J. S. 1996. Late glacial airtemperature, oceanographic and ice sheet inter-actions in the southern Barents Sea region. In:ANDREWS, J. T.,AUSTIN,W. E. N., BERGSTEN, H. &JENNINGS, A. E. (eds) Late Quaternary Palaeo-ceanography of the North Atlantic Margins.Geological Society, London, Special Publi-cations, 111, 303–321.

VORREN, T. O., HALD, M. EDVARDSEN, M. & LIND-HANSEN, O-W. 1983. Glaciogenic sediments andsedimentary environments on continentalshelves: General principles with a case studyfrom the Norwegian shelf. In: EHLERS, J. (ed.)Glacial Deposits in North-west Europe. Balkema,Rotterdam, 61–73.

VORREN, T. O., KRISTOFFERSEN, Y. & ANDREASSEN, K.1986. Geology of the inner shelf west of NorthCape, Norway. Norsk Geologisk Tidsskrift, 66,99–105.

VORREN, T. O., HALD, M. & LEBESBYE, E. 1988. LateCenozoic environments in the Barents sea. Pale-oceanography, 3, 601–612.

VORREN, T. O., LEBESBYE, E., ANDREASSEN, K. &LARSEN, K. B. 1989. Glacigenic sediments on apassive continental margin as exemplified by theBarents Sea. Marine Geology, 85, 251–272.

VORREN, T. O., LEBESBYE, E. & LARSEN, K. B. 1990.Geometry and genesis of the glacigenic sedi-ments in the southern Barents Sea. In:DOWDESWELL, J. A. & SCOURSE, J. D. (eds)Glacimarine Environments: Processes and Sedi-ments. Geological Society, London, SpecialPublications, 53, 269–288.

VORREN, T. O., RICHARDSEN, G., KNUTSEN, S.-M. &HENRIKSEN, E. 1991. Cenozoic erosion and sedi-mentation in the western Barents Sea. Marineand Petroleum Geology, 8, 317–340.



Geomorphology of buried glacigenic horizons in the Barents ... et al 2002.pdf · Geomorphology of...

Documents

Transcript of Geomorphology of buried glacigenic horizons in the Barents ... et al 2002.pdf · Geomorphology of...