Earth Surface Processes and Landforms 526 M. Jakobsson Earth …research.bpcrc.osu.edu/geo/publications/jakobsson_etal_ESPL08.pdf · Glacial geomorphology of the Central Arctic Ocean:

526 M. Jakobsson et al.

Copyright © 2008 John Wiley & Sons, Ltd. Earth Surf. Process. Landforms 33, 526–545 (2008)DOI: 10.1002/esp

Earth Surface Processes and LandformsEarth Surf. Process. Landforms 33, 526–545 (2008)Published online in Wiley InterScience(www.interscience.wiley.com) DOI: 10.1002/esp.1667

Glacial geomorphology of the Central Arctic Ocean:the Chukchi Borderland and the Lomonosov RidgeMartin Jakobsson,1* Leonid Polyak,2 Margo Edwards,3 Johan Kleman4 and Bernard Coakley5

1 Department of Geology and Geochemistry, Stockholm University, Sweden2 Byrd Polar Research Center, Ohio State University, Ohio, USA3 The Department of Geology and Geophysics, University of Hawaii, Hawaii, USA4 Department of Physical Geography and Quaternary Geology, Stockholm University, Sweden5 Department of Geology and Geophysics, University of Alaska Fairbanks, Alaska, USA

AbstractThe last decade of geophysical seafloor mapping in the Arctic Ocean from nuclear sub-marines and icebreakers reveals a wide variety of glaciogenic geomorphic features at waterdepths reaching 1000 m. These findings provide new and intriguing insights into the Quatern-ary glacial history of the Northern Hemisphere. Here we integrate multi- and single beambathymetric data, chirp sonar profiles and sidescan images from the Chukchi Borderlandand Lomonosov Ridge to perform a comparative morphological seafloor study. This invest-igation aims to elucidate the nature and provenance of ice masses that impacted the ArcticOcean sea floor during the Quaternary. Mapped glaciogenic bedforms include iceberg keelscours, most abundant at water depths shallower than ~350–400 m, flutes and megascaleglacial lineations extending as deep as ~1000 m below the present sea level, small drumlin-like features and morainic ridges and grounding-zone wedges. The combination of thesefeatures indicates that very large glacial ice masses extended into the central Arctic Oceanfrom surrounding North American and Eurasian ice sheets several times during the Quaternary.Ice shelves occupied large parts of the Arctic Ocean during glacial maxima and ice rises wereformed over the Chukchi Borderland and portions of the Lomonosov Ridge. More geophys-ical and sediment core data combined with modeling experiments are needed to reconstructthe timing and patterns of these events. Copyright © 2008 John Wiley & Sons, Ltd.

Keywords: Arctic Ocean; seafloor morphology; geophysical mapping; ice shelves; icebergs;glaciogenic bedforms; Pleistocene

*Correspondence to: MartinJakobsson, Department ofGeology and Geochemistry,Stockholm University,Sweden. E-mail:[email protected]

Received 10 April 2007;Revised 6 August 2007;Accepted 7 August 2007

Introduction

It is well established that large areas of Polar continental margins as well as some submarine ridges in the adjacentdeep oceans have been repeatedly impacted by grounded ice sheets (see, e.g., Davies et al., 1997; Dowdeswell and ÓCofaigh, 2002). Ice-produced bedforms preserved on the seafloor in such areas provide the keys to our understandingof the marine extension and history of these ice sheets. Over the last two decades, glaciogenic morphology has beeninvestigated at many sites on formerly glaciated seafloors on the continental margins of North America (Josenhans andZevenhuizen, 1990; Todd et al., 1999), northern Eurasia (Ottesen et al., 2005; Polyak et al., 1997; Solheim et al.,1990; Vogt et al., 1994) and around the Antarctic continent (Anderson, 1999; Canals et al., 2002, 2000; Ó Cofaighet al., 2002; Shipp et al., 1999; Wellner et al., 2006). New advances in geophysical mapping technologies such ashigh-resolution multibeam bathymetric echo sounders and sub-bottom profilers can provide an even clearer view ofglaciogenic morphology than remote sensing imaging of land areas due to the effects of post-glacial land erosion andobscuring vegetation cover. For example, a variety of glaciogenic bedforms such as de Geer moraines, flutes, drumlinsand iceberg scours have been mapped using multibeam sonar with exceptional detail on the Svalbard, Antarctic (RossSea) and Nova Scotia continental margins (Ottesen and Dowdeswell, 2006; Shipp et al., 1999; Todd et al., 1999;Wellner et al., 2006).

Glacial geomorphology of the Central Arctic Ocean 527


Table I. Expeditions when data used in the present study were collected

Cruise

Arctic Ocean 96SCICEX 98/99

HLY0302HLY0501/0503

*The bathymetric and sidescan sonar data acquisition was only partly working during the 1998 cruise.

Vessel

Swedish icebreaker OdenUS Navy nuclear submarinesUSS Pargo and HawkbillUSCGC HealyUSCGC Healy

Geophysical mapping/geological sampling

chirp sonar profiling/sediment coringchirp sonar profiling/swath bathymetry andsidescan with interferometric sonar*chirp sonar profiling/multibeam bathymetrychirp sonar profiling/multibeam bathymetry/sediment coring

References (e.g.)

Backman et al. (1997)Edwards and Coakley (2003)

Jakobsson et al. (2005)Darby et al. (2005)

The perennially sea ice covered Arctic Ocean stands out as one critical area for which geomorphic data collection isless extensive, despite its obvious importance for understanding the Quaternary history of both the Arctic Ocean deepcentral basin and the shallow shelf areas, where large ice sheets extended during glacial maxima. Nevertheless, the useof seafloor mapping equipment from submarines and icebreakers during the last decade has generated a considerableamount of geomorphic data from the Arctic Ocean, only a minor portion of which has been published (Engels et al.,2008; Jakobsson, 1999; Jakobsson et al., 2005; Kristoffersen et al., 2004; Polyak et al., 2007, 2001). Not surprisingly,the interpretation of reported glaciogenic features allows for different viewpoints, and the reconstruction of the historyand dynamics of ice masses invading the Arctic Ocean remain poorly understood. The hypothesis of the pan-Arctic iceshelf proposed by Mercer (1970) and further advocated notably by Grosswald and Hughes (1999) postulates theexistence of a kilometre-thick floating ice shelf over the entire Arctic Ocean, which probably implies the initialformation of this ice shelf by thickening of sea ice. The opposite point of view is that all deep-draft erosion found inthe Arctic Ocean was generated by large icebergs carried in ‘armadas’ by currents (Kristoffersen et al., 2004). There isa need for a synthesis and analysis of the accumulated data in order to further test these hypotheses. Here we reportand discuss a compilation of seafloor mapping data acquired during the 1998–1999 SCICEX nuclear-submarineexpeditions and several icebreaker cruises including the Oden 1996, the USCGC Healy 2003, and the 2005 Healy–Oden Trans-Arctic Expedition (HOTRAX) (Table I; Figure 1). We focus on the two principal areas that contain themajority of mapped glaciogenic bedforms in the Arctic Ocean – the Chukchi Borderland and the central LomonosovRidge. A comparative analysis of glaciogenic morphologies from these areas allows a characterization of the impact ofdeep-draft ice on the central Arctic Ocean.

Study Areas

There are several specifics of the Arctic Ocean that make it different from other oceanic regions that have beenaffected by large ice masses. The Arctic Ocean basin is nearly landlocked and relatively small, approximately9·5 × 106 km2 (Jakobsson, 2002), with only one deep-water connection to the world ocean via the Fram Strait (Figure 1).No other ocean experienced such dramatic changes between glacial and interglacial periods; during periods of low sealevel during glacial stages the area of the Arctic Ocean shrank by more than 50 per cent (Jakobsson, 2002) and abouthalf of the exposed shallow continental margins were occupied by continental ice sheets (Dyke et al., 2002; Svendsenet al., 2004). The ice and melt-water discharge from these ice sheets had an exceptionally high impact on the ArcticOcean circulation and sedimentation regimes. Indeed, seafloor mapping data indicate bedforms resulting from glacialerosion and moulding of sea floor at depths reaching 1000 m. These glacially impacted areas are not always immedi-ately adjacent to former ice sheets, but are located inside (Lomonosov Ridge) or protrude into (Chukchi Borderland,Yermak Plateau) the Arctic Ocean basin.

The Chukchi Borderland is comprised of a group of generally less than 1000 m deep, north-trending ridges thatsurround the extensional Northwind Basin (Figure 2) (Hall, 1990). The easternmost of these ridges is the NorthwindRidge, which is deeper than its western neighbours and is characterized by an exceptionally steep slope towards theCanada Basin and a gently rounded to flat topped ridge crest. The Chukchi Spur–Chukchi Plateau composite ridge lieson the western side of the Northwind Basin and has a wide (>140 km at 78° N) flat topped crest mainly shallower than600 m (Figure 2(a)). In addition to the large Northwind Ridge and Chukchi composite ridge, several much smallerridges rise above the floor of the Northwind Basin.

