in the level of European lakes indicate that during this event
the climate was wetter between 50 and 43° N, while north
and south of this zone, it was drier. Pollen records rarely
show this event, perhaps because these climate variations
were not significant enough to modify the vegetation in a
major way, or because it did not affect the growing period of
the main plants characteristic of these paleoenvironments, as
was shown in Northern Europe and North America.
The Mechanisms
The scale of changes observed as well as the multiplicity of
their impacts (precipitation, temperature, oceanic circulation), very quickly pointed to the Atlantic thermohaline
circulation as a major actor in this abrupt change in climate.
The experience of Heinrich events (see above) suggested
that here too, a sudden influx of fresh water to the ocean
could have altered the thermohaline circulation and the climate around the North Atlantic. However, 8200 years ago,
only a small part of the Laurentide ice sheet could have been
involved, with the disappearance at this time of the dome
covering what is currently the Hudson Bay. Traces of
paleo-shorelines of lakes also indicate the disappearance of
two large pro-glacial lakes (the Ojibway and Agassiz lakes)
around this period, which could have contributed an influx
of fresh water to the ocean. However, the first
14 C datings
did not enable a precise chronology of these events. More
precise dating, as well as an evaluation of the volume
released, make it possible to pinpoint the massive draining of
Lake Agassiz as a cause of this cold event (Clarke et al.
2004). These data indicate a draining of about 100,000 km
3
of water in about a year, thus causing a massive influx to the
ocean.
Modeling and the Global and Hemispheric
Consequences
Nevertheless, can the consequences for the climate of this
extremely abrupt but yet short-lived influx be replicated in
coupled climate models? Although the impact of this huge
influx to the ocean was quickly identified, the first simulation
of the 8200-year event using boundary conditions consistent
with the time period was achieved in the early 2000s
(Renssen et al. 2001). These authors were able to show that
by forcing an ocean general circulation model coupled with
a simplified atmospheric model using an idealized water flux
of 0.75 Sv for twenty years, they produced a slowing of the
thermohaline circulation in the Atlantic for about three
centuries, with a temperature drop of 1–5 °C on the continents bordering the North Atlantic. This result fits well with
the data, although the freshwater forcing is relatively larger
than is accepted today. Even more importantly, it revealed
disparities in the seasonal climate response, with a greater
anomaly in summer than in winter over northern Europe and
a north-south bipolarity, with cooling in the north and
warming at 60° S. Subsequent studies confirmed that a
freshwater influx, even one lasting only one year, was sufficient to obtain a climate response consistent with the one
obtained from the analysis of paleoclimate indicators. The
climate impact was not immediate, as can be seen in
Fig. 29.6. For example, in Greenland, the maximum point
was reached after about thirty years.
Other authors have also highlighted the role of climate
‘noise’ in the duration of the ocean’s response to the
freshwater forcing. Indeed, they show that by adding noise
(with an average value of zero) to the freshwater forcing, the
duration of the response by the model can be changed from a
few decades to two or three centuries. This highlights the
importance of the initial climate state to the response to a
given forcing, as the response can have a more or less global
effect as was the case for the 8200-year event, or have a local
impact on a decadal scale. The 8200-year event was also
reproduced in a coupled general circulation model incorporating water isotopes (Legrande et al. 2006), demonstrating
that the results were consistent not only with temperature
estimates but also with the paleoclimate indicators themselves, directly simulated in the model.
These authors were able to correctly predict the variations
in d
18 O associated with the 8200-year event from the calcite
of marine sediments before the publication of the first data
showing the 8200-year event in marine sediment cores.
It is important to note that although the 8200-year event is
considered an abrupt event during the interglacial period,
this classification is not quite accurate. In fact, it is a final
manifestation of the glacial climate through the draining of a
periglacial lake, the result of a melting ice sheet . To date, no
abrupt temperature changes involving glacial ice sheet
behavior have been observed during the Holocene.
Outlook
Global Connections
The work of the last thirty years has made it possible to
demonstrate a variability in the climate system on millennial
time scales, characterized by transitions occurring, in some
cases, over only a few decades. This variability was first
observed in North Atlantic marine sediment cores for
Heinrich events, and in Greenland ice cores for DansgaardOeschger events. Gradually, climate records have been
analyzed at increasingly fine temporal resolutions and variations with temporal characteristics similar to those in the
North Atlantic and Greenland have been discovered in areas
29 Rapid Climate Variability: Description and Mechanisms
417
the climate was wetter between 50 and 43° N, while north
and south of this zone, it was drier. Pollen records rarely
show this event, perhaps because these climate variations
were not significant enough to modify the vegetation in a
major way, or because it did not affect the growing period of
the main plants characteristic of these paleoenvironments, as
was shown in Northern Europe and North America.
