GEOPHYSICAL RESEARCH LETTERS, VOL. 28, NO. 11, PAGES 2221-2224, JUNE 1, 2001
A simulation of the Last Glacial Maximum with a
coupled atmosphere-ocean GCM
Akio Kitoh, Shigenori Murakami
Meteorological Research Institute, Tsukuba, Ibaraki 305-0052, Japan
Hiroshi Koide
Japan Meteorological Agency, Chiyoda, Tokyo 100-8122, Japan
Abstract. A simulation of the Last Glacial Maximum
(LGM) climate has been performed by a global coupled
atmosphere-ocean general circulation model (AOGCM). Two
simulations are conducted for the LGM with different initial
conditions: one with the present-day initial condition and
the other with a pre-conditioning of fresh water flux over
the North Atlantic. After more than 200 year integration
both LGM simulations attained a similar quasi-equilibrium
state. The global mean surface air temperature dropped 3.9
◦
C and precipitation decreased 11% compared to the present
day simulation. The sea surface temperature dropped 1.7◦ C
in the tropics, with larger decreases in the Atlantic than in
the Indo-Pacific. There is a region with warmer than present
sea surface temperatures over the subtropical eastern Pacific. The thermohaline circulation in the North Atlantic in
the LGM simulation is slightly stronger than in the present
day simulation.
Introduction
Simulations of paleoclimate with general circulation models (GCM) are useful to examine the mechanism of climate
and its variability, and to validate climate models under
different climate conditions from the present by comparing with available geological evidence. Since the geological
evidences are relatively abundant for recent 20,000 years,
many paleoclimate simulations targeting for this period have
been conducted with atmospheric GCMs and mixed-layer
ocean coupled GCMs within the Paleoclimate Modeling
Intercomparison Project (PMIP) [Joussaume and Taylor,
1995]. However these simulations do not include ocean dynamics explicitly. Recently, there emerge some experiments
with intermediate coupled models [Ganopolski et al., 1998;
Weaver et al., 1998] and a coupled atmosphere-ocean GCM
[Bush and Philander, 1999]. In this study, we conduct a
simulation of Earth’s climate in the Last Glacial Maximum
(LGM: 21,000 years ago) with the Meteorological Research
Institute (MRI) global coupled atmosphere-ocean GCM. We
conducted two simulations with different initial conditions
for LGM: one with the present-day initial condition and the
other with a pre-conditioning of fresh water flux over the
North Atlantic.
Copyright 2001 by the American Geophysical Union.
Paper number 2000GL012271.
0094-8276/01/2000GL012271$05.00
Model and experiment
We use the MRI CGCM1 [Tokioka et al., 1995] for this
simulation. The atmospheric part of the model has horizontal resolution of 4◦ latitude by 5◦ longitude, and 15 vertical
levels with the model top at 1 hPa. This atmospheric GCM
was used in the PMIP [Pinot et al., 1999a; Kageyama et al.,
2001]. The oceanic part adopted variable resolution 0.5◦ ∼
2.0◦ for latitudinal direction with finer grid in the tropics and
fixed 2.5◦ for longitudinal direction. There are 21 vertical
levels with the bottom at depth 5000 m. A realistic sea ice
model is included. Seasonal and diurnal cycles are included
in model integration. The model uses flux adjustments for
heat and fresh water.
For the present-day simulation (control run), the present
Earth’s orbital parameters and CO2 concentration value of
345 ppm are used as the boundary conditions. The control
run is integrated for 150 years with flux adjustment of heat
and fresh water flux at atmosphere-ocean boundary. For
the LGM, the orbital parameters are changed to the corresponding values and CO2 concentration is reduced to 200
ppm. The orographic difference from the present with ice
sheet distribution [Peltier, 1994] and the albedo over the ice
sheet are considered. The effect of sea level drop in LGM is
reflected on the altitude of land. The change in land-sea distribution is ignored to use the same flux adjustment values
with the control run.
The first LGM experiment started from January of the
31st year of the control run without any pre-conditioning of
sea ice, temperature and salinity in the oceans. The model
is integrated for 70 years with the maximum 24-times acceleration in the deep oceans (equivalent to 1680 years integration), followed by further 140 years’ integration without
the acceleration. In this experiment, the model could not
show the vast sea ice extent over the North Atlantic at that
era. The reason of this might be that the initial conditions
of the sea ice and oceans taken from the control run are not
suitable for the LGM experiments.
