The Role of Hudson Strait Outlet in Younger Dryas Sedimentation in the Labrador Sea - Abrupt Climate Change: Mechanisms, Patterns, and Impacts - page 106

corresponding retreat of the ice margin was continuous rather
than catastrophic.
At the time of H0, ice-contact deposits were formed on the
inner shelf, seaward of the sill at the eastern end of Hudson
Strait [Andrews and MacLean, 2003; Rashid and Piper,
2007]. Ice then retreated to behind the sill (Figure 9b). The
shallow depth of the sill and the narrowness of the outlet
(Figure 1b) would have greatly reduced the supply of large
icebergs. However, as suggested by Johnson and Lauritzen
[1995], unstable ice near the sill would be susceptible to
catastrophic freshwater outbursts, as a result of freshwater
ponding subglacially or against the northward moving Un-
gava ice stream, which was active during H0 as well as
during the younger Gold Cove and Noble Inlet advances
[Andrews and MacLean, 2003]. A higher proportion of fresh
water might also result from glaciological processes at the
end of H events. Studies of sediment supply through the
Trinity Trough ice stream to Orphan Basin, offshore New-
foundland, show that supply of subglacial detritus to glaci-
genic debris flows on the slope was on at least three
occasions followed by larger amounts of fresh water that

foot of slope. Alternatively, freshwater outbursts may have
supplied surface freshwater plumes with sediment in much
greater concentration than during the main H event. Such
energetic flowed as a jet from the sill outlet

across the shelf, producing nepheloid-flow layer deposits
[Rashid et al., 2003a] or midwater plumes [Hesse et al.,
2004].
In summary, we propose that the contrast between the
sedimentological signatures of H1 and H0 in the Labrador
Sea results from freshwater outbursts. During H0, the source
of most carbonate-rich glacial sediment input into the Lab-
rador Sea was the Hudson Strait, which was kept by the sill
before the outburst. A similar process took place at the end of
H1 when ice had retreated to the sill; in contrast, during the
main depositional phase of H1, when ice extended farther
seaward, sediment supply was of lower concentration and
spread over a longer time, resulting in a much longer dura-
tion of distal plume deposition. The distal plume deposits
were thus thicker and hence recognizable over a much larger
area.

flows would have

flowed hyperpycnally [Tripsanas and Piper, 2008b]. The
data are not available to distinguish between these two hy-
potheses, and indeed, both processes may have occurred
concurrently. We suggest that such enhanced freshwater
supply at the end of H0 and H1 accounts for the high
proportion of gravity

5.6. Where Is the H0 Freshwater Signal?

18 O signal in Labrador Sea cores is in part a proxy for
the supply of fresh water from the former LIS outlet through
Hudson Strait. The δ

The δ

18 O Nps shows a varied signature in H0,
H1, and H2 in cores Hu97-09 and Hu97-16, with depleted
δ

flow deposits in H0 and at the end of
H1. Freshwater outbursts would have had the velocity to
erode and transport previously deposited carbonate rock

18 O values in H2 (Figures 2 and 3). There was a regional
depletion of 1.5 ‰ sometime just before the beginning of H1
(Figures 2 and 3), which might be attributed to the deglacial
changes of the Northern Hemisphere. Since the carbonate
and δ

flour. These outbursts had a much higher instantaneous dis-
charge of water and sediment than that which occurred
during the main H event in H1 or H2 when ice extended
across the shelf. As a result, the duration of surface plume
supply was much shorter, but the power of the discharge to
transport sediment across the shelf to deep water was much
greater. This is why the distribution of sediment types during
H0 more closely resembles that from a classic freshwater
outburst, the 19 ka Laurentian Channel outburst (Table 2)
[Piper et al., 2007], than a typical H event, with much greater
duration and smaller peak freshwater

18 O were measured in the same sample, there are no

arti cial leads or lags arising from the tuning process. This

advantageous method of data acquisition allowed us to sug-
gest that the changes in δ

18 O Nps precede the carbonate in-

crease by ~350 600 years (see also Stoner et al. [2000] in

MD95-2024). Despite the enormous

-

flux of icebergs released
through the strait during H1 [Dowdeswell et al., 1995] and
the evidence of a freshwater plume as far south as the Nova
Scotian margin [Piper and Skene, 1998], the

18 O Nps during
H1 shows only ~1.3 ‰ depleted isotopic signature in core
Hu97-16 and ~2

ows.
We can only speculate about the nature of sediment trans-
port process across the shelf because of the lack of core
penetration on the shelf to the H0 horizon. The hydraulic
head from the sill to the deep continental shelf might have
allowed sediment-laden outbursts to

δ

in core Hu97-09. (Lack of adequate
numbers of N. pachyderma (s) between 840 and 580 cm
prevented a full assessment of the magnitude of fresh waters
in H1.) The

‰

18 O Nps perturbation during H0 is only <0.3

flow hyperpycnally as a in

core Hu-97-16. Farther downstream, the only other high-
resolution record that reveals a minor perturbation in δ

δ ‰

sediment gravity flow across the continental shelf [cf. Russell

and Arnott, 2003; Addington et al., 2007]. Because of the
Coriolis effect, the

18 O Nps
at the YD horizon is at site Hu91-94 in the southern Labrador
Sea. In this core,

flow would tend to occupy the southern
part of the transverse trough [Chapman, 2000]. On reaching
the shelf edge, it would have accelerated down the slope [cf.
Piper and Normark, 2009], depositing mud turbidites at the

18 O Nps is ~0.3

δ ‰ depleted in the YD com-

18 O Nps depleted values during H1 [Hillaire-
Marcel et al., 1994]. We cannot evaluate the relative

pared to ~2 ‰ δ