circulation model, the ocean surface temperatures and the
extent of sea ice have to be defined for all ocean-type points
in the atmospheric model. In practice, this is extremely
restrictive since at the very least, reconstructions of ocean
surface conditions with a spatial resolution similar to the
model for a typical seasonal cycle would be required. These
numerical experiments are almost always based on strong
assumptions as to the ocean surface conditions as well as the
land-ocean distribution, topography, and extent and altitude
of the ice sheets. The other forcings are better known, at least
for the relatively recent periods of the Quaternary: insolation
(Berger 1978; Laskar 2004) and atmospheric concentration
of greenhouse gases, with measurements from Antarctic ice
cores now dating as far back as 800,000 years.
Thus, a ‘realistic’ simulation of a paleoclimate is based on
several series of assumptions: those related to the design of
the model itself and those related to the fixing of boundary
conditions for a specific experiment. So, this simulation
summarizes both our knowledge of climate characterization
(incorporated into the climate model), and our understanding
of the forcings of this climate (expressed in the forcing and
boundary conditions). It is also constrained by the limitations of computers and technology of its time.
Take the example of the climate simulations for the Last
Glacial Maximum (LGM, about 21,000 years ago). The first
simulations of the climate of this period (Gates 1976) were
carried out shortly after the first reconstructions of the ocean
surfaces (sea surface temperature, extent of sea ice) and of
land (ice caps) were produced (CLIMAP 1976). At this time,
the LGM atmospheric concentration of CO 2 was not known.
It was only in the 1980s (Manabe and Broccoli 1985) that
the first numerical simulation was carried out taking into
Fig. 25.2 History of climate
models used in successive IPCC
reports. From the 4th IPCC report
(2007), Solomon et al. (Ed.),
Cambridge University Press,
http://www.ipcc.ch/publications_
and_data/ar4/wg1/en/contents.
html
25 Modeling and Paleoclimatology
329
extent of sea ice have to be defined for all ocean-type points
in the atmospheric model. In practice, this is extremely
restrictive since at the very least, reconstructions of ocean
surface conditions with a spatial resolution similar to the
model for a typical seasonal cycle would be required. These
numerical experiments are almost always based on strong
assumptions as to the ocean surface conditions as well as the
land-ocean distribution, topography, and extent and altitude
of the ice sheets. The other forcings are better known, at least
for the relatively recent periods of the Quaternary: insolation
(Berger 1978; Laskar 2004) and atmospheric concentration
of greenhouse gases, with measurements from Antarctic ice
cores now dating as far back as 800,000 years.
Thus, a ‘realistic’ simulation of a paleoclimate is based on
several series of assumptions: those related to the design of
the model itself and those related to the fixing of boundary
conditions for a specific experiment. So, this simulation
summarizes both our knowledge of climate characterization
(incorporated into the climate model), and our understanding
of the forcings of this climate (expressed in the forcing and
boundary conditions). It is also constrained by the limitations of computers and technology of its time.
Take the example of the climate simulations for the Last
Glacial Maximum (LGM, about 21,000 years ago). The first
simulations of the climate of this period (Gates 1976) were
carried out shortly after the first reconstructions of the ocean
surfaces (sea surface temperature, extent of sea ice) and of
land (ice caps) were produced (CLIMAP 1976). At this time,
the LGM atmospheric concentration of CO 2 was not known.
It was only in the 1980s (Manabe and Broccoli 1985) that
the first numerical simulation was carried out taking into
Fig. 25.2 History of climate
models used in successive IPCC
reports. From the 4th IPCC report
(2007), Solomon et al. (Ed.),
Cambridge University Press,
http://www.ipcc.ch/publications_
and_data/ar4/wg1/en/contents.
html
25 Modeling and Paleoclimatology
329