A radiocarbon-dated marine sediment core retrieved in Kane Basin, central
Nares Strait, was analysed to constrain the timing of the postglacial opening
of this Arctic gateway and its Holocene evolution. This study is based on a
set of sedimentological and geochemical proxies of changing sedimentary
processes and sources that provide new insight into the evolution of ice
sheet configuration in Nares Strait. Proglacial marine sedimentation at the
core site initiated ca. 9.0 cal ka BP following the retreat of
grounded ice. Varying contributions of sand and clasts suggest unstable sea
ice conditions and glacial activity, which subsisted until ca. 7.5 cal ka BP under the combined influence of warm atmospheric temperatures and
proglacial cooling induced by the nearby Innuitian (IIS) and Greenland (GIS)
ice sheets. An interval rich in ice-rafted debris (IRD) is interpreted as the collapse of the ice
saddle in Kennedy Channel ca. 8.3 cal ka BP that marks the complete
opening of Nares Strait and the initial connection between the Lincoln Sea
and northernmost Baffin Bay. Delivery of sediment by icebergs was
strengthened between ca. 8.3 and ca. 7.5 cal ka BP
following the collapse of the buttress of glacial ice in Kennedy Channel that
triggered the acceleration of GIS and IIS fluxes toward Nares Strait. The
destabilisation in glacial ice eventually led to the rapid retreat of the GIS
in eastern Kane Basin at about 8.1 cal ka BP as evidenced by a noticeable
change in sediment geochemistry in our core. The gradual decrease in
carbonate inputs to Kane Basin between
The Holocene history of Nares Strait, Northwest Greenland, has remained somewhat cryptic despite investigations during the past 4 decades (e.g. Blake Jr., 1979; Kelly and Bennike, 1992; Mudie et al., 2004; Jennings et al., 2011.). Nares Strait is a key gateway for Arctic seawater and ice toward the Atlantic Ocean, contributing to up to half of the volume of water transported through the Canadian Arctic Archipelago (CAA), which provides fresh water to the Labrador Sea and influences deep water formation (Belkin et al., 1998; Münchow et al., 2006; McGeehan and Maslowski, 2012). Nares Strait supplies one of the most productive regions of the Arctic, the North Water Polynya (NOW), with nutrient-rich Pacific water (Jones et al., 2003; Jones and Eert, 2004) and maintains its very existence by trapping sea and calved glacial ice in ice arches in the north and south of the strait (Melling et al., 2001; Mundy and Barber, 2001).
Despite the importance of Nares Strait, intrinsic investigations into its
late Pleistocene history, which is intimately linked with the dynamics of
the bordering Innuitian (IIS) and Greenland (GIS) ice sheets, are relatively
sparse and much of the knowledge relies on land-based studies. Debate
initially surrounded early studies into glacial configuration in the CAA
with some authors concluding that the CAA channels were not blocked during
the Last Glacial Maximum (LGM) (Franklin Ice Complex theory; e.g. England,
1976), while others argued that the IIS coalesced with the bordering
Greenland and Laurentide ice sheets (e.g. Blake Jr., 1970). The presence of
erratic boulders originating from Greenland on Ellesmere Island (England,
1999), cosmogenic nuclide surface-exposure dating (Zreda et al., 1999), and
radiocarbon dating on mollusc shells (e.g. Bennike et al., 1987; Blake Jr. et
al., 1992; Kelly and Bennike, 1992) finally settled the argument in favour
of the latter narrative by supporting the coalescence of the IIS and GIS
along Nares Strait between 19 and ca. 8
Here we present sedimentological, geochemical, and geochronological data obtained from a 4.25 m long marine sediment core (AMD14-Kane2b) retrieved in Kane Basin, central Nares Strait. This core provides a continuous sedimentary record spanning the last 9.0 kyr, i.e. from the inception of the Early Holocene retreat of the GIS and IIS in Nares Strait to modern times. Our set of sedimentological and geochemical records derived from this study presents the first offshore evidence of an ice-free environment in Kane Basin in the Early Holocene and offers a unique opportunity to explore the local dynamics of ice sheet retreat leading to the opening of the strait and the establishment of the modern oceanographic circulation pattern.
