three-dimensional ocean model coupled with a simplified
atmospheric model. The simulated annual temperature
changes are very low, in accordance with the low amplitude
of the change in insolation. Between 60° N and 70° N, these
studies show cooling during the Holocene in response to the
reduction in summer insolation relatively in line with
observations (Fig. 30.3). The increase in greenhouse gas
concentrations since the Middle Holocene partly offsets the
reduction in temperature by about 0.5 °C. The beginning of
the Holocene is also a time when the ocean was still disturbed by the recent deglaciation and when there were larger
ice caps than are currently present in North America and
Greenland. Simulations (Renssen et al. 2005) indicate that
the main effect of the Fennoscandian cap at the beginning of
the Holocene was to delay the climate optimum in eastern
Canada and Greenland (Fig. 30.3). These studies also suggest an increase in variability during the Holocene, mainly
around the northern seas, resulting from feedbacks between
atmospheric circulation, ocean and sea ice.
The simulations show no major changes in thermohaline
circulation of the ocean during the Holocene, except perhaps
a slight increase between 9000 and 8000 years before the
present. It also appears that the initial conditions and the
long-term variations of the intermediate and deep ocean have
had little impact on the characteristics of the climate during
the Holocene. Thus, it can be taken that the different periods
are more or less in equilibrium with the forcings. Hence,
equilibrium simulations for a particular period of the Holocene can be used to correctly identify the main characteristics of the climate. This result is important because the
ignorance of oceanic conditions at the global scale at the
beginning of the Holocene is a source of uncertainty for
numerical simulations. Nevertheless, the equilibrium is not
necessarily verified over the last 4000 years because of a
200-year lag between the forcing and the climate response of
the reorganization of the boreal forest in the northern
hemisphere (Crucifix et al. 2002).
Reference Periods and Analysis of Feedbacks
Simulations from intermediate complexity models describe
the major trends and main feedbacks that have shaped the
Holocene in the mid and high latitudes of the northern
hemisphere but do not allow for a detailed analysis of
feedbacks or changes in variability. Some periods, such as
the beginning of the Holocene (9500 years ago) and the
Middle Holocene (6000 years ago), have been the subject of
numerous modeling studies and benefit from international
coordination (Joussaume and Taylor 1995; Joussaume et al.
1999; Kutzbach 1988; Braconnot et al. 2007a).
In particular, there is an effort to compile data at the
global scale for the Holocene (Prentice and Webb 1998), and
this helps, along with simulations, to improve our understanding of the mechanisms and the different feedbacks
involved and to assess the ability of climate models used for
climate projections for the next century to represent a different climate than the current one (Cane et al. 2006; Braconnot et al. 2012).
Feedbacks from Snow, Vegetation and from Sea
Ice in High Latitudes
Feedbacks from ocean, vegetation and snow and ice cover
are the main factors that have conditioned the evolution of
temperature in the different regions of the northern hemisphere during the Holocene. When forest is replaced by
grass, the reflective power of the combination of snow and
vegetation is more marked. This results in a positive loop of
increasing snow, cooling and withdrawal of the forest. This
mechanism in particular describes entry into the last
glaciation (de Noblet 1996). The ocean response leads to a
phase shift in the seasonal cycle due to its high thermal
inertia and evaporation capacity, which limits surface heating. Thus, spring and autumn are more affected than summer
and winter by oceanic feedbacks. These effects combine
with the feedback from sea ice and snow to reduce seasonal
contrasts at high latitudes and to amplify the cooling during
the Holocene (Wohlfahrt et al. 2004).
The feedback from snow and sea ice is manifested in a
modification of the surface albedo. It is therefore possible to
quantify it in terms of energy (W/m
2 ), using a simple
approach to climate sensitivity analysis (Taylor 2007). This
allows us to define this effect as:
Albedo effect ¼ ðSWncs pal À SWncs Ok ÞÀSWf
where SWf = (1 − a Ok )DSWi represents the radiative forcing
related to the change in insolation and SWncs represents the
net solar radiation (SW: short wave) in a clear sky (ncs) for
the paleoclimatological (pal) simulation and for the control
simulation (Ok). The radiative forcing takes into account the
model characteristics via the albedo of the control simulation. Indeed, although the same insolation disturbance SWi is
applied to different models, the difference in net incident
solar radiation depends on the selected model as a function
of the surface characteristics, the representation of clouds
etc. which characterize the planetary albedo in the model
(Fig. 30.6).
As an example, Fig. 30.6 shows that between February
and April, the estimated radiative forcing for two different
models ranges from −12 to −4 W/m
2 between 20° N and
50° N. The cooling associated with this insolation deficit
allows the snow to remain longer on land.