The Lomonosov Ridge is a continental sliver that separated from the Barents–Kara Sea continental margin at about57 Ma during the opening of the Eurasian Basin through sea floor spreading (Brozena et al., 2003; Jokat et al., 1995;



Figure 1. Bathymetric map of the Arctic Ocean showing track lines of the cruises that collected the geophysical data studied inthis work (orange, Healy 2003; yellow, Healy (HOTRAX) 2005; red, SCICEX 1999; purple, SCICEX 1998; light grey, Oden 1996).The types of data collected during each of these cruises are listed in Table I. The black boxes outline the locations of the mainstudy areas of the Chukchi Borderland (Figure 2) and the Lomonosov Ridge (Figure 3). The two main surface circulation patternsthat govern the present sea ice drift, the Beaufort Gyre and Transpolar Drift, are marked with grey bold arrows. AG, AmundsenGulf; AR, Alpha Ridge; CP, Chukchi Plateau; LR, Lomonosov Ridge; MS, M’Clure Strait; MR, Mendeleev Ridge; MV, MackenzieValley; NR, Northwind Ridge; StA, Saint Anna Trough; YP, Yermak Plateau; FS, Fram Strait. The 1000 m isobath is shown with a blackline.

Vogt et al., 1979). The ridge extends 1650 km across the central Arctic Ocean, from the continental margin of northernGreenland all the way to the shallow shelf of the Laptev Sea. The ridge crest is generally deeper than 1000 m,although there are shallower portions in both the eastern and western ends and in the central part between about83°50″ and 87°30″ N; 140° and 160° E (Figure 3). The morphology is here characterized by a flat topped ridge crest,with the exception of the smaller areas at about 84°20′ N; 148° E and 86° N; 158° E just extending above the 1000 misobath.

The first evidence of deep-draft ice impact in the Arctic Ocean was provided by single-beam bathymetric linesshowing the extensively scoured surface of the Chukchi Plateau at depths to more than 400 m (Hunkins et al., 1962).Deeper scours, to at least 850 m, were found on the Yermak Plateau, indicating the passage of numerous exceptionallylarge icebergs exceeding in thickness the biggest icebergs found today off Antarctic coasts (Vogt et al., 1994). The1996 Oden expedition collected sub-bottom profiler records and sediment cores from the Lomonosov Ridge, at thevery centre of the Arctic Ocean (Backman et al., 1997). The surveyed portion of the ridge crest was shown to betruncated to water depths of ~1000 m, with a large stratigraphic hiatus and a layer of over-compacted sedimentscharacterizing the truncated surface (Jakobsson, 1999; Jakobsson et al., 2001).

The SCICEX 1998–1999 submarine expeditions with USS Pargo and Hawkbill included a relatively dense survey(sub-bottom and/or sidescan sonar profiling with approximately 10 km between track lines) of the central LomonosovRidge as well as a number of tracks on the Chukchi Borderland (Edwards and Coakley, 2003) (Figure 1). These dataprovided convincing evidence that the eroded areas on both the Lomonosov Ridge and the Chukchi Borderland



Figure 2. (a) Bathymetric map of the Chukchi Borderland and Mendeleev Ridge (see the methods section for an explanation ofthe bathymetric compilation). Major glaciogenic features and mapped limits of erosion are indicated. The directional features referto drumlins (Northwind Ridge) and ice-dragged sediment blocks or boulders (Chukchi Plateau). Locations of sub-bottom profilesin Figures 4(a) and 8 are indicated with bold yellow lines; also shown are locations of Figures 6 and 7. The 1000 m isobath is shownwith a white line and the 600 m with a black line. (b) Bathymetric profile along the Northwind Ridge. Mapped depths of erosionare shown by dashed lines. Also shown is the location of chirp sonar fragments and locations of figures with multibeam andsidescan images.

contain widespread, diverse glaciogenic morphology such as sets of parallel lineations (flutes), transverse ridges anddrumlins (Polyak et al., 2001). The combination of these features depicts a dramatic impact of large ice masses thatonce invaded the central Arctic Ocean from both the Laurentide and Eurasian ice sheets. Later surveys performedfrom the USCGC Healy detailed and expanded the SCICEX results and revealed a complex pattern of ice impact onthe eroded areas (Darby et al., 2005; Jakobsson et al., 2005). In particular, combined with stratigraphic investigationof sediment cores (Polyak et al., 2007), these data indicate multiple episodes of glacial erosion on the ChukchiBorderland with different orientations of ice movement.



Figure 3. (a) Bathymetric map of the central Lomonosov Ridge showing the ice eroded areas, the main directions of the erodingglacier ice inferred from mapped glaciogenic features and re-deposited sediment. Locations of sub-bottom profiles in Figures 5 and9 are indicated with bold yellow lines. The 1000 m isobath is shown with a white line. (b) Bathymetric profile from S to N shownin (a). The deepest erosion mapped along this profile is at ~940 –950 m. Below ~850 m the seafloor morphology is characterizedby flutes with a less rough and more regular appearance than above this water depth.

Methods

USCGC Healy Multibeam bathymetry and chirp sonar sub-bottom profilingThe USCGC Healy is equipped with a 12 kHz Seabeam 2112 multibeam bathymetric sonar. It has 151 beams andproduces a swath width of about 2·5–3·5 times the water depth. Multibeam bathymetry was collected from theAlaskan continental shelf and Chukchi Borderland during the HLY0302, HLY0501 and HLY0503 cruises (Table I andFigure 1). The Lomonosov Ridge was mapped during the HLY0503 cruise, but only at water depths deeper than thedeepest observed ice erosion, i.e. depths greater than 1000 m. Sound speed profiles of the water column for depthcalibration were acquired at least once a day either using the ship’s CTD (SeaBird SBE-911) or expendable probes(XBT, XSV, XCTD). During heavy ice breaking in the central Arctic Ocean (9/10 sea ice cover), the quality of theacquired bathymetric data was substantially reduced. The initial post-processing of the multibeam data was doneonboard USCGC Healy directly after data acquisition using the software Caris HIPS™ and Fledermaus™. For thispresent study, further post-processing and data analysis has been carried out at Stockholm University using the samesoftware set-up as onboard USCGC Healy. Grids with optimized resolutions (25–100 m) depending on the generaldepths of mapped areas were created for the analysis of glaciogenic bedforms.

In addition to a multibeam system, the Healy has two sub-bottom profilers: (1) Knudsen 320B/R dual frequency(centre frequencies 4·5 and 12 kHz) and (2) ODEC Bathy-2000 (centre frequency 3·5 kHz). Both systems are ‘chirpsonar’ sub-bottom profilers operating using frequency modulated (FM) signals and apply match filtering signal processing



to compress and remove the source signature (Schock et al., 1989). The sub-bottom data in the two systems’ acquisi-tion software can be stored both in native (ODEC and Knudsen respectively) and SEG-Y formats. The SEG-Y fileswere post-processed in this study using the public domain seismic software package Sioseis (http://sioseis.ucsd.edu/).Envelopes of the sub-bottom traces (the correlates) were computed through standard signal processing procedures,automatic gain control (AGC) was applied when needed, and the sub-bottom profiles were displayed in grey scales.

SCICEX bathymetry and sub-bottom profilesThe Seafloor Characterization and Mapping Pod (SCAMP) was installed on the nuclear submarine USS Hawkbill forthe SCICEX 1998–1999 expeditions and contained a sidescan swath bathymetric sonar and a modified ODEC Bathy-2000 sub-bottom profiler (Chayes et al., 1998). The ODEC sub-bottom profiling data were stored in SEG-Y formatand post-processed with Sioseis in this present work, applying the same set-up as described above for the USCGCHealy data. The SCICEX sidescan and swath bathymetry was initially processed using the open source softwareMB-System and Generic Mapping Tools (GMT) (Caress and Chayes, 2001; Wessel and Smith, 1991) by the HawaiiMapping Research Group (Edwards and Coakley, 2003). Bathymetric grids with resolutions of 250 m × 250 m on apolar stereographic projection were included in the bathymetric regional compilations described below and used tocalibrate the depths of the SCICEX sub-bottom profiles.

Bathymetric regional compilationsDigital bathymetric models (DBMs) with grid resolution of 1250 m × 1250 m on a polar stereographic projectionwere compiled for the central Lomonosov Ridge and the Chukchi Borderland with adjacent areas using all availabledepth information (Figures 2 and 3). In addition to the multibeam bathymetry acquired during the USGCG Healyand SCICEX expeditions, the depth information includes the entire International Bathymetric Chart of the ArcticOcean (IBCAO) database (Jakobsson et al., 2000) and multibeam bathymetry not yet included in IBCAO from theHLY0405 cruise with USCGC Healy (Mayer, 2004) and R/V Nathaniel B Palmer cruises within the Shelf BasinInteractions (SBI) programme (http://www.eol.ucar.edu/projects/sbi/). In order to compile the DBMs at a resolution of1250 m × 1250 m, interpolation between the multibeam, single beam and digitized bathymetric contours from bathymetricmaps was performed applying the same methodology and algorithms as used to compile the IBCAO DBM (Jakobssonet al., 2000).