Schematic circulation in the Canadian and northern Greenland
sectors of the Arctic Ocean
Nares Strait is a long (530 km) and narrow channel separating Northwest
Greenland from Ellesmere Island, Arctic Canada, connecting the Arctic Ocean
to the Atlantic Ocean in Baffin Bay (Fig. 1). Kane Basin is the central,
wide (120 km large at its broadest point, totalling an area of approximately
27 000 km
The oceanographic circulation in Nares Strait consists of a generally
southward-flowing current driven by the barotropic gradient between the
Lincoln Sea and Baffin Bay (Kliem and Greenberg, 2003; Münchow et al.,
2006), while the baroclinic temperature balance generates strong, northerly
winds that affect surface layers (Samelson and Barbour, 2008; Münchow
et al., 2007; Rabe et al., 2012). The relative influence of the barotropic
vs. baroclinic factors that control the currents in Nares Strait is highly
dependent on the presence of sea ice that inhibits wind stress when landfast
(Rabe et al., 2012; Münchow, 2016). Long-term ADCP measurements of flow
velocity record average speeds of 20–30 cm s
Sea ice concentration in Nares Strait is usually over 80 % from September to June (Barber et al., 2001). The state of the ice varies between mobile (July to November) and fast ice (November to June). The unique morphology of the strait leads to the formation of ice arches in Nares Strait when sea ice becomes landfast in the winter. The ice arches are a salient feature in the local and regional oceanography of Nares Strait: they not only block sea ice from drifting southward in the strait, sustaining the existence of the NOW Polynya (Barber et al., 2001), but they also control the export of low-salinity Arctic water into Baffin Bay (Münchow, 2016). The main iceberg sources for the strait are Petermann Glacier in Hall Basin and Humboldt Glacier in Kane Basin, both outlets of the GIS.
Geology of Northwest Greenland and Ellesmere Island along Nares Strait. Adapted from Harrison et al. (2011).
The Greenland coast bordering Kane Basin is relatively flat. In Inglefield Land the Precambrian basement is exposed, displaying supracrustal crystalline rocks and metamorphic rocks, essentially reported as aluminous metasediments and gneisses or granitoid gneisses, with some references to quartzite (Fig. 2, Koch, 1933; Dawes, 1976; Dawes and Garde, 2004; Harrisonet al., 2006, and references therein). Dawes and Garde (2004) postulated that this Precambrian basement also underlies the 100 km wide Humboldt Glacier, a claim that is supported by the dominance of crystalline material delivered in modern glacimarine sediments in front of the Humboldt Glacier (Fig. 2, Kravitz, 1976). To the north, the Precambrian basement in Washington Land is overlaid by Cambrian, Ordovician, and Silurian dolomites, limestones, and evaporites (Koch, 1929a, b; Harrison et al., 2006, and references therein). The Ellesmere shore of Kane Basin rises abruptly from sea level and is punctured by narrow fjords, penetrating inland for nearly 100 km (Kravitz, 1982). In southern Kane Basin, the same Precambrian crystalline rocks outcrop to form the Ellesmere–Inglefield Precambrian Belt. The central and northern sectors of Ellesmere Island's coast mainly comprise Cambrian to Devonian carbonates and evaporates. Fluviodeltaic quartz sandstone, volcanistic sandstone, minor arkose, and sometimes coal are found in the Paleogene Eureka Sound sequence that occurs along the western coast of Kane Basin, on the Ellesmere Island flank of Kennedy Channel, and on Judge Daly Promontory (Christie, 1964, 1973; Kerr, 1967, 1968; Miall, 1982; Oakey and Damaske, 2004). Coal-bearing Paleogene clastics also occur along the coast of Bache Peninsula and in morainic deposits on Johan Peninsula in southwestern Kane Basin (Fig. 2, Kalkreuth et al., 1993).
Kravitz (1976) described modern sedimentation in Kane Basin according to three main provinces defined on the basis of mineralogical and grain size characteristics. The first province covers the eastern, central, and southern part of the basin in which the predominant crystalline clay and silt sediments are water-transported off Humboldt Glacier and Inglefield Land. The second province, in the west of the basin, includes a higher fraction of ice-transported materials, mostly carbonates with clastic debris occurring in the deeper trough. Northern Kane Basin makes up the third province in which water-transported, mostly carbonate sediments from Washington Land are deposited in its northernmost part, while ice-transported crystalline particles are more common in the southern part of this province.
