records. In addition, knowing that the residence time of
14 C
in the ocean (reservoir age variations) varies during abrupt
events makes absolute dating based on this single variable
even more difficult. However, these variations are interesting
because they give us information on changes in ocean
ventilation and ocean circulation (Chap. 21).
However, it is possible to synchronize the different
records with each other using parameters known to be the
same or to vary in a similar manner in different locations on
the planet. This allows us to study the chronology of events
recorded in different places and to propose hypotheses about
the mechanisms behind glacial abrupt climate variability.
Specific events, such as volcanic eruptions, whose falling
ash appears in ice, marine and lake sediments, can help to
synchronize records (Rasmussen et al. 2014). Other parameters can be used depending on the type of record.
One example is the use of variations in the atmospheric
concentration of methane to place Greenland and Antarctic
ice records on the same time scale. Methane is well mixed in
the atmosphere and so variations in its concentration are the
same everywhere on the globe. Members of the EPICA
community (2006) used this method to compare the
expression of abrupt events at the two poles. Comparisons
show an anti-correlation between isotopic signals in
Greenland and Antarctica. In addition, the strength of the
warming in Antarctica is related to the length of the cooling
in Greenland.
Another example is the use of magnetic properties to synchronize marine records (see Chap. 7). At the scale of the
ocean basin (e.g., the North Atlantic, Elliot 2002), variations in
rock magnetism can be used to put marine records on a common timescale. At the interhemispheric scale, paleomagnetic
intensity can be used, especially for episodes of sudden and
intense variation, such as around the Laschamp event.
Mechanisms
The climate reconstructions described above demonstrate the
existence of rapid climate variability of great amplitude that
involve all the components of the climate system. Understanding all aspects of this type of variability is still a
challenge for the scientific community. An approach to test
our understanding of these phenomena is to build models
that include processes that are thought to contribute in a
critical way to rapid variability. Well-informed choices from
the possible forcings and processes can highlight the
importance of each one for a particular type of event or for a
level of variability (see Chap. 25, volume 2). Models have
thus been constructed to explain the iceberg armadas (the
Heinrich events themselves, strictly speaking) or to estimate
the inflow of freshwater to the North Atlantic that could
explain the observations or even to establish the connection
between the inflow of freshwater to the North Atlantic and
the climate consequences. Each study focuses on one aspect
of the rapid climate variability. For the time being, no study
has tried to include all of the factors of rapid climate variability in a single model, which would require a representation of the ice sheets, climate, ocean, chemical composition
of the atmosphere, conditions of land surfaces, etc. constituting an almost complete model of the Earth system, which
does not currently exist. The following sections present
some of the suggested mechanisms behind Heinrich events
and how they impact on global climate as well as on the
Dansgaard-Oeschger cycles.
Heinrich Events
Ice-Sheet Instabilities
One of the first explanations of the cycles of spectacular
armadas of icebergs discharged from the North American ice
sheet was proposed by MacAyeal in 1993. He constructed a
model of the part of the Laurentide ice sheet positioned on
what is today the Hudson Bay, that is, on a sedimentcovered bedrock. When the base of the ice sheet reaches its
melting point, causing liquid water to be present there, the
ice slides much more easily than it would if the base of the
ice sheet and the sediments were frozen or if it was on a
bedrock not covered with sediment. The ice basal layer can
warm up through the input of geothermal energy when the
ice sheet is thick enough, because the temperature at the base
of an ice column is dependent on the pressure exerted by this
column and because of the ice insulating its base from the
cold conditions at its top. Based on these properties,
MacAyeal proposes the following cycle: (1) the ice sheet
grows due to the accumulation of snow on the surface (assumed to be constant over time in this model) and the base of
the ice is frozen; (2) once the ice sheet is thick enough, its
base melts and a layer of liquid water forms at the
base-sediment-ice interface; (3) the ice sheet then quickly
slides towards the ocean and its elevation decreases; (4) the
elevation of the ice sheet decreases sufficiently so that the
basal layer freezes again, thus slowing down the movement
of the ice, and returning to step 1. MacAyeal calls this model
‘binge/purge’. In this model, the time between two “purge”
events (i.e. between two episodes of iceberg break-up) is
determined by the characteristic time necessaryfor the ice
sheet to ticken and hence depends on the configuration of
the ice sheet (size, distance to margins, characteristics of the
underlying surface) and on its surface mass balance. It
should be stressed that this mechanism works even for a
constant climate. MacAyeal demonstrates an inherent
oscillation in the ice sheets and shows that for the Laurentide
410
M. Kageyama et al.
14 C
in the ocean (reservoir age variations) varies during abrupt
events makes absolute dating based on this single variable
even more difficult. However, these variations are interesting
because they give us information on changes in ocean
ventilation and ocean circulation (Chap. 21).
