Data from Vostok and EPICA ice cores in Antarctica,
showed that atmospheric CO 2 varied by about 100 ppm
during the transition from an ice age to an interglacial period,
and that these variations were correlated with changes in air
temperatures, suggesting a link between CO 2 and climate.
Today, this link is commonly accepted and attention is
focused on global warming due to the increase in greenhouse
gases in the atmosphere. However almost forty years ago, in
1972, the climatological community met in the United States
to discuss the imminence of the next ice age. This question
arose because geological data showed that for about a million
years, the Earth had alternated between cold episodes corresponding to the glaciation phases, and the warmer interglacial periods, with a pseudo-periodicity of 100,000 years.
Data available in 1972 showed that the previous two
interglacial periods had lasted about 10,000 years. Yet, the
Holocene, the interglacial period we are currently experiencing, has been going on for 10,000 years. It seemed
therefore reasonable to think that this warm phase would
soon end and give way to a new ice age. This reasoning was
based on the assumption that all interglacial periods are of
equal duration. However, although the first numerical
experiments carried out with statistical models indicated that
the cooling initiated 6000 years ago (after the Holocene
climate optimum) would continue in the future, Oerlemans
and Van der Veen (1984) of the University of Utrecht
showed, with an ice-sheet model, that the transition to a new
glacial phase would not occur before 50,000 years.
Based on data from marine sediments and ice cores, it is
now well established that the duration of an interglacial
period varies considerably from one climatic cycle to the
next. The same models used to predict the future were used
to simulate the last interglacial-glacial transition. Indeed,
taking into account the atmosphere-ocean feedbacks, these
models produce perennial snow in the Canadian archipelago
(Khodri et al. 2001). It is also known that the Earth’s orbital
parameters (see Chap. 28) that govern the seasonal and latitudinal distribution of solar energy can vary considerably
from one cycle to another. For example, geological data tell
us that the interglacial period occurring 400,000 years ago
(marine isotopic stage 11) was exceptionally long. This situation corresponds to a weak eccentricity where the seasonal
and latitudinal distribution of solar energy varied very little.
Celestial mechanics tell us that a similar situation should
recur within 20,000 years, and for this reason the marine
isotopic stage 11 is often considered to be one of the best
analogues for the future climate. Many internal feedbacks
generated by the different components of the climate system
amplify or reduce the effect of the latitudinal and seasonal
distribution of insolation. Thus, climate projections on the
scale of a few hundreds of thousands of years require the
variations of both insolation and atmospheric CO 2 (and other
greenhouse gases) to be taken into account in models
including representations of the atmosphere, ocean, cryosphere, lithosphere and vegetation.
One of the first models of this type was developed for the
northern hemisphere at the University of Louvain-la-Neuve,
Belgium (Gallée 1992). It successfully reproduced the main
characteristics of the current climate, as well as the variations
in ice volume during glacial-interglacial cycles. Since the
model has been shown to provide reasonable simulations of
the 100,000-year cycle, it has been applied as a second step
to the simulation of future climates. Several tests have thus
been carried out on the climate of the next 130,000 years,
either by keeping the CO 2 constant at different levels
(290 ppm, 200 ppm and 250 ppm), or by using the CO 2
variations of the last glacial-interglacial cycle (Loutre and
Berger 2000). The results of these simulations suggest, on
the one hand, that the climate of the next 50,000 years is
particularly sensitive to the level of atmospheric CO 2 concentration and that our current interglacial period will be
much longer than any other one in the past. It could last
more than 55,000 years with CO 2 levels between 230 and
290 ppm: the first glacial stage would appear around 60,000
AD, and the next glacial maximum (in terms of ice volume)
would be around 100,000 AD, followed by a deglaciation
phase that would end around 120,000 AD. These results
suggest that the marked differences between our interglacial
period (present and future) and the previous Quaternary
warm periods are due to the small insolation variations that
characterize the former. Based on another scenario designed
to reproduce the natural variations of CO 2 and not the
anthropogenic contribution, other simulations have been
carried out, either with a more elaborate version of the
Loutre and Berger’s model, or with a climate model of
intermediate complexity coupled with a more sophisticated
model of the evolution of polar ice sheets (Ganopolski et al.
2016). The results of these simulations are in agreement with
those presented previously, and show that the next glacial
inception will be postponed by at least 100,000 years.
Although, the variations in summer insolation at 65°N
have long been considered as the pacemaker of
glacial-interglacial transitions, the above results show that
the level of atmospheric CO 2 will be a key parameter in the
future. While the natural variations of CO 2 along with the
insolation forcing exclude the possibility of a glacial episode
in the future for at least 50,000 years, it is possible that the
impact of the anthropogenic contribution will lead to a
complete disappearance of polar ice sheets, and that a return
to a glacial phase could occur only in 100,000 years. Coupled climate-ice sheet models, validated on the last
glacial-interglacial cycle, will make it possible to explore the
threshold values of CO 2 which could have long-term effects
on the fate of ice sheets.
