Quaternary sedimentation in the Arctic Ocean : Recent advances and further challenges

USAGE Permission is granted to copy this article for use in teaching and research. Republication, systematic reproduction, or collective redistribution of any portion of this article by photocopy machine, reposting, or other means is permitted only with the approval of The Oceanography Society. Send all correspondence to: info@tos.org or The Oceanography Society, PO Box 1931, Rockville, MD 20849-1931, USA. Oceanography THE OffICIAL MAGAzINE Of THE OCEANOGRAPHY SOCIETY

The photo shows a working moment during the 2007 lOMrOg (lomonosov ridge off greenland) expedition, where the scientific part was performed from the Swedish icebreaker Oden (at the photo's bottom), and breaking heavily ridged ice was aided by the russian nuclear-powered icebreaker 50 Years of Victory (on the left).More specifically, this moment was an attempt to salvage the seismic hydrophone streamer jammed in ice.
reCeNT adVaNCeS aNd FurTher ChalleNgeS (see Weber andRoots, 1990, andStein, 2008, for overviews).However, the mostly small volume and large spacing of sediment-core samples in these early works, combined with laboratory methods cruder than today's, resulted in often confusing results and a lack of stratigraphic coherency.
The new phase of Arctic Ocean research began in the 1980s with the regular use of icebreakers capable of collecting larger-diameter and longer (up to 10-15 m) sediment cores from locations selected based on geophysical mapping (see Stein, 2008, for an overview).The accumulation of this higher-quality core material, together with advances in chronostratigraphic and proxy-based methods, eventually led to the development of a new stratigraphy based on the apparent cyclicity of paleoclimate-related proxies and refinement of paleomagnetic data (Jakobsson et al., 2000).The new approach, largely constraining the last ca.250,000 years (Marine Isotope Stages [MIS] 1-7), became widely accepted with some variations by all research groups working with Arctic Ocean sediments (e.g., Backman et al., 2004;Spielhagen et al., 2004;Polyak et al., 2004).These Ocean glaciations (Vogt et al., 1994;Jakobsson, 1999;Polyak et al., 2001).(Moran et al., 2006;Backman and Moran, 2009).(Stein et al., 2010a,b) and a series of seafloor mapping cruises to the Chukchi Borderland (Mayer, 2003(Mayer, , 2004;;Mayer andArmstrong, 2007, 2008).

SediMeNT COre STr aTigr aphY aNd SediMeNTarY eNVirONMeNTS
Although age constraints for Arctic Ocean sediments are still tentative, especially for older strata, the general impression is that the age of sediment cores collected during HOTRAX, LOMROG, and comparable cruises generally does not extend to the base of the Quaternary (e.g., Polyak et al., 2009;Sellén et al., 2010;Stein et al., 2010a,b).This result suggests that the cores represent the that it was more than 50% smaller during peak glacial times (Jakobsson, 2002).
Such contrasting environments inevitably affected hydrographic, biotic, and sedimentary conditions in the ocean.
Sediment size and mineralogical composition suggest that interglacial sediments are primarily deposited by melt out from sea ice and may be partially redistributed by near-bottom currents (e.g., Darby et al., 2009;Polyak et al., 2009).Based on the upper portions of cores investigated with reasonably developed stratigraphy to estimated MIS 7, these sediments were deposited at low to moderate sedimentation rates, from several millimeters to 1-2 cm kyr -1 (Backman et al., 2004;Polyak et al., 2009;Stein et al., 2010b).Some of these intervals in the western Arctic, labeled PW (pink-white) layers in the earlier literature (e.g., Clark et al., 1980), have a characteristically high detrital carbonate content, mostly dolomites that can be traced to Canadian Shield rocks (Bischof et al., 1996;Phillips and Grantz, 2001;Polyak et al., 2009;Stein et al., 2010a,b;Yurco et al., 2010) Ice Sheet (e.g., Spielhagen et al., 2004;O'Regan et al., 2008).
Differences in sedimentation patterns between the eastern and western Arctic Ocean are not only reflected in sediment provenance and sedimentation rates but also result in different stratigraphies related to the histories of Eurasian and North American ice sheets.Notably, massive fluxes of sediment from Eurasian icebergs initiated at MIS 6, from ca. 130,000 to 190,000 years ago (Jakobsson et al., 2001;Spielhagen et al., 2004;O'Regan et al., 2008O'Regan et al., , 2010)), when the united Barents-Kara Ice Sheet expanded to the shelf break and began to shed large volumes of icebergs directly into the Arctic Ocean basin (Svendsen et al., 2004).Although the age control for older sediments is still provisional, it is clear that in the western Arctic Ocean, a comparable influx of coarse sediment, originating in this case from North American ice sheets, started several hundred thousand years earlier than MIS 6, possibly at or soon after MIS 16 (Figure 3; Polyak et al., 2009;Stein et al., 2010a,b).This difference indicates that the high sea ice concentration inferred from the low content of coarse icerafted debris in the Middle Pleistocene The existence of major sedimentological events in Arctic Ocean history, combined with long-range transport by sea ice and icebergs, allows for a correlation of sediment cores across large areas of the seafloor using easy-tomeasure proxies such as bulk density and magnetic susceptibility logs (e.g., Sellén et al., 2010).This approach facilitates extensive spatial reconstruction of paleoceanographic and related paleoclimatic conditions using a limited number of reference cores with well-developed age control.However, caution must be taken  (Spielhagen et al., 2004;Polyak et al., 2009;Stein et al., 2010b).Figure 4 illustrates the distribution of detrital carbonates and estimated average Middle-Late Quaternary sedimentation rates based on these correlations, with new data added from the LOMROG area (Hanslik, 2011).We note that although age estimates are mostly tentative, especially for the older parts of the stratigraphy, they do not affect the relative spatial difference in sedimentation rates.; adler et al., 2009; polyak et al., 2009) exemplifying the Trans-polar drift and Beaufort gyre circulation systems, respectively (Figure 1).lr cores show an abrupt increase of coarse sediment from the Barents-kara ice Sheet at Marine isotope Stage (MiS) 6 (~ 130,000-190,000 years ago), whereas Mr and other western arctic Ocean cores show a much earlier increase in coarse sediment from the North american ice sheets (tentatively bracketed between MiS 12 and 16, ~ 0.5-0.7 million years ago).The latter increase is marked notably by high concentrations of calcium primarily from detrital carbonates, mostly dolomites of Canadian Shield provenance (see also Stein et al., 2010a,b;Yurco et al., 2010).Yellow fill on the Mr graph shows slump within MiS 6. able detail (e.g., Mayer andArmstrong, 2007, 2008;Dowdeswell et al., 2010;Jakobsson et al., 2010b;Stein et al., 2010b).The continuing growth of these data combined with earlier, mostly opportunity-based mapping efforts (e.g., Edwards and Coakley, 2003) provides an invaluable asset for compre-  These features indicate that expansive ice shelves and/or ice rises capped a large part or the entirety of the basin during some of the Pleistocene glaciations (Vogt et al., 1994;Jakobsson, 1999;Polyak et al., 2001Polyak et al., , 2007;;Jakobsson et al., 2005Jakobsson et al., , 2008cJakobsson et al., , 2010b;;Mayer andArmstrong, 2007, 2008;Engels et al., 2008;Dowdeswell et al., 2010).This extensive ice cover was first suggested for the central Arctic Ocean, when Oden mapped a portion of the Lomonosov Ridge with a towed CHIRP subbottom profiler in 1996 (Jakobsson, 1999).and Coakley, 2003).Erosion was found to be widespread; eroded sediment had been displaced onto the Amerasian flank of the ridge, and iceberg plowmarks and glacigenic lineations were associated with the eroded surface.
Similar data were also collected on this cruise from the Chukchi Borderland accompanied by armadas of megabergs (Polyak et al., 2001).
During the HOTRAX´05 cruise, Healy collected multibeam imagery that confirmed and detailed earlier findings of glacigenic bedforms on the Chukchi Borderland (Jakobsson et al., 2008c).
Glacial landforms on Yermak Plateau as well as ice scours on Morris Jessup Rise were observed previously (Vogt et al., 1994;Kristoffersen et al., 2004;Spielhagen et al., 2004), but the new multibeam bathymetry provided a more comprehensive picture and, in combination with retrieved sediment cores, enabled stratigraphic evaluation of the timing of the glacial impact.Geophysical data combined with stratigraphically constrained cores from HOTRAX, LOMROG, and other expeditions (Vogt et al., 1994;Polyak et al., 2001Polyak et al., , 2007;;Jakobsson et al. 2001Jakobsson et al. , 2005Jakobsson et al. , 2008c;;Engels et al., 2008;Stein et al., 2010b) suggest the possibility of multiple deepdraft erosional events during several glacial intervals.An extensive marine ice sheet complex, probably including ice shelves and ice rises, existed in the Arctic Ocean during MIS 6 at least in the Amerasia Basin (Jakobsson et al., 2010b).
This conclusion is consistent with reconstruction of a supersized Barents-Kara Ice Sheet that likely extended all the way to the shelf edge along the entire Barents-Kara and part of the Laptev Sea margin during MIS 6 (Svendsen et al., 2004).The Arctic margin of North American ice sheets predating the Last Glacial Maximum cannot be well constrained from terrestrial studies; more research is needed on the glacial history of the Chukchi Borderland and adjacent North American margin, where sediment stratigraphy indicates at least three episodes of glacial erosion (Polyak et al., 2007;Engels et al., 2008;Jakobsson et al., 2010b).

SuMMarY
This review of recent icebreaker-based, paleochanges in sea ice (e.g., Belt et al., 2007;Műller et al., 2009) and adjacent terrestrial environments (Cooke et al., 2009).There is also much need for T h e C h a N g i N g a r C T i C O C e a N | S p e C i a l i S S u e O N T h e i N T e r N aT i O N a l p O l a r Y e a r ( 2 0 0 7 -2 0 0 9) Quaternary Sedimentation in the arctic Ocean B Y l e O N i d p O lYa k a N d M a r T i N J a k O B S S O N Oceanography | Vol.24, No.3 52 past fundamental processes such as ocean circulation, water exchange between ocean basins, biological production, and the marine cryosphere (primarily sea ice).The Arctic Ocean sedimentary archive holds the long-time perspective on sea ice evolution, a component required to understand dramatic Arctic sea ice retreat over the last decades (Figure 1) and projection of its future change (see Polyak et al., 2010, and Jakobsson et al., 2010a, for reviews).However, the history of sedimentation in the Arctic Ocean and even its modern sedimentary processes and patterns are only fragmentarily understood.This limited knowledge is due to a combination of difficulties that include collecting sediment cores and mapping the seafloor in a still perennially ice-bound ocean, and complications with interpreting Arctic sedimentary records, which seem to be intrinsically related to the presence of sea ice cover.Stratigraphic and sedimentological investigation of Arctic Ocean seafloor sediments began more than 50 years ago when multiple, small-diameter and fairly short sediment cores started to be collected throughout the basin in the 1950s to 1970s-initially from Soviet drifting ice camps and then from their US and Canadian counterparts These data highlighted the importance of combining more thorough Arctic seafloor mapping with sediment coring in key locations.This paper presents the major results aBSTr aCT.This paper reviews current knowledge of sedimentation patterns in the Arctic Ocean during the pronounced climatic cycles of the last several hundred thousand years, an especially relevant time period that provides long-term context for present climate change.The review is largely based on data collected during recent research icebreaker cruises to the Arctic Ocean, with a focus on the 2005 Healy-Oden TransArctic Expedition (HOTRAX) and 2007 Lomonosov Ridge Off Greenland (LOMROG) expedition.The sediment cores and geophysical seafloor mapping data collected enable reconstruction of past oceanic environments.Evaluation of these data suggests that the two major Arctic Ocean circulation systems, the Trans-Polar Drift and the Beaufort Gyre, persisted throughout most of the Late to Middle Quaternary, approximately the last 0.5 to 0.7 million years.Extreme conditions, nonanalogous to modern environments, also occurred in the past, especially during Pleistocene glacial intervals.Some of these intervals likely featured much thickened and/or concentrated sea ice and incursions of ice shelves and armadas of megasized icebergs from the margins to the center of the Arctic Ocean.In contrast, much warmer conditions with reduced sea ice extent existed during interglacial periods.Characterization of ice conditions during these intervals is critical for evaluating the present and projected future reduction of Arctic sea ice. of the most recent collection of sediment cores and seafloor mapping data from the Arctic Ocean that have not been covered by earlier review papers.The focus of this review is on sediment-core and geomorphic data that help comprehend Quaternary sedimentary environments pertinent to understanding the trajectory and consequences of the present Arctic change.reCeNT aChieVeMeNTS During the last few years, concerns about the abruptness of climate change in the Arctic stimulated advances in the collection of Arctic Ocean sediments (Figure 1).The first development was the ability to conduct scientific drilling in the central Arctic with sea ice present.The Integrated Ocean Drilling Program's Arctic Coring Expedition (ACEX; Backman et al., 2006) recovered the first long Cenozoic sedimentary sequence from the central Arctic Ocean.The recovered core greatly expanded our understanding of Earth's long-term climate evolution and also the Arctic's tectonic, paleogeographic, and climatic settings going back an estimated 56 million years of the US Coast Guard cutter Healy and the Swedish icebreaker Oden (Figure 2).Twenty-one large-diameter cores up to 15 m in length and accompanying multicore sediment samples and geophysical records were collected from a number of morphological structures across the Arctic Ocean (Figure 1).The focus was to collect data on submarine ridges and plateaus, where sediments are only minimally subjected to redeposition by downslope processes.In addition, eight long cores were raised from a higher-sedimentation continental shelf and slope setting at the Chukchi Alaskan margin.It is important that HOTRAX cores were collected across a wide range of modern sea ice conditions, from nearly open water to heavily ice-covered areas with climatological ice concentrations of almost 90% (Figure 2).This geographic coverage makes the collection especially valuable for reconstructing past ice conditions.LOMROG 2007 was the first scientific icebreaker expedition to reach the virtually unexplored part of the Arctic Ocean north of Greenland (Figure 1).This area is characterized by the most severe sea ice conditions in the entire Arctic and is projected to be the last to become seasonally ice free in a continued global warming scenario (Wang and Overland, 2009).Except for LOMROG, marine geological, geophysical, and oceanographic research north of Greenland has only been carried out from ice camps established on drifting sea ice, with limited capabilities for collection of sediment cores and geophysical seafloor data (e.g., Nørgaard-Pedersen et al., 2007).LOMROG was a two-ship operation, with the Russian nuclear icebreaker Fifty Years of Victory assisting Oden in taking sediment cores and mapping the seafloor with its hull-mounted multibeam echosounder and CHIRP sonar subbottom profiler along the expedition route (Figure 2).Ten cores were raised from the southern Lomonosov Ridge and Morris Jessup Rise at sites selected following interpretation of newly acquired geophysical mapping data (Figure 1).The sediment cores retrieved during LOMROG complement and significantly expand the HOTRAX collection.Most important are coring locations with the potential to reveal Arctic Ocean sea ice dynamics.If studies of the LOMROG cores indicate past periods of seasonally ice-free waters north of Greenland, this will indicate the possibility of mostly open-water conditions for the entire Arctic Ocean.
overall frigid, but vigorously fluctuating climate of the last couple of million years, with large ice sheets growing and waning at the Arctic periphery on multimillennial (Milankovitch) time scales (see Fitzpatrick et al., 2010, and companion papers for an overview).Sea level falls accompanying glaciations repeatedly turned the Arctic Ocean into a much smaller basin, with broad and shallow continental margins exposed or covered by ice sheets, riverine fluxes diminished, and connections to other oceans limited to Fram Strait alone (Figure 1).In fact, the wide and relatively shallow Arctic shelves occupy more than half of the Arctic Ocean area, implying
Deglacial intervals stand out by high content of sand and coarser sediment indicative of deposition from icebergs, with elevated sedimentation rates of several centimeters per thousand years.
. These layers likely indicate catastrophic discharges of icebergs from the Laurentide Ice Sheet, similar to Heinrich events in the North Atlantic, and make useful stratigraphic markers for core correlation across the western Arctic.In the area north of Greenland, some of these layers have a peculiar high-magnesium calcite composition, probably reflecting contributions from proximal sources on Ellesmere Island and Greenland (Nørgaard-Pedersen et al., 2007).A much bigger difference is displayed by iceberg events in the eastern Arctic Ocean (Eurasia Basin including the Eurasian and central segments of the Lomonosov Ridge), which primarily carry material from the Barents-Kara

(
pre-MIS6) section of ACEX (O'Regan et al., 2010) may be characteristic of the eastern, but not the western, Arctic Ocean.It is notable that the sharp rise in sediment content of Laurentide provenance at or near the onset of the Middle Pleistocene is consistent with the broadly accepted view that the Mid-Pleistocene Transition in Earth's response to orbital variability (Milankovitch cycles) was related to the growth in the volume of the Laurentide Ice Sheet (Clark et al., 2006, and references therein).
lines of evidence from independent proxies.More robust correlations are enabled by the combined use of textural, geochemical, paleobiological, and paleomagnetic data, including unique events such as pronounced swings in paleomagnetic inclination, specific foraminiferal assemblages, and a distinct increase in detrital carbonates with increasing Laurentide fluxes

Figure 3 .
Figure 3.Comparison of sample sediment records from the central lomonosov ridge (lr; from O'reganet al., 2010)  and Mendeleev ridge(Mr; adler et al., 2009;polyak et al., 2009) exemplifying the Trans-polar drift and Beaufort gyre circulation systems, respectively (Figure1).lr cores show an abrupt increase of coarse sediment from the Barents-kara ice Sheet at Marine isotope Stage (MiS) 6 (~ 130,000-190,000 years ago), whereas Mr and other western arctic Ocean cores show a much earlier increase in coarse sediment from the North american ice sheets (tentatively bracketed between MiS 12 and 16, ~ 0.5-0.7 million years ago).The latter increase is marked notably by high concentrations of calcium primarily from detrital carbonates, mostly dolomites of Canadian Shield provenance (see alsoStein et al.,  2010a,b;Yurco et al., 2010).Yellow fill on the Mr graph shows slump within MiS 6.
the eastern and western parts of the Arctic Ocean due to principal differences in ocean circulation and sediment provenance, further complicated by severe fossil dissolution in Eurasia Basin sediments.This difference necessitates the development of robust independent age controls for reference records from both basins, a challenging task due to various chronostratigraphic complications in Arctic environments (e.g., Polyak et al., 2009, for an overview) and a virtual lack of long stratigraphic records other than ACEX.l arge-SCale SediMeNTarY paT TerNS deriVed FrOM SeaFlOOr MappiNg While sediment cores are undoubtedly critical for proxy-based paleoceanographic studies, the usefulness of a core is limited without knowledge of seafloor processes influencing its sedimentary environments on both local and regional scales.Geophysical mapping, including multibeam swath bathymetry and highresolution subbottom profiling, will, under most circumstances, provide the much needed spatial context for a sediment core record.For example, geophysical mapping may reveal patterns of sediment erosion and redeposition by currents, glacial processes, or mass wasting-features that are often difficult to identify in a sediment core.Major research icebreakers that currently operate in the Arctic Ocean, such as Polarstern, Healy, and Oden, are each equipped with a modern multibeam echosounder and a subbottom profiler capable of mapping the seafloor morphology and uppermost ~ 30-100 m of sediment stratigraphy in consider- hending the history of the Arctic Ocean and related climatic changes.The major types of geomorphic forms on the ocean floor (bedforms) are usually related to erosional and depositional activities of downslope mass-wasting processes and bottom currents (e.g., Figure 5a).In addition, polar areas both in the Arctic and around the Antarctic feature numerous bedforms generated by deep-draft ice.These features include iceberg scours (plowmarks) and glacial sole markings, primarily flutes or megascale lineations, but sometimes also other glacigenic forms such as drumlins and morainic

Figure 4 .
Figure4.geographic distribution of detrital carbonates in sediment cores (violet semitransparent fill) and estimated average long-term (Middle-late Quaternary) sedimentation rates in the western arctic Ocean (cm kyr -1 , white lines).Carbonates in the amerasia Basin sediments are mostly composed of dolomite (pink semitransparent fill for major provenance), although north of greenland they also have a significant high-magnesium calcite component.Sedimentation rates are based on age estimates at least to the bottom of MiS 7 (ca.250,000 years ago) and, where recovered, to the initial peak of high laurentide sediment inputs (Figure3).Core sites are shown in red (hOTraX), violet (lOMrOg), yellow (Polarstern:Stein et al., 2010b), and grey (other collections).dotted white lines show reconstructed maximal extent of pleistocene ice sheets(dyke et al., 2002;Svendsen et al., 2004); white arrows indicate major ice streams at the northern laurentide margin.Mr, Nr, ar, and lr are for Mendeleev, Northwind, alpha, and lomonosov ridges, respectively.
Although difficult ice conditions only allowed towing the fragile CHIRP over short stretches, the acquired subbottom profiles revealed a pronounced erosional surface on the ridge crest close to the North Pole.More comprehensive evidence of this erosion was provided in 1999 when the US nuclear submarine USS Hawkbill systematically mapped this section of the Lomonosov Ridge using both a CHIRP and a side-scan sonar for detailed imaging of the seafloor surface (SCICEX program; Edwards

Figure 5 .
Figure 5. Multibeam sonar images and bathymetric cross sections of megascale seafloor geomorphic features: (a) mud waves in the alpha ridge area, (b) iceberg scours on lomonosov ridge off greenland, and (c) glacial lineations on Yermak plateau (see Figure 1 for locations).Note differences between the three bedform types in both areal and bathymetric cross-section patterns.
which are limited to bathymetric highs, bedforms related to current activities occur over a much broader and deeper depth range.Remarkable mud waves were discovered on Alpha Ridge at four distinctly separate areas along a stretch of 180 km of the HOTRAX track in water depths from 1,900 to 2,300 m (Figures 1, 5).In one area, the mud waves show a pattern of two generations interfering at an oblique angle.Wave heights are between 10 and 50 m, and wave lengths (crest to crest) are between 300 and 1000 m.Buried sediment waves with similar dimensions are known from seismic reflection studies in this area(Hall, 1970), but HOTRAX multibeam data demonstrated for the first time that these features occur on the seafloor surface.Due to the limited data set, notably a lack of high-resolution seismostratigraphy, the origin and age of these mud waves is difficult to constrain, which has given ground to speculative interpretations such as a relationship to a shock wave from an asteroid impact(Kristoffersen et al., 2008).Sediment cores collected in the mud wave area do not show evident signs of erosional events, indicating that they either predated the sediments recovered or were relatively slow and had only a minimal impact on depositional processes.Contour or turbidity currents, the most common agents of deep-sea sediment wave formation(Wynn and Stow, 2002), are not likely for the middle of the Amerasia Basin.The intersecting pattern of wave generations suggests that they may be related to basin-scale tidal processes.Deepwater currents induced from internal tides and internal waves are known to be capable of forming large, kilometer-scale mudwaves on the seafloor(He et al., 2008).Recent observations indicate the likelihood of vertical motions with amplitudes of 10-20 m in the deep Amerasia Basin (Timmermans complex, geological/geophysical expeditions to the Arctic Ocean, with a focus on HOTRAX'05 and LOMROG'07, shows that the data collected significantly expand our understanding of sedimentary and related paleoceanographic/ paleoclimatic processes during the Quaternary (estimated age limit for sediment cores retrieved).This time interval is especially important for evaluating natural climatic variability, sensitivity, and feedbacks-critical knowledge for understanding present climate change and projecting its future course.Two principal types of data collected are sediment cores and geophysical mapping of the seafloor, including multibeam bathymetry and subbottom profiling.Sediment core records are critical for reconstructing paleoenvironments at a specific location and constraining their ages.Geophysical mapping provides spatial context and can be especially helpful where sedimentation is strongly variable and includes erosional or nondepositional events.Such settings are common in polar areas due to the effects of the marine cryosphere (sea ice, icebergs, and marine portions of ice sheets), including potential blockage of sediment delivery by an ice canopy and direct impact of deep-draft ice on the seafloor.One principal conclusion from available sediment core data is that the Trans-Polar Drift and the Beaufort Gyre were likely the robust major features of Arctic Ocean circulation throughout the time period covered, estimated to extend to at least the beginning of the Middle Quaternary (ca.650,000 years ago), possibly excepting glacial maxima.This circulation system resulted in considerably different sedimentary patterns in the eastern and western parts of the Arctic Ocean (roughly corresponding to the Eurasia and Amerasia Basins, respectively).A practical implication of this conclusion is that comprehensive stratigraphies and paleoceanographic histories need to be developed for both basins as no single site can represent the entire Arctic Ocean.Meanwhile, to date, only one long paleoceanographic record (ca.56 million years, with discontinuities) has been recovered from the central Arctic Ocean.Despite the inferred general longterm stability of Arctic Ocean circulation, large deviations from historically observed conditions occurred repeatedly during Quaternary climatic fluctuations, driven by Earth's orbital cyclicity.Arctic environments were especially extreme during the major ice ages that were likely characterized not only by much thicker and more solid sea ice cover but also by invasions of vast Pleistocene ice sheets from the margins into the central part of the Arctic Ocean.Although the exact patterns and timing of these ice sheet expansions into the oceanic realm have yet to be investigated, available data reveal erosion and sculpting of submarine ridges and plateaus by ice throughout the ocean at modern water depths reaching ~ 1,000 m.These events were succeeded by episodes of ice sheet disintegration, likely sometimes abrupt, when armadas of icebergs with drafts reaching several hundred meters were ubiquitous in the Arctic Ocean.Icebergs were exceptionally potent agents of sediment delivery to remote corners of the ocean from the areas eroded by ice sheets.This process is exemplified by the dispersal of detrital carbonates from the hinterland of the Canadian Shield into and throughout the western Arctic Ocean.During some interglacial periods, climatic conditions were warmer than at present.These intervals are not yet sufficiently investigated, but some data, such as the abundance of paleobiota, especially the presence of subarctic species, indicate the possibility of considerably reduced sea ice cover.Both cold and warm time extreme events encompassed nonanalog conditions, which must be considered in paleoclimatic modeling.Another aspect of Arctic paleoceanography that requires in-depth investigation is distinguishing impacts of sea ice from those of ice sheets in sedimentary records.Arctic paleoceanography problems reviewed here help define the priorities and strategies of future data collection and research.One obvious, although very challenging, goal is to drill more long boreholes to characterize the long-time history of the Arctic Ocean.Another objective is to increase the coverage of multibeam seafloor mapping, especially in key areas, such as the sites of ice-shelf grounding, in order to better understand the ocean's geological and paleoclimatic history.Development and refinement of paleoenvironmental proxies is also critical for characterizing past conditions, especially those that can be applied to evaluation of modern climate change.For example, the growing field of biomarker research shows potential for evaluating Adler, R.E., L. Polyak, J.D. Ortiz, D.S. Kaufman, J.E.T. Channell, C. Xuan, A.G. Grottoli, E. Sellen, and K.A. Crawford.2009.Sediment record from the western Arctic Ocean with an improved Late Quaternary age resolution: HOTRAX core HLY0503-8JPC, Mendeleev Ridge.Global and Planetary Change 68:18-29, http://dx.doi.org/10.1016/j.gloplacha.2009.03.026.Anderson, J.B., S.S. Shipp, A.L. Lowe, J.S. Wellner, and A.B. Mosola.2002.The Antarctic Ice Sheet during the Last Glacial Maximum and its subsequent retreat history: A review.Quaternary Science Reviews 21:49-70, http://dx.doi.org/10.1016/S0277-3791(01)00083-X. reFereNCeS