Diachronous evolution of sea surface conditions in the Labrador Sea and Baffin Bay since the last deglaciation

Abstract

Assessing changes in sea surface conditions due to the effects of past freshwater outflow through Baffin Bay and Davis Strait to the Labrador Sea, hereafter referred to as the Baffin Bay corridor, is relevant in understanding the variability in Labrador Sea Water (LSW) formation. Here, regional changes in oceanographic circulation and sea surface conditions are reconstructed based on organic-walled dinoflagellate cyst (dinocyst) assemblages from four cores collected from deep, central sites of the Baffin Bay corridor. All cores exhibit a major shift in dinocyst assemblages since the late glacial period. This shift consists of a change from a polar–subpolar heterotrophic species assemblage tolerating cold and near permanent ice-covered conditions, to assemblages characterized by a higher diversity and the occurrence of phototrophic taxa associated with mild conditions. Sea surface reconstructions from the modern analogue technique display a shift from harsh, quasi-perennial ice cover to warmer summer sea surface temperatures and a seasonal sea ice. South of the Davis Strait sill, this regime shift occurred at ca. 11.9 cal ka BP due to the influence of North Atlantic waters. Baffin Bay, however, remained densely sea ice covered until about 7.4 cal ka BP, when these warmer waters penetrated into Baffin Bay and mixed with the West Greenland Current (WGC). This mixing was facilitated by the retreat of the Greenland and Laurentide Ice Sheet (LIS) margins. A major change in Labrador Sea surface conditions occurred nearly at about the same time (~7.6 cal ka BP) when the strong stratification of surface waters weakened because of the reduction in meltwater supplies from the LIS that allowed winter convection and the inception of LSW formation. All these new records demonstrate large amplitude fluctuations in sea surface conditions tightly controlled by the relative strengths and shifts of the warmer WGC and colder Baffin Island Current.

Keywords

Baffin Bay Davis Strait dinoflagellate cysts Labrador Sea last deglaciation sea surface conditions

Introduction

Arctic outflow through the Baffin Bay corridor, consisting of Baffin Bay, Davis Strait, and the Labrador Sea, provides a significant contribution of freshwater to the North Atlantic, as shown by both observations and modeling studies (Aksenov et al., 2010; Curry et al., 2011; Serreze et al., 2006). Changes in the Arctic’s hydrological cycle due to ongoing global warming are expected to modify the export of freshwater from the Arctic to the Labrador Sea, leading to a more stratified upper ocean, which may inhibit convective mixing and the formation of an intermediate water mass (Cheng and Rhines, 2004; Goosse et al., 1997; Wadley and Bigg, 2002). From this viewpoint, documenting late- to post-glacial changes in sea surface conditions and productivity, linked to variations in freshwater flow and oceanic circulation through the Baffin Bay corridor, may help comprehend their evolution in areas downstream from the Arctic Ocean.

Monitoring past changes in sea surface conditions through the Baffin Bay corridor since the deglaciation has been undertaken almost exclusively using paleoceanographic records from nearshore areas such as Disko Bugt (e.g. Andresen et al., 2010), Nares Strait (e.g. Levac et al., 2001), and Lancaster Sound (e.g. Ledu et al., 2008). Such records, unlike those of offshore sites, have high temporal resolution due to high sedimentation rates influenced by glacial erosion and mainly reflect local atmospheric and hydrographic conditions (e.g. Erbs-Hansen et al., 2013; Seidenkrantz et al., 2008). As illustrated here, deeper sites off the shelf break yield records representative of large-scale changes in surface water conditions in response to the regional ocean dynamics rather than to coastal processes, but they offer a much lesser time resolution because of lower sedimentation rates.

Baffin Bay is a difficult environment for paleoceanographic studies because of inherent challenges of the setting of a chronostratigraphy in deep-sea sedimentary sequences. Low sedimentation rates characterize interglacial stages (<10 cm ka⁻¹) in comparison to high rates during glacial periods (e.g. Simon et al., 2012; Srivastava et al., 1989). Additionally, at depths greater than 900 m, calcium carbonate dissolution occurs in Holocene sediments because of low carbonate saturation states (Azetsu-Scott et al., 2010) and oxidation of organic matter related to high productivity thus preventing radiocarbon dating of calcareous microfossils (Aksu 1983; de Vernal et al., 1992; Osterman and Nelson, 1989; Schröder-Adams and Van Rooyen, 2011). Carbonate dissolution also makes it impossible to use foraminiferal assemblages as paleoceanographic proxies, and thus to derive oxygen and carbon stable isotope records for potential chronological correlations.

Cores from central Baffin Bay (TWC16), eastern Baffin Bay (CC70), southern Davis Strait (TWC08), and the northwest Labrador Sea (CC04) were collected along the Baffin Bay corridor during the HU2008029 cruise (Figure 1; Campbell and de Vernal, 2009). Despite the above methodological difficulties, they were analyzed with the aim to provide an assessment of past sea surface conditions since the last glaciation, and more specifically, to document the effects of freshwater fluxes from surrounding ice sheets and the Arctic on sea surface temperatures (SSTs) and sea ice cover between the Arctic and North Atlantic oceans.

Figure 1.

Map indicating sampling location for cores CC04, TWC08, TWC16, and CC70, which are used in this study, as well as other cores mentioned in the text (008P, 012P, DA00-04P, DA00-06P, DA04-31P, DA04-41P, MSM343300, P009, P021). Surface currents are also displayed including the North Atlantic Current (NAC), Irminger Current (IC), East Greenland Current (EGC), West Greenland Current (WGC), Baffin Island Current (BIC), and Labrador Current (LC). The approximate location of the West Greenland Current Front (WGCF) and the Labrador Shelf-Slope Front (LSSF) are shown as dotted lines (after Belkin et al., 2009).

Modern hydrographic setting and location of core sites

The surface and subsurface waters of the Baffin Bay corridor form a counter clockwise gyre influenced by northward flowing, warm high-salinity Atlantic waters, and by southward flowing cold low-salinity Arctic waters (Figure 1) (see Buch, 1990/2000; Cuny et al., 2002, 2005; Ribergaard et al., 2008; Tang et al., 2004 for details listed below). Along the West Greenland shelf, the West Greenland Current (WGC) carries cool, less saline Arctic water from the East Greenland Current (EGC) that has been slightly modified by warmer, more saline Atlantic water. The Atlantic water is carried through the western branch of the Irminger Current (IC), which flows below ~100 m down to ~800 m along the West Greenland shelf and slope.

Upon reaching northern Baffin Bay, the surface component of the WGC (upper 300 m) mixes with cold, less saline Arctic water that enters via Nares Strait, Lancaster Sound, and Jones Sound and becomes the Baffin Island Current (BIC). Some warmer Arctic Intermediate water of Atlantic origin passes over the sills of Nares Strait and combines with the Irminger component of the WGC that has recirculated in Baffin Bay to form the Baffin Bay Intermediate Water. The BIC follows the Baffin Island coast on the continental shelf and slope. Baffin Bay Deep and Bottom waters cool with depth and cover the deepest areas of the Bay. The BIC flows south along Baffin Island through Davis Strait into the northwest Labrador Sea. Finally, it mixes with colder, fresher Arctic water exiting Hudson Strait to form the Labrador Current (LC). Along the eastern Canadian shelf and upper slope, the surface layer of the LC overlies the WGC branch that has extended westward near Davis Strait and circulated through the gyre. Below the LC lies the Labrador Sea Water (LSW), which is formed by the sinking of dense waters when cooled in winter (Lazier, 1973). LSW reaches depths up to 2500 m and lies above the North East Atlantic Deep Water and the Denmark Strait Overflow Water (Yashayaev, 2007).

Sea ice cover along the Baffin Bay corridor is variable, ranging from 0 to >10 months yr⁻¹, thus with a concentration ranging from 0% to about 90%. Ice starts forming in northwest Baffin Bay in September and extends southwestward to form a complete sea ice cover by March (Tang et al., 2004; Wang et al., 1994), occasionally reaching the northwest Labrador Sea. The warm WGC prevents sea ice formation along the southwestern Greenland coast and reduces the sea ice extent in eastern Baffin Bay. In opposition, the cold and stratified LC fosters sea ice growth along the Newfoundland and Labrador shelf. In Baffin Bay, sea ice begins to melt in April in the North Water Polynya and along the Greenland coast, moving westward until ice-free conditions occur, normally in August and September. A large interannual variability of sea ice cover is coupled to the strong seasonality in air temperatures and wind patterns (Tang et al., 2004).

An oceanic front is a constricted zone of increased productivity and nutrients due to a significant horizontal gradient in water mass properties such as sea surface temperature and salinity (Belkin et al., 2009). Two polar fronts are present in the study area (Figure 1), both separating cold, low-salinity polar current from warmer, more saline current. The Labrador Shelf-Slope Front (LSSF), which is situated on the NW Labrador/SW Baffin Island slope, is associated with the LC, and the West Greenland Current Front (WGCF) follows the western edge of the WGC (Belkin et al., 2009, supplementary material, available online). The location of these polar fronts is dependent on the strength of the LC and WGC.

Methods

The four sites used in this study were cored during the HU2008029 cruise aboard the CCGS Hudson (Figure 1). Site information, core lengths, and sampling intervals are listed in Table 1. Trigger cores TWC04, TWC08, and TWC70 were analyzed for this study and core TWC16 was analyzed by Steinhauer (2012). Two core composites were created in order to extend the Holocene trigger core records into the deglacial interval: CC04 in the northwest Labrador Sea combines TWC04 with the data from PC04 reported in Gibb et al. (2014), and CC70 in eastern Baffin Bay combines TWC70 with data from PC70 reported in Jennings et al. (2014). Core descriptions, photographs, and onboard measurements including magnetic susceptibility are found in the cruise report (Campbell and de Vernal, 2009).

Table 1.

List of cores used in this study with core location and water depth (m), as well as the modern sea surface conditions (sea surface temperature (SST) in winter and summer, sea surface salinity (SSS), months per year of sea ice cover, and gC m⁻² of annual productivity) for each location provided by the NODC and NSIDC.

