Laurentide Ice Sheet Meltwater and the Atlantic Meridional Overturning Circulation During the Last Glacial Cycle: A View From the Gulf of Mexico - Abrupt Climate Change: Mechanisms, Patterns, and Impacts - page 48

18 O records pro-
vide some evidence for meltwater during Heinrich events
and some interstadials. Data from Deep-Sea Drilling Project
609 in the northeast Atlantic reveal a signi

storage in the low-latitude Atlantic Ocean during the preced-
ing H1 stadial [Vellinga and Wu, 2004; Schmidt et al., 2004;
Weldeab et al., 2006; Carlson et al., 2008].
During MIS 3, there is also a correspondence between
inferred LIS meltwater input, Heinrich stadials, and NADW
reduction (Figure 8). In particular, the largest and longest
inferred meltwater spike coincides with the greatest NADW
reduction during the Heinrich 4 stadial. LIS meltwater ap-
pears to start during the Heinrich 4 stadial and continue
through the subsequent D-O interstadial 8. Similar to the last
deglaciation, the persistence of meltwater through interstadial
8 seems contradictory to evidence for strong resumption of
NADW [Curry and Oppo, 1997] and appears to require salt
buildup in the low-latitude Atlantic Ocean to offset the
meltwater effect [Vellinga and Wu,2004;Schmidt et al.,
2006; Jaeschke et al., 2007]. The cessation of this meltwater
spike may coincide with the following stadial, but age con-
trol is insuf

debris records and planktic foraminiferal δ

18 O mini-
mum recorded in Neogloboquadrina pachyderma (sinistral)
associated with Heinrich event 1 [Bond et al., 1993]. Other
North Atlantic cores also exhibit signi

cant δ

18 O excursions
and other parameters during Heinrich stadials, particularly
those associated with H1 and H4 [Sarnthein et al., 1995;
Maslin et al., 1995; Zahn et al., 1997; Cortijo et al., 1997;
Vidal et al., 1997; Labeyrie et al., 1999; Mangerud et al.,
2004; Peck et al., 2006; Toucanne et al., 2010]. These

cant δ

18 O
excursions and associated increases in ice-rafted debris are
found across the North Atlantic, generally consistent with

δ

cient to evaluate this possibility. Nevertheless,
the number and timing of meltwater events does not match
the nine D-O events in this interval, as discussed above, yet
they do match the inferred decreases in NADW production.
The following pattern emerges from comparison of proxy
data for LIS melting and AMOC reduction during the last
glacial cycle (Figure 8): (1) LIS meltwater begins during
Heinrich stadials and lasts through the subsequent D-O
event. (2) LIS meltwater appears to coincide with major AIM
events. (3) LIS meltwater is associated with distinct changes
in deep circulation during Heinrich events. The likely trigger
for LIS meltwater during the last deglaciation was summer
insolation, but insolation changes were much too slow dur-
ing MIS 3 to initiate LIS melting. The trigger for LIS melting
and Heinrich events during MIS 3 remains unknown [Hem-
ming, 2004] but may have been enhanced by CO 2 rise via the
bipolar seesaw [Knorr and Lohmann, 2007; Barker et al.,
2009] or via the displacement of the Southern Hemisphere
westerlies [Anderson et al., 2009; Denton et al., 2010].
To what extent did LIS meltwater to the Gulf of Mexico
directly affect the AMOC? Modeling work has found differ-
ent sensitivities. A classic paper determined that input to the
Gulf of Mexico had four to five times less impact on the

thermohaline circulation versus the North Atlantic [Manabe
and Stouffer, 1997]. A recent paper found that the sensitivity
was about half that of northern North Atlantic input; 0.28 Sv
to the Gulf of Mexico was sufficient to reduce AMOC by
about 40% [Otto-Bliesner and Brady, 2010]. Clearly, melt-
water input to the northern North Atlantic is more effective at
reducing the AMOC.
Accordingly, a key unresolved issue is the contribution of
the other circum-North Atlantic ice sheets. Data are limited
on the individual Greenland, Iceland, and Fennoscandian ice
sheets and other outlets of the LIS. However, ice-rafted

Figure 9. Core CH69-K09 records from offshore Newfoundland
[Labeyrie et al., 1999]. Winter SST based on planktic foraminiferal
assemblage data is plotted with

18 ObasedonGlobigerinoides
bulloides (solid line) and Neogloboquadrina pachyderma (s.)
(dashed line). Heinrich events based on percent ice-rafted debris
are shaded.

δ