Glaciogenic Bedforms

Iceberg scoursThe most common type of seafloor feature in polar areas is ice keel scours, also commonly called plowmarks, furrowsor gouges (Davies et al., 1997; Duncan and Goff, 2001; Meyer et al., 2000). Sea ice can only scour the seafloor towater depths of less than approximately 50 m, while iceberg erosion can extend to depths of several hundred metres(Dowdeswell and Bamber, 2007; Dowdeswell and Whittington, 1992). Understanding of the formation of icebergscours has been enhanced by comparison with modern iceberg drift patterns (Tchernia and Jeannin, 1984) and bydirect field observations (Woodworth-Lynas et al., 1991). On the Chukchi Borderland and adjacent Alaskan marginscours are very abundant at depths shallower than ~350–400 m (Figure 4) (Engels et al., 2008; Hunkins et al., 1962;Jakobsson et al., 2005; Polyak et al., 2001). These are mostly sinuous, chaotically oriented, overlapping incisions witha v-shaped cross-section, less than 50 m wide, a few metres deep and less than 5 km long – analogous to ‘gouge-typefurrows’ at the Antarctic margin (Meyer et al., 2000; Wellner et al., 2006). This is the most common type of moderniceberg scouring around sea-bound glaciers and ice sheets. At greater water depths on the Chukchi Borderland scoursare more sparsely distributed, but they are larger and extend linearly for much longer distances. Individual scoursreaching water depths of ~500 m are up to 700 m wide, 30 m deep and several kilometres long. They display internalgroove-and-ridge topography as characteristic of the ‘plane furrow’ type (Meyer et al., 2000; Wellner et al., 2006).These features must have been produced by very large, possibly tilted tabular icebergs (Dowdeswell and Bamber,2007). A similar pattern in the distribution of iceberg scours has been reported from the St. Anna Trough at theglaciated Eurasian Arctic margin (Polyak et al., 1997).

Exceptionally deep seafloor areas (to ~850 m below present sea level) displaying multiple, large iceberg scours arefound on the Yermak Plateau and the central Lomonosov Ridge (Figures 3 and 5) (Jakobsson, 1999, 2000; Polyaket al., 2001, 1999; Vogt et al., 1994). These scours are closely spaced and have a sub-parallel orientation, most

http://sioseis.ucsd.edu/

http://www.eol.ucar.edu/projects/sbi/



Figure 4. (a) Chirp profile CS2 from the ramp to the Northwind Ridge (see Figure 2 for location). (b) Multibeam swath from thesame profile as shown in (a). Data were collected during the HLY0503 expedition (Table I). (c) Sidescan image (SCICEX 1999 data)from the same location as CS2.



Figure 5. (a)–(c) Chirp profiles LR1 and LR2 along the iceberg-scoured area on the Lomonosov Ridge and LR3 from a deeperfluted area. The transition from an ice-scoured to lineated seabed at about 850 m water depth is clearly seen in LR1. Data werecollected during the SCICEX 1999 expedition (Table I). The redeposited sediment leeward of the eroded ridge top is seen in LR2and LR3. (d) Sidescan image of iceberg scours (SCICEX 1999 data) from the same location with sub-bottom profile LR1.

probably indicating the passage of (an) armada(s) of large icebergs (Figure 5(d)). The prevalent orientation of closelyaligned scours on the Lomonosov Ridge allows for two opposite interpretations of iceberg drift direction – from theEurasian versus the Amerasian basin of the Arctic Ocean (Jakobsson, 2000; Kristoffersen et al., 2004; Polyak et al.,2001). Accordingly, the mapped scours on the Yermak Plateau may indicate an ice drift direction either from or intothe Arctic Ocean.

Flutes and megascale lineationsAnother type of bedform common on glacially eroded polar continental margins areas is formed by fields of linear,evenly spaced, ridge-and-groove features – in both the Arctic (Jakobsson et al., 2005; Ó Cofaigh et al., 2002; Ottesenet al., 2005; Polyak et al., 2001, 1997; Solheim et al., 1990) and the Antarctic (Anderson, 1999; Gilbert et al., 2003;Shipp et al., 1999; Wellner et al., 2006). Such features, referred to as flutes or lineations, are characteristic of glaciatedterrains with a relatively soft bed (Benn and Evans, 1998; Dowdeswell et al., 2004). Although there is a variety ofopinions on the significance of erosional versus depositional mechanisms in the formation of flutes, most authorsagree on their relation to relatively fast flowing ice such as at surging ice margins and ice streams, with the length offlutes roughly proportional to ice velocity (Clark et al., 2003; Stokes and Clark, 1999).

In the Arctic Ocean, lineations are found on the Lomonosov Ridge and the Chukchi Borderland with the adjacentAlaskan margin (Figures 4–7) (Engels et al., 2008; Jakobsson, 1999; Jakobsson et al., 2005; Polyak et al., 2001), aswell as in the deep troughs traversing the formerly glaciated Arctic margins (Ottesen et al., 2005; Polyak et al., 1997;Stokes et al., 2006). Lineations are typically visible at water depths below the range of intense iceberg scouring andare observed to ~1000 m depth on the Lomonosov Ridge and ~900 m on the Chukchi Borderland (Northwind Ridge)(Figures 2 and 3). Deep-water Arctic Ocean lineations are mostly spaced between less than 50 and 200 m (crest tocrest) and have low amplitudes (under 5–10 m) and lengths of more than 15 km. However, it should be noted thatthese are probably longer as measurements are limited by swath data coverage and/or the width of eroded ridge tops.The largest lineations mapped on the Chukchi Borderland (on the ramp to the Northwind Ridge) are spaced between 200and 1000 m apart and have a relief (from the ridge top to the groove bottom) as high as 40 m (Figure 6) (Jakobssonet al., 2005). These features are similar to megascale glacial lineations reported from the Antarctic continental margin,



Figure 6. Multibeam lines at the ramp to the Northwind Ridge showing megascale lineations (see Figure 2 for location). Themultibeam data were acquired during the HLY0503 and HLY0302 expeditions (Table I).

where they have been related to palaeo-ice streams (Anderson, 1999; Ó Cofaigh et al., 2002; Shipp et al., 1999;Wellner et al., 2006).

Seafloor lineations comparable to those formed by glacial processes are found in non-glaciated regions as well(Bryant et al., 2007; Flood, 1983; Kuijpers et al., 2002; Poppe et al., 2006). These features typically occur in areasaffected by persistent currents, either tidal or topographically steered flows through straits or along slopes. However,despite the general similarity, current-generated lineations have marked differences from their glaciogenic counter-parts. While the latter are composed of ridges and troughs with similar cross-sectional shape and dimensions, currentlineations are commonly formed by distinct furrows cut into generally flat seafloor and separated with interspacesseveral times wider than the furrows. Furthermore, these furrows have some peculiar features such as ‘tuning forks’that are characteristic of the action of flowing water, but not ice (Flood, 1983; Poppe et al., 2006). In addition tomorphology, critical information is also provided by the seafloor settings in which the lineations are found. Deep-seacurrent furrows are typically located along the slopes of continental margins and ridges, whereas glaciogenic flutes canoccur in a variety of geomorphological settings, from troughs to ridge crests, and can be oriented at various angles tobathymetric obstacles. Moreover, glacial lineations are always found in an orderly combination with other glaciogenicbedforms including iceberg scours, morainic ridges and drumlins, as well as large-scale erosional features such astruncated strata. The combination of the above allows a confident identification of the glaciogenic origin of thelineations found on the Lomonosov Ridge and Chukchi Borderland.

Unlike chaotic or sub-parallel iceberg scours, subglacial lineations retain their parallel fabric for long distances.This, combined with the width of lineation fields (tens of kilometres or possibly more), indicates that they have beengenerated by very large, coherently moving ice masses such as ice streams or grounded portions of ice shelves knownas ice rises. Lineated areas are typically associated with large-scale seafloor erosion, accentuated by a conspicuoustruncation of pre-glacial strata capped with a mostly thin layer of diamict sediment, in which the lineations are formed(Dowdeswell et al., 2004). It must be noted that sub-bottom sparker or sonar records across iceberg-scoured andlineated areas may look similar due to abundant hyperbolic reflections; therefore, swath imagery (sidescan sonar ormultibeam bathymetry) is critical for a definitive identification of lineations (Figures 4–7).

DrumlinsDrumlins are not common features of the Arctic Ocean seafloor, unlike on some glaciated continental margins, wheredrumlins are characteristic of the transitional zone between crystalline bed on the inner shelf and soft sedimentary



Figure 7. Sidescan images from the northern Northwind Ridge (SCICEX 1999 data) (see Figure 2 for location). (a) A complex oflineations and morainic ridges; some of the latter are drumlinized.

strata further offshore (Anderson et al., 2001; Wellner et al., 2006, 2001). Seafloor drumlins can also be formed fromice advances across morainic ridges (Fader et al., 1997). The reason for the rarity of drumlins in the ice eroded areasof the Arctic Ocean may be the soft pre-glacial bed comprised by poorly consolidated Meso-Cenozoic strata (Grantzet al., 1990; Moran et al., 2006). Nevertheless, small drumlin-like forms 300–600 m long are found at some sites,notably on the Northwind Ridge (Figure 7(a)). These bedforms appear to be formed by overriding and overprinting ofmorainic ridges by a younger ice movement. The presence of drumlins probably indicates a warm based ice sheet, andtheir extension is particularly important for identification of the direction of ice flow (Hättestrand et al., 2004).