Core ID	Core location	Latitude (°N)	Longitude (°W)	Water depth (m)	Modern sea surface conditions					Core sampling
	Core location	Latitude (°N)	Longitude (°W)	Water depth (m)	SST (summer)	SST (winter)	SSS (summer)	Sea ice (months yr⁻¹)	Productivity (gC m⁻²)	Core ID	Length (cm)	Interval
CC04	Northwest Labrador Sea	61.4639	58.0365	2674	7.3	3.7	34.2	0.7	167.9	TWC04	56	1 cm
TWC08	Southern Davis Strait	64.3931	58.1347	857	4.2	0.2	32.6	4.4	53.3	TWC08	156	2 to 43 cm
TWC16	Central Baffin Bay	70.4619	64.6578	2063	3.2	−1.10	29.9	9	56.8	TWC16	155	1 to 8 cm, 2 to 64 cm
CC70	Western Greenland	68.2279	57.6175	444	3.6	−1.7	32.8	6.2	113.6	TWC70	208	4 cm

The length (cm) and sampling interval for the cores analyzed in this study (trigger sections) are also listed.

The chronostratigraphy was established based on radiocarbon dates from biogenic remains. Cores TW70 and PC70 were marked by poor biogenic carbonate preservation and very rare occurrence of foraminifer shells but contained mollusk shells and seaweed which permitted to obtain radiocarbon dates (Jennings et al., 2014). The radiocarbon dates from cores TWC04, TWC08, and TWC16 were derived from planktonic foraminiferal populations that consist of >95% Neogloboquadrina pachyderma left-coiled (Npl). Radiocarbon ages were calculated using the Libby half-life of 5568 years and normalized to a δ¹³C of −25‰. The ages were then converted to calibrated years BP based on OxCal 4.2 (Bronk Ramsey, 2008), using the Marine13 calibration curve of Reimer et al. (2013). No additional correction (ΔR) was applied.

The number of radiocarbon dates from cores TWC04 and TWC70 was sufficient to create age–depth models. The calibrated ages in core TWC70 were combined with those of PC70 (Jennings et al., 2014) then modeled in OxCal to create the age–depth model for the composite CC70 sequence (Table 2, Figure 2). The core collection date of 2008 (−58 cal a BP) was assigned to −15 cm in TWC70, which is offset from its corresponding box core based on physical properties. For consistency with CC04 and other records from deeper sites, the ΔR used by Jennings et al. (2014) was not used in the age–depth model for CC70. Cores TWC04 and PC04 (Gibb et al., 2014) were modeled individually then spliced at 8.3 cal ka BP for the composite CC04 sequence (Table 2, Figure 2). For the purpose of this paper, data from CC04 will only be reported for the last 15 cal ka BP, but core PC04 extends to ~36 cal ka BP (Gibb et al., 2014). Physical properties including diffuse spectral reflectance (CIE L*a*b*) values and photographs from the box core and trigger core (Campbell and de Vernal, 2009) were used to determine that the TWC04 core top was undisturbed, assigning an age of AD 2008 (−58 cal a BP) to 0 cm of CC04. The calibrated ages were reported as the modeled median cal. age BP. Linear interpolation between each modeled date was used to calculate the age–depth relationship when plotting proxy data against age and sedimentation rates.

Table 2.

Radiocarbon dates for the cores in this study. Refer to reference list for more details on previously published dates.

Core name and depth interval (cm)	Laboratory identifiera	AMS-¹⁴C ages (yr BP)b	Calibrated age interval (2σ) (cal yr BP)c	Calibrated median age (cal yr BP)	Modeled age interval (2σ) (cal yr BP)d	Modeled median age (cal yr BP)	Reference
TWC04
8–9	CAMS-155456	1690 ± 35	1170–1314	1250	1170–1315	1249	This paper
20–21	OS-96929	2380 ± 25	1921–2102	2008	1923–2101	2009	This paper
32–33	CAMS-155457	5085 ± 35	5323–5552	5448	5323–5551	5447	This paper
44–45	OS-96930	6790 ± 35	7245–7403	7318	7245–7403	7317	This paper
56–57	CAMS-155458	7315 ± 30	7680–7860	7776	7681–7860	7775	This paper
PC04
*84–85	CAMS-155462	7735 ± 30	8416–8181	8319	8441–8196	8335	Gibb et al. (2014)
96–97	CAMS-153492	8780 ± 60	9544–9302	9451	9540–9300	9447	Gibb et al. (2014)
116–117	CAMS-155463	9290 ± 30	10,218–10,033	10,150	10,218–10,009	10,147	Gibb et al. (2014)
216–217	CAMS-153493	10,990 ± 45	12,630–12,380	12,520	12,620–12,375	12,473	Gibb et al. (2014)
244–245	CAMS-153494	11,400 ± 120	13,128–12,636	12,881	13,071–12,635	12,816	Gibb et al. (2014)
388–389	CAMS-153495	12,700 ± 60	14,878–13,952	14,226	14,871–14,010	14,363	Gibb et al. (2014)
TWC08
104–105	GRL 1897-S	16,460 ± 110	19,043–19,651	19,370			Andrews et al. (2014)
136–137	GRL 1898-S	17,480 ± 80	20,351–20,850	20,597			Andrews et al. (2014)
TWC16
56–57	CAMS-146826	11,800 ± 40	13,167–13,383	13,276			Steinhauer (2012)
TWC70
97–98	AA-84709	3252 ± 37	2760–3021	2887	2757–3014	2882	Jennings et al. (2014)
134–135	SUERC-25670	5913 ± 39	6035–6295	6198	6033–6286	6190	Jennings et al. (2014)
172	AA-84710	7898 ± 54	8072–8359	8229	8061–8346	8218	Jennings et al. (2014)
194–195	SUERC-3058	8464 ± 40	8693–9019	8887	8694–9008	8879	Jennings et al. (2014)
PC70
55–56	SUERC-30590	7501 ± 41	9238–9470	9363	9222–9462	9345	Jennings et al. (2014)
142–144.5	AA-84711	7501 ± 42	10,438–10,733	10,590	10,430–10,705	10,574	Jennings et al. (2014)
188–189	AA-84713	7501 ± 43	10,834–11,175	11,035	10,885–11,175	11,054	Jennings et al. (2014)
209–211	AA-84714	7501 ± 44	11,250–11,775	11,497	11,240–11,731	11,460	Jennings et al. (2014)

The analyses were made at the following institutions:

CAMS = Lawrence Livermore National Laboratory’s Center for Accelerator Mass Spectrometry.

OS = The National Ocean Sciences Accelerator Mass Spectrometry Facility, Woods Hole Oceanographic Institute.

AA = NSF-Arizona Accelerator Mass Spectrometry Laboratory.

SUERC = Scottish Universities Environmental Research Centre.

Radiocarbon age was calculated using the Libby half-life of 5568 years and corrected with a δ¹³C of −25‰.

A marine reservoir correction of 400 years was applied with no additional correction (ΔR), and the ages converted to calibrated years using Oxcal 4.2 (Bronk Ramsey, 2008) and the Marine13 (Reimer et al., 2013) calibration curve.

The ages were modeled using Oxcal 4.2 (Bronk Ramsey, 2008).

The ¹⁴C date used in the calibration and age-model is an average of three dates (7852 ± 55 yr BP) representing the final drainage of glacial lake Agassiz (refer to Gibb et al., 2014 for further details).

Figure 2.

Age versus depth relationship for cores CC04 (a) and CC70 (b) based on the calibrated ¹⁴C ages listed in Table 2. The age–depth model is shown in purple and blue and the sedimentation rates are represented by the black lines. The age–depth model for core CC04 (a) is in two sections, with core PC04 in blue and TWC04 in purple.

Due to the absence of radiocarbon chronologies in the Baffin Bay/Davis Strait records, the geochemical properties and micropaleontological content of the representative cores TWC16 and TWC08 were tentatively correlated with cores CC70 and CC04 from eastern Baffin Bay and northern Labrador Sea respectively. Ice rafted debris (IRD), calcium carbonate content, dinoflagellate cyst (dinocyst) assemblages, and the dinocyst-based sea surface reconstructions are used for correlations.

Each sample used for dinocyst analysis consists of a 5-cm³ subsample of sediment from a 1-cm thick section. Core length and sampling interval are listed in Table 1. The subsamples were rinsed through 106- and 10-µm mesh sieves. Weight percent coarse fraction was calculated from the dried >106-µm size fraction from the dinocyst preparations. This makes a representative proxy for IRD because it mainly consists of detrital material and low abundance of light biological remains (microfossils). The 10- to 106-µm size fraction was processed according to the method described by de Vernal et al. (1999). The organic residue was mounted onto a slide with Kaiser’s glycerol gelatin and examined for dinocysts. Dinocyst species were identified following the nomenclature provided by de Vernal et al. (2001), Head et al. (2001), Radi et al. (2013), and Rochon et al. (1999). Dinocysts were tabulated and concentrations were calculated using the marker-grain method (Matthews, 1969), which provides an accuracy of ±10% for a 95% confidence interval (de Vernal et al., 1987). Relative percent abundance of the dinocyst taxa in assemblage for each sample was calculated based on identifying a minimum of 300 cysts whenever possible. These counts were also used for further statistical analyses including principal component analysis (PCA) on percentage data to determine assemblage zones in each core, and the application of the modern analogue technique (MAT) to reconstruct sea surface conditions. SST in summer and winter, sea surface salinity (SSS) in summer, sea ice cover expressed as the number of months per year with a sea ice concentration of >50%, and primary productivity (gC m⁻²) were reconstructed with MAT following the procedure described by de Vernal et al. (2005). MAT was applied using the updated ‘modern’ dinocyst database of the Northern Hemisphere that includes 1492 sites and 66 taxa (see database at http://www.geotop.ca; de Vernal et al., 2013b), using scripts prepared by Guiot (CEREGE, France) for the software R (http://cran.r-project.org/). We calculated the most probable conditions from the average of the five best analogues weighted inversely by their distance, and the variance provided upper and lower boundaries. Methods for the validation tests and error calculations were made after splitting of the data sets as described by de Vernal et al. (2013b). The error of prediction was established at ±1.4 months yr⁻¹ for sea ice cover, ±1.2°C and ±1.6°C for winter and summer SSTs, respectively, and ±2.6 for SSS. The error of prediction of SSS is high because the database includes low-salinity environments (down to 5 psu), where surface salinity is particularly variable. When considering only the >30 and >33 salinity domains, the summer SSS errors of prediction are ±1.3 and ±0.8 psu, respectively.

Results

The four cores used in this study were collected from very distinct regions as shown by their water depth and corresponding modern sea surface conditions (Table 1). Summarized core descriptions are listed below; however, greater detail and photographs can be found in the cruise report (Campbell and de Vernal, 2009). The radiocarbon dates, their calibrated ages, publication reference, and laboratory identification are listed for each core in Table 2. The modeled age intervals and modeled median age (cal ka BP) for CC04 and CC70 are also listed in Table 2; their age–depth models and sedimentation rates are shown in Figure 2. The chronology, dinocyst assemblage, and PCA (Figures 3a, 4a, 5a, 6a), and MAT reconstruction (Figures 3b, 4b, 5b, 6b) for each core are described below.

Figure 3.

(b) Core CC04 – reconstruction of sea surface conditions from dinocyst assemblages and the weight percent coarse fraction (>106 µm, mostly representative of ice rafted debris – IRD) plotted versus age (cal ka BP). Sea surface temperatures (SSTs) in winter (w) and summer (s) are represented by blue and pink curves respectively. The lighter blue and pink curves correspond to possible maximum and minimum SSTs calculated from a set of five modern analogues. Sea surface salinity, sea ice cover, and annual productivity are represented by thin black lines (most probable values) and gray shading (minimum and maximum). The thick black lines represent a 5-point running mean. Ecostratigraphic units identified by the principal component (PC) analyses and sea surface reconstructions are separated by a horizontal dashed line. (a) Core CC04 – concentrations of dinocysts expressed as number of specimens per cm³ of sediment, relative abundance (percentage) of the main dinocyst taxa, and scores for PCs 1 and 2 are plotted versus age (cal ka BP). The complete dinocyst dataset can be found on the GEOTOP website (http://www.geotop.ca). Ecostratigraphic units identified by the PC analyses are separated by a dashed line, and marked with ages.

Figure 4.

(b) Core TWC08 – reconstruction of sea surface conditions from dinocyst assemblages and the weight percent coarse fraction (>106 µm, mostly representative of ice rafted debris – IRD) plotted versus depth (cm). Sea surface temperatures (SSTs) in winter (w) and summer (s) are represented by blue and pink curves respectively. The lighter blue and pink curves correspond to possible maximum and minimum SSTs calculated from a set of five modern analogues. Sea surface salinity, sea ice cover, and annual productivity are represented by thin black lines (most probable values) and gray shading (minimum and maximum). The thick black lines represent a 5-point running mean. Ecostratigraphic units identified by the principal component (PC) analyses and sea surface reconstructions are separated by a horizontal dashed line and identified by an inferred age (?) in cal ka BP. (a) Core TWC08 – concentrations of dinocysts expressed as number of specimens per cm³ of sediment, relative abundance (percentage) of the main dinocyst taxa, and scores for PCs 1 and 2 are plotted versus depth (cm). The complete dinocyst dataset can be found on the GEOTOP website (http://www.geotop.ca). Ecostratigraphic units identified by the PC analyses are separated by a dashed line.

Figure 5.

(b) Core CC70 – reconstruction of sea surface conditions from dinocyst assemblages and the weight percent coarse fraction (>106 µm, mostly representative of ice rafted debris – IRD) plotted versus age (cal ka BP). Sea surface temperatures (SSTs) in winter (w) and summer (s) are represented by blue and pink curves respectively. The lighter blue and pink curves correspond to possible maximum and minimum SSTs calculated from a set of five modern analogues. Sea surface salinity, sea ice cover, and annual productivity are represented by thin black lines (most probable values) and gray shading (minimum and maximum). The thick black lines represent a 5-point running mean. Ecostratigraphic units identified by the principal component (PC) analyses and sea surface reconstructions are separated by a horizontal dashed line. (a) Core CC70 – concentrations of dinocysts expressed as number of specimens per cm³ of sediment, relative abundance (percentage) of the main dinocyst taxa, and scores for PCs 1 and 2 are plotted versus age (cal ka BP). The complete dinocyst dataset can be found on the GEOTOP website (http://www.geotop.ca). Ecostratigraphic units identified by the PC analyses are separated by a dashed line.

Figure 6.

(b) Core TWC16 – reconstruction of sea surface conditions from dinocyst assemblages and the weight percent coarse fraction (>106 µm, mostly representative of ice rafted debris – IRD) plotted versus depth (cm). Sea surface temperatures (SSTs) in winter (w) and summer (s) are represented by blue and pink curves respectively. The lighter blue and pink curves correspond to possible maximum and minimum SSTs calculated from a set of five modern analogues. Sea surface salinity, sea ice cover, and annual productivity are represented by thin black lines (most probable values) and gray shading (minimum and maximum). The thick black lines represent a 5-point running mean. Ecostratigraphic units identified by the principal component (PC) analyses and sea surface reconstructions are separated by a horizontal dashed line and identified by an inferred age (?) in cal ka BP. (a) Core TWC16 – concentrations of dinocysts expressed as number of specimens per cm³ of sediment, relative abundance (percentage) of the main dinocyst taxa, and scores for PCs 1 and 2 are plotted versus depth (cm). The complete dinocyst dataset can be found on the GEOTOP website (http://www.geotop.ca). Ecostratigraphic units identified by the PC analyses are separated by a dashed line.

CC04 – The northwest Labrador Sea

The CC04 record presented here extends to ~15.0 cal ka BP with sedimentation rates varying between 3.5 and 93.1 cm ka⁻¹ (Figure 2). The piston core PC04 contains massive, dark gray silty clay from ~15 to 8.3 cal ka BP (436 to 60 cm), with diffuse laminations from ~15 to 12.9 cal ka BP (436 to 251 cm), and a mottled texture with bioturbation from 251 to 70 cm (i.e. after 12.9 cal ka BP). A layer of pebbles has been found and dated at ~14.5 cal ka BP (410 to 390 cm) (Gibb et al., 2014). The trigger core (TWC04), which spans about the last 7.8 cal ka BP, consists of massive, hemipelagic dark gray silty clay. Core CC04 contains up to 9 weight percent coarse fraction (>106 µm; considered here as IRD), prior to ~11.3 cal ka BP (Figure 3b).

The dinocyst concentrations and assemblages of CC04 are plotted relative to age from ~15 cal ka BP to present (Figure 3a). The dinocyst concentrations vary from as low as 5 up to 10,000 cysts cm⁻³ through the core. From ~15 to 11.9 cal ka BP, the assemblages are dominated by up to 100% Brigantedinium spp., but may include up to 18% Islandinium minutum. From ~11.9 to 7.6 cal ka BP, Brigantedinium spp. is significantly reduced (<41%) whereas I. minutum increases up to 29%, and several other species occur in significant numbers. They include Operculodinium centrocarpum (8–40%), Nematosphaeropsis labyrinthus (7–14%), as well as the cysts of Pentapharsodinium dalei (11–58%), which becomes the dominant species by ~8.5 cal ka BP. I. minutum nearly disappears after ~7.6 cal ka BP and the cysts of P. dalei (4–36%) also decrease, while O. centrocarpum (43–75%), accompanied by N. labyrinthus (5–33%) becomes the dominant species until ~2.4 cal ka BP. After ~2.4 cal ka BP, O. centrocarpum (down to 20%) is replaced by N. labyrinthus (up to 63%) as the dominant species. The other accompanying taxa have low relative abundance (<10%) throughout the core.

The first component (PC1) of the PCA identified four assemblage zones in CC04, which account for 79.2% of the total variance, while PC2 accounted for 8.9% (Figure 3a). The lower interval, from ~15 to 11.9 cal ka BP, is characterized by the quasi-exclusive dominance of Brigantedinium spp., a heterotrophic taxon often associated with harsh, ice-covered conditions (de Vernal et al., 1997, 2001, 2013a, 2013b; Gibb et al., 2014; Rochon et al., 1999). The second interval, from ~11.9 to 7.6 cal ka BP, marks the increase in I. minutum and the occurrence of phototrophic species such as O. centrocarpum and N. labyrinthus, which are often associated with the North Atlantic Drift (Rochon et al., 1999). The cyst of P. dalei, as the dominant species in this interval, suggests large seasonal gradients of temperature due to strongly stratified surface waters (Rochon et al., 1999; Solignac et al., 2006). The interval from ~7.6 to 2.4 cal ka BP is dominated by O. centrocarpum indicating warmer, temperate conditions. The upper interval from 2.4 cal ka BP to present is characterized by the dominance of N. labyrinthus.

The sea surface conditions in CC04 appear to vary significantly throughout the record (Figure 3b). To clearly display the trends in variability, a 5-point running mean was added to the sea surface reconstructions. Three of the four PCA assemblage zones are identifiable in these reconstructions. Overall, winter SST ranged from −1.8°C to 5.3°C, summer SST from −0.8°C to 13.3°C, salinity from 28.4 to 35.0, sea ice cover between 0 and 11.4 months yr⁻¹, and productivity between 61 and 374 gC m⁻². The interval between ~15 and 11.9 cal ka BP records cold, nearly perennially ice-covered conditions. Between ~11.9 and 7.6 cal ka BP, winter SSTs increase slightly to an average of 0.8°C, while summer SSTs greatly increase to an average of 9°C. This interval recorded (1) the lowest salinity with a minimum of 28.4 at around 10.4 cal ka BP, (2) the highest annual productivity, and (3) a reduced sea ice cover averaging about 3 months yr⁻¹. After ~7.6 cal ka BP, the salinity increases to ~34 and winter SSTs increase slightly to an average of 3.5°C. Summer SSTs decrease slightly, yet fluctuate greatly between 6°C and 10°C.

TWC08 – Davis Strait

Core TWC08 consists of massive dark grayish brown silty clays with some mud clasts and high proportions of IRD, which peak to 50% at 35 cm (Figure 4b). The core has a calibrated age at 104–105 cm of 19.4 cal ka BP and one at 136–137 cm of 20.6 cal ka BP, suggesting high sedimentation rates during the glacial period. This core is affected by carbonate dissolution above these intervals and therefore no Holocene age-model was derived. Chronological interpretation was made from the piston core collected with TWC08 (PC08) by Andrews et al. (2014) who discussed in detail the geochemical data used to correlate TWC08 and PC08 with other dated records from the area. The cores are primarily correlated using sediment provenance of mineralogical and carbonate contents. The most recent detrital carbonate event (DC0) found between 60 and 25 cm in TWC08 and at the core top of PC08 was tentatively correlated with other records and assigned an age from approximately 13 to 10.5 cal ka BP (Andrews et al., 2012, 2014; Simon et al., 2014). The maximum percentage of IRD in TWC08 at 34 cm (Figure 4b) indicates a large influx of icebergs into southern Davis Strait. This peak occurs during DC0, suggesting icebergs were exiting Baffin Bay between about 13 and 10.5 cal ka BP. Considering the lack of biogenic carbonate and suggested chronology by Andrews et al. (2014), it is reasonable to assume that the Holocene is comprised within the upper 34 cm of TWC08.

Dinocyst concentrations are very low (<500 cysts cm⁻³) in the lower interval analyzed, which likely precedes the Holocene (Figure 4a). They increase from 35.5 to 22 cm to reach values of 5000 cysts cm⁻³ and fluctuate between 5000 and 10,000 cysts cm⁻³ in the upper part of the core. The dinocyst assemblages are dominated by the heterotrophic species I. minutum (24–57%) and Brigantedinium spp. (10–65%), which have their highest abundances below 35.5 cm. In the upper 35.5 cm of the core, there is a change in relative abundance in the accompanying taxa. The cysts of P. dalei (0–30%) and O. centrocarpum (0–17%) are more abundant between 35.5 and 24 cm, and N. labyrinthus (0–34%) is more abundant above 24 cm. The other accompanying taxa have low relative abundance (<10%) throughout the core.

The PC1 scores change from positive to negative at 35.5 cm and account for 79.9% of the variance (Figure 4a). Similarly to core CC04, the PCA scores are influenced by the shift from heterotrophic to phototrophic species indicating a shift toward warmer conditions. PC2, explaining 11.1% of the variance, is primarily reflecting the change in assemblage at 24 cm from P. dalei and O. centrocarpum to N. labyrinthus.

The MAT results show cold conditions and dense sea ice cover for about 8 months yr⁻¹ below 33.5 cm (Figure 4b). A shift to warmer SSTs (up to 11°C), reduced salinity (down to 30), seasonal sea ice (3.0–6.5 months yr⁻¹), and increased productivity (up to 270 gC m⁻²) is recorded between 33.5 and 24 cm. Above 24 cm, recurring cold summer conditions and a slight increase in sea ice cover are recorded.

CC70 – Eastern Baffin Bay

The CC70 record extends to ~12.8 cal ka BP with sedimentation rates ranging between 11.4 and 101.6 cm ka⁻¹ (Figure 2). The core was previously described by Jennings et al. (2014) for sedimentological content and by St-Onge and St-Onge (2014) for magnetic properties. It depicts four distinct intervals. The base of the composite core (piston core) at ~12.1 cal ka BP (421 cm) to ~11.2 cal ka BP (376 cm) consists of dark gray muds with high amounts of pebbles and sand, and IRD (>106 µm) amounting up to 85% (Figure 5b). Between ~11.2 and 10.6 cal ka BP (376 and 326 cm), the sediment consists of bioturbated silty mud with shells, pebbles, vertical burrows, and Fe-rich dolomites from northern Baffin Bay (see Jennings et al., 2014). IRD and bioturbation are not visible in the interval from ~10.6 to ~9.2 cal ka BP which consists of massive silty clay. After ~9.2 cal ka BP, from 223 cm to the top of the composite core, the sediment consists of bioturbated mud. Two sand intervals are present at ~2.7 cal ka BP (100–110 cm) and ~1.2 cal ka BP (48–52 cm).

The concentrations of dinocysts in core CC70 range from very low (35 cysts cm⁻³) to very high (56,000 cysts cm⁻³) with the lowest values measured at the bottom of the core (Figure 5a). Prior to ~7.4 cal ka BP, the dinocyst assemblage mainly consists of the heterotrophic species I. minutum and Brigantedinium spp. (>90%). O. centrocarpum (50–75%) dominates after ~7.4 cal ka BP with accompanying taxa that include N. labyrinthus, Spiniferites elongates, and the cysts of P. dalei.

Based on the scores of the first two components of the PCA in CC70, two assemblage zones have been identified (Figure 5a). The scores for PC2 (21.5% of the total variance) identify a shift in the relative abundance of I. minutum and Brigantedinium spp. at ~9.5 cal ka BP. The PC1 scores (77.3% of the total variance) mark a shift at ~7.4 cal ka BP, representing the change in dominant species from heterotrophic to O. centrocarpum and other phototrophic taxa. Hence, the change indicates a shift from species which tolerate cold conditions with dense sea ice (de Vernal et al., 1997, 2001, 2013a, 2013b; Gibb et al., 2014; Rochon et al., 1999) to milder conditions associated with the North Atlantic Current (Rochon et al., 1999).

The two shifts identified by the PCA are distinguishable in the sea surface reconstructions, particularly through the salinity and sea ice cover records (Figure 5b). Before ~7.4 cal ka BP, sea surface conditions are reconstructed as cold, with extensive sea ice cover. The shift at ~9.5 cal ka BP distinguishes a regime marked by low salinity (28.0 on average) and quasi-perennial ice cover from a regime with higher salinity (31.7 on average) and seasonal sea ice cover of about 8 months yr⁻¹, without clear change in SSTs. After ~7.4 cal ka BP, there is an increase in summer SST from 2.5°C up to 9.5°C, an increase in salinity toward values between 32 and 34, while sea ice cover decreases to seasonal extent from 1 to 6 months yr⁻¹. Productivity increased slightly to an average of 160 gC m⁻². Although the records show highly variable values, the smoothed data suggest a trend toward milder conditions throughout the Holocene.

TWC16 – Central Baffin Bay

Core TWC16 contains laminated silty clays with abundant IRD (Figure 6b) and mud clasts from the bottom to 16 cm downcore. Above 16 cm, the occurrence of IRD and clasts is significantly diminished. A color change from gray (bottom) to brown (top) occurs at 50 cm. Core TWC16 yielded one calibrated age of 13.3 cal ka BP at 56–57 cm. This was the uppermost possible date due to carbonate dissolution above. Therefore, it is assumed that the upper 56 cm of TWC16 spans the deglacial and postglacial periods in central Baffin Bay.

The samples contain rare dinocysts except in the upper 10 cm, where their concentrations gradually increase to 1260 cysts cm⁻³ (Figure 6a). The dinocyst assemblage is dominated by 70–80% Impagidinium pallidum from the core bottom to 16 cm, where it is replaced by O. centrocarpum. When either of these two species dominates an assemblage marked by low species diversity, the assemblage is associated with cool-cold open ocean conditions (Rochon et al., 1999). In the lower part of the record, I. pallidum is accompanied by I. minutum. The overall assemblage reflects low productivity and harsh conditions (cf. Steinhauer, 2012). In the upper 17 cm of the sequence, the disappearance of heterotrophic species and increase in diversity of phototrophic taxa, which include Spiniferites spp., N. labyrinthus, and Impagidinium sphaericum, illustrate the onset of milder conditions (Rochon et al., 1999).

Two assemblage zones have been identified based on the scores from PC1, which accounts for 92.7% of the total variance (Figure 6a). The zones are divided at 17.5 cm by the shift in dominant species, from I. pallidum to O. centrocarpum, in addition to the occurrence of accompanying taxa.

The sea surface reconstructions demonstrate a regime shift at 17.5 cm (Figure 6b). The lower zone corresponds to cold conditions with about 8 months yr⁻¹ of sea ice cover. The conditions warm throughout the upper interval, with a significant reduction in sea ice cover and summer SST and salinity reaching maximums of 9°C and 34.7, respectively.

As pointed out by Steinhauer (2012), the sea surface reconstructions of the upper part of the core do not reflect conditions of recent decades. There is a large discrepancy between reconstructed sea surface conditions from the topmost sample and the modern hydrographic data, which average 3.2°C and −1.1°C in summer and winter respectively, with salinity of 29.9, sea ice cover for 9 months yr⁻¹, and 57 gC m⁻² of productivity (Table 1). Actually, the top centimeter of core TWC016 possibly integrates several hundred years of sedimentation if we take into account the very low postglacial sedimentation rates (Simon et al., 2012) and mixing by bioturbation as shown from ²¹⁰Pb and ¹³⁷Cs data (Steinhauer, 2012). Alternatively, the preservation of dinocysts in the upper part of the core might be questioned since the assemblages are lacking the taxa that are most susceptible to degradation because of oxidation such as Brigantedinium spp. The aerobic degradation of the more labile cysts of heterotrophic species has been documented by Zonneveld et al. (1997, 2001, 2007). However, deep Baffin Bay is currently a low [O₂] environment (3.2–3.5 ml l⁻¹; Aksu, 1983; Schröder-Adams and Van Rooyen, 2011). Although we cannot totally discard the possibility of selective preservation of dinocysts, the occurrence of subpolar taxa leading to the reconstruction of mild conditions is significant. Hence, we do not think that there is a major bias in the reconstruction and we rather consider that the top centimeter does not represent modern conditions as it likely integrates at least several centuries of sedimentation. In this case, the discrepancy between estimated sea surface conditions from dinocyst assemblages and the recent hydrography would reflect a cooling trend during the latest Holocene.

Discussion

Dinocyst assemblages and the application of MAT allowing reconstruction of sea surface conditions reflect the shift from cold, quasi-perennial sea ice cover to warmer conditions with seasonal ice cover at all four sites. According to the chronology in radiocarbon dated cores CC04 and CC70 (Figure 2), this shift toward mild postglacial conditions occurred in a time transgressive manner, around 11.9 cal ka BP in the northwest Labrador Sea (Figure 3b) and 7.4 cal ka BP in Baffin Bay (Figure 5b). The northern Labrador Sea and Davis Strait cores CC04 and TWC08 (Figure 4b) also recorded variation in salinity reflecting changes in freshwater–meltwater discharges during the deglaciation and resulting in changes in oceanographic regimes. These changes are discussed below in reference to the available chronostratigraphic information, to assess the timing of events in the Baffin Bay–Labrador Sea corridor during the Holocene.

Deglaciation and breakup of quasi-perennial sea ice cover

In the northwest Labrador Sea core CC04, very low dinocyst concentrations and the almost exclusive occurrence of the heterotrophic species Brigantedinium spp. and I. minutum characterize the record prior to ~11.9 cal ka BP (Figure 3a). Such an assemblage can be associated with harsh conditions (de Vernal et al., 1997, 2001, 2013a, 2013b; Gibb et al., 2014; Rochon et al., 1999) and led to the reconstruction of very cold sea surface conditions with quasi-perennial sea ice cover (Figure 3b). After 11.9 cal ka BP, the appearance of phototrophic taxa suggests milder conditions that can be associated with the influence of North Atlantic waters. In particular, the occurrence of cysts of P. dalei suggests large seasonal gradients of temperature, from freezing in winter to mild in summer, due to low thermal inertia in stratified surface waters (Rochon et al., 1999; Solignac et al., 2006). The increase in summer SSTs coincides with the onset of relatively warm air temperatures on Baffin Island as reconstructed from chironomid and pollen in lake records (e.g. Briner et al., 2006; Fréchette and de Vernal, 2009). This early Holocene thermal optimum inland occurred during the high insolation phase of the Holocene (Berger and Loutre, 1991) which was also marked by meltwater discharges from the retreating Laurentide Ice Sheet (LIS) that probably resulted in low surface water salinity and stratification of the upper water mass. Deglaciation in the Baffin and Labrador sectors of the LIS accelerated after the Younger Dryas (YD), as recorded regionally at ~12.9–11.6 cal ka BP (Andrews et al., 1995a; Dyke, 2004; Jennings et al., 1996; Kaplan and Miller, 2003). It was accompanied by rapid retreat of the Baffin Island coastal outlet glaciers between 12 and 10 cal ka BP (Briner et al., 2009). During the deglaciation of the northeast LIS, there were a few ice advances including the Gold Cove advance (~11.3–11.0 cal ka BP; Kaufman et al., 1993) and the Noble Inlet advance (~10.0–9.5 cal ka BP; Stravers et al., 1992) in Hudson Strait. There were also ice advances on Baffin Island during the Cockburn substage (~9.5–8.5 cal ka BP; Andrews and Ives, 1978) and possibly the 8.2-ka event (Miller et al., 2005; Young et al., 2012). During the deglaciation, meltwater discharges no doubt played a role in the variability of sea surface conditions notably with regard to salinity and its role on surface water stratification, especially after ~11.9 cal ka BP. The regime shift at about 11.9 cal ka BP is thus represented by the change from a quasi-perennially ice-covered glacial phase to a seasonally ice-covered deglacial phase.

The changes in sea surface conditions reconstructed from the Davis Strait core TWC08 (Figure 4b) are very similar to those of CC04 (Figure 3b), which is consistent given the location of core TW08 on the southern side of the Davis Strait sill in the northernmost part of the Labrador Sea. The TWC08 record can thus be correlated with that of CC04 based on the carbonate peak and IRD corresponding to DC0 (60–25 cm ≈ 13–10.5 cal ka BP; cf. Andrews et al., 2014). On these grounds, it is reasonable to propose that the transition at 34 cm in TWC08 may correspond to the major regime shift recorded at ~11.9 cal ka BP in core CC04.

The sedimentological record of the eastern Baffin Bay core CC70 suggests that the site was proximal to the ice margin from 12.2 to about 11.4 cal ka BP, with high IRD content reflecting ablation by ice calving from the adjacent Greenland Ice Sheet (GIS) margin (Figure 5b; Jennings et al., 2014). Low dinocyst concentration, exclusive dominance of heterotrophic species, the reconstruction of low SST and salinity together suggest low productivity and quasi-perennial sea ice cover likely due to the ablating GIS margin and related meltwater discharge (Figure 5a and b). At ~9.5 cal ka BP, the reconstruction of higher salinity and slight decrease in sea ice cover correspond to the reduction of ice ablation material (cf. Jennings et al., 2014), which indicates the diminution of proximal GIS ablation activities in eastern Baffin Bay. This change occurred after the rapid retreat of the GIS ice margin at the head of Disko Bugt by ~10.3 cal ka BP (Lloyd et al., 2005; Long and Roberts, 2003).

There are other paleoceanographic records from Baffin Bay that permit to estimate the timing of the breakup of permanent sea ice cover. Unfortunately, these records are restricted to the surrounding shelves where sea ice dynamics may differ because of local influences related to the GIS and LIS margins, or to the North Water polynya (e.g. Hamel et al., 2002). The dinocyst assemblages in core P009 from Lancaster Sound in northwest Baffin Bay (Figure 1) recorded the breakup of quasi-perennial ice at ~9.6 cal ka BP (Ledu et al., 2008) while cores 008P and 012P from northern Baffin Bay recorded it around 10.5 cal ka BP (Levac et al., 2001). Indirect proxies other than dinocysts were used for inferences about the transition from very dense to seasonal sea ice cover in Baffin Bay. Benthic foraminifera from northern (cores 008P and 012P) and western Baffin Bay suggest change from perennial to seasonal ice cover between 10.9 and 9.2 cal ka BP (Knudsen et al., 2008; Osterman and Nelson, 1989). Perennial sea ice breakup between 10.3 and 9.2 cal ka BP has been inferred from mollusks, driftwood, and midges (Bennike, 2004; Dyke et al., 1996; Funder 1990; Funder and Weidick, 1991). Finally, glacial geomorphology indicates that the flow of Arctic water through Nares Strait commenced by ~10.6 cal ka BP which fully deglaciated between 9.0 and 8.3 cal ka BP (England et al., 2006; Jennings et al., 2011), and through Lancaster Sound channeling the Canadian Arctic Archipelago starting after the YD to ~9.0 cal ka BP (Dyke, 1999). Although not directly related to sea surface conditions, these proxies suggest that the Baffin Bay coastal regions experienced the breakup of permanent sea ice between 10.5 and 9.0 cal ka BP. The CC70 record marks the breakup in eastern Baffin Bay at ~9.5 cal ka BP which fits within the time frame set by other records.

Transition toward full interglacial conditions around 7.5 cal ka BP

The dominance of phototrophic taxa (O. centrocarpum, N. labyrinthus, and the cysts of P. dalei), which can be associated with a strengthening of the warmer North Atlantic Drift (Rochon et al., 1999), dates from ~7.6 cal ka BP in core CC04 and possibly core TWC08 (Figures 3a and 4a), and from ~7.4 cal ka BP in core CC70 (Figure 5a). The sea surface reconstructions from core CC04 in the northwest Labrador Sea indicate an increase in winter SST, decrease in summer SST, and increase in SSS (Figure 3b). The reduced seasonal gradients in temperature and increase in salinity suggest reduced stratification of the upper water masses likely related to decreased meltwater discharge from the LIS. The regime shift at ~7.6 cal ka BP occurs when more than 90% of the LIS had retreated (Dyke, 2004), after the Hudson Strait deglaciation (Andrews et al., 1995b; Kerwin, 1996) and the subsequent drainage of glacial lakes Agassiz and Objiway (Barber et al., 1999) at 8.4 cal ka BP. However, the shift slightly predates the collapse of the ice cap of Foxe Basin at ~7.0 cal ka BP (Briner et al., 2009; Miller et al., 2005).

The main change in salinity recorded in the northern Labrador Sea that corresponds to reduced stratification coupled with increased advection of North Atlantic water would have fostered convective mixing of the upper water column during winter initiating LSW formation and modern circulation in the northwest North Atlantic (cf. also de Vernal and Hillaire-Marcel, 2006; Hillaire-Marcel et al., 2001). The advection of Atlantic water and reorganization within the Labrador Sea also apparently started at 7.7 cal ka BP off southwest Greenland in cores DA04-41P and -31P (Seidenkrantz et al., 2013), and ~7.6 cal ka BP south of Davis Strait in cores CC04, TWC/P021 (Figure 1; de Vernal et al., 2001, 2013a), P094 (Orphan Knoll; Solignac et al., 2004), and possibly TWC08. Hence, the overall data from northern Labrador Sea and Davis Strait suggest an important regional shift in sea surface conditions at ~7.6 cal ka BP with a transition toward modern postglacial conditions.

The species assemblages and MAT reconstructions also permit to identify a shift at ~7.4 cal ka BP in eastern Baffin Bay core CC70 as an increase in SSTs and further reduction in seasonal extent of sea ice cover (Figure 5a and b). The sea surface conditions after the shift are related to the strengthened Atlantic inflow through Davis Strait into Baffin Bay with the WGC. At this time, the GIS reached maximum postglacial retreat within the Isfjord (Briner et al., 2010; Young et al., 2011), possibly related to the increased strength of the WGC (Holland et al., 2008; Young et al., 2011). Also in core CC70, Jennings et al. (2014) identified foraminiferal species indicative of warmer intermediate water from ~7.5 cal ka BP, which they attributed to an increased influence of Atlantic water and retreat of the GIS and LIS. There are now a few records that have revealed a shift due to surface and subsurface warming related to a strengthened WGC in Baffin Bay around 7.5 cal ka BP, many focusing around Disko Bugt. Similarly to CC70, core MSM343300 at the southwestern edge of Disko Bugt (Figure 1) contains heterotrophic species that dominate the assemblages prior to ~7.3 cal ka BP followed by the appearance of species associated with the North Atlantic water influence (Ouellet-Bernier et al., 2014). Benthic foraminiferal assemblages also recorded the strengthening of the North Atlantic component in subsurface waters in Disko Bugt area, at 7.3 cal ka BP in core MSM343300 (Perner et al., 2013), and between 7.7 and 7.5 cal ka BP in the nearshore cores DA00-04P and -06P (Seidenkrantz et al., 2013) (Figure 1).

The extremely low dinocyst concentrations that characterize most of central Baffin Bay core TWC16 indicate low productivity due to quasi-perennial ice cover (Figure 6a and b). Heterotrophic species that tolerate heavy sea ice formation are present. The most striking feature is the dominance of I. pallidum, which is an oligotrophic species currently found at high abundance in cold, open ocean environments such as the eastern Greenland Sea (Bonnet et al., 2010; de Vernal et al., 2001; Rochon et al., 1999). Accordingly, the sea surface reconstructions with analogues selected from this area reflect very cold conditions and low productivity which is coupled with nearly perennial sea ice cover because of the presence of heterotrophic taxa throughout most of the interval. The shift at 17.5 cm toward warmer SSTs with ice-free conditions at core top is the result in the change from an assemblage dominated by I. pallidum to an assemblage with higher diversity of subpolar taxa. From 17.5 to 8 cm, the change in dinocyst assemblages reflects warming sea surface conditions. For the upper 8 cm of the core, likely because of the occurrence of I. sphaericum, most modern analogues are from the Barents and southeastern Greenland Seas, which led to reconstruct relatively warm and saline sea surface conditions, with summer SSTs of about 8°C, salinity around 34, and sea ice restricted to winter. Such assemblages mark the establishment of full interglacial conditions in Baffin Bay. The change recorded from 17.5 cm in central Baffin Bay might correlate with the shift recorded at 7.4 cal ka BP in eastern Baffin Bay.

Changes in sea surface conditions during the mid- and late-Holocene

In the Baffin Bay cores, the dinocyst assemblage composition above the transition recorded at ca. 7.5 cal ka BP is relatively uniform. However, in core CC04 from the northern Labrador Sea, there is a significant shift in dinocyst assemblage at ~2.4 cal ka BP as clearly shown from the PCA (Figure 3b). It is associated with a change in species dominance, from O. centrocarpum to N. labyrinthus (Figure 3a). However, it does not correspond to a major change in reconstructed sea surface conditions, which only show a slight diminution in summer SSTs and seasonal gradients of temperatures (Figure 3b). In modern sediments of the reference database, N. labyrinthus appears to be cosmopolitan in the North Atlantic, often dominating the assemblage along the IC (Boessenkool et al., 2001; Marret et al., 2004; Rochon et al., 1999). It has also been associated with the mixing of cold Arctic waters of EGC or LC with warm North Atlantic waters of the IC or WGC (Rochon et al., 1999), and was positively correlated with nutrient levels (Devillers and de Vernal, 2000). These conditions correspond to those recorded along the LSSF, which is currently located northwest of CC04 (Figure 1; Belkin et al., 2009). Therefore, the change in the dinocyst assemblages might reflect southward and/or eastward shift of the polar front during the course of the late Holocene with a strengthened BIC. A late Holocene increase in Arctic water outflow through the Canadian Arctic Archipelago into Baffin Bay, including Nares Strait and Lancaster Sound, is supported by slight changes in dinocyst assemblages corresponding to a cooling trend (Ledu et al., 2008; Levac et al., 2001). Therefore, an increase in N. labyrinthus and dinocyst concentrations coupled with fluctuations in sea surface conditions in the northwest Labrador Sea (and possibly other areas along the IC), may be interpreted as reflecting the core’s proximity to the polar front linked to an increase in Arctic water outflow and a strengthened BIC and LC.

One of the most prominent features of late Holocene paleoclimate in the northwest Atlantic is a cooling trend that has been associated with Neoglaciation. Ice advance of the GIS occurred during the late Holocene (Briner et al., 2010; Long and Roberts, 2003). Surface and subsurface cooling along coastal West Greenland over the last ~4 ka have been attributed to varying strength in the Atlantic (IC) versus Arctic (EGC) components of the WGC (Andresen et al., 2010; Erbs-Hansen et al., 2013; Lloyd et al., 2007; Møller et al., 2006; Moros et al., 2006; Ouellet-Bernier et al., 2014; Perner et al., 2013; Seidenkrantz et al., 2007, 2008). Temperature fluctuations in those records were associated with the ‘Medieval Warm Period’ and ‘Little Ice Age’. The sea surface conditions reconstructed at our sites also show fluctuations during the last 3 ka, but they do not necessarily correlate with other records, nor are their trends consistent. In contrast to Neoglacial cooling, the average sea surface conditions in core CC70 increase by approximately 1.5°C after ~3.1 cal ka BP (Figures 6b and 7), suggesting an increased contribution of the Atlantic component to the WGC. Coupled with lower salinities between ~1.9 and 0.9 cal ka BP, the slightly higher SSTs can also suggest summer warming due to increased stratification and reduced thermal inertia in the surface layer. Similarly, reconstructions based on dinocyst assemblages in core MSM343300 indicate phases of warmer temperatures and lower salinities, which are attributed to increased meltwater from the GIS during intervals of warmer WGC flux (Figure 7; Ouellet-Bernier et al., 2014). The variations in sea surface conditions in CC70 do not seem to be perfectly in phase with those recorded in core MSM343300, which is more directly influenced by the WGC.

Figure 7.

Reconstructions of sea surface temperature and salinity in summer, sea ice cover, and annual productivity from dinocyst assemblages of cores CC70 (blue) and MSM343300 (red; Ouellet-Bernier et al., 2014) plotted versus age (cal ka BP). The thick black lines represent a 5-point running mean. Ecostratigraphic units for core CC70 identified by the principal component analyses and sea surface reconstructions are separated by a horizontal dashed line.

Actually, the SSS records from these two areas appear to be negatively correlated (Figure 7), suggesting that fluctuations in the salinity nearshore may have little influence offshore. Core CC70 is located ~155 km west of the Greenland coast within the BIC–WGC frontal zone (Figure 1; Belkin et al., 2009; Curry et al., 2011; Fisheries and Oceans Canada, 2012) with a modern summer SST and salinity of 3.6°C and 32.8 respectively, and 6.2 months yr⁻¹ of sea ice cover (Table 1). Its postglacial assemblages are characterized by phototrophic taxa and dominant O. centrocarpum (Figure 6a), which is typical of open ocean conditions. In contrast, core MSM343300 is located ~15 km west of the Greenland coast on the southern flank of the opening of Disko Bugt (Figure 1) with modern summer SST and salinity of 4.4°C and 32.9 respectively and about 4 months yr⁻¹ of sea ice cover. Hence dinocyst assemblages are largely dominated by I. minutum and reflect a neritic environment. The reconstructed sea surface conditions from core CC70 throughout the postglacial appear to be about 5°C colder, 1 unit of salinity higher, and characterized by slightly more extensive sea ice cover than those from core MSM343300 (Figure 7). Thus, the surface waters at the coastal site (MSM343300) show greater warming in summer because of lower thermal inertia and the more direct influence of the WGC (and EGC) than sites offshore toward central Baffin Bay (CC70), which are under the influence of the cyclonic ocean gyre mixing BIC with WGC and being responsible for lower stratification.

Conclusion

Despite difficulties to establish chronostratigraphical schemes, poor preservation due to biogenic carbonates, and low postglacial sedimentation rates, the Holocene records from Baffin Bay, Davis Strait, and northern Labrador Sea presented here provide some insight into the postglacial paleoceanographical reorganization in the northwest North Atlantic. These records are mostly based on dinocyst assemblages. They illustrate strong regionalism in the dinocyst distribution, but yield a consistent picture of the main paleoceanographical changes. South of Davis Strait, the breakup of perennial sea ice and onset of relatively warm sea surface conditions began ~11.9 cal ka BP with dinocyst records that indicate the strengthening of the Atlantic component of the WGC. In Baffin Bay, the breakup of perennial sea ice began after ~9.5 cal ka BP and the regime shift to warmer SSTs occurred later, at about 7.5 cal ka BP. The onset of warmer conditions in Baffin Bay seems also to be a response to a strengthened North Atlantic component of the WGC and to relate with the final retreat of the LIS and GIS. Therefore, the early to mid-Holocene optimum, identified as the intensification of the North Atlantic Current (de Vernal and Hillaire-Marcel, 2006), occurred diachronously north of the Labrador Sea, in Baffin Bay, the delayed establishment of full postglacial conditions being likely related to late glacial retreat in the northeast LIS possibly caused by limited exchanges through the narrow Davis Strait. The timing of changes in Baffin Bay is coherent with reduction of meltwater–freshwater export to the Labrador Sea, which likely contributed to higher salinity and density favorable for convective mixing and the formation of LSW. Beyond the major change recorded at around 7.5 cal ka BP, paleoceanographic assessment from dinocyst assemblages in cores from the Baffin Bay corridor suggests that sea surface conditions experienced spatial and temporal variability throughout the postglacial, responding to large-scale changes in the open oceanic regime. This variability seems also to be related to fluctuations in the position of the polar front that constitutes a boundary between inflowing North Atlantic water and outflowing Arctic water.

Footnotes

Acknowledgements

This paper is a contribution to the Past4Future project of the 7th Framework Program of the European Commission. Special thanks to Maryse Henry for her help and expertise in the Micropaleontology and Marine Palynology Laboratory – GEOTOP, and to Helen Gillespie, Susan Fudge, and Danny Boyce at Memorial University for their in kind support of laboratory equipment. We are also grateful to the reviewers who provided useful and constructive comments on the manuscript.

Funding

Support from the Ministère du Développement Économique, Innovation et Exportation (MDEIE) and Fonds Québécois de Recherche sur la Nature et les Technologies (FQRNT) is acknowledged. Special thanks to the Canadian Foundation for Climate and Atmospheric Sciences (CFCAS), Natural Resources Canada (NRCan), and the Natural Sciences and Engineering Research Council of Canada (NSERC) for their financial support of the HU2008029 expedition in the Labrador Sea.

References

Aksenov

Bacon

Coward

. (2010) Polar outflow from the Arctic Ocean: A high resolution model study. Journal of Marine Systems 83(1): 14–37.

Aksu

(1983) Holocene and Pleistocene dissolution cycles in deep-sea cores of Baffin Bay and Davis Strait: Palaeoceanographic implications. Marine Geology 53(4): 331–348.

Andresen

McCarthy

Dylmer

. (2010) Interaction between subsurface ocean waters and calving of the Jakobshavn Isbræ during the late Holocene. The Holocene 21(2): 211–224.

Andrews

Ives

(1978) ‘Cockburn’ nomenclature and the Late Quaternary history of the eastern Canadian Arctic. Arctic Alpine Research 10: 617–633.

Andrews

Barber

Jennings

. (2012) Varying sediment sources (Hudson Strait, Cumberland Sound, Baffin Bay) to the NW Labrador Sea slope between and during Heinrich events 0 to 4. Journal of Quaternary Science 27: 475–484.

Andrews

Gibb

Jennings

. (2014) Variations in the provenance of sediment from ice sheets surrounding Baffin Bay during MIS 2 and 3 and export to the Labrador Shelf Sea: Site HU2008029-0008 Davis Strait. Journal of Quaternary Science 29(1): 3–13.

Andrews

Jennings

Kerwin

. (1995a) A Heinrich-like event, H-0 (DC-0): Source (s) for detrital carbonate in the North Atlantic during the Younger Dryas Chronozone. Paleoceanography 10(5): 943–952.

Andrews

MacLean

Kerwin

. (1995b) Final stages in the collapse of the Laurentide Ice Sheet, Hudson Strait, Canada, NWT: 14 C AMS dates, seismics stratigraphy, and magnetic susceptibility logs. Quaternary Science Reviews 14(10): 983–1004.

Azetsu-Scott

Clarke

Falkner

K et al.

(2010) Calcium carbonate saturation states in the waters of the Canadian Arctic Archipelago and the Labrador Sea. Journal of Geophysical Research: Oceans 115(C11). Available at: http://onlinelibrary.wiley.com/doi/10.1029/2009JC005917/abstract.

10.

Barber

Dyke

Hillaire-Marcel

. (1999) Forcing of the cold event of 8,200 years ago by catastrophic drainage of Laurentide lakes. Nature 400(6742): 344–348.

11.

Belkin

Cornillon

Sherman

(2009) Fronts in large marine ecosystems. Progress in Oceanography 81(1): 223–236.

12.

Bennike

(2004) Holocene sea-ice variations in Greenland: Onshore evidence. The Holocene 14(4): 607–613.

13.

Berger

Loutre

(1991) Insolation values for the climate of the last 10 million years. Quaternary Science Reviews 10(4): 297–317.

14.

Boessenkool

Van Gelder

Brinkhuis

. (2001) Distribution of organic-walled dinoflagellate cysts in surface sediments from transects across the Polar Front offshore southeast Greenland. Journal of Quaternary Science 16(7): 661–666.

15.

Bonnet

de Vernal

Hillaire-Marcel

. (2010) Variability of sea-surface temperature and sea-ice cover in the Fram Strait over the last two millennia. Marine Micropaleontology 74: 59–74.

16.

Briner

Davis

Miller

(2009) Latest Pleistocene and Holocene glaciation of Baffin Island, Arctic Canada: Key patterns and chronologies. Quaternary Science Reviews 28(21): 2075–2087.

17.

Briner

Michelutti

Francis

. (2006) A multi-proxy lacustrine record of Holocene climate change on northeastern Baffin Island, Arctic Canada. Quaternary Research 65(3): 431–442.

18.

Briner

Stewart

HAM

Young

. (2010) Using proglacial-threshold lakes to constrain fluctuations of the Jakobshavn Isbræ ice margin, western Greenland, during the Holocene. Quaternary Science Reviews 29(27): 3861–3874.

19.

Bronk Ramsey

(2008) Deposition models for chronological records. Quaternary Science Reviews 27(1): 42–60.

20.

Buch

(1990/2000). A Monograph on the Physical Oceanography of the Greenland Waters. Greenland Fisheries Research Institute (Report) (reissued in 2000 as Danish Meteorological Institute, Scientific report, 00-12), 405 pp.

21.

Campbell

de Vernal

(2009) CCGS Hudson expedition 2008029: Marine geology and paleoceanography of Baffin Bay and adjacent areas Nain NL to Halifax NS 28 August–23 September. Open file 5989. Geological Survey of Canada. Available at: http://www.geotop.ca/upload/files/publications/rapport_mission/Hudson_2008-029_cruise_report2.pdf.

22.

Cheng

Rhines

(2004) Response of the overturning circulation to high-latitude fresh-water perturbations in the North Atlantic. Climate Dynamics 22: 359–372.

23.

Cuny

Rhines

Kwok

(2005) Davis Strait volume freshwater and heat fluxes. Deep Sea Research Part I: Oceanographic Research Papers 52(3): 519–542.

24.

Cuny

Rhines

Niiler

. (2002) Labrador Sea boundary currents and the fate of the Irminger Sea Water. Journal of Physical Oceanography 32(2): 627–647.

25.

Curry

Lee

Petrie

(2011) Volume Freshwater and Heat Fluxes through Davis Strait, 2004–05*. Journal of Physical Oceanography 41(3): 429–436.

26.

De Vernal

Hillaire-Marcel

(2006) Provincialism in trends and high frequency changes in the northwest North Atlantic during the Holocene. Global Planetary Change 54(3): 263–290.

27.

de Vernal

Henry

Bilodeau

(1999) Technique de préparation et d’analyse en micropaléontologie. Les Cahiers du GEOTOP Université du Québec à Montréal 3, unpublished report.

28.

de Vernal

Larouche

Richard

PJH

(1987) Evaluation of palynomorph concentrations: Do the aliquot and the marker-grain methods yield comparable results? Pollen et Spores 29(2–3): 291–303.

29.

de Vernal

Bilodeau

Hillaire-Marcel

. (1992) Quantitative assessment of carbonate dissolution in marine sediments from foraminifer linings vs shell ratios: Davis Strait northwest North Atlantic. Geology 20(6): 527–530.

30.

de Vernal

Eynaud

Henry

. (2005) Reconstruction of sea surface conditions at middle to high latitudes of the Northern Hemisphere during the Last Glacial Maximum (LGM) based on dinoflagellate cyst assemblages. Quaternary Science Reviews 24(7): 897–924.

31.

de Vernal

Henry

Matthiessen

. (2001) Dinoflagellate cyst assemblages as tracers of sea surface conditions in the northern North Atlantic Arctic and sub-arctic seas: The new ‘n=677’ database and application for quantitative paleoceanographical reconstruction. Journal of Quaternary Science 16(7): 681–699.

32.

de Vernal

Hillaire-Marcel

Rochon

. (2013a) Dinocyst-based reconstructions of sea ice cover concentration during the Holocene in the Arctic Ocean the northern North Atlantic Ocean and its adjacent seas. Quaternary Science Reviews 79: 111–121.

33.

de Vernal

Rochon

Fréchette

. (2013b) Reconstructing past sea ice cover of the Northern hemisphere from dinocyst assemblages: Status of the approach. Quaternary Science Reviews 79: 122–134.

34.

de Vernal

Rochon

Turon

. (1997) Organic-walled dinoflagellate cysts: Palynological tracers of sea-surface conditions in middle to high latitude marine environments. Geobios 30(7): 905–920.

35.

Devillers

de Vernal

(2000) Distribution of dinoflagellate cysts in surface sediments of the northern North Atlantic in relation to nutrient content and productivity in surface waters. Marine Geology 166(1): 103–124.

36.

Dyke

(1999) Last glacial maximum and deglaciation of Devon Island, Arctic Canada: Support for an Innuitian Ice Sheet. Quaternary Science Reviews 18(3): 393–420.

37.

Dyke

(2004) An outline of the deglaciation of North America with emphasis on central and northern Canada. In: Ehlers

Gibbard

(eds) Quaternary Glaciations: Extent and Chronology – Developments in Quaternary Science, vol 2b (part II: North America). Amsterdam: Elsevier, pp. 373–424.

38.

Dyke

Dale

McNeely

(1996) Marine molluscs as indicators of environmental change in glaciated North America and Greenland during the last 18, 000 years. Géographie physique et Quaternaire 50(2): 125–184.

39.

England

Atkinson

Bednarski

. (2006) The Innuitian Ice Sheet: Configuration, dynamics and chronology. Quaternary Science Reviews 25(7): 689–703.

40.

Erbs-Hansen

Knudsen

Olsen

. (2013) Paleoceanographical development off Sisimiut, West Greenland, during the mid-and late Holocene: A multiproxy study. Marine Micropaleontology 102: 79–97.

41.

Fisheries and Oceans Canada (2012) Ice Navigation in Canadian Waters Icebreaking Program. Maritime Services Canadian Coast Guard Fisheries and Oceans Canada 53 p.

42.

Fréchette

de Vernal

(2009) Relationship between Holocene climate variations over southern Greenland and eastern Baffin Island and synoptic circulation pattern. Climate of the Past 5: 347–359.

43.

Funder

(ed.) (1990) Late Quaternary Stratigraphy and Glaciology in the Thule Area – Meddelelser om Gronland: Geoscience, vol. 22, 63 pp.

44.

Funder

Weidick

(1991) Holocene boreal molluscs in Greenland – Palaeoceanographic implications. Palaeogeography, Palaeoclimatology, Palaeoecology 85(1): 123–135.

45.

Gibb

Hillaire-Marcel

de Vernal

(2014) Oceanographic regimes in the northwest Labrador Sea since Marine Isotope Stage 3 based on dinocyst and stable isotope proxy records. Quaternary Science Reviews 92: 269–279.

46.

Goosse

Fichefet

Campin

J-M

(1997) The effects of the water flow through the Canadian Archipelago in a global ice–ocean model. Geophysical Research Letters 24(12): 1507–1510.

47.

Hamel

de Vernal

Gosselin

. (2002) Organic-walled microfossils and geochemical tracers: Sedimentary indicators of productivity changes in the North Water and northern Baffin Bay (High Arctic) during the last centuries. Deep Sea Research II 49: 5277–5295.

48.

Head

Harland

Matthiessen

(2001) Cold marine indicators of the late Quaternary: The new dinoflagellate cyst genus Islandinium and related morphotypes. Journal of Quaternary Science 16(7): 621–636.

49.

Hillaire-Marcel

de Vernal

Bilodeau

. (2001) Absence of deep water formation in the Labrador Sea during the last interglacial period. Nature 410: 1073–1077.

50.