Sediment core AMD14-Kane2b was retrieved at 217 m of water depth in Kane Basin,
Nares Strait (79
The description of the various lithofacies was based on the visual
description of the core and high-resolution images using a computed
tomography (CT) scanner (Siemens SOMATOM Definition AS
High-resolution (0.5 cm) X-ray fluorescence (XRF) scanning was conducted
along the archive using an Avaatech XRF core scanner. The semi-quantitative
elemental composition of the sediment was measured throughout the whole
archive with the exception of two units, which contain large clasts.
Measurements were acquired with generator settings of 10, 30, and 50 kV in
order to detect elements in the range of Al to Ba. Elemental ratios or
normalisation to the sum of all elements except Rh and Ag, whose counts are
biased during data acquisition, were used to minimise the effects of grain
size and water content on elemental counts (Tjallingii et al., 2007; Weltje and Tjallingii, 2008).
XRF core-scanner-derived elemental ratios have been used as a time-efficient
method to assess down-core variations in grain size (e.g. Guyard et al.,
2013; Mulder et al., 2013; Bahr et al., 2014) and/or sediment sources for
detrital material in similar high-latitude locations (e.g. Møller et al.,
2006; Bervid et al., 2017). The applicability of this approach in Kane Basin
is tested in the present study by using
The chronology is based on a set of 18 radiocarbon ages obtained from mixed
benthic foraminifera samples and unidentified mollusc shells. The core top
is dated at
Reservoir ages in Nares Strait are difficult to assess owing to the scarcity
of pre-bomb specimens in collections of marine shells from the area. Only
three molluscs were dated in Nares Strait with
Core AMD14-Kane2b age model
AMS radiocarbon ages on selected carbonate material. Asterisks indicate data that were not used in the age model. MBF: mixed benthic foraminifera, MS: unidentified mollusc shell.
According to our chronology, core AMD14-Kane2b covers approximately the last 9.0 kyr (Fig. 4). The comparison of the
Major changes in depositional environments, most particularly during the
time interval corresponding to the lower half of our sediment core, explain
the wide range of sedimentation rates. High sedimentation rates are observed
between the base and
In modern sediments, the spatial variability of sediment geochemistry in
Kane Basin is likely related to their provenance. Heavy crystalline minerals
(e.g. garnet and orthopyroxene) occur in the eastern province of the basin
with provenance from the Humboldt Glacier and Inglefield Land, whereas
carbonates in its western sector are sourced from Ellesmere Island or from
Washington Land in its northern sector (Fig. 2, Kravitz, 1976). The
geochemical composition of modern sediments varies likewise, with
notably high concentrations of Fe and Zn in the eastern sector of Kane
Basin (Kravitz, 1994). Although the exact chemical variability of the source
geological units is not known at present, we consider the sedimentary
rocks from eastern Kane Basin and northern Nares Strait likely to be rich in
Ca, whereas higher concentrations of Fe, Si, and K presumably characterise
the crystalline rocks of the Ellesmere–Inglefield Precambrian Belt. We
propose the use of
XRF counts in core AMD14-Kane2b are largely dominated by Ca and Fe, which are
anticorrelated. Our records show a good correlation between normalised K
counts and clay content in the < 2 mm fraction (laser diffraction
grain size data) with a correlation factor of
3.5 kHz chirp profile across the coring location. Core
AMD14-Kane2b is represented by the orange box. Vertical scale in s (TWT)
with depth conversion assuming 100 ms (TWT)
The chirp 3.5 kHz sub-bottom profile obtained prior to core recovery is
shown in Fig. 5. Given the good recovery of recent sediments at the top of
core AMD14-Kane2b (Fig. 3), we place the top of core AMD14-Kane2b at the
sediment–water interface on this profile. Assuming an acoustic velocity of
1500 m s
Sedimentological results and elemental signature of the detrital fraction of core AMD14-Kane2b. Laser diffraction grain size repartition (< 2 mm fraction) is shown as % sand, silt, and clay. Normalised Zr counts are not shown but their profile is similar to that of Ti.
Details of CT scans and thin sections for each lithologic unit of core AMD14-Kane2B and summarised descriptions and interpretations. The paleo-environmental implications discussed in this study have been outlined here.