Morainic ridges and wedgesIn several areas on Chukchi Borderland, the glacially influenced seafloor features nested sets of slightly sinuous orarcuate ridges, transverse or oblique to flutes (Figure 7(a)). Ridges appear to be less than 10 m high (their exact heightis difficult to assess due to the insufficient resolution of the SCICEX bathymetry), reach 100+ m in width and areindividually spaced between 100 and several hundred metres. Similar features have been mapped on many glaciatedshelves and interpreted as transverse morainic ridges (grounding zone wedges) formed by stepwise retreat of anice-sheet margin or grounding line (Davies et al., 1997; Ottesen and Dowdeswell, 2006; Ottesen et al., 2005; Polyak



Figure 7. (b) Cross-cutting relationship between two generations of lineations.

et al., 1997; Shipp et al., 2003; Todd et al., 1999; Wellner et al., 2006). Ridges mapped on the southeastern ChukchiPlateau are remarkably parallel to bathymetric contours (Polyak et al., 2001). This may indicate that the formation ofthese ridges was controlled by sea level and occurred in relation to the grounding line of a retreating ice cap (lift-offmoraines) (Ottesen and Dowdeswell, 2006).

At several locations, larger ridges or wedges reach thicknesses of ~30 m. An example is the prominent asymmetric,nested wedges at the northernmost eroded bathymetric high of the Northwind Ridge (Figure 8(c)). These wedges havean acoustically transparent character with virtually no visible internal reflectors, and coring during the HOTRAXexpedition revealed a diamict sediment composition. Similar wedges are characteristic of former grounding lines atthe maximal ice-sheet extension where significant volumes of sediment have been accumulated (Anderson, 1999;Elverhøi and Maisey, 1983; Shipp et al., 1999). Fairly large ridges are also found at several segments of the ChukchiPlateau perimeter at water depths of 600–700 m (Figures 2 and 8(a)), probably marking the limits of an ice rise thatonce capped the plateau.

Redeposited sediment accumulationsSeveral seafloor depressions adjacent to eroded areas on both the Lomonosov Ridge and Chukchi Borderland containaccumulations of acoustically transparent material forming lenses occasionally as much as 75 m thick and several



Figure 8. Sub-bottom profiles from the Chukchi Borderland. Figure 2 shows the locations of the profiles. (a) Morainic ridgesfrom the Chukchi Plateau. (b), (e) Glacially carved bathymetric benches in the Chukchi continental shelf. (c), (f) Smooth seabedwith no signs of glacial erosion on the Northwind Ridge at 620 m water depth and on the Chukchi Shelf at 450 m. (d) Morainicwedge on the Northwind Ridge.

kilometres across, sometimes with a convex relief (Figures 2–5). These accumulations, particularly pronounced on theLomonosov Ridge, are coupled with an ice eroded seafloor, and we assume that these are bodies of diamict sedimentformed by downslope debris flow (creep) of till in grounding-line proximal areas, similar to inferred ‘marine flow-till’accumulations in the depressions of the Barents–Kara margin that reach the thickness of almost 100 m (Gataullinet al., 1993; Polyak et al., 2002). Coring of these deposits on the Lomonosov Ridge and Chukchi Borderland mayprovide the necessary material to study their exact formation mechanism.

Large-scale bathymetric featuresSeafloor areas that feature small to medium-sized subglacial bedforms such as flutes and morainic ridges also oftenbear evidence of erosion over large areas. For example, this is expressed in the carving of broad troughs and bathymetricbenches on the continental shelves and slopes (Blasco et al., 1990; Canals et al., 2002) and planation of ridge andplateau tops (Jakobsson, 1999; Polyak et al., 2001). Troughs that traverse glaciated continental margins in both theAntarctic and the Arctic typically contain multiple flutes or megascale lineations that indicate that these troughsoperated as conduits for fast-flowing ice streams (Canals et al., 2000; Polyak et al., 1997; Stokes et al., 2006). A less



common case of continental slope erosion is a flattened bathymetric bench that is evident on the Alaskan margin atwater depths between 400 and 550 m (Figure 8(b), (e)) (Engels et al., 2008). The broadest part of this bench, almost40 km wide, can be traced for ca. 100 km along the Beaufort/Alaskan slope, and its narrower continuation extends allthe way to the Chukchi Borderland, with small additional benches visible at some segments. Lineations found on theAlaskan slope are concentrated within this bench, which, together with its clearly erosional nature (truncation of slopestrata), suggest that the bench has been formed by a large ice mass sliding along the margin.

A spectacular large-scale erosional feature that is intrinsic to the interior of the Arctic Ocean is the planation ofridge and plateau tops. This erosion is especially conspicuous atop relative bathymetric highs of the Northwind andLomonosov Ridges, which feature flat, somewhat tilted surfaces with a pronounced unconformity (Figures 5(c) and8(d)). The geometry of unconformity allows an estimate of the thickness of eroded strata, which amounts to more than50 m on the Lomonosov Ridge (Jakobsson, 1999).

Discussion

Regional glacial historiesChukchi Borderland. Numerous and diverse glaciogenic bedforms indicative of an impact(s) of very large and coher-ent ice masses are mapped on the Chukchi Borderland, including the Chukchi Plateau, Chukchi Spur and NorthwindRidge, at water depths reaching locally more than 900 m (Figure 2). Sets of bedforms are commonly superimposedupon one another, forming a geomorphic palimpsest indicative of a complex erosion history that probably involvesseveral glacial events. This inference is consistent with the stratigraphic evidence of at least two asynchronous diamicthorizons recovered in sediment cores from the eroded area at the ramp to the Northwind Ridge (Polyak et al., 2007).Based on these data, the youngest and least extensive ice-grounding event occurred during the Last Glacial Maximum(LGM), whereas the more widespread glacial impact is dated to Marine Isotope Stage (MIS) 4 or older (Polyak et al.,2007). The preliminary results of the HOTRAX’05 expedition suggest the presence of yet another diamict horizon,tentatively attributed to MIS 6.

Except for iceberg scours that largely obliterate pre-existing bedforms at water depths to 350–400 m, the mostcommon glaciogenic features are lineations or megascale lineations running in a variety of orientations (Figure 2).Lineations are found on all bathymetric highs of the Northwind Ridge that rise above the depth limit for the glacialerosion, henceforth referred to as the ‘trimline’, and on the flanks of the Chukchi Plateau. At least two of the lineatedfields have directional markings such as dragged sediment blocks or boulders on the Chukchi Plateau (Polyak et al.,2001) and drumlinized ridges at the northern Northwind Ridge (Figure 7(a)). These markings indicate the SE–NWdirection of eroding ice at these sites. An overprinting of one generation of lineations by another is clearly visible inseveral locations (Figure 7(b)). These features may indicate the impact of separate glacial events or a complex patternof ice flow during just one event (Dowdeswell et al., 2006; Wellner et al., 2006).

Sets of morainic ridges are also common for the Chukchi Plateau and some sites on the Northwind Ridge (Figures 2,7(a) and 8(a)). The depth-contour alignment of ridges (lift-off moraines) mapped at the SE edge of the ChukchiPlateau indicates that they probably encircle the plateau (Polyak et al., 2001). This is corroborated by the consistentpresence of ridges on sub-bottom sonar lines crossing its edges (Figures 2 and 8(a)). The placement of ridges suggeststhat they mark the limits of an ice cap that once covered the plateau, whereas the nested, bathymetry-controlled ridgepattern follows the retreat of this ice cap during sea-level rise. On the Northwind Ridge morainic ridges are abundantat the northernmost eroded mount where they are generally transverse to flutes and are partially drumlinized (Figure7(a)). Furthermore, a stack of thick morainic wedges resides at the northern edge of this shallow area (Figure 8(d)),and large morainic accumulations are found atop eroded highs further south. In addition to lineations and ridges,lenses of displaced diamict material are common for the depressions in the Chukchi Plateau and slopes of NorthwindRidge bathymetric highs (Figures 2 and 4).

The bathymetric distribution of the eroded areas at the Chukchi Borderland is somewhat asymmetric (Figure 2). Theeroded surface (trimline) dips to the largest water depths of more than 900 m at the southern part of the NorthwindRidge, from where it shallows to ~620 m along the ridge northwards and to between 760 and 800 m southwardstowards the continental margin. Chirp sonar data show a yet shallower trimline at 550–700 m along the western edgeof the Chukchi Plateau, whereas less than 200 km SW of the plateau the Chukchi margin appears to be uneroded atdepths as shallow as 450 m (Figure 8(f)). This trimline distribution indicates that the impact of eroding ice masses onthe sea floor had a complex pattern probably controlled by the interaction of ice geometry and dynamics with thelarge-scale morphology of the continental slope and the borderland. The upper limit of the distribution of erosionalbedforms towards the continental margin appears to be at a depth of ca. 400 m south of the Chukchi Cap; however,



their identification at shallower depths may not be possible because of intense iceberg scouring at deglaciation andwave erosion during sea-level rise.

