variations that may have been experienced contemporaneously in parts of the globe very distant from each other.
Finding the explanation for these connections is also a
challenge faced by modellers, because attempting to model
these events means testing our knowledge of climate as it is
interpreted in the models which are also used to predict
future climates. In terms of modeling, the first experiments
testing the sensitivity of the ocean-atmosphere system to
freshwater discharges in the North Atlantic in a glacial
context have been carried out in the years 2000.
These experiments, if they are to be carried out with a
resolution sufficiently fine to enable comparison between
model results and reconstructions, require a calculation time
at the forefront of the capability of the most powerful
modern computers. For example, Kageyama et al. (2009),
using an ocean-atmosphere general circulation model, show
that the link between the cooling in the North Atlantic and
the decrease in the Indian monsoon can be explained by a
decrease in the summer temperature gradient in the upper
half of the troposphere between the Indian Ocean and the
Tibetan plateau. However, these results are obtained for
conditions corresponding to the LGM and not to Marine
Isotopic Stage 3, and further experiments will be required to
better understand the signal transmission mechanisms
between the North Atlantic and Asia. Furthermore, although
these experiments reproduce certain climate events contemporaneous with the Heinrich events (southward migration of the Intertropical Convergence Zone, reduction of the
Indian monsoon, cooling and drying in Western Europe),
they do not manage to reproduce others, such as variations in
the Southeast Asian monsoon. The modeling and understanding of the millennial variability within the climate
system thus remains in many respects a challenge for
modellers.
We have, until now, attempted to describe climate variability on the millennial scale mainly in terms of the
“physical” Earth system, that is, including the ocean, the
atmosphere and the cryosphere. However, the records contain more information than this simple description. Take, for
example, the case of methane concentrations. We have
shown that they are useful to synchronize the records from
Greenland and Antarctica but this important greenhouse gas
emitted mainly from the wetlands in the tropics and high
latitudes shows significant variations. The role of biogeochemical phenomena in millennial climate variability also
needs to be better understood and modeled.
Abrupt Event Interactions—Large Climate
Transitions
Although abrupt events are the result of internal variability
in the climate system, they are nonetheless sensitive to its
large longer term variations. In particular, these events are
more numerous and of greater amplitude in glacial periods
than in interglacial periods. This suggests a different
expression of millennial variability depending on the size of
the ice sheets in the northern hemisphere. The
high-resolution records of the most recent climate transitions
(the deglaciation between LGM and the Holocene as well as
the entry into the last glaciation) may provide information on
the conditions favoring events of larger amplitude. However,
these transitions are punctuated by abrupt climate events.
This raises the question of the role that these events might
play in the transition itself, knowing that a rapid event can
have, in the case of Dansgaard-Oeschger events in the
Greenland cores, an amplitude equivalent to half the difference between glacial and interglacial states.
Let us consider the case of the last deglaciation
(Fig. 29.7). A few thousand years after the LGM came the
Heinrich H1 event. Climate conditions returned to an almost
glacial level. This event was followed by a warm phase, with
the transition between these two events showing up as abrupt
in many records, particularly around the North Atlantic. This
is the Bølling-Allerød phase, whose climate is almost
interglacial. However, this period was followed by the cold
period of the Younger Dryas, which is sometimes considered
to be the most recent Heinrich event (H0), which is wrong
because there is no corresponding layer of detritic elements
in the Ruddiman belt. It is after this last cold phase that the
climate of the current interglacial, the Holocene, became
definitively established (apart from the 8200-year event).
This shows that the last deglaciation was not a
smooth transition. On the contrary, it is a series of abrupt
events, as if the climate system ‘hesitated’ between two
equilibria, one glacial and the other interglacial. The role of
abrupt events during this climate transition is therefore
important, but they still need to be understood and modeled.
Entry into the last glaciation at the end of the Eemian is
also characterized by the appearance of abrupt events,
*110 ka before today, in a context where glacial ice sheets
had already developed over Canada. One might think that
these cold events would promote entry into glaciation, but
this assumption ignores the fact that cooler air at high latitudes also contains less water and is therefore less able to
supply the water needed to build the ice sheets up at a significant rate. These compensating factors need to be assessed. Here again, the influence of abrupt events on the
evolution of the ice sheets has yet to be assessed, in comparison with other mechanisms and feedbacks, such as the
slower changes in state of the ocean, cryosphere, vegetation,
atmospheric concentrations of greenhouse gases, as well as
external changes such as changes in insolation. Models of
the Earth system can lead to a better understanding of the
reconstructed signals by conducting sensitivity experiments
for each of these factors.