Based on CT scans and grain size records, five lithological units were defined for core AMD14-Kane2b, each corresponding to specific depositional environments (Fig. 6, Table 2). The sedimentological processes at play will be examined here, while their environmental significance will be considered in the discussion section of this paper.
Unit 1 (425–394 cm, ca. 9.0 cal ka BP) encompasses three subunits of distinct lithological nature.
Subunit 1A (425–416 cm) consists of high-density, occasionally sorted coarse
sediment in a clayey matrix, interbedded with thinner layers of lower-density silty clay (Table 2).
The base of the coarser laminations shows
erosional contact with the underlying finer beds (thin sections in Table 2).
Grain size analysis reveal large amounts of sand (26 %–39 %) and silt
(24 %–32 %) in the < 2 mm fraction in this interval. The relative
weight of the 125–800
These laminated deposits display all the characteristic of ice-proximal
deposits (List, 1982; Ó Cofaigh and Dowdeswell, 2001). Unit 1A was most
likely deposited at the ice sheet margin some
Subunit 1B (416–410 cm) displays a sharp decrease in sediment density with
the replacement of sand by finer material (60 % clay in the < 2 mm
fraction; Fig. 6, Table 2). While subunit 1B encompasses some clasts
(CT scan in Table 2), the amount of sand and silt actually present in 1B may
be lower than reflected by the laser diffraction and wet-sieving data, as
the analysed samples likely included coarser material from the overlying and
underlying subunits 1A and 1C. XRF data for subunit 1B show high
The finer grain size in this subunit is indicative of a change from an ice
margin to an ice-proximal glacimarine environment in which suspended matter
settling from turbid meltwater plumes is likely the main depositional
process (Elverhøi et al., 1980; Syvitstki, 1991, Dowdeswell et al., 1998;
Hogan et al., 2016), although the limited thickness (4 cm) of subunit 1B is
rather unusual for this process. The geochemical grain size tracers
Subunit 1C (410–394 cm) interrupts the fine-grained sedimentation with a
sharp increase in the occurrence of outsized clasts. The coarser fractions
account for a significant part of the sediment (up to 18 % for both the
> 800 and 800–125
Given the high gravel content in subunit 1C, we consider the clasts to have been predominantly iceberg-rafted to the core location rather than sea-ice-rafted (Pfirman et al., 1989; Nürnberg et al., 1994). These large amounts of IRD among very poorly sorted material can be interpreted as (1) increased iceberg calving rates, (2) changes in the delivery of sediment by icebergs (increased melting of or dumping from icebergs), or (3) a severe decrease in the delivery of finer particles that increases the apparent contribution of clasts to the sediment (Hogan et al., 2016, and references therein).
Unit 2 (394–320 cm, 9.0–8.3 cal ka BP) can be divided into two subunits based on grain size and density. The relative weight of the coarse fraction varies throughout unit 2 with a generally decreasing trend.
Subunit 2A (394–370 cm, 9.0–8.8 cal ka BP) is composed of poorly sorted,
bioturbated sediment (
The dominance of fine particles in subunit 2A with occasional clasts points
to a delivery by meltwater plumes and iceberg rafting. The decreasing
The sediments of subunit 2B (370–320 cm, 8.9–8.3 cal ka BP) have a lower
density and a lower sand and silt content than those of subunit 2A, while
clay content reaches maximum values averaging 63 %. Scarce lonestones
occur in this subunit and the sediment appears to be faintly laminated. Four
biogenic carbonate samples, both mollusc and mixed benthic foraminifera
samples, were dated in subunit 2B, and high sedimentation rates of
These high sedimentation rates, substantial concentrations of clay, and the
slightly laminated aspect of subunit 2B indicate that these sediments were
mainly delivered by meltwater plumes in a more distal glacial setting (Ó
Cofaigh and Dowdeswell, 2001). Relatively high
Unit 3 (320–300 cm, 8.3 cal ka BP) stands out as a clast-rich interval. The high density of this unit is comparable to that of subunits 1A and 1C. CT scans and thin sections reveal the presence of a finer-grained horizon enclosed between coarser material, dividing this interval into three subunits (Table 2).