Overall, the orientation of glaciogenic bedforms suggests the possibility of two mechanisms of glacial impact on theChukchi Borderland. One orientation of lineations and the spatial pattern of large-scale erosion indicate the impact ofice masses overriding the Chukchi Borderland from the east and southeast (Figure 2). We infer that they were formedby grounding portions of a coherent floating ice masses (ice shelves) propelled by ice streams of North-AmericanPleistocene ice sheets (Dyke et al., 2002; Kleman and Glasser, 2007). The nearest possible source was the NW sectorof the Laurentide Ice Sheet, which contained very large ice streams discharging voluminous amounts of ice into theadjacent Arctic Ocean (Bigg and Wadley, 2001; Stokes et al., 2005, 2006). The east–west oriented lineations suggestthe direct flow of ice from these ice streams, whereas the NW-trending flutes and megascale lineations indicate a morecurved pattern of ice flow, possibly steered by the continental slope and constrained by very thick ice in the interior ofthe Arctic Ocean. This pathway for the flow of eroding ice is consistent with the erosional bench at the Alaskan slopebetween the NW Laurentide margin and the Chukchi Borderland (Engels et al., 2008).

The other pattern that can be inferred from the distribution of glaciogenic bedforms such as the concentric ridges atthe edges of the Chukchi Plateau and the orientation of many fluted fields is the impact of an ice cap centred over theplateau or the entire borderland. Such an ice cap may have developed from an ice rise, such as those found around theAntarctic, where they are formed when an ice shelf grounds on a shallow portion of seafloor (Fahnestock et al., 2000).After its inception, an ice rise may start to act as an independent, local ice cap expanding from the centre outwards,with the ice shelf flowing around. From the available data we cannot elucidate conclusively the complex interactionbetween these potential sources of erosion. More detailed mapping and stratigraphic investigation is needed to under-stand the spatial and temporal relationship of bedforms formed by each of these ice masses.

Lomonosov Ridge. The major highs of the central Lomonosov Ridge, extending above approximately 1000 mpresent water depth, bear clear evidence of an impact of grounded ice masses, which extensively eroded the ridgecrest and left their imprint in the form of glaciogenic bedforms including lineations, iceberg scours and associatedlenses of redeposited sediment (Figure 3). The uncomformable truncation of the strata revealed by the sub-bottomprofiles provides an unambiguous erosional explanation for the overall flattened ridge crest morphology seen inbathymetric compilations (Figures 3 and 5) (Jakobsson et al., 2000; Naryshkin, 1999). The seafloor of the truncatedareas contains multiple flutes and ice scours. There is a marked morphological transition at approximately 850 mwater depth seen in the shallower southern eroded area around 86° N (Figures 3 and 5(a), (b)). Above this depth, theseabed is dominated by irregular scours with widths exceeding 100 m, suggesting rough under-ice topography. Sidescanimages of the scoured surface show subparallel iceberg scours with a prevalent orientation across the ridge (Figure5(d)) (Jakobsson, 2000; Polyak et al., 2001). Below the transition, the seafloor appears much more even and displaysa fluted fabric on sidescan images. We suggest that this change in morphology marks the limit for deep draft icebergs,whereas the deeper fluted seabed (down to 1000 m present water depth) requires larger, coherent ice masses such asmoving ice shelves that ground over bathymetric highs and form ice rises.

Icebergs that scoured the Lomonosov Ridge at depths shallower than 850 m could have been clustered together inarmadas (Kristoffersen et al., 2004) during periods of intense iceberg discharge from disintegrating ice sheets. The samedepth of relict iceberg scours with a similar orientation is found on the Yermak Plateau (Vogt et al., 1994). Icebergscours down to 850 m below present water depth imply iceberg drafts of about 700–730 m during glacial maxima(120–150 m lower sea level). Today, icebergs off Antarctica and Greenland reach depths of 500–550 m (Barnes and Lien,1988; Dowdeswell and Bamber, 2007; Dowdeswell et al., 1993), but relict scours off Greenland have been found to reachsubstantially larger water depths (Kuijpers et al., 2007). The iceberg dimensions required to ground on the LomonosovRidge are thus not exceptional. However, deep draft icebergs are usually found close to the grounding line of a glacier,whereas the distance between the scoured areas of the Lomonosov Ridge and the nearest continental margin that couldhave provided hundreds of metres thick icebergs during glacial maxima was more than 700 km away.

The position of redeposited sediments on only one side of the Lomonosov Ridge indicates that the eroding icemoved across the ridge from the Eurasian to the Amerasian side (Figures 3 and 5). The Barents and Kara Sea icesheet, which reached the shelf edge at least during Marine MIS 2, 4 and 6 (Svendsen et al., 2004), seems the mostlikely source for this ice. We note also that this continental margin features very large erosional troughs such as theSt. Anna Trough that probably contained major ice streams during glacial maxima (Siegert and Dowdeswell, 2004).Sediment core data show that the undisturbed 2–3 m sedimentary drape atop the ice-eroded surface of the LomonosovRidge below ca. 800 m water depth has been deposited since Marine Isotope Stage (MIS) 5·5 (Jakobsson et al., 2001).This result indicates that the deepest erosional event in this area occurred during MIS 6, which was the time of thelargest known glaciation on the Barents–Kara shelf (Svendsen et al., 2004). One core collected during the ArcticOcean 96 expedition with Oden at a shallower water depth of approximately 600 m has a shorter post-erosionalstratigraphy, thus indicating a younger erosional event caused by a thinner ice.



Figure 9. Sub-bottom profiles from the Lomonosov Ridge showing uneroded seafloor above 1000 m. Critical seafloor depthsused in our interpretations of the glacial morphology are shown in the profiles. The depths have been derived by depth calibrationof the sub-bottom profiles using multibeam bathymetry.

We conclude that the least contradictory interpretation of the existing data is that the major erosion that impactedthe central Lomonosov Ridge to 1000 m water depth was related to the extension of an ice shelf from the Barents–Kara ice sheet during MIS 6. After grounding on the proximal side of the ridge crest, the ice shelf could have formedan ice rise, so that further ice dynamics were controlled by the interaction of an ice shelf and ice rise. This mightexplain the lower thickness of the eroding ice on the distal (Amerasian) side of the ridge as indicated by the unerodedsurface in this area at water depth as shallow as ca. 900 m (Figures 3 and 9). Subparallel scours above 850 m depthwere probably gouged during deglaciation when armadas of icebergs formed upon disintegration of the ice shelf. Thedirection of iceberg movement is not apparent. Kristoffersen et al. (2004) hypothesized that icebergs were pushedacross the ridge from the Eurasian into Amerasian basin by currents due to intermediate Atlantic water influx.However, such a circulation pattern is not observed in the Arctic Ocean at present (Rudels et al., 1994), while glacial-time circulation is not understood and needs to be modelled. We infer that icebergs, as opposed to the ice shelf, mayhave rather moved towards the Fram Strait during disintegration of circum-Arctic ice sheets. This pattern is consistentwith the orientation of scours on the Yermak Plateau (Vogt et al., 1994).

Pan-Arctic perspectiveThe widespread occurrence of conspicuous glaciogenic bedforms on ridges and plateaus in the interior of the ArcticOcean indicates that very large masses of thick ice covered considerable portions or the entirety of this basin at severaltimes during the Pleistocene. The exact mechanisms, timing and provenance of these ice masses are not yet wellunderstood, but some inferences can be made. We find both the hypothesis of a pan-Arctic ice shelf as first proposedby (Mercer, 1970) and the opposite position that all deep-draft glacial erosion was generated by large ‘armadas’ oficebergs (Kristoffersen et al., 2004) difficult to reconcile with the observed patterns of seafloor erosion described inthis paper (Figures 2 and 3). The diversity of mapped glaciogenic bedforms, their orderly distribution over large areasand association with large-scale seafloor erosion indicate that most of these features were formed by voluminous,cohesive ice masses grounded on the seafloor rather than by disparate icebergs. On the other hand, the distributionpattern of the bedforms is not consistent with an extensive kilometre-thick ice shelf covering the entire basin, butsuggests a more complex erosional history involving ice shelves flowing from the ice sheet margins at the ArcticOcean continental margin and the formation of ice rises on the impacted ridges. This pattern is especially explicit onthe Chukchi Borderland, where the uneven depth distribution of erosional trimline and a complex superposition ofvarious bedforms indicate a repetitive impact of ice shelves protruding from the northern margin of the Laurentide icesheet combined with an ice rise formed over the Chukchi Plateau or the entire borderland (Figure 10). Based on cross-sectional area and the size of the associated trough-mouth fans, McClure Strait and Amundsen Gulf ice streams



Figure 10. Bathymetric map of the Arctic Ocean, modified to show conditions during maximum glaciation. Sea level is lowered by120 m. Ice sheet limits on the Eurasian and North American sides are shown, respectively, after Svendsen et al. (2004) and Dykeet al. (2002). Ice streams are shown after De Angeles and Kleman (2005) and Kleman and Glasser (2007). Blue arrows show mainice streams. Red arrows show ice stream tributaries and episodic diversions of ice stream drainage within main ice streamcorridors. Projected flow lines of ice shelves and limits of ice rises are mapped according to the authors’ interpretation ofobserved glaciogenic seafloor bedforms.