Holland

Thomas

de Young

. (2008) Acceleration of Jakobshavn Isbrae triggered by warm subsurface ocean waters. Nature Geoscience 1(10): 659–664.

51.

Jennings

Sheldon

Cronin

. (2011) The Holocene history of Nares Strait: Transition from glacial bay to Arctic-Atlantic throughflow. Oceanography 24(3): 26–41.

52.

Jennings

Tedesco

Andrews

. (1996) Shelf erosion and glacial ice proximity in the Labrador Sea during and after Heinrich events (H-3 or 4 to H-0) as shown by foraminifera. Geological Society, London, Special Publications 111(1): 29–49.

53.

Jennings

Walton

Ó Cofaigh

. (2014) Paleoenvironments during Younger Dryas-Early Holocene retreat of the Greenland Ice Sheet from outer Disko Trough central west Greenland. Journal of Quaternary Science 29(1): 27–40.

54.

Kaplan

Miller

(2003) Early Holocene delevelling and deglaciation of the Cumberland Sound region, Baffin Island, Arctic Canada. Geological Society of America Bulletin 115: 445–462.

55.

Kaufman

Miller

Stravers

. (1993) Abrupt early Holocene (9.9–9.6 ka) ice-stream advance at the mouth of Hudson Strait, Arctic Canada. Geology 21(12): 1063–1066.

56.

Kerwin

(1996) A regional stratigraphic isochron (ca. 8000¹⁴C yr BP) from final deglaciation of Hudson Strait. Quaternary Research 46(2): 89–98.

57.

Knudsen

Stabell

Seidenkrantz

M-S

. (2008) Deglacial and Holocene conditions in northernmost Baffin Bay: Sediments foraminifera diatoms and stable isotopes. Boreas 37(3): 346–376.

58.

Lazier

JRN

(1973) The renewal of Labrador Sea water. Deep Sea Research 20(4): 341–353.

59.

Ledu

Rochon

de Vernal

. (2008) Palynological evidence of Holocene climate change in the eastern Arctic: A possible shift in the Arctic oscillation at the millennial time scale. Canadian Journal of Earth Science 45: 1363–1375.

60.

Levac

de Vernal

Blake

(2001) Sea surface conditions in northernmost Baffin Bay during the Holocene: Palynological evidence. Journal of Quaternary Science 16(4): 353–363.

61.

Lloyd

Kuijpers

Long

. (2007) Foraminiferal reconstruction of mid- to late-Holocene ocean circulation and climate variability in Disko Bugt West Greenland. The Holocene 17: 1079–1091.

62.

Lloyd

Park

Kuijpers

. (2005) Early Holocene palaeoceanography and deglacial chronology of Disko Bugt west Greenland. Quaternary Science Reviews 24(14): 1741–1755.

63.

Long

Roberts

(2003) Late Weichselian deglacial history of Disko Bugt, West Greenland, and the dynamics of the Jakobshavns Isbrae ice stream. Boreas 32(1): 208–226.

64.

Marret

Eirı́ksson

Knudsen

. (2004) Distribution of dinoflagellate cyst assemblages in surface sediments from the northern and western shelf of Iceland. Review of Palaeobotany and Palynology 128(1): 35–53.

65.

Matthews

(1969) The assessment of a method for the determination of absolute pollen frequencies. New Phytologist 68(1): 161–166.

66.

Miller

Wolfe

Briner

. (2005) Holocene glaciation and climate evolution of Baffin Island, Arctic Canada. Quaternary Science Reviews 24(14): 1703–1721.

67.

Møller

Jensen

Kuijpers

. (2006) Late Holocene environmental and climatic changes in Ameralik Fjord Southwest Greenland – evidence from the sedimentary record. The Holocene 16: 685–695.

68.

Moros

Jensen

Kuijpers

(2006) Mid-to late-Holocene hydrological and climatic variability in Disko Bugt central West Greenland. The Holocene 16(3): 357–367.

69.

Osterman

Nelson

(1989) Latest Quaternary and Holocene paleoceanography of the eastern Baffin Island continental shelf Canada: Benthic foraminiferal evidence. Canadian Journal of Earth Sciences 26: 2236–2248.

70.

Ouellet-Bernier

M-M

de Vernal

Hillaire-Marcel

. (2014) Paleoceanographic changes of Disko Bugt area West Greenland during the Holocene. The Holocene 24(11): 1573–1583.

71.

Perner

Moros

Jennings

. (2013) Holocene palaeoceanographic evolution off West Greenland. The Holocene 23(3): 374–387.

72.

Radi

Bonnet

Cormier

M-A

. (2013) Operational taxonomy and (paleo-) autecology of round brown spiny dinoflagellate cysts from the Quaternary of high northern latitudes. Marine Micropaleontology 98: 41–57.

73.

Reimer

Bard

Bayliss

. (2013) IntCal13 and Marine13 radiocarbon age calibration curves 0–50000 years cal BP. Radiocarbon 55(4): 1869–1887.

74.

Ribergaard

Olsen

Mortensen

(2008) Oceanographic Investigations off West Greenland 2007. In: Scientific Council Meeting, June 2008, NAFO SCR Doc. 08/3. Available at: http://archive.nafo.int/open/sc/2008/scr08-003.pdf.

75.

Rochon

de Vernal

Turon

J-L

. (1999) Distribution of dinoflagellate cyst assemblages in surface sediments from the North Atlantic Ocean and adjacent basins and quantitative reconstruction of sea surface parameters. American Association of Stratigraphic Palynologists Contribution Series No 35.

76.

Schröeder-Adams

Van Rooyen

(2011) Response of recent benthic foraminiferal assemblages to contrasting environments in Baffin Bay and the Northern Labrador Sea, Northwest Atlantic. Arctic 317–341.

77.

Seidenkrantz

M-S

Aagaard-Sørensen

Sulsbrück

. (2007) Hydrography and climate of the last 4400 years in a SW Greenland fjord: Implications for Labrador Sea palaeoceanography. Holocene 17(3): 387–401.

78.

Seidenkrantz

M-S

Ebbesen

Aagaard-Sørensen

. (2013) Early Holocene large-scale meltwater discharge from Greenland documented by foraminifera and sediment parameters. Palaeogeography, Palaeoclimatology, Palaeoecology 391: 71–81.

79.

Seidenkrantz

M-S

Roncaglia

Fischel

. (2008) Variable North Atlantic climate seesaw patterns documented by a late Holocene marine record from Disko Bugt West Greenland. Marine Micropaleontology 68: 66–83.

80.

Serreze

Barrett

Slater

. (2006) The large-scale freshwater cycle of the Arctic. Journal of Geophysical Research 111(C11): C11010.

81.

Simon

Hillaire-Marcel

St-Onge

. (2014) North-eastern Laurentide, western Greenland and southern Innuitian ice stream dynamics during the last glacial cycle. Journal of Quaternary Science 29(1): 14–26.

82.

Simon

St-Onge

Hillaire-Marcel

(2012) Late Quaternary chronostratigraphic framework of deep Baffin Bay glaciomarine sediments from high-resolution paleomagnetic data. Geochemistry, Geophysics, Geosystems 13(11). Available at: http://onlinelibrary.wiley.com/doi/10.1029/2012GC004272/abstract.

83.

Solignac

de Vernal

Hillaire-Marcel

(2004) Holocene sea surface conditions in the North Atlantic-contrasted trends and regimes between the eastern and western sectors (Labrador Sea vs Iceland Basin). Quaternary Science Reviews 23(3): 319–334.

84.

Solignac

Giraudeau

de Vernal

(2006) Holocene sea surface conditions in the western North Atlantic: Spatial and temporal heterogeneities. Paleoceanography 21(2): PA2004.

85.

Srivastava

Arthur

Clement

. (eds) (1989) Proceedings of the Ocean Drilling Program Scientific Results, vol. 105. College Station, TX: Ocean Drilling Program.

86.

Steinhauer

(2012) Postglacial paleoceanography of central Baffin Bay from palynological tracers. Master’s Thesis, Université du Québec à Montréal.

87.

St-Onge

M-P

St-Onge

(2014) Environmental changes in Baffin Bay during the Holocene based on the physical and magnetic properties of sediment cores. Journal of Quaternary Science 29(1): 41–56.

88.

Stravers

Miller

Kaufman

(1992) Late glacial ice margins and deglacial chronology for southeastern Baffin Island and Hudson Strait, eastern Canadian Arctic. Canadian Journal of Earth Sciences 29(5): 1000–1017.

89.

Tang

CCL

Ross

Yao

. (2004) The circulation water masses and sea-ice of Baffin Bay. Progress in Oceanography 63: 183–228.

90.

Wadley

Bigg

(2002) Impact of flow through the Canadian Archipelago and Bering Strait on the North Atlantic and Arctic circulation: An ocean modelling study. Quarterly Journal of the Royal Meteorological Society 128(585): 2187–2203.

91.

Wang

Mysak

Ingram

(1994) Interannual variability of sea-ice cover in Hudson Bay Baffin Bay and the Labrador Sea. Atmosphere-ocean 32(2): 421–447.

92.

Yashayaev

(2007) Hydrographic changes in the Labrador Sea 1960–2005. Progress in Oceanography 73: 242–276.

93.

Young

Briner

Rood

. (2012) Glacier extent during the Younger Dryas and 8.2-ka event on Baffin Island, Arctic Canada. Science 337(6100): 1330–1333.

94.

Young

Briner

Stewart

. (2011) Response of Jakobshavn Isbræ, Greenland, to Holocene climate change. Geology 39(2): 131–134.

95.

Zonneveld

Bockelmann

Holzwarth

(2007) Selective preservation of organic-walled dinoflagellate cysts as a tool to quantify past net primary production and bottom water oxygen concentrations. Marine Geology 237(3): 109–126.

96.

Zonneveld

KAF

Versteegh

GJM

de Lange

(1997) Preservation of organic walled dinoflagellate cysts in different oxygen regimes: A 10000 years natural experiment. Marine Micropaleontology 29: 393–405.

97.

Zonneveld

KAF

Versteegh

GJM

de Lange

(2001) Palaeoproductivity and post-depositional aerobic organic matter decay reflected by dinoflagellate cyst assemblages of the Eastern Mediterranean S1 sapropel. Marine Geology 172: 181–195.