Evidence for an ice shelf covering the central Arctic Ocean during the penultimate glaciation

Abstract

The hypothesis of a km-thick ice shelf covering the entire Arctic Ocean during peak glacial conditions was proposed nearly half a century ago. Floating ice shelves preserve few direct traces after their disappearance, making reconstructions difficult. Seafloor imprints of ice shelves should, however, exist where ice grounded along their flow paths. Here we present new evidence of ice-shelf groundings on bathymetric highs in the central Arctic Ocean, resurrecting the concept of an ice shelf extending over the entire central Arctic Ocean during at least one previous ice age. New and previously mapped glacial landforms together reveal flow of a spatially coherent, in some regions >1-km thick, central Arctic Ocean ice shelf dated to marine isotope stage 6 (∼140 ka). Bathymetric highs were likely critical in the ice-shelf development by acting as pinning points where stabilizing ice rises formed, thereby providing sufficient back stress to allow ice shelf thickening.

Differences between the last two glacial maxima and implications for ice-sheet, δ18O, and sea-level reconstructions

“…If PGM and LGM deep-sea temperatures were similar (within about 0.5 C), then a fundamental PGMeLGM offset is implied in the relationships between mean global-ocean d sw and sea level or ice volume. This might arise from: (a) ocean circulation differences among glacial cycles, filling different volumes of the deep sea with waters of different mean d sw values, although such contrasts are thought to be averaged over multi-millennial periods (e.g., Siddall et al, 2010); (b) different moisture pathways feeding contrasting glacial configurations (e.g., Ullman et al, 2014;Colleoni et al, 2016), with impacts on atmospheric vapour isotopic fractionation and, hence, ice d 18 O; or (c) potential development of massive, largely floating Arctic ice shelves during certain glacials (e.g., PGM), and not during others (e.g., LGM) e these would cause negligible sea level change, but considerable mean ocean d 18 O change (Niessen et al, 2013;Jakobsson et al, 2016). Below, we discuss these three options in turn.…”

“…In the Arctic, erosional features have been found at depths of~1 km on the Lomonosov Ridge, Chukchi Borderlands, Yermak Plateau, East Siberian Margin, Baffin Bay, and Hovgaard Ridge (Fram Strait) (e.g., Polyak et al, 2001;Kuijpers et al, 2007;Dowdeswell et al, 2010;Gebhardt et al, 2011;Niessen et al, 2013;Arndt et al, 2014), while other portions of the Lomonosov and Mendeleev Ridges are largely devoid of glaciogenic features, which may suggest ice-free conditions . Bathymetric highs in the Arctic may have acted as pinning points, allowing ice-rise formation that stabilised and facilitated ice-shelf thickening (Vogt et al, 1994;Grosswald and Hughes, 1999;Jakobsson et al, 2016).…”

“…(Largely) floating Arctic ice shelves during glacial intervals were proposed in the 1970s and 1980s (Mercer, 1970;Hughes et al, 1977;Broecker, 1975;Grosswald, 1980;Denton and Hughes, 1981;Williams et al, 1981;Chappell and Shackleton, 1986), but then were overlooked due to difficulties in obtaining data from the region and a lack of direct evidence for such shelves during the LGM (for an overview, see Jakobsson et al, 2016). Recent geophysical mapping in the Arctic, however, has led to a re-evaluation of large floating Arctic ice shelves during the Pleistocene.…”

“…Age control for many of the features remains enigmatic, often relying on stratigraphic correlation and biostratigraphy; for example, identification of diagnostic MIS 5e nannofossils gives a likely, or minimum, MIS 6 age estimate for features on Morris Jesup Rise, Lomonosov Ridge, Yermak Plateau, Hovgaard Ridge, Mendeleev Ridge, and Arlis Plateau (Jakobsson, 1999;Polyak et al, 2001;Matthiessen and Knies, 2001;Kristoffersen et al, 2004;Spielhagen et al, 2004;Jakobsson et al, 2008Jakobsson et al, , 2010Jakobsson et al, , 2016Arndt et al, 2014;L€ owemark et al, 2016). Debate continues not only about the age of the submarine features (e.g., Flower, 1997;Niessen et al, 2013), but also about the scale of any ice shelves, from Arcticwide as proposed by Hughes et al (1977) and more recently by Jakobsson et al (2016), to (much) more limited extents (Engels et al, 2008;Jakobsson et al, 2010;Niessen et al, 2013;Stein et al, 2017).…”

“…The Lomonosov Ridge may have acted as a topographic barrier to Atlantic water circulation in the Amerasian Basin, possibly promoting ice-shelf growth in this region . Extensive MIS 6 ice-shelf/shelves covering the central Arctic (e.g., Hughes et al, 1977;Jakobsson et al, 2016) suggest partial contact with warmer Atlantic waters in cavities under the ice shelf, analogous to modern Antarctic ice shelves (e.g., Kirschner et al, 2013), with a potential for continued exchange of warmer waters across Lomonosov Ridge below the grounded ice shelf. Alternatively, Kristoffersen et al (2004) suggested that advection of warmer Atlantic waters to Lomonosov Ridge during MIS 6 was associated with surges or collapses of Saalian ice-sheets, which may have facilitated northward drift of deep-draft icebergs across Lomonosov Ridge from their discharge areas in the northern Barents-Kara region.…”