In order to create more plausible oceanic initial conditions, we put freshwater input over the North Atlantic. Intermittent massive freshwater injection can cause a reduction of sea surface salinity and a weakening of the thermohaline circulation (THC). It forms a density stratification
of sea water in a near surface layer, resulting in a suitable
condition for generation of sea ice. We adopt an experimental design similar to that used by Manabe and Stouffer [1995]. For the first 55 years in this simulation, we use
a preconditioning which simulates the freshwater input to
the North Atlantic with the maximum value at the North
2221
2222
KITOH ET AL.: CGCM SIMULATION OF THE LGM CLIMATE
ice boundary in the North Atlantic is larger than the present
(Fig. 2a), but is significantly underestimated compared to
the CLIMAP [1981] reconstruction. However, a recent reconstruction by Pinot et al. [1999b] suggests much warmer
LGM SST than the CLIMAP, and the simulated sea ice distribution is closer to that obtained by Pinot et al. [1999b].
Without a continuous freshwater injection, the once weakened THC recovers and the expanded sea ice retreats. At
the LGM quasi-equilibrium, a strong cooling at the sea ice
boundary results in a stronger meridional overturning in the
Atlantic than the present (Fig. 3a). The stronger THC at
the LGM simulation seems to be inconsistent with the reconstructed THC [e.g. Duplessy et al., 1988]. This result suggests that the experimental conditions used (i.e. the orbital
parameters, CO2 concentration and continental ice sheets)
may not be enough to fully reproduce the LGM climate.
Figure 1. Time series of (a) global mean SST [◦ C] and (b)
meridional overturning in the Atlantic Ocean at 38◦ N, 1100m
[Sv]. Closed circles are for the LGM simulation without fresh
water preconditioning, and open circles are for the LGM simulation with fresh water preconditioning. A thin solid line denotes
the value corresponding to the last 40 years of the present-day
simulation.
American coast. We regard this period of time integration
as a preconditioning. After then, we removed this additional
freshwater input, and integrated the model without the acceleration for 215 years. Hereafter we refer to this run with
freshwater preconditioning simply as the LGM run.
Results
Figure 1 shows the time series of the global mean sea
surface temperature (SST) and the meridional overturning
in the North Atlantic at 38 ◦ N, 1100m. Freshwater input
dilutes sea surface salinity, causes a weakening of the THC
and generates large amount of sea ice (Fig. 2a and b). After
the preconditioning period, the weak THC recovers slowly
throughout 200 years. At the end of this integration, a quasiequilibrium state is obtained, at least in the upper oceans.
This state is very similar to the quasi-equilibrium state without a preconditioning. In the following analyses we use the
average of the last 40 years of this LGM run and compared
with the present-day run.
Figure 2c shows the sea ice distribution in February and
Fig. 3b shows the meridional overturning in the Atlantic
Ocean at the LGM quasi-equilibrium stage. Simulated sea
Figure 4 shows the 40-year mean differences (LGM minus Present) of (a) 500 hPa geopotential height, (b) surface
air temperature, (c) surface wind vectors, and (d) precipitation. Surface air temperature falls in LGM almost all
over the earth, except for small areas in the subtropical Pacific (Fig. 4b). Global average surface air temperature is
3.9◦ C colder in the LGM simulation than in the control.
The largest anomaly is found in the high latitudes of the
Northern Hemisphere. Temperature falls about 20◦ C and
10◦ C over the Laurentide ice sheet and over the Fennoscandian ice sheet, respectively. Both the albedo and lapse rate
effects are contained in these differences. Over the oceans,
temperature decrease is larger in the northwestern North
Pacific and North Atlantic. The large temperature drop
in the northwester North Pacific is due to an expanse of
sea ice. In the tropics (20◦ S-20◦ N), the LGM SST is 1.7◦ C
lower than the present-day simulation. The model simulates
more pronounced cooling in the Atlantic (3.0 ◦ C) than in the
Indian-Pacific Ocean (1.4◦ C). This inter-basin difference in
the SST drop is in close agreement with paleotemperature
proxies [e.g. Rosell-Melé et al., 1998].
Precipitation generally decreased due to cooler climate
(Fig. 4d). Annually averaged global mean precipitation
in LGM (2.59 mm d−1 ) is 11% less compared to that in
the present (2.91 mm d−1 ). Weakened Afro-Asian summer
monsoon resulted in significantly less precipitation there.