probably delivered most of the ice to ice shelves in the western Arctic. The three ice stream catchments that drainedthe Innuitian Ice Sheet northwards (Figure 10) had much smaller catchments than the McClure and Amundsen Gulfice streams. The extent to which Mackenzie ice streams reached the marine environment is poorly understood. Thelineation swarms on land cannot be reliably traced onto the shelf. As shown in Figure 10, there is direct landformevidence for large shifts in catchment location and size and the spatial arrangement of ice streams within the McClureStrait, Amundsen Gulf and Mackenzie valley ice stream corridors. The important implication is that ice dischargefrom the northern Laurentide Ice Sheet to the Arctic Ocean was highly variable in space and time, possibly onmuch shorter timescales than the major stadial/interstadial pacing of terrestrial ice volume variations. This shiftingdynamics, where the entrance point of major ice influx in the western Arctic probably shifted drastically in location,may have had important consequences for ice shelf thickness distribution, and may be an explanation for the shiftingmovement pattern indicated by the mapped glacial morphology on the Chukchi plateau. The implications from such ascenario immediately raise the question of how an ice shelf extending into the Canada Basin and fed from the majorice streams of the western Canadian Arctic Archipelago could reach the Chukchi Borderland. We hypothesize that a



strong Beaufort Gyre could help by pressing sea ice and the ice shelf towards the Beaufort margin and, thus, the iceshelf would flow along the Alaskan continental margin as suggested by Engels et al. (2008). An ice shelf covering thesouthern part of the Canada Basin up to ca. 76° N would be of the order of 500 × 103 km2, which is similar in size tothe currently 487 × 103 km2 Ross Ice Shelf of Antarctica. It should be noted that the initial grounding on the ChukchiBorderland may not necessarily have been from a coherent ice shelf extending all the way across the Canada Basin.Instead, large floating icebergs could have grounded on the Chukchi Plateau, which is partly shallower than 300 m andthus facilitated the growth of an ice cap.

The erosion mapped on the Lomonosov Ridge displays a simpler pattern, which is dictated by the narrowness of theridge that was oriented nearly transverse to the impacting ice flow. The principal question arising from these results ishow to get an Arctic Ocean ice shelf thick and large enough in order to ground on the central Lomonosov Ridge.Floating ice shelves fed by large ice sheets thin out with increasing distance from the grounding line due to theinternal ice shelf strain rate (Dowdeswell and Bamber, 2007). The mechanism of calving involves the propagation ofa fracture through the ice and, typically, icebergs calve from an ice shelf when the thickness has been reduced toaround a couple of hundred metres (Dowdeswell and Bamber, 2007). In order to test the hypothesis that an ice shelfonce grounded on the Lomonosov Ridge, a numerical ice sheet model was applied (Jakobsson et al., 2006). Theexperiments, albeit not including the complex dynamics of floating ice shelves, indicate that free-flowing ice emanat-ing from the Barents–Kara shelf could not have been thick enough to ground on the Lomonosov Ridge if the ice strainrate is not reduced. However, if this ice shelf was supported or ‘buttressed’ by a counter-force such as a substantiallythickened sea ice, it could become grounded on the ridge in the time frame of major glacial maxima (Jakobsson et al.,2006). A circum-Arctic modelling experiment shows similar preliminary results (Ritz et al., 2007). It should alsobe noted that surface mass balance and the basal melt/freeze-on of Arctic ice shelves during glacial maximaare completely unknown, but would have been strong controls on their ability to reach for example the LomonosovRidge.

An accurate age constraint for the erosional events and their correlation between various parts of the Arctic Oceanis critical for incorporating these events into the overall paleoglaciological context for the Northern Hemisphere. Theage of the Lomonosov Ridge erosion has been estimated as MIS 6 (Jakobsson et al., 2001) and on the ChukchiBorderland two erosional events have been dated to the LGM and MIS 4 or older (Polyak et al., 2007). Additionally,the new results from HOTRAX’05 core logs (Darby et al., 2005) reveal a third diamicton on the NorthwindRidge, which may be MIS 6 in age. The latter suggests the possibility of a synchronous expansion of ice shelvesfrom the Eurasian (Barents–Kara) and North American ice sheets into the central Arctic Ocean during MIS 6,which was possibly the most extensive Pleistocene glaciation in the Arctic, at least in Eurasia (Svendsen et al., 2004).This synchronicity would make it possible for the ice shelves to coalesce somewhere in the Arctic Ocean andconsequently expand over the entire basin. Limited evidence from other areas of the circum-Arctic margins pro-vides some support for this scenario. Notably, recently re-investigated glaciotectonic features in pre-Quaternarysediments and in buried ice at the northern part of new Siberian Islands indicate the advance of an ice sheet fromthe north-east (Grosswald, 1980; Tumskoy and Basilyan, 2007). This pattern of ice advance might be related tothe grounded edge of a thick floating ice shelf that extended to the periphery of the Arctic Ocean due to the interac-tion of major Eurasian and North-American ice shelves. Another piece of this pan-Arctic mosaic is located atthe northern coasts of Greenland and Ellesmere Island. There is evidence of glacier erosion from ice movingeastwards along the Greenland coast (Funder and Hansen, 1996; Funder and Kjær, 2007), but the source of this iceis not understood. If this source turns out to be from the Canadian Archipelago west of the Nares Strait, thiswould indicate the likelihood of a thick ice shelf over the entire Canada Basin that deflected ice streams of theInnuitian Ice Sheet eastwards. This question should be elucidated by future mapping of the adjacent part of theLomonosov Ridge, which shows evidence of deep-draft ice erosion on seismic profiles (Kristoffersen and Mikkelsen,2006) . Such mapping is planned for the 2007 LOMROG (Lomonosov Ridge off Greenland) Expedition (Jakobsson,2007).

AcknowledgementsThe work on this paper was supported by the Royal Swedish Academy of Sciences through a grant financed by the Knut and AliceWallenberg Foundation, the Swedish Research Council, Stockholm University Climate Research Environment through a grantfinanced by FORMAS and the US National Science Foundation. Data collection was also supported by several awards from the USNational Science Foundation, National Ocean and Atmospheric Administration, Swedish Research Council and Swedish PolarResearch Secretariat (award to Jan Backman). We thank Dennis Darby, Julian Dowdeswell and an anonymous reviewer for provid-ing useful comments on the manuscript and Paul Henkart for support on the signal processing of the SCICEX sub-bottom profiling data.We greatly appreciate that the Center for Coastal and Ocean Mapping, University of New Hampshire, made available their datacollected from the Chukchi Borderland.



References

Anderson JB. 1999. Antarctic Marine Geology. Cambridge University Press: Cambridge.Anderson JB, Wellner JS, Lowe AL, Mosola AB, Shipp SS. 2001. The footprint of the expanded West Antarctic Ice Sheet: ice stream history

and behavior. GSA Today 11: 4 –9.Backman J, Jakobsson M, Knies J, Knutsen JO, Kristoffersen Y, Lif A, Musatov E, Stein R. 1997. Geological Coring and High Resolution

Chirp Sonar Profiling. Swedish Polar Research Secretariat.Barnes PW, Lien R. 1988. Iceberg rework shelf sediments to 500 m off Antarctic. Geology 16: 1130–1133.Benn DI, Evans DJA. 1998. Glaciers and Glaciation. Arnold: London.Bigg GR, Wadley MR. 2001. The origin and flux of icebergs into the Last Glacial Maximum Northern Hemisphere Oceans. Journal of

Quaternary Science 16: 565–573.Blasco SM, Fortin G, Hill PR, O’Connor MJ, Brigham-Grette J. 1990. The late Neogene and Quaternary stratigraphy of the Canadian

Beaufort continental shelf. In The Geology of North America, Vol. L, The Arctic Ocean Region, Grantz A, Johnson L, Sweeney JF (eds).Geological Society of America: Boulder, CO; 491–501.

Brozena JM, Childers VA, Lawver LA, Gahagan LM, Forsberg R, Faleide JI, Eldholm O. 2003. New aerogeophysical study of the EurasiaBasin and Lomonosov Ridge: implications for basin development. Geology 31(9): 825–828.

Bryant W, Bean D, Slowey N, Dellapenna T, Scott E. 2007. Deepwater currents form mega-furrows near US Gulf’s Sigsbee Escarpment.Offshore 61(7): 94–98.

Canals M, Casamor JL, Urgeles R, Calafat AM, Domack EW, Baraza J, Farran M, De Batist M. 2002. Seafloor evidence of a subglacialsedimentary system off the northern Antarctic Peninsula. Geology 30(7): 603–606.

Canals M, Urgeles R, Calafat AM. 2000. Deep sea-floor evidence of past ice streams off the Antarctic Peninsula. Geology 28(1): 31–34.Caress DW, Chayes DN. 2001. Improved Management of Large Swath Mapping Datasets in MB-System Version 5, American Geophysical

Union, EOS Transactions, San Francisco; OS11B-0373.Chayes DN, Kurras G, Edwards M, Anderson RM, Coakley BJ. 1998. Swath mapping the Arctic Ocean from US Navy submarines;

installation and performance analysis of SCAMP operation during SCICEX 1998. EOS Transactions, American Geophysical Union79(45): F854.

Clark CD, Tulaczyk M, Stokes CR, Canals M. 2003. A groove-ploughing theory for the production of mega-scale glacial lineations, andimplications for ice-stream mechanics. Journal of Glaciology 49: 240–256.

Darby D, Jakobsson M, Polyak L. 2005. Icebreaker expedition collects key Arctic sea floor and ice data. EOS Transactions, AmericanGeophysical Union 86(52): 549–556.

