184
evidence that a nutrient depleted intermediate water was formed instead
of deep water during the Last Glacial Maximum (Duplessy et aI, 1988).
Observations also show that atmospheric 14 0 was quite steady during the
Younger Dryas period (Broecker, 1994). Since NADW is a major downward
pathway for 14C, 14C would be expected to go up in the atmosphere when
N ADW was shut off, other things being equal, so it remains to explain 14 0
variability in the NADW hypothesis.
In spite of this ambiguity in the paleo-record, the NADW hypothesis
remains attractive because numerous modeling results show that the opposition of thermal and haline forcing involved in N ADW formation is
capable of producing multiple equilibria and internal variability. The most
complete of these studies was performed by Manabe and Stouffer (1988)
who showed that a coupled ocean-atmosphere model has two equilibria:
one with and one without NADW formation. The difference in the two
climates was most pronounced in the North Atlantic region where warmer
sea surface and surface air temperatures and reduced sea ice accompany
NADW production. Manabe and Stouffer show that the freshening of the
North Atlantic in the NADW off mode is due to the increased residence
time in the region of freshening rather than changes in the hydrological
cycle - in other words, it is due to internal ocean processes rather than a
coupled ocean-atmosphere interaction. If we accept that these two climate
modes underlie the paleo- record, we are faced with two questions: (1)
how do transitions between the two modes occur?, and (2) why have such
transitions not been observed since the end of the Younger Dryas, about
11,000 years ago, or in the last warm period 120,000 years ago?
The most obvious place to look for answers to these questions is in the
freshwater flux into the North Atlantic. During glacial times, ice sheets
enhanced the capacity for freshwater storage on land as well as the potential
for pulse-like release. There is evidence for such impulsive freshwater inputs
but it is hard to reconcile with direct forcing of mode switches. First
consider the period of deglaciation that began 14,000 years ago. The sea
level record from Barbados corals shows that the warm periods before
and after the Younger Dryas contained meltwater spikes with maximum
melting rates near 0.4 Sv (Fairbanks, 1989). The melting was considerably
reduced during the cold Younger Dryas period. This suggests a negative
feedback: reduced ocean heat transport leads to reduced melting allowing
increased overturning. Impulsive freshwater inputs also occur during the
heart of the glacial period. Recently, layers of ice-rafted debris associated
evidence that a nutrient depleted intermediate water was formed instead
of deep water during the Last Glacial Maximum (Duplessy et aI, 1988).
Observations also show that atmospheric 14 0 was quite steady during the
Younger Dryas period (Broecker, 1994). Since NADW is a major downward
pathway for 14C, 14C would be expected to go up in the atmosphere when
N ADW was shut off, other things being equal, so it remains to explain 14 0
variability in the NADW hypothesis.
In spite of this ambiguity in the paleo-record, the NADW hypothesis
remains attractive because numerous modeling results show that the opposition of thermal and haline forcing involved in N ADW formation is
capable of producing multiple equilibria and internal variability. The most
complete of these studies was performed by Manabe and Stouffer (1988)
who showed that a coupled ocean-atmosphere model has two equilibria:
one with and one without NADW formation. The difference in the two
climates was most pronounced in the North Atlantic region where warmer
sea surface and surface air temperatures and reduced sea ice accompany
NADW production. Manabe and Stouffer show that the freshening of the
North Atlantic in the NADW off mode is due to the increased residence
time in the region of freshening rather than changes in the hydrological
cycle - in other words, it is due to internal ocean processes rather than a
coupled ocean-atmosphere interaction. If we accept that these two climate
modes underlie the paleo- record, we are faced with two questions: (1)
how do transitions between the two modes occur?, and (2) why have such
transitions not been observed since the end of the Younger Dryas, about
11,000 years ago, or in the last warm period 120,000 years ago?
The most obvious place to look for answers to these questions is in the
freshwater flux into the North Atlantic. During glacial times, ice sheets
enhanced the capacity for freshwater storage on land as well as the potential
for pulse-like release. There is evidence for such impulsive freshwater inputs
but it is hard to reconcile with direct forcing of mode switches. First
consider the period of deglaciation that began 14,000 years ago. The sea
level record from Barbados corals shows that the warm periods before
and after the Younger Dryas contained meltwater spikes with maximum
melting rates near 0.4 Sv (Fairbanks, 1989). The melting was considerably
reduced during the cold Younger Dryas period. This suggests a negative
feedback: reduced ocean heat transport leads to reduced melting allowing
increased overturning. Impulsive freshwater inputs also occur during the
heart of the glacial period. Recently, layers of ice-rafted debris associated