This snow introduces a local feedback of −4 to
−16 W/m
2 on the continent which reinforces the insolation
deficit and the cooling. Both simulations produce the same
type of feedback, but with different amplitudes and geographical localization. On the other hand, in summer, the
30 An Introduction to the Holocene and Anthropic Disturbance
431
atmospheric model. The simulated annual temperature
changes are very low, in accordance with the low amplitude
of the change in insolation. Between 60° N and 70° N, these
studies show cooling during the Holocene in response to the
reduction in summer insolation relatively in line with
observations (Fig. 30.3). The increase in greenhouse gas
concentrations since the Middle Holocene partly offsets the
reduction in temperature by about 0.5 °C. The beginning of
the Holocene is also a time when the ocean was still disturbed by the recent deglaciation and when there were larger
ice caps than are currently present in North America and
Greenland. Simulations (Renssen et al. 2005) indicate that
the main effect of the Fennoscandian cap at the beginning of
the Holocene was to delay the climate optimum in eastern
Canada and Greenland (Fig. 30.3). These studies also suggest an increase in variability during the Holocene, mainly
around the northern seas, resulting from feedbacks between
atmospheric circulation, ocean and sea ice.
The simulations show no major changes in thermohaline
circulation of the ocean during the Holocene, except perhaps
a slight increase between 9000 and 8000 years before the
present. It also appears that the initial conditions and the
long-term variations of the intermediate and deep ocean have
had little impact on the characteristics of the climate during
the Holocene. Thus, it can be taken that the different periods
are more or less in equilibrium with the forcings. Hence,
equilibrium simulations for a particular period of the Holocene can be used to correctly identify the main characteristics of the climate. This result is important because the
ignorance of oceanic conditions at the global scale at the
beginning of the Holocene is a source of uncertainty for
numerical simulations. Nevertheless, the equilibrium is not
necessarily verified over the last 4000 years because of a
200-year lag between the forcing and the climate response of
the reorganization of the boreal forest in the northern
hemisphere (Crucifix et al. 2002).
Reference Periods and Analysis of Feedbacks
Simulations from intermediate complexity models describe
the major trends and main feedbacks that have shaped the
Holocene in the mid and high latitudes of the northern
hemisphere but do not allow for a detailed analysis of
feedbacks or changes in variability. Some periods, such as
the beginning of the Holocene (9500 years ago) and the
Middle Holocene (6000 years ago), have been the subject of
numerous modeling studies and benefit from international
coordination (Joussaume and Taylor 1995; Joussaume et al.
1999; Kutzbach 1988; Braconnot et al. 2007a).
In particular, there is an effort to compile data at the
global scale for the Holocene (Prentice and Webb 1998), and
this helps, along with simulations, to improve our understanding of the mechanisms and the different feedbacks
involved and to assess the ability of climate models used for
climate projections for the next century to represent a different climate than the current one (Cane et al. 2006; Braconnot et al. 2012).
Feedbacks from Snow, Vegetation and from Sea
Ice in High Latitudes
Feedbacks from ocean, vegetation and snow and ice cover
are the main factors that have conditioned the evolution of
temperature in the different regions of the northern hemisphere during the Holocene. When forest is replaced by
grass, the reflective power of the combination of snow and
vegetation is more marked. This results in a positive loop of
increasing snow, cooling and withdrawal of the forest. This
mechanism in particular describes entry into the last
glaciation (de Noblet 1996). The ocean response leads to a
phase shift in the seasonal cycle due to its high thermal
inertia and evaporation capacity, which limits surface heating. Thus, spring and autumn are more affected than summer
and winter by oceanic feedbacks. These effects combine
with the feedback from sea ice and snow to reduce seasonal
contrasts at high latitudes and to amplify the cooling during
the Holocene (Wohlfahrt et al. 2004).
The feedback from snow and sea ice is manifested in a
modification of the surface albedo. It is therefore possible to
quantify it in terms of energy (W/m
2 ), using a simple
approach to climate sensitivity analysis (Taylor 2007). This
allows us to define this effect as:
Albedo effect ¼ ðSWncs pal À SWncs Ok ÞÀSWf
where SWf = (1 − a Ok )DSWi represents the radiative forcing
related to the change in insolation and SWncs represents the
net solar radiation (SW: short wave) in a clear sky (ncs) for
the paleoclimatological (pal) simulation and for the control
simulation (Ok). The radiative forcing takes into account the
model characteristics via the albedo of the control simulation. Indeed, although the same insolation disturbance SWi is
applied to different models, the difference in net incident
solar radiation depends on the selected model as a function
of the surface characteristics, the representation of clouds
etc. which characterize the planetary albedo in the model
(Fig. 30.6).
As an example, Fig. 30.6 shows that between February
and April, the estimated radiative forcing for two different
models ranges from −12 to −4 W/m
2 between 20° N and
50° N. The cooling associated with this insolation deficit
allows the snow to remain longer on land.
This snow introduces a local feedback of −4 to
−16 W/m
2 on the continent which reinforces the insolation
deficit and the cooling. Both simulations produce the same
type of feedback, but with different amplitudes and geographical localization. On the other hand, in summer, the
30 An Introduction to the Holocene and Anthropic Disturbance
431