Davies TA, Bell T, Cooper AK, Josenhans H, Polyak L, Solheim A, Stoker MS, Stravers JA. (eds). 1997. Glaciated Continental Margins: anAtlas of Acoustic Images. Chapman and Hall: London.

De Angeles H, Kleman J. 2005. Palaeo-ice streams in the northern Keewatin sector of the Laurentide Ice Sheet. Annals of Glaciology 42:135–144.

Dowdeswell JA, Bamber JL. 2007. Keel depths of modern Antarctic icebergs and implications for sea-floor scouring in the geologicalrecord. Marine Geology 243(1– 4): 120–131.

Dowdeswell JA, Ó Cofaigh C. 2002. Glacier-influenced sedimentation on high-latitude continental margins: introduction and overview.In Glacier-Influenced Sedimentation on High-Latitude Continental Margins, Dowdeswell JA, Ó Cofaigh C (eds). Geological Society:London; 1–9.

Dowdeswell JA, Ó Cofaig C, Pudsey CJ. 2004. Thickness and extent of the subglacial till layer beneath an Antarctic paleo-ice stream.Geology 32(1): 13–16.

Dowdeswell JA, Ottesen D, Rise L. 2006. Flow switching and large-scale deposition by ice streams draining former ice sheets. Geology34(4): 313–316.

Dowdeswell JA, Villinger H, Whittington RJ, Marienfeld P. 1993. Iceberg scouring in Scoresby Sund and on the East Greenland continentalshelf. Marine Geology 111: 37–53.

Dowdeswell JA, Whittington RJ. 1992. The sizes, frequencies and freeboards of East Greenland icebergs observed using ship radar andsextant. Journal of Geophysical Research 97: 3515–3528.

Duncan CS, Goff JA. 2001. Relict iceberg keel marks on the New Jersey outer shelf, southern Hudson apron. Geology 29(5): 411–414.Dyke AS, Andrews JT, Clark PU, England JH, Miller GH, Shaw J, Veillette JJ. 2002. The Laurentide and Innuitian ice sheets during the last

glacial maximum. Quaternary Science Reviews 21: 9–31.Edwards MH, Coakley BJ. 2003. SCICEX investigations of the Arctic Ocean System. Chemie der Erde 63(4): 281–328.Elverhøi A, Maisey G. 1983. Glacial erosion and morphology of the eastern and southeastern Weddell Sea shelf. In Fourth International

Symposium Antarctic Earth Science, Olivier RL, James PR, Jago JB (eds). Australian Academy of Science: Adelaide; 483–487.Engels JL, Edwards MH, Polyak L, Johnson PD. 2008. Seafloor evidence for ice shelf flow across the Alaska/Beaufort margin of the Arctic

Ocean. Earth Surface Processes and Landforms. DOI: 10.1002/esp1601Fader GBJ, Stea RR, Courtney RC. 1997. A seabed drumlin field on the inner Scotian Shelf, Canada. In Glaciated Continental Margins: an

Atlas of Acoustic Images, Davies TA et al. (eds). Chapman and Hall: London; 50–51.Fahnestock MA, Scambos TA, Bindschadler RA, Kvaran G. 2000. A millennium of variable ice flow recorded by the Ross Ice Shelf,

Antarctica. Journal of Glaciology 46(155): 652–664.Flood R. 1983. Classification of sedimentary furrows and a model for furrow initiation and evolution. Geological Society of America

Bulletin 94(5): 630–639.

10.1002/esp1601



Funder S, Hansen L. 1996. The Greenland ice sheet – a model for its culmination and decay during and after the last glacial maximum.Bulletin of the Geological Society of Denmark 42: 137–152.

Funder S, Kjær KH. 2007. Longterm 2006, the first APEX activity in Greenland and its continuation. First Conference on Arctic Palaeoclimateand its Extremes: Planning of International Polar Year Activities and Preliminary Results. Stockholm University, Royal Swedish Acad-emy of Sciences: Stockholm.

Gataullin VN, Polyak LV, Epstein OG, Romanyuk BF. 1993. Glacigenic deposits of the Central Deep: a key to the Late Quaternaryevolution of the eastern Barents Sea. Boreas 22: 47–58.

Gilbert R, Domack EW, Camerlenghi A (eds). 2003. Deglacial History of the Greenpeace Trough: Ice Sheet to Ice Shelf Transition in theNorthwestern Weddell Sea. Antarctic Peninsula Climate Variability 79. American Geophysical Union: Washington, DC.

Grantz A, May SD, Taylor PT, Lawver LA. 1990. Canada Basin. In The Arctic Ocean Region, Geology of North America, Grantz A, JohnsonGL, Sweeney JF (eds). Geological Society of America: Boulder, CO; 379–402.

Grosswald MG. 1980. Late Weichselian ice sheets of northern Eurasia. Quaternary Research 13: 1–32.Grosswald MG, Hughes TJ. 1999. The case for an ice shelf in the Pleistocene Arctic Ocean. Polar Geography 23: 23–54.Hall JK. 1990. Chukchi Borderland. In The Arctic Ocean Region. Geology of North America, Grantz A, Johnson GL, Sweeney JF (eds).

Geological Society of America: Boulder, CO; 337–350.Hättestrand C, Götz S, Näslund J-O, Fabel D, Stroeven AP. 2004. Drumlin formation time: evidence from Northern and Central Sweden.

Geografiska Annaler Series A: Physical Geography 86(2): 155–167.Hunkins K, Herron T, Kutschale H, Peter G. 1962. Geophysical studies of the Chukchi Cap, Arctic Ocean. Journal of Geophysical Research

67(1): 235–247.Jakobsson M. 1999. First high-resolution chirp sonar profiles from the central Arctic Ocean reveal erosion of Lomonsov Ridge sediments.

Marine Geology 158: 111–123.Jakobsson M. 2000. Mapping the Arctic Ocean: Bathymetry and Pleistocene Paleoceanography. Stockholm University: Stockholm.Jakobsson M. 2002. Hypsometry and volume of the Arctic Ocean and its constituent seas. Geochemistry, Geophysics, Geosystems 3(5): 1–

18.Jakobsson M. 2007. Lomonosov Ridge off Greenland 2007 (LOMROG): geophysical mapping and coring for the Arctic Ocean glacial and

sea ice history. First Conference on Arctic Palaeoclimate and its Extremes: Planning of International Polar Year Activities and Prelimi-nary Results. Stockholm University, Royal Swedish Academy of Sciences: Stockholm.

Jakobsson M, Cherkis N, Woodward J, Macnab R, Coakley B. 2000. New grid of Arctic bathymetry aids scientists and mapmakers. EOS,Transactions American Geophysical Union 81: 89, 93, 96.

Jakobsson M, Gardner JV, Vogt P, Mayer LA, Armstrong A, Backman J, Brennan R, Calder B, Hall JK, Kraft B. 2005. Multibeambathymetric and sediment profiler evidence for ice grounding on the Chukchi Borderland, Arctic Ocean. Quaternary Research 63: 150–160.

Jakobsson M, Løvlie R, Arnold EM, Backman J, Polyak L, Knutsen J-O, Musatov E. 2001. Pleistocene stratigraphy and paleoenvironmentalvariation from Lomonosov Ridge sediments, central Arctic Ocean. Global and Planetary Change 31(1– 4): 1–22.

Jakobsson M, Siegert M, Paton M. 2006. Is grounding of an ice shelf possible in the central Arctic Ocean? A modelling experiment. InFourth International Conference on Arctic Margins, ICAM IV, Scott RA, Thurston DK (eds). US Department of the Interior MineralManagement Services: Dartmouth, Nova Scotia; 141–156.

Jokat W, Weigelt E, Kristoffersen Y, Rasmussen T, Schöne T. 1995. New insights into the evolution of the Lomonsov Ridge and theEurasaian Basin. Geophysical Journal International 122: 378–392.

Josenhans HW, Zevenhuizen J. 1990. Dynamics of the Laurentide Ice Sheet in Hudson Bay Canada. Marine Geology 92: 1–26.Kleman J, Glasser NF. 2007. The subglacial thermal organisation (STO) of ice sheets. Quaternary Science Reviews 26: 585–597.Kristoffersen Y, Coakley B, Jokat W, Edwards M, Brekke H, Gjengedal J. 2004. Seabed erosion on the Lomonosov Ridge, central Arctic

Ocean: a tale of deep draft icebergs in the Eurasia Basin and the influence of Atlantic water inflow on iceberg motion? Paleoceanography,19: PA3006.

Kristoffersen Y, Mikkelsen N. 2006. On sediment deposition and nature of the plate boundary at the junction between the submarineLomonosov Ridge, Arctic Ocean and the continental margin of Arctic Canada/North Greenland. Marine Geology 225(1– 4): 265–278.

Kuijpers A, Dalhoff F, Brandt MP, Hümbs P, Schott T, Zotova Z. 2007. Giant iceberg plow marks at more than 1 km water depth offshoreWest Greenland. Marine Geology 246(1): 60–64.

Kuijpers A, Hansen B, Huehnerbach V, Larsen B, Nielsen T, Werner F. 2002. Norwegian Sea overflow through the Faroe–Shetland gatewayas documented by its bedforms. Marine Geology 188(1/2): 147–164.