However, it is possible to synchronize the different
records with each other using parameters known to be the
same or to vary in a similar manner in different locations on
the planet. This allows us to study the chronology of events
recorded in different places and to propose hypotheses about
the mechanisms behind glacial abrupt climate variability.
Specific events, such as volcanic eruptions, whose falling
ash appears in ice, marine and lake sediments, can help to
synchronize records (Rasmussen et al. 2014). Other parameters can be used depending on the type of record.
One example is the use of variations in the atmospheric
concentration of methane to place Greenland and Antarctic
ice records on the same time scale. Methane is well mixed in
the atmosphere and so variations in its concentration are the
same everywhere on the globe. Members of the EPICA
community (2006) used this method to compare the
expression of abrupt events at the two poles. Comparisons
show an anti-correlation between isotopic signals in
Greenland and Antarctica. In addition, the strength of the
warming in Antarctica is related to the length of the cooling
in Greenland.
Another example is the use of magnetic properties to synchronize marine records (see Chap. 7). At the scale of the
ocean basin (e.g., the North Atlantic, Elliot 2002), variations in
rock magnetism can be used to put marine records on a common timescale. At the interhemispheric scale, paleomagnetic
intensity can be used, especially for episodes of sudden and
intense variation, such as around the Laschamp event.
Mechanisms
The climate reconstructions described above demonstrate the
existence of rapid climate variability of great amplitude that
involve all the components of the climate system. Understanding all aspects of this type of variability is still a
challenge for the scientific community. An approach to test
our understanding of these phenomena is to build models
that include processes that are thought to contribute in a
critical way to rapid variability. Well-informed choices from
the possible forcings and processes can highlight the
importance of each one for a particular type of event or for a
level of variability (see Chap. 25, volume 2). Models have
thus been constructed to explain the iceberg armadas (the
Heinrich events themselves, strictly speaking) or to estimate
the inflow of freshwater to the North Atlantic that could
explain the observations or even to establish the connection
between the inflow of freshwater to the North Atlantic and
the climate consequences. Each study focuses on one aspect
of the rapid climate variability. For the time being, no study
has tried to include all of the factors of rapid climate variability in a single model, which would require a representation of the ice sheets, climate, ocean, chemical composition
of the atmosphere, conditions of land surfaces, etc. constituting an almost complete model of the Earth system, which
does not currently exist. The following sections present
some of the suggested mechanisms behind Heinrich events
and how they impact on global climate as well as on the
Dansgaard-Oeschger cycles.
Heinrich Events
Ice-Sheet Instabilities
One of the first explanations of the cycles of spectacular
armadas of icebergs discharged from the North American ice
sheet was proposed by MacAyeal in 1993. He constructed a
model of the part of the Laurentide ice sheet positioned on
what is today the Hudson Bay, that is, on a sedimentcovered bedrock. When the base of the ice sheet reaches its
melting point, causing liquid water to be present there, the
ice slides much more easily than it would if the base of the
ice sheet and the sediments were frozen or if it was on a
bedrock not covered with sediment. The ice basal layer can
warm up through the input of geothermal energy when the
ice sheet is thick enough, because the temperature at the base
of an ice column is dependent on the pressure exerted by this
column and because of the ice insulating its base from the
cold conditions at its top. Based on these properties,
MacAyeal proposes the following cycle: (1) the ice sheet
grows due to the accumulation of snow on the surface (assumed to be constant over time in this model) and the base of
the ice is frozen; (2) once the ice sheet is thick enough, its
base melts and a layer of liquid water forms at the
base-sediment-ice interface; (3) the ice sheet then quickly
slides towards the ocean and its elevation decreases; (4) the
elevation of the ice sheet decreases sufficiently so that the
basal layer freezes again, thus slowing down the movement
of the ice, and returning to step 1. MacAyeal calls this model
‘binge/purge’. In this model, the time between two “purge”
events (i.e. between two episodes of iceberg break-up) is
determined by the characteristic time necessaryfor the ice
sheet to ticken and hence depends on the configuration of
the ice sheet (size, distance to margins, characteristics of the
underlying surface) and on its surface mass balance. It
should be stressed that this mechanism works even for a
constant climate. MacAyeal demonstrates an inherent
oscillation in the ice sheets and shows that for the Laurentide
410
M. Kageyama et al.