Northern Brazil is also the region with significantly less precipitation. On the other hand there are regions with more
precipitation such as southwestern North America and the
Andes. They are generally in accord with reconstruction
data [Farrera et al., 1999]. Over the oceans decreased precipitation is simulated where the model climate (control)
shows heavy precipitation such as the maritime continent,
Figure 2. Sea ice concentration in February: (a) present, (b) at the end of preconditioning (year 55), (c) LGM (years 251-290).
KITOH ET AL.: CGCM SIMULATION OF THE LGM CLIMATE
2223
Figure 3. Meridional streamfunction integrated over the Atlantic sector: (a) present, (b) LGM (years 251-290). Contour interval
is 3 Sv.
ITCZ and SPCZ. Compensating increased precipitation is
found over the subtropical Pacific where local SST anomaly
is positive.
Surface air temperature in the model is warmer in LGM
over the subtropical Pacific region around 20◦ N, 120◦ ∼
180◦ W. The surface wind converges toward this warm SST
anomaly region (Fig. 4c). Larger precipitation anomalies
(Fig. 4d) are found over these regions. This feature is in
controversy as the CLIMAP [1981] reconstruction shows a
similar warming but recent work by Lee and Slowey [1999]
shows a cooling. Other GCM experiments so far did not reproduce this warming feature [e.g. Dong and Valdes, 1998].
A weaker subtropical anticyclone (Fig. 4c) is in favor of
dynamically warming the upper oceans there. Kitoh [1997]
obtained a similar SST anomaly pattern as an effect of orography and discussed the mechanism in detail.
In the Northern Hemisphere mid-latitudes, the Aleutian
Low is much stronger in the LGM, accompanied by an anticyclonic circulation over Alaska. These circulation anomalies are primarily caused by the orographic effect of the Lau-
Figure 4. Differences between the LGM and the present-day simulation. (a) 500 hPa geopotential height [m], (b) surface air
temperature [◦ C], (c) surface wind vector, and (d) precipitation [mm/day].
2224
KITOH ET AL.: CGCM SIMULATION OF THE LGM CLIMATE
rentide ice sheet that blocks the mid-latitude westerly winds.
Over the central North Pacific, the westerly jet is stronger
and southward-shifted in accord with the stronger meridional thermal contrast there, and the jet stream circumvents the ice sheet northward over the northern North America. Therefore, temperature decrease is attenuated from the
southern Alaska to off California by the warm southwesterly wind. Hall et al. [1996] reported qualitatively similar features using the AGCM with the CLIMAP prescribed
SST, while the same model with computed SSTs with a simple slab ocean obtained different results [Dong and Valdes,
1998].
Concluding remarks
The atmospheric circulation, precipitation, sea water
salinity and sea ice distribution in the LGM simulation with
the MRI coupled GCM share many similarities with the observed evidences and the previous modeling results. The
simulated sea ice distribution in the LGM over the North
Atlantic, however, is much smaller than that indicated by
the CLIMAP [1981] reconstruction. The quasi-equilibrium
state started from the initial conditions generated by a preconditioning is similar to that started from the present day
initial conditions. This fact shows that the quasi-equilibrium
state is independent of initial conditions such as the initial
state of the THC, within the framework of this experimental configuration. Certainly a much longer integration with
non flux-adjusted models is desirable, but more work can
be done with current models to investigate climate change
mechanisms for the LGM. Our result suggests that some
mechanism causing frequent weakening of the THC, such as
intermittent fresh water input, may be needed to reproduce
the climate of the LGM.
Acknowledgments. Computations are performed at the
Center for Global Environmental Research of the National Institute for Environmental Studies.
References
Broccoli, A.J., Tropical cooling at the Last Glacial Maximum: An
atmosphere-mixed layer ocean model simulation. J. Climate,
13, 951-976, 2000.
Bush, A.B.G. and S.G.H. Philander, The climate of the Last
Glacial Maximum: Results from a coupled atmosphere-ocean
general circulation model. J. Geophys. Res., 104, 24509-24525,
1999.
CLIMAP, Seasonal reconstructions of the earth’s surface at the
Last Glacial Maximum. Geological Soc. America, Map Chart
Ser., MC-36, 1981.
Dong, B. and P.J. Valdes, Simulations of the Last Glacial Maximum climates using a general circulation model: prescribed
versus computed sea surface temperatures. Clim. Dyn., 14,
571-591, 1998.
Duplessy, J.C., N.J. Shackleton, R.G. Fairbanks, L. Labeyrie, D.
Oppo and N. Kallel, Deepwater source variations during the
last climatic cycle and their impact on the global deepwater
circulation. Paleoceanography, 3, 343-360, 1988.
Farrera, I., S.P. Harrison, I.C. Prentice, G. Ramstein, J. Guiot,
P.J. Bartlein, R. Bonnefille, M. Bush, W. Cramer, U. von
Grafenstein, K. Holmgren, H. Hooghiemstra, G. Hope, D.
Jolly, S.-E. Lauritzen, Y. Ono, S. Pinot, M. Stute and G. Yu,
Tropical climates at the Last Glacial Maximum: a new synthesis of terrestrial palaeoclimate data. I. Vegetation, lake-levels
and geochemistry. Clim. Dyn., 15, 823-856, 1999.
Ganopolski, A., S. Rahmstorf, V. Petoukhov and M. Claussen,
Simulation of modern and glacial climate with a coupled model
of intermediate complexity. Nature, 391, 351-356, 1998.
Hall, N.M.J., P.J. Valdes and B. Dong, The maintenance of the
last great ice sheets: A UGAMP GCM study. J. Climate, 9,
1004-1019, 1996.
Joussaume, S. and K.E. Taylor, Status of the Paleoclimate Modeling Intercomparison Project (PMIP). Proceedings of the First
International AMIP Scientific Conference., Monterey, USA,
WCRP-92, WMO/TD-No.732, 425-430, 1995.
Kageyama, M., O. Peyron, S. Pinot, P. Tarasov, J. Guiot, S. Joussaume and G. Ramstein, The Last Glacial Maximum climate
over Europe and western Siberia: a PMIP comparison between
models and data. Clim. Dyn., 17, 23-43, 2001.
Kitoh, A., Mountain uplift and surface temperature changes. Geophys. Res. Lett., 24, 185-188, 1997.
Lee, K.E. and N.C. Slowey, Cool surface waters of the subtropical
North Pacific Ocean during the last glacial. Nature, 397, 512514, 1999.
Manabe, S. and R.J. Stouffer, Simulation of abrupt climate
change induced by freshwater input to the North Atlantic
ocean. Nature, 378, 165-167, 1995.
Peltier, W.R., Ice age paleotopography. Science, 265, 195-201,
1994.
Pinot, S., G. Ramstein, S.P. Harrison, I.C. Prentice, J. Guiot,
M. Stute and S. Joussaume, Tropical paleoclimates at the
Last Glacial Maximum: comparison of Paleoclimate Modeling
Intercomparison Project (PMIP) simulations and paleodata.
Clim. Dyn., 15, 857-874, 1999a.
Pinot, S., G. Ramstein, I. Marsiat, A. de Vernal, O. Peyron, J.C.
Duplessy and M. Weinelt, Sensitivity of the European LGM
climate to North Atlantic sea-surface temperature. Geophys.
Res. Lett., 26, 1893-1896, 1999b.
Rosell-Melé, A., E. Bard, K.-C. Emeis, P. Farrimond, J. Grimalt,
P.J. Müller and R.R. Schneider, Project takes a new look at
past sea surface temperatures. EOS Trans., 79, 393-394, 1998.
Tokioka T., A. Noda, A. Kitoh, Y. Nikaidou, S Nakagawa, T. Motoi, S. Yukimoto and K. Takata, A Transient CO2 Experiment
with the MRI CGCM –Quick report–. J. Meteor. Soc. Japan,
73, 817-826, 1995.
Weaver, A.J., M. Eby, A.F. Fanning and E.C. Wiebe, Simulated
influence of carbon dioxide, orbital forcing and ice sheets on the
climate of the Last Glacial Maximum. Nature, 394, 847-853,
1998.
A. Kitoh and S. Murakami, Climate Research Department, Meteorological Research Institute, 1-1 Nagamine,
Tsukuba, Ibaraki 305-0052, Japan. (e-mail: [email protected],
[email protected])
H. Koide, Climate Prediction Division, Japan Meteorological
Agency, 1-3-4 Otemachi, Chiyoda, Tokyo 100-8122, Japan. (email: [email protected])
(Received August 29, 2000; revised March 27, 2001;
accepted March 27, 2001.)