Mayer LA. 2004. Cruise Report HE0405: U.S. Law of the Sea Cruise to Map the Foot of the Slope and 2500-m Isobath of the US ArcticOcean Margin. Center for Coastal and Ocean Mapping/Joint Hydrographic Center, University of New Hampshire: Durham, NH.

Mercer JH. 1970. A former ice sheet in the Arctic Ocean? Palaeogeography Palaeoclimatology Palaeoecology 8: 19–27.Meyer C, Anderson J, Skipp S. 2000. Iceberg furrow patterns in the Ross Sea: implications for the retreating ice margin. West Antarctic Ice

Sheet Initiative Seventh Annual Workshop Meeting, Sterling, VA.Moran K, Backman J, Brinkhuis H, Clemens SC, Cronin T, Dickens GR, Eynaud F, Gattacceca J, Jakobsson M, Jordan RW, Kaminski M,

King J, Koc N, Krylov A, Martinez N, Matthiessen J, McInroy D, Moore TC, Onodera J, O’Regan M, Pälike H, Rea B, Rio D, SakamotoT, Smith DC, Stein R, John KS, Suto I, Suzuki N, Takahashi K, Watanabe M, Yamamoto M, Farrell J, Frank M, Kubik P, Jokat W,Kristoffersen Y. 2006. The Cenozoic palaeoenvironment of the Arctic Ocean. Nature 441: 601–605.

Naryshkin G. 1999. Bottom Relief of the Arctic Ocean, bathymetric contour map. Head Department of Navigation and Oceanography andAll-Russia Research Institute for Geology and Mineral Resources of the World, Russian Academy of Sciences.



Ó Cofaigh C, Pudsey CJ, Dowdeswell JA, Morris P. 2002. Evolution of subglacial bedforms along a paleo-ice stream, Antarctic Peninsulacontinental shelf. Geophysical Research Letters 29: 1– 4.

Ottesen D, Dowdeswell JA. 2006. Assemblages of submarine landforms produced by tidewater glaciers in Svalbard. Journal of GeophysicalResearch 111(F01016).

Ottesen D, Dowdeswell JA, Rise L. 2005. Submarine landforms and the reconstruction of fast-fl owing ice streams within a large Quaternaryice sheet: the 2500-km-long Norwegian–Svalbard margin (57°–80° N). Geological Society of America Bulletin 117(7/8): 1033–1050.

Polyak L, Darby D, Bischof J, Jakobsson M. 2007. Stratigraphic constraints on late Pleistocene glacial erosion and deglaciation of theChukchi margin, Arctic Ocean. Quaternary Research 67: 234 –245.

Polyak L, Edwards MH, Coakley BJ, Jakobsson M. 2001. Ice shelves in the Pleistocene Arctic Ocean inferred from glaciogenic deep-seabedforms. Nature 410(6827): 453–459.

Polyak L, Edwards MHE, Jakobsson M, Coakley B. 1999. Glacial scouring in the deep Arctic Ocean: evidence from the Chukchi Plateauand the Lomonosov Ridge, American Geophysical Union Fall Meeting. EOS Transactions. American Geophysical Union: San Francisco.

Polyak L, Forman SL, Herlihy FA, Ivanov G, Krinitsky P. 1997. Late Weichselian deglacial history of the Svyataya (Saint) Anna Trough,northern Kara Sea, Arctic Russia. Marine Geology 143: 169–188.

Polyak L, Gataullin V, Gainanov V, Gladysh V, Goremykin Y. 2002. Kara Sea expedition yields insight into LGM ice sheet extent. EOSTransactions, American Geophysical Union 83(46): 525–529.

Poppe LJ, Ackerman SD, Doran EF, Moser MS, Stewart HF, Forfinski NA, Gardner UL, Keene JA. 2006. Geologic Interpretation andMultibeam Bathymetry of the Sea floor in Southeastern Long Island Sound, 2006-1059, US Geological Survey.

Ritz C, Peyaud V, Krinner G. 2007. Modelling the impact of ice shelves dynamics on the evolution of Northern Hemisphere ice sheets. FirstConference on Arctic Palaeoclimate and its Extremes: Planning of International Polar Year Activities and Preliminary Results. StockholmUniversity: Stockholm.

Rudels B, Jones EP, Anderson LG, Kattner G. 1994. On the intermediate depth waters of the Arctic Ocean. The Polar Oceans and TheirRole in Shaping the Global Environment. American Geophysical Union: Washington, DC; 33–46.

Schock SG, LeBlanc LR, Mayer LA. 1989. Chirp subbottom profiler for quantitative sediment analysis. Geophysics 54: 445–450.Shipp S, Anderson JB, Domack EW. 1999. Late Pleistocene–Holocene retreat of the West Antarctic ice-sheet system in the Ross Sea; Part 1,

Geophysical results. Geological Society of America Bulletin 111(10): 1468 –1516.Shipp SS, Wellner JS, Anderson JB. 2003. Retreat Significance of a Polar Ice Stream: Sub-Glacial Geomorphic Features and Sediments

from the Ross Sea, Antarctica, Geological Society of London Special Publication 203; 277–304.Siegert MJ, Dowdeswell JA. 2004. Numerical reconstructions of the Eurasian Ice Sheet and climate during the Late Weichselian. Quater-

nary Science Reviews 23(11–13): 1273–1283.Solheim A, Russwurm L, Elverhøi A, Nyland Berg M. 1990. Glacial Geomorphic Features: Direct Evidence for Grounded Ice in the

Northern Barents Sea and Implications for the Pattern of Deglaciation and Late Glacial Sedimentation, Geological Society of LondonSpecial Publication 53; 253–268.

Stokes CR, Clark CD. 1999. Geomorphological criteria for identifying Pleistocene ice streams. Annals of Glaciology 28: 67–74.Stokes C, Clark C, Darby D, Hodgson D. 2005. Late Pleistocene ice export events into the Arctic Ocean from the M’Clure Strait Ice Stream,

Canadian Arctic Archipelago. Global and Planetary Change 49: 139–162.Stokes CR, Clark CD, Winsborrow MCM. 2006. Subglacial bedform evidence for a major paleo-ice stream and its retreat phases in

Amundsen Gulf, Canadian Arctic Archipelago. Journal of Quaternary Science 21: 399–412.Svendsen JI, Alexanderson H, Astakhov VI, Demidov I, Dowdeswell JA, Henriksen M, Hjort C, Houmark-Nielsen M, Hubberten HW,

Ingólfson O, Jakobsson M, Kjær K, Larsen E, Lokrantz H, Lunkka JP, Lyså A, Mangerud J, Maslenikova O, Matioushkov A, Murray A,Möller P, Niessen F, Saarnisto M, Siegert C, Stein R, Siegert MJ, Spielhagen R. 2004. The late Quaternary ice sheet history of NorthernEurasia, Quaternary Science Reviews special QUEEN volume. Quaternary Science Reviews 23(11–13): 1229–1271.

Tchernia P, Jeannin PF. 1984. Circulation in Antarctic waters as revealed by iceberg tracks 1973–1983. Polar Research 22: 263–269.Todd BJ, Fader GBJ, Courtney RC, Pickrill RA. 1999. Quaternary geology and surficial sediment processes; Brown Bank, Scotian Shelf,

based on multibeam bathymetry. Marine Geology 162(1): 165–214.Tumskoy V, Basilyan A. 2007. Geological, permafrost and glacial interactions in Siberian Arctic: an idea for an new international project.

First Conference on Arctic Palaeoclimate and its Extremes: Planning of International Polar Year Activities and Preliminary Results.Stockholm University, Royal Swedish Academy of Sciences: Stockholm.

Vogt PR, Crane K, Sundvor E. 1994. Deep Pleistocene iceberg plowmarks on the Yermak Plateau: sidescan and 3·5 kHz evidence for thickcalving ice fronts and a possible marine ice sheet in the Arctic Ocean. Geology 22(5): 403–406.

Vogt PR, Taylor PT, Kovacs LC, Johnson GL. 1979. Detailed aeromagnetic investigation of the Arctic Basins. Journal of GeophysicalResearch 84(B3): 1071–1089.

Wellner JS, Heroy DC, Anderson JB. 2006. The death mask of the Antarctic ice sheet: comparison of glacial geomorphic features across thecontinental shelf. Geomorphology 75: 157–171.

Wellner JS, Lowe AL, Shipp SS, Anderson JB. 2001. Distribution of glacial geomorphic features on the Antarctic continental shelf andcorrelation with substrate: implications for ice behavior. Journal of Glaciology 47(158): 397– 411.

Wessel P, Smith WHF. 1991. Free software helps map and display data. EOS Transactions, American Geophysical Union 72(41): 441, 445–446.

Woodworth-Lynas CMT, Josenhans HW, Barrie JV, Lewis CFM, Parrott DR. 1991. The physical process of seabed disturbance duringiceberg grounding and scouring. Continental Shelf Research 11: 939–961.

Documents

Earth Surface Processes and Landforms 526 M. Jakobsson Earth …research.bpcrc.osu.edu/geo/publications/jakobsson_etal_ESPL08.pdf · Glacial geomorphology of the Central Arctic Ocean: