Sea-Surface Characteristics of the Newfoundland Basin of the Northwest Atlantic Ocean during the Last 145,000 Years: A Study Based on the Sedimentological and Paleontological Proxies

: Dramatic changes occur in the sea-surface characteristics (i.e., temperature and salinity) and freshwater input due to the interaction of cold and fresh Labrador Current and warm and salty North Atlantic Current (NAC) on the southeast Grand Banks. As a result, the biological productivity and seasonal stratiﬁcation of the upper water masses are intensiﬁed. Such changes must have been more dramatic during the glacial times due to the penetration of the Polar and Arctic fronts and southward migration of the Gulf Stream/NAC. However, the extent to which such changes impacted the sea-surface characteristics in the Newfoundland Basin is poorly known. We report changes in the sea-surface characteristics using a piston core (Hu9007-08) collected from the Milne seamount during the last 145,000 years. Heinrich layers H1, H2, H4, and H5 and H11 within the MIS3 and at the penultimate deglaciation were identiﬁed by the ice-rafted detritus (IRD) and Neogloboquadrina pachyderma peaks and lighter oxygen isotopes. Rapid turnover by the foraminiferal species with distinct depth habitats and ecological niches in the mixed-layer and thermocline suggests an interplay between the polar and subpolar water masses during the Heinrich and non-Heinrich periods. Only two North Atlantic-wide cooling events, C24 and C21, in which the latter event linked to the minor IRD event during the marine isotope stage (MIS) 5 in Hu90-08, compared to the eight events in the eastern subpolar gyre (e.g., ODP site 984). Millennial-scale N. pachyderma variability in the western subpolar gyre appears to be absent in the eastern subpolar gyre during the MIS3 suggesting the occasional presence of salty and warm water by the NAC inﬂow, implying a different climate state between the western and eastern subpolar gyre. Although T. quinqueloba data are fragmentary, there are differences between the western and eastern subpolar gyre in addition to the differences within the western subpolar gyre during MIS5 that might imply a variable inﬂuence by the subpolar water. This ﬁnding suggests that the inﬂuence by the NAC outweighs the impact of cold and fresh polar water in the northern northwest Atlantic during the MIS5.


Introduction
Due to the periodic expansions and contractions of the Laurentide Ice Sheet (LIS) during the last glacial cycle, enormous amounts of icebergs, sediments, and meltwater were discharged to the neighboring Labrador Sea and the North Atlantic [1]. These discharge events, known as the Heinrich events (H-events), are recorded as ice-rafted detritus (IRD) layers in the northwest Labrador Sea [2,3] and open North Atlantic [4,5], were the extreme manifestation of cold, abrupt climate events [6][7][8]. Employing a suite of sedimentological

Oceanographic Setting
The Newfoundland Basin lies on the southeastern Grand Banks of the Northwest Atlantic Ocean (Figure 1), a key area where the exchange between subpolar and subtropical gyres takes place. The Labrador Current brings southward flowing cold and freshwater, which is seasonally replaced by northward-flowing warm and salty water of the North Atlantic Current (NAC). These two powerful surface currents traverse on the west and east side of the Newfoundland Basin. The spatial interplay of these currents determines the SST and sea-surface salinity of the Basin. The Labrador Current is the continuation of the Western Greenland and Baffin Island currents, with average annual flow speeds of 25-35 cm/s [28]. The cold Labrador Current sweeps the entire northwest Atlantic continen-tal margin ( Figure 1) [1,29]). In contrast, the NAC is the continuation of the Gulf Stream (e.g., [30]) that flows through the Newfoundland Basin and transports warm and salty water to the northeast Atlantic, typically the strongest between 42 • and 39.5 • W [31]. At the southeastern tip of the Grand Banks of Newfoundland, part of the NAC recirculates or continues in an eastward direction, while the remaining fraction turns northward. It is also a critical region where the modern Labrador Current and NAC frequently interact.
The SST and salinity of the Northwest Atlantic Ocean, particularly in the Newfoundland Basin, greatly vary. The annual SSTs vary between 7 • and 17.5 • C in the Newfoundland Basin. When the influence of the cold Labrador Current increases, the impact on the southeast Grand Banks of the northwestern edge of the Newfoundland Basin expands, and the SSTs significantly decrease ( Figure 1). Conversely, when the influence of the warm and salty NAC increases, the SSTs on the southeastern edge of the Newfoundland Basin increase resulting in a decrease in temperatures between the eastern and western sides of the Basin. It can be seen that the sea-surface of the Newfoundland Basin is mainly affected by the warm and salty NAC in the summer, and the cold Labrador Current has a smaller range. Down-depth SST plot ( Figure 2b) suggests that the temperature rapidly decreases to 7 • C at~500 m in the west, whereas the 7 • C isotherm crops out on the southeastern Grand Banks reflecting the spatial extent of the Labrador Current. The spatial impact of the Labrador Current and North Atlantic Current is also demonstrated by the salinity gradient ( Figure 2c) between the eastern and western Newfoundland Basin. General oceanographic features of the modern northwest Atlantic Ocean. The mean annual sea-surface temperature with 5 • C contour demarcates between the subpolar and subtropical gyres. The location of the cores used in this study is listed in Table 1. Note that "white star" denotes the location of core Hu90-08 (this study). The tabular box with the discontinuous line shows the location of  Basin. The location of the published records of a few cores used in the study is shown by the yellow dots, whereas the position of the sediment core Hu9007-08 (this study) is shown by the red star. (b) sea-surface temperature distribution on the SE Grand Banks on the X-Y transect is shown in Figure 2a. (c) sea-surface salinity along the X-Y section, which demonstrates the low salinity Labrador Current (west), whereas the position of the North Atlantic Current is shown by the high salinity (east). Note LC = Labrador Current; GS = Gulf Stream; NAC = North Atlantic Current; FC = Flemish Cap; and FP = Flemish Pass. Figure 2b,c were plotted using the Ocean Data View software [32].

Sediment Samples
The core Hu9007-08 (hereafter Hu90-08; 44.5329 • N and 40.3855 • W) was retrieved at 3522 m water depth top of Milne seamount from the eastern Newfoundland Basin of the northwest Atlantic Ocean by the joint Memorial University of Newfoundland and Geological Survey of Canada cruise. The core site is located near the western edge of the IRD-belt (Figure 1), too far from the eastern Canadian continental margin to be influenced by the mass-transport deposits [1]. Therefore, the core site is an ideal location to study past changes in the sea-surface conditions. Core Hu90-08 is 11 m long, but the top 841.5 cm was used for this study, which was sampled at 10 cm intervals. However, the MIS3 period was sampled at 5 cm intervals to improve the temporal resolution (i.e., 400 to 1000 years). Sediments were freeze-dried, weighed, and soaked in deionized water for 24 h to disperse. Samples were then ultrasonicated (up to 5 s) while stirring, then rinsed through a 63 µm diameter sieve. The washed residues were dried using an oven at 60 • C and stored in glass vials. Sediments <63 µm were allowed to stand for 24 h, and the supernatant was poured out and dried, and stored for future analysis.

IRD and Planktonic Foraminifer Count
The >63 µm fractions were sieved again with a 150 µm sieve and weighed and were used to identify and count planktonic foraminifera and IRD. Each sample was divided equally about 7-9 times by a splitter when the total number of planktonic foraminifer complete shells is within the range of 300-600. Using a micropaleontological tray, foraminifers were identified and counted, and IRD was also counted [5,40]. After completing foraminiferal assemblages count, the statistical treatment such as absolute abundance, percentage of each species, trends, etc., was carried out.

Oxygen and Carbon Isotopes Determination
Five to 10 clean and unstained N. pachyderma in the 150-250 µm size fractions were used to determine δ 18 O and δ 13 C following the method described by Simstich et al. [41] and Wu and Hillaire-Marcel [42]. A Finnigan MAT253 isotope ratio mass spectrometer with a Kiel III device was used for isotopic analysis at the University of Florida [43]. The overall analytical reproducibility, as determined from replicate measurements on carbonate standards NBS-18, NBS-19, and an internal standard, is routinely better than ±0.08‰ (±1σ) for δ 18 O [1]. Vital effects were corrected following the method of Stangeew [44] in which the equation, 0.7‰; ∆δ 18 O (vital effect) = δ 18 O Eq − δ 18 O shell , was used for the modern oceanographic setting; however, accurate corrections must be different due to the imposition of the Arctic and Polar fronts during the glacial periods. In any case, the δ 18 O were used to construct the stratigraphy in which the marine isotope stages (MISs) were identified.

14 C-AMS Dating
Five samples at different depths (Table 2) were selected in which each sample weigh >5000 µg of intact and clean N. pachyderma for 14 C Accelerator Mass Spectrometry (AMS) dating. Samples were sent to the W. M. Keck Carbon Cycle AMS Facilities at the University of California in Irvine, USA. The 14 C-AMS dates were calibrated to the calendar year before the present (1950) using the latest version of the radiocarbon calibration program CALIB 8.10 [45] in which the Marine20 reservoir ages were used [46].

Age Model Construction
The age model of the core Hu90-08 was constructed by employing two methods: (1) Using five 14 C-AMS dates between 10.5 and 311 cm, and (2) constructing the oxygen isotope stratigraphy using the δ 18 O in N. pachyderma (Figure 3a) in which MISs were clearly identified. Fourteen tie points older than 40 ka ( Figure 3) were obtained through matching with the deep North Atlantic benthic δ 18 O stack of Lisiecki and Stern [47]. A total of 19 control points were used to construct the age model in which linear sedimentation rates between the 14 C-AMS dates/tie points are considered. As a result, it appears that the core Hu90-08 contains climate records between 3.54 and 145 ka for the depth interval between 0 and 841.5 cm (Figure 4). The estimated sedimentation rates vary between 1.62 and 8.5 cm/year, in which the Holocene period records the lowest sedimentation rates.  (Table 2) were obtained by visually aligning the δ 18 O values of Hu90-08 with those of the δ 18 O values of Lisiecki and Stern [47] and five 14 C-Accelerator Mass Spectrometry (AMS) dates (Table 2) to construct the age model of core Hu90-08. Note that the upward triangles (b) denote the position of 14 C-AMS dates.  Table 2. A linear fit (discontinuous green line) with the R 2 = 0.949815 suggests a robust depth-age relationship, and the core contains climate records between 3.54 and 145 ka from 0 to 841.5 cm. Sedimentation rates range from 1.62 cm/kyr to 8.5 cm/kyr in which the Holocene period shows the lowest sedimentation rates.

Results
The IRD concentration varies between 24.75/g and 6483/g of dry sediments from 3.6 to 127.34 ka broadly divided into two periods (Figure 5b): The MIS5 (76-128 ka) in which the IRD is minimal (758/g); and the MIS2-4, high-frequency variable IRD with high IRD/g most of which are identified as the H-layers [1,48]. pachyderma and the concentration of (b) ice-rafted detritus (IRD/g) and (c) planktonic foraminifera (Foram/g) per gram of dry sediments in core Hu90-08 (this study). The percentage of six dominant foraminifera is plotted as shown from (d-i). Note the y-axis for %N. pachyderma is reversed (see detailed text). Downward filled-headed arrows represent the 14 C-AMS dates used to construct the stratigraphy between 6.34 and 42.90 ka, whereas the simple-headed arrows show the tie points as illustrated in Figure 3. Note the yellow bar represents the MIS4. Hx = Heinrich events; Cx = Cooling events.
The total planktonic foraminiferal concentration varies from 6.1 to 20,018 shells/g (Figure 5c) between the cold MIS2-4 and warm MIS1 and 5 periods. The Neogloboquadrina pachyderma, Neogloboquadrina incompta, Turborotalita quinqueloba, Globigerina bulloides, and Globorotalia inflata are the important planktonic foraminifers comprising nearly 68.53% of the foraminifers' concentration ( Figure 5). Both white and pink varieties of Globigerinoides ruber were also identified; however, due to the low concentration, i.e., 2.03%, it is not further discussed in detail in this study. The %N. pachyderma varies from 1.34 to 85.53% (Figure 5d) during the MIS2-4 with an increasing trend from 21.76 to 84.93%, whereas it is almost stable with <5% throughout the MIS5. The %N. incompta (Figure 5e) is high during the MIS5 and exhibits high-frequency variability in MIS3. Further, it appears that the low %N. incompta mostly matches with the high %N. pachyderma. The %G. bulloides (Figure 5f) and %G. inflata (Figure 5g) co-vary throughout the interval; however, their concentration remained relatively stable between 16 and 39 ka, although the concentration of the former is low compared to the latter and remained closed to the average of 5.45% (Figure 5g). The concentration of T. quinqueloba (Figure 5i H1, H2, H4, and H5 layers were identified by the IRD/g and %N. pachyderma peak and lighter δ 18 O in N. pachyderma in core Hu90-08 ( Figure 5). The identification of H-layers is constrained by the 14 C-AMS dates ( Table 2). H11 was confidently identified at 128-134 ka by the high IRD/g and %N. pachyderma peak [4,25]. The minor IRD peak associated with the minor %N. pachyderma peak at 86-88 ka tentatively correlated to the cooling event C21 ( Figure 5) of the North Atlantic [27,49].

Discussion
Changes in the subpolar gyre appear to have a tenuous link to the subtropical gyre of the North Atlantic during the past [50,51]. Depending on the thermal conditions of the mixed-layer and thermocline, this apparent link might have varied between weak and strong states. Data from core Hu90-08 illuminate on this link, which is broadly divided into four periods: Penultimate deglaciation including the termination II, MIS5, MIS3, and the last deglacial period including the Holocene. Five published records (Table 1), namely CH69-K09 [23], EW9303-37JPC (hereafter 37JPC; [24]), MD03-2664 [25,26], SU90-24 [19], and the Ocean Drilling Program (ODP) site 984 [27] in the subpolar gyre are plotted with Hu90-08 in Figures 6 and 7 as well, allowing us to provide an integrated assessment of broader changes in the sea-surface characteristics during the last glacial cycle. It should be stated a priori that some of the published records used in the study uses dated chronology due either to unavailability of the initial depth-scale data or 14 C-dates/tie points or the unwillingness of the authors to share those data for unknown reasons. A two-pronged approach was employed to use those published data: publications in which depth-scale data are unavailable, for example, published δ 18 O from core 37JPC and benthic foraminifera δ 18 O for SU90-24, we used the dated age models. In contrast to cores 37JPC and SU90-24, data from cores CH69-K09, 984, and MD03-2664 are updated following our age-model construction protocol (see Section 3.2.5).
In terms of choosing proxies used to correlate among the records illustrated in Figures 6 and 7, two approaches were employed. The most commonly used proxies that reflect changes in the sea-surface, namely the IRD/g, %N. pachyderma, and %T. quinqueloba is used for the MIS5-4 period ( Figure 6). Unavailability of N. incompta, G. bulloides, and G. inflata census data from cores 37JPC, MD03-2664, and 984 prevented us from making a similar correlation for MIS5-4 as was carried out for MIS3-2 ( Figure 7). It should be stated that the IRD/g and %N. pachyderma are used as a common guide for both warm (MIS5 and I) and cold (MIS4-2) in assessing cold and freshwater input into the subpolar gyre of the North Atlantic [27,52].   Figure 5; however, %N. pachyderma (c-e), %N. incompta (f-h) and % G. inflata (i,j) of cores Hu90-08, CH69-K09, and SU90-24 are plotted. %G. bulloides (k) from core SU90-24 is independently plotted. Downward filled and unfilled headed triangles exhibit the 14 C-AMS dates and tie points, respectively, used to construct the age model of core Hu90-08 (this study). Two published records, namely SU90-24 [19] and CH69-K09 [23], were also plotted to provide an integrated assessment of the western subpolar gyre. Note the vertical discontinuous line represents the termination I (TI), whereas the Hx = Heinrich events.

The penultimate Deglaciation (129.5 to 125 ka) in the Newfoundland Basin
The penultimate deglacial sequence, including the H11 in core Hu90-08, is similar to the sequence of events on core 37JPC [24]. The concentration of N. pachyderma is >80%, with traces of T. quinqueloba, N. incompta, G. bulloides, and G. inflata during H11 (Figures 5 and 6). The rapid heavy to light δ 18 O shift in core Hu90-08 ( Figure 4) is consistent with the δ 18 O change in N. pachyderma of cores CH69-K09 [23] and MD03-2664 [52]. The shift in the δ 18 O reflects changes in the sea-surface freshening, suggesting the reorganization of the deep-water masses through North Atlantic Deep-Water formation, as recorded by the benthic foraminiferal carbon isotopes from the nearby records [53,54]. A few lighter δ 18 O in N. pachyderma within the penultimate deglaciation on the southern Greenland margin [35,55] were attributed to the Greenland Ice Sheet (GIS) meltwater discharge. However, the lighter δ 18 O events appear to be absent on the high-resolution Eirik Drift records [25,52], which might be either due to the limited spatial influence of the meltwater or a mismatch between the age models. In any case, if these events were not recorded on the high-resolution Eirik Drift sites (Figure 1), it is implausible that those events will be recorded either in core 37JPC or Hu90-08 (this study) due to the coarse temporal resolution or dissipation of the meltwater signatures at those sites. However, Winsor et al. [55] and Carlson et al. [35] provided an alternative explanation suggesting that the sustained cold and fresh East Greenland Current may have suppressed the δ 18 O from~128 ka onward at site MD99-2227 (Figure 1), which cannot be tested at site Hu90-08 at present due to coarse temporal resolution.
The high IRD/g and %N. pachyderma peak with the concomitant decrease in %N. incompta in H11 of core Hu90-08 ( Figure 6) suggest that the ice-rafting and meltwater were discharged. The precipitous decrease from 14.54% to near absent N. incompta and traces of G. inflata and T. quinqueloba suggests the impact of prevailing meltwater and the absence of seasonal subpolar water. In core 37JPC [~300 km from the SE Grand Banks coast], the 95% N. pachyderma and 683 lithics/g with the trace T. quinqueloba in H11 suggest that widespread ice-rafting during H11 in the Newfoundland Basin [24]. Carlson et al. [35] reported a CaCO 3 (%) peak and Fe and Ti trough using the XRF-scanner data for H11, where the authors attributed it to the GIS discharge. An IRD/g and %N. pachyderma peak with the lighter δ 18 O in N. pachyderma and Cibicidoides wuellerstorfi was also reported for H11 in core MD03-2664 [25,52,54]. The authors suggested that such changes in the planktonic and benthic foraminifers reflect the meltwater impact. However, such δ 18 O shift was not found in cores 37JPC or Hu90-08, most likely due to the coarse temporal resolution of the latter records or spatial difference in response to the meltwater.

Sea-Surface Characteristics during the MIS5
The lightest δ 18 O in N. pachyderma of 1.41‰ at 125.81 ka is considered the onset of MIS5e in which both N. incompta and G. bulloides present the highest values and the concentration of N. pachyderma is <5% (Figure 6). Throughout the MIS5, two prominent cold events labeled as C24 and C21 [49], in which a minor IRD peak was tied to the latter cold event. Mokeddam et al. [27] identified six cooling events between H11 and C24, including three minor events within C27 (a, b and c) on the eastern subpolar gyre. The authors used the N. pachyderma, T. quinqueloba, and N. incompta to assess migration of the Polar and Arctic fronts from 102 to 134 ka. The %N. pachyderma at sites Hu90-08 and 984 during MIS5e-c is identical; however, the extent and magnitude of concentration of T. quinqueloba and N. incompta vary between 0 and 71% and 0.38 and 45.31%, respectively, at site 984, reflecting a strong influence by the subpolar water mass. In contrast to site 984, the %T. quinqueloba and %N. incompta at site Hu90-08 vary between 0.86 and 9.28% and 4.33 and 24.92%, respectively, suggesting a strong influence by the warm and salty NAC. Irvali et al. [52] also reported the %T. quinqueloba and %N. incompta from 110 and 135 ka at MD03-2664 (Figure 6), in which the millennial-scale variability recorded in the eastern subpolar gyre (at site 984) appears to be subtle or absent. Further, an IRD peak centered at 117 ka in the western subpolar gyre (at MD03-2664) is also absent at sites Hu90-08 and 984, suggesting the regional nature of the 117-ka event. In short, there are heterogeneities between the eastern and western subpolar gyre and intra-western subpolar gyre (i.e., between the SE Grand Banks and eastern Newfoundland Basin) during the MIS5e-c. These differences most likely reflect the extent to which the Labrador Current and NAC injected cold and freshwater versus warm and salty water, respectively, to the subpolar gyre.
A millennial-scale cooling event at 125 ka (between~126 and 124 ka) within the warmest MIS5e was identified by the %N. pachyderma from 18 to 76%, decrease in %N. incompta, and %G. bulloides on the Eirik Drift [25,52]) suggesting the influence by the polar water compared to the subpolar water. The SST was cooled by 3.4 • C in the Mg/Catemperature in N. pachyderma during this event, which could be attributed to ice-raftingderived meltwater; however, there was no abrupt increase in the IRD associated with this event. Irvali et al. [25,52] also reported a~0.8% decrease in δ 18 O in N. pachyderma, which implied to the authors the existence of fresher near-surface conditions. Galaasen et al. [53] reported lighter benthic foraminiferal δ 13 C during this event and suggested a slowing down of the North Atlantic Deep-Water formation. This cooling event appears to be absent in core Hu90-08, which might be due to the coarse data resolution, or the cooling event was minimized by the interference of the warm NAC inflow, consistent with the absence of cooling signature in core 37JPC (Figure 6). A minor dip in the δ 18 O in core Hu90-08 could be correlative to that cooling event; however, we refrained from interpreting this single datum any further.

Sea-Surface Characteristics during the MIS4-MIS3
The C24 cooling event centered at 106.5 ka in core Hu90-08 corresponds to the sharp rise in the %N. pachyderma in the eastern subpolar gyre ( Figure 6). A similar peak in %N. pachyderma at 109.5 ka also found in core 37JPC, although there appears to be an offset in age between the two records (i.e., 37JPC and Hu90-08), most likely due to the use of outdated age model [56] in core 37JPC. An intermediate IRD peak at 88 ka with a concomitant rise in the %N. pachyderma ( Figure 6) is correlated to the cooling event C21 for the first time in the western subpolar gyre. The lighter δ 18 O event appears to follow the IRD event suggesting freshening of the sea-surface.
The IRD/g and %N. pachyderma peaks identify H2, H4, and H5 [57] with the concomitant light δ 18 O, although there are differences in δ 18 O during each H-event (Figure 7). One of the striking features is the covariation of the foraminifers/g with that of the abrupt rise in IRD/g, consistent with the hypothesis of a near-collapse of the productivity [5,58]. The highest concentration (85%) of N. pachyderma but significant reduction in N. incompta, T. quinqueloba, G. bulloides, and G. inflata during H-events ( Figure 5) suggests an incursion by the polar water. Moreover, the decline of N. pachyderma but an increase in N. incompta immediately after the H-events reflects rapid changeover by the subpolar to transitional water at site Hu90-08. The lighter N. pachyderma δ 18 O suggests a sea-surface freshening during the H2, H4, and H5, consistent with the records at other regional sites (Figures 1 and 7). de Vernal et al. [38] reported SSTs, sea-surface salinities, and sea-ice covers during the H0, H1, and H2, including the LGM (Figure 1; Table 1). The authors reported 0 • C and <5 • C SSTs during February and August, respectively, with sea-ice cover ranging from 4 to 12 months/year. Our data are consistent with the broader temperature aspect of de Vernal et al. [38] when the transfer function of %N. pachyderma to SST [59] is applied. Elliot et al. [19] also reported the concentration of N. pachyderma (Figure 7e), which remained 95% during the MIS3 in the Irminger Basin, suggesting an absolute dominance by the polar water mass. By developing a Heinrich variable (RM), Hiscott et al. [10] reported the intensity of ice-rafting on the sea-surface, matching with the summer and winter SSTs. However, the exact mechanism by which the ice-rafting (i.e., meltwater discharge) and the rise in the SSTs were linked remains unclear.
The IRD records at Hu90-08 ( Figure 7) are consistent with the H1, H2, H4, and H5 ice-rafting records from the nearby core SU90-11 [60], except for two minor IRD peaks within the MIS4 in Hu90-08. An increasing trend from 20 to 85.5% in N. pachyderma between 54 and 10 ka (Figure 7), a mirror image to the %N. incompta might reflect a gradual seasurface cooling and more frequent incursions by the polar water during the winter and spring times [10,59]. This longer-scale inverse covariation between the %N. pachyderma and %N. incompta was frequently interrupted by the H-events. In contrast to the incursion by the polar water, the increase in %N. incompta ( Figure 5) suggests frequent incursions by the warm and salty water due to the NAC inflow, most likely during the late spring and summer. The high-frequency variability in %T. quinqueloba ranging from 0.86 to 19.55% suggests occasional intrusions by the subpolar water [27] at site Hu90-08, most likely due to the penetration of the glacial Labrador Current [1,61].
In contrast to the records at site Hu90-08, the %N. pachyderma peaks in the millennialscale records of CH69-K09 due to its higher temporal resolution are narrower, implying that the impact of H-events-induced meltwater lasted for a short period. Granted that there might be a mismatch between the age models of cores, however, a close inspection reveals much more subtle changes. For example, the core top sample of CH69-K09 contains 15.80% polar species (N. pachyderma), 71.46% transitional subarctic species (N. incompta, G. bulloides, and G. inflata), and~10% tropical and subtropical species (Orbulina universa, G. ruber pink and white, and G. sacculifer). The presence of these species most likely reflects occasional influence by the warm NAC inflow. In contrast to core CH69-K09, subtropical and subtropical species constitute 4.47%, 37.06% transitional species, and <7% N. pachyderma in the core top of site Hu90-08. Further, the overall %N. pachyderma is higher in Hu90-08 than that of CH69-K09 during the MIS3 (Figure 7). Lighter δ 18 O during H4, H2, and H1 in G. bulloides and N. pachyderma, respectively, reflect the impact of meltwater on the mixed-layer and thermocline at site CH69-K09. Such evaluation in δ 18 O between the G. bulloides and N. pachyderma during H-events cannot be assessed due to the unavailability of data in Hu90-08. On the basis of the IRD and N. pachyderma peaks and δ 18 O, it could be summarized that the sudden invasion by the meltwater freshened the sea-surface during the H-events in the western subpolar gyre (Newfoundland Basin).