Subunit 3A (320–313 cm) corresponds to the lower clast-rich subunit. A
significant portion of the bulk sediment is attributed to 800–125
The high clast content and absence of grading suggest that the sediments
forming subunit 3A were ice-rafted and deposited at the core location (Ó
Cofaigh and Dowdeswell, 2001). The predominant carbonate (low
Faint laminations are visible on the CT scan images of subunit 3B (313–305 cm). The sediment of this subunit is composed essentially of clay and silt
(47 % and 43 %, respectively) with a relatively low sand content (< 10 % in the < 2 mm
fraction and each of the coarser fractions
represents less than 3 % of the sediment weight).
The poor sorting of sediments in subunit 3B could possibly indicate that
they were ice-transported, but the near absence of clasts (e.g. in contrast
to the overlaying and underlying subunits of interval 3) contradicts this
hypothesis. The modest contribution of clay along with the relatively high
silt content rather points to the transport and deposition of these
sediments by a high-velocity current. The elemental signature of this
subunit (low
Subunit 3C (305–300 cm) contains large amounts of coarse material with an
average of 44 % sand and only 32 % clay in the < 2 mm
fraction. The sand in this subunit is coarser than in 3A with the 800–125
The very high clast content of subunit 3C along with high
Unit 4 (300–280 cm, 8.3–8.1 cal ka BP) has a similar density
and clay content (
The high clay content of unit 4 suggests that delivery from meltwater plumes
was the dominant sedimentary process at play during this time interval. The
substantial amount of sand in this unit indicates that a significant
proportion of the sediment was also ice-rafted to the location. As
previously mentioned, the increase in ice-rafted debris can indicate (1) increased
calving rates when originating from iceberg rafting, (2) changes
in iceberg delivery of sediment (increased melting or dumping of icebergs),
or (3) a decrease in the delivery of finer particles that increases the
apparent contribution of clasts to the sediment (Hogan et al., 2016, and
references therein). The high sedimentation rates (
Unit 5 (280–0 cm, 8.1–0 cal ka BP) clearly differs from underlying units with regard to the < 2 mm grain size fraction (Fig. 6). The clay content drops to steady, lower values (49 % on average) and the CT scans show a generally homogenous sediment with frequent traces of bioturbation. Changes in grain size divide unit 5 into two subunits.
The sediments in subunit 5A (280–250 cm, 8.1–7.5 cal ka BP) contain a
relatively high proportion of sand peaking at 12 % in the < 2 mm
fraction, while the combined contribution of the coarser fractions averages
at
The significant decrease in clay particles in subunit 5A compared to units 4
and 2B suggests that delivery from meltwater plumes was reduced in this
interval, either in relation to a decrease in glacial melting rates or to a
more ice-distal setting.
The sharp decrease in the
Subunit 5B (250–0 cm, 7.5–0 cal ka BP) is generally homogenous with
lonestones occurring sporadically throughout. The silt content increases
gradually from
A sample of mixed benthic foraminifera yielded a radiocarbon age some 2 kyr older than expected at 238.5 cm. This sample probably contains a mixture of coeval and remobilised foraminifera (either by bioturbation or by water–ice transport from another location).
The overall limited contribution of the coarser fractions to the sediment of
subunit 5B in comparison to the underlying lithologic units indicates that
ice delivery of sediment was reduced during this interval. Furthermore, the
relatively low amounts of clay imply that meltwater delivery was also
weakened. The sediments of subunit 5B were likely primarily
water-transported to the core site (Gilbert, 1983; Hein and Syvitski, 1992).
The increase in silt and
Comparison of sieved grain size data from AMD14-Kane2b and
paleoceanographic proxies from HLY03-05CG in Hall Basin (Jennings et al.,
2011). Radiocarbon ages presented in Jennings et al. (2011) were calibrated
with
Our study of core AMD14-Kane2b has enabled us to reconstruct a succession of depositional environments in Kane Basin following the retreat of the formerly coalescent GIS and IIS in Nares Strait (Fig. 8). Here we discuss our reconstructions in light of other paleoceanographic and paleoclimatic studies to provide a broader view of the Holocene history of Nares Strait (Figs. 6, 7, and 8, Table 2).
GIS and IIS retreat in Nares Strait. Adapted from England (1999)
and including data from Bennike (2002) for Washington Land. Locations for
core AMD14-Kane2b in Kane Basin and HLY03-05 (Jennings et al., 2011) in Hall
Basin are marked by crosses. All mollusc ages from England (1999) were
calibrated with
Previous studies have shown that the presence of erratic Greenland boulders
on Ellesmere Island from Kennedy Channel to the northern entrance of Nares
Strait attest to the coalescence of the IIS and GIS along the western side
of northern Nares Strait during the Last Glacial Maximum (LGM) (e.g.
England, 1999). The absence of such erratics along the western and southern
coasts of Kane Basin implies that the confluence of the two ice sheets was
further at sea in the southern half of the strait (England, 1999).
Radiocarbon dating on samples from raised beaches provides minimum ages for
marine ingress in Nares Strait. These ages are older in the northern and
southern extremities of the strait, while only younger ages are yielded by
samples in northern Kane Basin and Kennedy Channel, implying that a central
(grounded) ice saddle persevered longer in the shallower sector of the
strait (England, 1999, and references therein; Bennike, 2002). In addition to
providing minimum ages for ice sheet retreat,
Our archive demonstrates that marine sedimentation took place in Kane Basin
as early as
The increasingly finer particles that compose unit 2 suggest a growing
distance between the core site and the ice margin. The dominant sedimentary
process at play is settling from meltwater plumes, which is typically
responsible for high sedimentation rates, along with frequent delivery of
IRD (Table 2). The Early Holocene was characterised by high atmospheric
temperatures during the Holocene Thermal Maximum (HTM) occasioned by greater
solar insolation (Bradley, 1990). The HTM has been defined for the eastern
sector of the CAA as the period between 10.7 and 7.8 cal ka BP based on the
Agassiz ice core record (Lecavalier et al., 2017). The high melting rates of
the ice sheets during the HTM (Fisher et al., 2011) likely enhanced the
delivery of particles by meltwater and contributed to the high sedimentation
rates observed in our core. More distant glacial ice from the site is also
in good agreement with the occurrence of molluscs dated between 8.8 and 8.4 cal ka BP on Ellesmere Island and Northwest Greenland (Fig. 8c, England, 1999). The elemental signature of subunit 2B may suggest, however, that
the GIS was still present in eastern Kane Basin and delivered material
derived from the gneiss basement to the core site. The volcanic clastics on
Ellesmere Island may have also contributed to Fe counts in our geochemical
record, but we consider their input marginal given the limited surface of
this geological unit compared to the gneiss and crystalline basement, which
outcrops in much of Inglefield Land and underlays Humboldt Glacier (Dawes and Garde,
2004). Furthermore, the IIS was a cold-base ice sheet (e.g. Tushingham,
1990; Dyke et al., 2002) and as such likely delivered overall less sediment from
meltwater than the warm-based GIS. The occurrence of IRD in unit 2 may imply
that relatively open water conditions occurred during this interval,
enabling icebergs to drift in Kane Basin, although this may simply be a
consequence of high calving rates as the GIS and IIS retreat. Reduced sea
ice occurrence in Kane Basin during the Early Holocene would be in good
agreement with low sea ice concentrations reported nearby in Lancaster Sound
(from 10 to 6 cal ka BP,
Unit 3 appears to be primarily iceberg-rafted, with an inclusion of a finer,
water-transported silty subunit (3B). A foraminifera-derived radiocarbon age
obtained from subunit 3B (Table 1) suggests sediment remobilisation within
this time interval. If we consider unit 3A to have been deposited by the passing
over Kane Basin of glacial ice having broken up in Kennedy Channel, then a
plausible origin for unit 3B could be the entrainment of sediment from
northern Nares Strait associated with the discharge of large amounts of
water as the connection was established. The absence of any molluscs in
Kennedy Channel predating 8.1 cal ka BP further suggests that Kennedy
Channel was still blocked until then, although this method can only provide
minimum ages for ice sheet retreat (Fig. 8e–g). This proposed age for the
opening of Kennedy Channel is only slightly younger than that proposed by
Jennings et al. (2011), i.e.
It has recently been demonstrated that the Humboldt Glacier retreated from a
previous position of stability ca.
The dominant depositional process in subunit 3C is iceberg rafting based on
the abundance of clasts in this interval. The elemental composition of
subunit 3C suggests that the sedimentary material likely originates from the
GIS in eastern Kane Basin (Fig. 2). Investigations into the internal
stratigraphy of the GIS and their comparison to north Greenland ice cores
have demonstrated that the collapse of the ice saddle in Kennedy Channel
triggered the acceleration of glacial fluxes along Nares Strait (MacGregor
et al., 2016). The destabilisation of the GIS following the collapse of the
ice saddle may have provoked intense calving that led to the deposition of
subunit 3C. In this regard, intense calving of the Humboldt Glacier as
recently dated by Reusche et al. (2018) at
However, the Reusche et al. (2018) findings also offer an alternative scenario
for the deposition of unit 3. Intense calving of the Humboldt Glacier may
have occurred as it retreated in eastern Kane Basin and abandoned a lateral
moraine in Washington Land ca.
The abundance of iceberg-rafted debris has increased considerably in unit 4 compared to unit 2. This is possibly the result of the aforementioned acceleration of the GIS and IIS along Nares Strait following the collapse of the ice saddle in Kennedy Channel (MacGregor et al., 2016), as well as the arrival of icebergs from new sources to Kane Basin situated in northern Nares Strait. The retreating GIS in eastern Kane Basin was likely a primary source of icebergs during this period. However, the high clay content in our record implies that the GIS was still relatively close to the core site and had not yet fully retreated in eastern Kane Basin, contributing to the high sedimentation rates recorded in this unit (Fig. 8e).
The abrupt decrease in clay content and sedimentation rates at 280 cm in our
record implies that the ice margin abruptly retreated ca. 8.1 cal ka BP (Fig.6,
Fig. 8f). The equal drop in
The low
Our investigation of core AMD14-Kane2b has provided, for the first time, a
paleo-environmental reconstruction in Kane Basin over the last ca. 9.0 kyr. The confrontation of our dataset with both land-based (England, 1999;
Bennike, 2002, Reusche et al., 2018) and marine (Jennings et al., 2011)
evidence offers several alternative paleo-environmental interpretations for
our record. Of particular interest is the determination of which of the two
IRD-rich units (unit 1c
While evolving from a short-lived ice-proximal depositional environment at
This archive provides a new viewpoint that has enabled us to propose a continuous timeline of the events related to the deglaciation of Kane Basin, which until now relied entirely on land-based studies. Our study suggests that the “bridge dipole” presented in Barber et al. (2001), in which warmer (colder) years exhibit more (less) sea ice in Smith Sound and less (more) ice in Nares Strait, may be extrapolated over the last 2 millennia. Future investigations into the Holocene variability of sea ice conditions in Kane Basin may provide a more comprehensive view on its controlling effect on the NOW Polynya. High productivity rates in the NOW Polynya, however, are also fuelled by the throughflow of nutrient-rich Pacific water via Nares Strait, and further investigation into how oceanographic circulation responded to postglacial changes in Nares Strait will provide more insight into the Holocene evolution of this highly productive area of the Arctic. Other than emphasising the need for further research into local reservoir age corrections, our study is inclined to contribute to future work on the export of low-salinity Arctic water and Holocene variations of deep water formation (Hoogakker et al., 2014; Moffa-Sanchez and Hall, 2017).
The data presented in this paper are available upon request (eleanor.georgiadis@u-bordeaux.fr).
The supplement related to this article is available online at:
EG, JG, and GM designed the research and carried out most of the analyses. GSO provided CT scan imagery. SS performed 210Pb measurements. EG, JG, and GM prepared the paper with contributions from all authors. EG, JG, GM, PM, PL, SS, and GSO contributed to the interpretation of the results.
The authors declare that they have no conflict of interest.
Eleanor Georgiadis' studentship is funded by both the Initiative d'Excellence
(IdEx) programme of the University of Bordeaux and the Natural Science and
Engineering Research Council of Canada (NSERC). We would like to thank Anne
Jennings, Sofia Ribeiro, Audrey Limoges, and Karen Luise Knudsen for
constructive conversations on the history of the North Water Polynya and Benoit
Lecavalier for having shared with us the much appreciated Agassiz ice core
temperature record. We also extend our gratitude to Sebastien Zaragosi and Bernard
Martin, who prepared and helped interpret thin sections,
and Isabelle Billy, Pascal Lebleu, Olivier Ther, and Caroline Guilmette for analytical support. This
work is supported by the Fondation Total, the
French Agence Nationale de la Recherche (GreenEdge project), the Network of
Centres of Excellence ArcticNet, and the European Research Council (StG
IceProxy). Finally, we wish to thank the CCGS