The Mechanisms
The scale of changes observed as well as the multiplicity of
their impacts (precipitation, temperature, oceanic circulation), very quickly pointed to the Atlantic thermohaline
circulation as a major actor in this abrupt change in climate.
The experience of Heinrich events (see above) suggested
that here too, a sudden influx of fresh water to the ocean
could have altered the thermohaline circulation and the climate around the North Atlantic. However, 8200 years ago,
only a small part of the Laurentide ice sheet could have been
involved, with the disappearance at this time of the dome
covering what is currently the Hudson Bay. Traces of
paleo-shorelines of lakes also indicate the disappearance of
two large pro-glacial lakes (the Ojibway and Agassiz lakes)
around this period, which could have contributed an influx
of fresh water to the ocean. However, the first
14 C datings
did not enable a precise chronology of these events. More
precise dating, as well as an evaluation of the volume
released, make it possible to pinpoint the massive draining of
Lake Agassiz as a cause of this cold event (Clarke et al.
2004). These data indicate a draining of about 100,000 km
3
of water in about a year, thus causing a massive influx to the
ocean.
Modeling and the Global and Hemispheric
Consequences
Nevertheless, can the consequences for the climate of this
extremely abrupt but yet short-lived influx be replicated in
coupled climate models? Although the impact of this huge
influx to the ocean was quickly identified, the first simulation
of the 8200-year event using boundary conditions consistent
with the time period was achieved in the early 2000s
(Renssen et al. 2001). These authors were able to show that
by forcing an ocean general circulation model coupled with
a simplified atmospheric model using an idealized water flux
of 0.75 Sv for twenty years, they produced a slowing of the
thermohaline circulation in the Atlantic for about three
centuries, with a temperature drop of 1–5 °C on the continents bordering the North Atlantic. This result fits well with
the data, although the freshwater forcing is relatively larger
than is accepted today. Even more importantly, it revealed
disparities in the seasonal climate response, with a greater
anomaly in summer than in winter over northern Europe and
a north-south bipolarity, with cooling in the north and
warming at 60° S. Subsequent studies confirmed that a
freshwater influx, even one lasting only one year, was sufficient to obtain a climate response consistent with the one
obtained from the analysis of paleoclimate indicators. The
climate impact was not immediate, as can be seen in
Fig. 29.6. For example, in Greenland, the maximum point
was reached after about thirty years.
Other authors have also highlighted the role of climate
‘noise’ in the duration of the ocean’s response to the
freshwater forcing. Indeed, they show that by adding noise
(with an average value of zero) to the freshwater forcing, the
duration of the response by the model can be changed from a
few decades to two or three centuries. This highlights the
importance of the initial climate state to the response to a
given forcing, as the response can have a more or less global
effect as was the case for the 8200-year event, or have a local
impact on a decadal scale. The 8200-year event was also
reproduced in a coupled general circulation model incorporating water isotopes (Legrande et al. 2006), demonstrating
that the results were consistent not only with temperature
estimates but also with the paleoclimate indicators themselves, directly simulated in the model.
These authors were able to correctly predict the variations
in d
18 O associated with the 8200-year event from the calcite
of marine sediments before the publication of the first data
showing the 8200-year event in marine sediment cores.
It is important to note that although the 8200-year event is
considered an abrupt event during the interglacial period,
this classification is not quite accurate. In fact, it is a final
manifestation of the glacial climate through the draining of a
periglacial lake, the result of a melting ice sheet . To date, no
abrupt temperature changes involving glacial ice sheet
behavior have been observed during the Holocene.
Outlook
Global Connections
The work of the last thirty years has made it possible to
demonstrate a variability in the climate system on millennial
time scales, characterized by transitions occurring, in some
cases, over only a few decades. This variability was first
observed in North Atlantic marine sediment cores for
Heinrich events, and in Greenland ice cores for DansgaardOeschger events. Gradually, climate records have been
analyzed at increasingly fine temporal resolutions and variations with temporal characteristics similar to those in the
North Atlantic and Greenland have been discovered in areas
29 Rapid Climate Variability: Description and Mechanisms
417