The Last Deglacial Climate (including the Holocene) in the Western Subpolar Gyre
The onset of the last deglacial period is marked by the heavy to lighter δ 18 O shift at 19 ka at site Hu90-08 (Figure 7). The prominent IRD/g and %N. pachyderma peak without the distinct lighter δ 18 O, as observed for other H-events in Hu90-08, is most likely due to the competing factors between the rise of the Northern Hemisphere insolation [62] and invasion by the warm and salty water due to the NAC inflow. The abrupt rise of %N. incompta and %G. inflata with the rapid fall in %N. pachyderma at 16.5 ka provides support for this hypothesis of climate amelioration in the western subpolar gyre. Such an invasion by the warm and salty waters must have been restricted to the western subpolar gyre, as revealed by the absence of a rise in %N. incompta, G. inflata, or G. bulloides in the Irminger Basin (at site SU90-24; Figure 7). The %N. pachyderma and the absence of an IRD peak centered at 12 ka (Figure 7) might correlate to the Younger Dryas (YD) cooling event, consistent with the other open North Atlantic paleo-proxy records of the YD [38,63,64]. It appears that changes related to the YD event were not recorded either in the Irminger Basin (at  or at the foot of the Newfoundland slope (at CH69-K09), consistent with the dynamics of the Coriolis force in the Northern Hemisphere and freshwater movement [65,66].
The Holocene record in Hu90-08 is short, amounting to~50 cm compared to the MIS3 or MIS5, consistent with the thickness of Holocene sediments to the broader deep Labrador Sea [34,48]. The IRD/g is very poor, similar to the concentration of IRD/g during the MIS5e-c, reflecting either very little icebergs appearance at the core site or less iceberg discharge from the GIS or less transport of Arctic sea-ice. The %N. pachyderma declines from 71.12 to 4.47%, whereas an increase in %T. quinqueloba compared to the MIS3/2 reflects changes in the Polar to Arctic Front ( Figure 5). This finding is similar to the eastern subpolar gyre (at site 984). A similar rising trend from 10.71 to 3.60 ka in %G. bulloides and G. inflata suggests changes from Polar to Arctic Front, consistent with the trends in %N. pachyderma and %T. quinqueloba (Figure 5i).
Bond et al. [67] used sediment cores west of core Hu90-08 to document changes in the surface water properties during the Holocene. Hoogakker et al. [68] applied the Mg/Ca-SST to the G. bulloides to report sub(surface) water conditions to assess the impact of freshwater on the Labrador Sea Water formation during the early Holocene (at MD95-2024). The absence of abundance data of G. bulloides data at site MD95-2024 raises the question about the extent to which sea-surface condition was recorded. For example, Stangeew [37] reported that N. pachyderma was the most abundant in the upper 0-50 m, whereas T. quinqueloba showed a variable distribution with a maximum in 0-200 m with traces of G. bulloides from the multinet plankton toes, very close to core MD95-2024. In addition, the %G. bulloides in Hu90-08 vary between 2.19 and 6.30% during the Holocene ( Figure 5), suggesting limited applicability to the use of G. bulloides in paleoceanographic reconstruction in the western subpolar gyre.

Conclusions
Changes in the sea-surface characteristics, namely the ice-rafting, meltwater discharge, and sea-surface temperatures, were reconstructed for the last 145,000 years using a sediment core Hu90-08 retrieved from the Milne seamount of the northwest Atlantic Ocean. Foraminiferal assemblage data in conjunction with the oxygen isotopes in Neogloboquadrina pachyderma suggest rapid turnover of the upper water masses, switching between the polar and subpolar waters. These changes were rapid during the abrupt climate events of the MIS3, whereas such changes are either subtle or absent during the MIS5.
Five Heinrich H1, H2, H4, H5, and H11 layers were identified by the high IRD/g and %N. pachyderma peak and traces of the subpolar species N. incompta, G. bulloides, and G. inflata suggest complete dominance by the polar water masses. The decline of N. pachyderma but the rise of N. incompta immediately after the H-events suggests a return of the subpolar water masses.
The anti-covariation of %N. pachyderm with the %N. incompta provides an indirect proxy for the sea-surface temperatures in which the former dominates in less than 8 • C, whereas the latter thrive at an optimal temperature of 12 • C. Such rapid changes in the western subpolar gyre reflect a close interaction between the cold and fresh Labrador Current and warm and salty North Atlantic Current.