29 Rapid Climate Variability: Description and Mechanisms
419
Finding the explanation for these connections is also a
challenge faced by modellers, because attempting to model
these events means testing our knowledge of climate as it is
interpreted in the models which are also used to predict
future climates. In terms of modeling, the first experiments
testing the sensitivity of the ocean-atmosphere system to
freshwater discharges in the North Atlantic in a glacial
context have been carried out in the years 2000.
These experiments, if they are to be carried out with a
resolution sufficiently fine to enable comparison between
model results and reconstructions, require a calculation time
at the forefront of the capability of the most powerful
modern computers. For example, Kageyama et al. (2009),
using an ocean-atmosphere general circulation model, show
that the link between the cooling in the North Atlantic and
the decrease in the Indian monsoon can be explained by a
decrease in the summer temperature gradient in the upper
half of the troposphere between the Indian Ocean and the
Tibetan plateau. However, these results are obtained for
conditions corresponding to the LGM and not to Marine
Isotopic Stage 3, and further experiments will be required to
better understand the signal transmission mechanisms
between the North Atlantic and Asia. Furthermore, although
these experiments reproduce certain climate events contemporaneous with the Heinrich events (southward migration of the Intertropical Convergence Zone, reduction of the
Indian monsoon, cooling and drying in Western Europe),
they do not manage to reproduce others, such as variations in
the Southeast Asian monsoon. The modeling and understanding of the millennial variability within the climate
system thus remains in many respects a challenge for
modellers.
We have, until now, attempted to describe climate variability on the millennial scale mainly in terms of the
“physical” Earth system, that is, including the ocean, the
atmosphere and the cryosphere. However, the records contain more information than this simple description. Take, for
example, the case of methane concentrations. We have
shown that they are useful to synchronize the records from
Greenland and Antarctica but this important greenhouse gas
emitted mainly from the wetlands in the tropics and high
latitudes shows significant variations. The role of biogeochemical phenomena in millennial climate variability also
needs to be better understood and modeled.
Abrupt Event Interactions—Large Climate
Transitions
Although abrupt events are the result of internal variability
in the climate system, they are nonetheless sensitive to its
large longer term variations. In particular, these events are
more numerous and of greater amplitude in glacial periods
than in interglacial periods. This suggests a different
expression of millennial variability depending on the size of
the ice sheets in the northern hemisphere. The
high-resolution records of the most recent climate transitions
(the deglaciation between LGM and the Holocene as well as
the entry into the last glaciation) may provide information on
the conditions favoring events of larger amplitude. However,
these transitions are punctuated by abrupt climate events.
This raises the question of the role that these events might
play in the transition itself, knowing that a rapid event can
have, in the case of Dansgaard-Oeschger events in the
Greenland cores, an amplitude equivalent to half the difference between glacial and interglacial states.
Let us consider the case of the last deglaciation
(Fig. 29.7). A few thousand years after the LGM came the
Heinrich H1 event. Climate conditions returned to an almost
glacial level. This event was followed by a warm phase, with
the transition between these two events showing up as abrupt
in many records, particularly around the North Atlantic. This
is the Bølling-Allerød phase, whose climate is almost
interglacial. However, this period was followed by the cold
period of the Younger Dryas, which is sometimes considered
to be the most recent Heinrich event (H0), which is wrong
because there is no corresponding layer of detritic elements
in the Ruddiman belt. It is after this last cold phase that the
climate of the current interglacial, the Holocene, became
definitively established (apart from the 8200-year event).
This shows that the last deglaciation was not a
smooth transition. On the contrary, it is a series of abrupt
events, as if the climate system ‘hesitated’ between two
equilibria, one glacial and the other interglacial. The role of
abrupt events during this climate transition is therefore
important, but they still need to be understood and modeled.
Entry into the last glaciation at the end of the Eemian is
also characterized by the appearance of abrupt events,
*110 ka before today, in a context where glacial ice sheets
had already developed over Canada. One might think that
these cold events would promote entry into glaciation, but
this assumption ignores the fact that cooler air at high latitudes also contains less water and is therefore less able to
supply the water needed to build the ice sheets up at a significant rate. These compensating factors need to be assessed. Here again, the influence of abrupt events on the
evolution of the ice sheets has yet to be assessed, in comparison with other mechanisms and feedbacks, such as the
slower changes in state of the ocean, cryosphere, vegetation,
atmospheric concentrations of greenhouse gases, as well as
external changes such as changes in insolation. Models of
the Earth system can lead to a better understanding of the
reconstructed signals by conducting sensitivity experiments
for each of these factors.
29 Rapid Climate Variability: Description and Mechanisms
419