474
S. Charbit et al.
showed that atmospheric CO 2 varied by about 100 ppm
during the transition from an ice age to an interglacial period,
and that these variations were correlated with changes in air
temperatures, suggesting a link between CO 2 and climate.
Today, this link is commonly accepted and attention is
focused on global warming due to the increase in greenhouse
gases in the atmosphere. However almost forty years ago, in
1972, the climatological community met in the United States
to discuss the imminence of the next ice age. This question
arose because geological data showed that for about a million
years, the Earth had alternated between cold episodes corresponding to the glaciation phases, and the warmer interglacial periods, with a pseudo-periodicity of 100,000 years.
Data available in 1972 showed that the previous two
interglacial periods had lasted about 10,000 years. Yet, the
Holocene, the interglacial period we are currently experiencing, has been going on for 10,000 years. It seemed
therefore reasonable to think that this warm phase would
soon end and give way to a new ice age. This reasoning was
based on the assumption that all interglacial periods are of
equal duration. However, although the first numerical
experiments carried out with statistical models indicated that
the cooling initiated 6000 years ago (after the Holocene
climate optimum) would continue in the future, Oerlemans
and Van der Veen (1984) of the University of Utrecht
showed, with an ice-sheet model, that the transition to a new
glacial phase would not occur before 50,000 years.
Based on data from marine sediments and ice cores, it is
now well established that the duration of an interglacial
period varies considerably from one climatic cycle to the
next. The same models used to predict the future were used
to simulate the last interglacial-glacial transition. Indeed,
taking into account the atmosphere-ocean feedbacks, these
models produce perennial snow in the Canadian archipelago
(Khodri et al. 2001). It is also known that the Earth’s orbital
parameters (see Chap. 28) that govern the seasonal and latitudinal distribution of solar energy can vary considerably
from one cycle to another. For example, geological data tell
us that the interglacial period occurring 400,000 years ago
(marine isotopic stage 11) was exceptionally long. This situation corresponds to a weak eccentricity where the seasonal
and latitudinal distribution of solar energy varied very little.
Celestial mechanics tell us that a similar situation should
recur within 20,000 years, and for this reason the marine
isotopic stage 11 is often considered to be one of the best
analogues for the future climate. Many internal feedbacks
generated by the different components of the climate system
amplify or reduce the effect of the latitudinal and seasonal
distribution of insolation. Thus, climate projections on the
scale of a few hundreds of thousands of years require the
variations of both insolation and atmospheric CO 2 (and other
greenhouse gases) to be taken into account in models
including representations of the atmosphere, ocean, cryosphere, lithosphere and vegetation.
One of the first models of this type was developed for the
northern hemisphere at the University of Louvain-la-Neuve,
Belgium (Gallée 1992). It successfully reproduced the main
characteristics of the current climate, as well as the variations
in ice volume during glacial-interglacial cycles. Since the
model has been shown to provide reasonable simulations of
the 100,000-year cycle, it has been applied as a second step
to the simulation of future climates. Several tests have thus
been carried out on the climate of the next 130,000 years,
either by keeping the CO 2 constant at different levels
(290 ppm, 200 ppm and 250 ppm), or by using the CO 2
variations of the last glacial-interglacial cycle (Loutre and
Berger 2000). The results of these simulations suggest, on
the one hand, that the climate of the next 50,000 years is
particularly sensitive to the level of atmospheric CO 2 concentration and that our current interglacial period will be
much longer than any other one in the past. It could last
more than 55,000 years with CO 2 levels between 230 and
290 ppm: the first glacial stage would appear around 60,000
AD, and the next glacial maximum (in terms of ice volume)
would be around 100,000 AD, followed by a deglaciation
phase that would end around 120,000 AD. These results
suggest that the marked differences between our interglacial
period (present and future) and the previous Quaternary
warm periods are due to the small insolation variations that
characterize the former. Based on another scenario designed
to reproduce the natural variations of CO 2 and not the
anthropogenic contribution, other simulations have been
carried out, either with a more elaborate version of the
Loutre and Berger’s model, or with a climate model of
intermediate complexity coupled with a more sophisticated
model of the evolution of polar ice sheets (Ganopolski et al.
2016). The results of these simulations are in agreement with
those presented previously, and show that the next glacial
inception will be postponed by at least 100,000 years.
Although, the variations in summer insolation at 65°N
have long been considered as the pacemaker of
glacial-interglacial transitions, the above results show that
the level of atmospheric CO 2 will be a key parameter in the
future. While the natural variations of CO 2 along with the
insolation forcing exclude the possibility of a glacial episode
in the future for at least 50,000 years, it is possible that the
impact of the anthropogenic contribution will lead to a
complete disappearance of polar ice sheets, and that a return
to a glacial phase could occur only in 100,000 years. Coupled climate-ice sheet models, validated on the last
glacial-interglacial cycle, will make it possible to explore the
threshold values of CO 2 which could have long-term effects
on the fate of ice sheets.
474
S. Charbit et al.