Simulations of Northern Hemisphere ice-sheet retreat:: sensitivity to physical mechanisms involved during the Last Deglaciation
Introduction
It is now commonly accepted that the succession of glacial–interglacial cycles are primarily driven by the changes in the seasonal distribution of insolation induced by variations of Earth's orbital parameters. However, this correlation is still subject to considerable uncertainty, such as the unexplained large amplitude of the 100 kyr period which corresponds to very small changes in insolation forcing (Berger, 1978). This suggests the existence of nonlinear amplification mechanisms (Hays et al., 1976; Imbrie et al., 1993; Paillard, 1998), such as atmospheric CO2 concentration (Verbitsky and Oglesby, 1992; Berger et al., 1999; Paillard, 2000; Shackleton, 2000), changes in the thermohaline circulation (Birchfield et al., 1994; Adkins et al., 1997), vegetation cover (Bonan et al., 1992; de Noblet et al., 1996; Gallimore and Kutzbach, 1996), or those related to the ice sheets. In particular, the ice-sheet configuration has been demonstrated to have a major impact on deep-water circulation (MacAyeal, 1993; Paillard and Labeyrie, 1994; Manabe and Stouffer, 1995; Paillard, 1995; de Noblet et al., 1996; Weaver et al., 1998), and atmospheric dynamics (Kutzbach and Wright Jr., 1985; Manabe and Broccoli, 1985; Broccoli and Manabe, 1987; Rind, 1987; Shinn and Barron, 1989; Ramstein and Joussaume, 1995; Felzer et al., 1996; Kageyama and Valdes, 2000). Moreover, internal processes related to the ice sheets, such as the strong positive feedbacks associated with the albedo of continental ice masses, the relationships between ice volume variations and isostatic rebound (Oerlemans, 1980; Pollard, 1982; Pollard, 1983; Hyde and Peltier, 1985; Lambeck, 1993; Crucifix et al., 2001) or the relationships between surface mass balance and both ice-sheet elevation and temperature, also have a strong influence on climate.
Numerical modeling is a valuable tool to investigate the influence of these factors and to better assess the mechanisms responsible for the succession of glacial–interglacial cycles, or for the millennial-scale climatic variability associated with the Heinrich events (Bond et al., 1992; Broecker et al., 1992) or the Dansgaard–Oeschger cycles (Bond et al., 1993; Dansgaard et al., 1993; Ganopolski and Rahmstorf, 2001; Paillard, 2001). Various models describing the dynamics of each component of the climate system (atmosphere, ocean, cryosphere, biosphere and lithosphere) exist. However, to account for the nonlinear interactions occurring between the different Earth's components, attention is now also focussed on the coupling of these different subsystems.
Due to their numerous interactions with other Earth's components, it is of key importance to analyze the response of the ice sheets to climatic conditions different from the present-day ones. To go a step forward, many studies have been devoted to the coupling between a simple climate model and an ice-sheet model to analyze the Northern Hemisphere ice-sheets response during an ice age (Deblonde and Peltier, 1991; Deblonde et al., 1992; Gallée et al., 1992; Berger et al., 1993; Deblonde and Peltier, 1993; Marsiat, 1994; Berger et al., 1996; Tarasov and Peltier, 1997; Berger et al., 1998; Li et al., 1998). These simple models allow for long-term simulations and hence, are able to provide some insights into the physics of the climate variability. However, they do not include any explicit representation of the atmospheric hydrological cycle which is indirectly responsible for snow accumulation over ice sheets. On the other hand, more realistic simulations are expected with the use of general circulation models (GCMs). Their resolution allows small-scale features to be captured, and therefore, they account for most of the dynamical processes at the origin of the water vapor transport and precipitation. In a recent study, Pollard (2000) compared the ice-sheet surface mass budgets, computed from climatic fields simulated by 17 GCMs and used the inferred results to drive a common snow–ice vertical column model, to predict net precipitation, melting and ablation over Antarctica and the Northern Hemisphere ice sheets for the present-day, the 6 kyr BP period and the Last Glacial Maximum (LGM). He found considerable scatter in the calculated mass balance of the Northern Hemisphere ice sheets, mainly caused by differences in the calculated annual melt rate due to GCM differences in summer air temperatures over the ablation areas. Earlier studies based upon the coupling of an atmospheric general circulation model (AGCM) with a 3D ice-sheet model (ISM) have been conducted to analyze the response of the Antarctic and Greenland ice sheets to an external forcing (Verbitsky and Oglesby, 1995). They found that the mean ice-sheet thickness was critically dependent on both ice viscosity and net snow accumulation. Fabre et al (1997), Fabre et al (1998) and Ramstein et al. (1997) examined whether the Northern Hemisphere ice-sheet configuration at the LGM, provided by a 3D thermomechanical ice-sheet model (Ritz et al., 1997), was consistent with polar ice-sheet reconstructions used to perform AGCM snapshots (Peltier, 1994). Satisfactory results were obtained with the use of the climate forcing simulated by the LMD5.3 AGCM. However, due to their high computational cost, the AGCMs can only provide snapshots of the climate; these latter experiments were thus performed under steady-state conditions, which is unrealistic since the climate did not remain constant during such a long period (10,000 yr integration for the ice-sheet model).
To analyze the relation between climate and ice sheets, we focussed on the Last Deglaciation. To simulate the evolution of the Northern Hemisphere ice sheets throughout this period more realistically, we use a time-evolving climatology to force a 3D-ice-sheet model (Ritz et al., 1997). This climatology is provided by the interpolation through time of climate snapshots simulated by the LMD5.3 AGCM at different periods of the Last Deglaciation (21, 15, 9, 6 kyr BP, and 0 kyr). The AGCM is forced by the astronomically derived insolation (Berger, 1978), the atmospheric CO2 content (Raynaud et al., 1993), the sea-surface temperatures, and the ice-sheet extent and topography and the associated sea-level changes (Peltier, 1994).
The aim of this paper is first to investigate whether the ice-sheet model, forced by the LMD5.3 AGCM produces a realistic deglaciation process. To evaluate the role of forcing mechanisms given by the set of AGCM boundary conditions, we performed a first experiment, called hereafter the standard experiment, with a linear time interpolation between AGCM snapshots. Although the marine sedimentary records (Bard et al., 1987; Lehman and Keigwin, 1992; Bond et al., 1993; Keigwin and Jones, 1995) and the ice-core data (Dansgaard et al., 1989; Johnsen et al., 1992; Grootes et al., 1993; Alley et al., 1993; Dansgaard et al., 1993) have demonstrated the existence of a large millennial-scale climatic variability, which is not accounted for by this linear interpolation, this approach is a first step to test the ISM response to a reasonable climatic evolution.
Secondly, we evaluate the impact of the forcing climate and of processes related to the internal dynamics of the Northern Hemisphere ice sheets. To do this, several sensitivity studies are carried out to examine the error induced by the time interpolation method and to investigate the influence of feedback mechanisms associated with the interactions between ice-sheet elevation and surface temperature. Finally, the response of ice sheets to the parameterizations of the ice flow (i.e. resulting from ice deformation or basal sliding) is also analyzed.
Section snippets
Description of the AGCM
The climate model used in this study is the LMD5.3 AGCM (Harzallah and Sadourny, 1995), developed at the Laboratoire de Météorologie Dynamique (CNRS, Paris). This version uses 64 points regularly spaced in longitude (resolution of 5.625°), 50 points regularly spaced in sine of latitude (each horizontal grid cell has a constant area), and 11 unevenly spaced vertical levels.
The atmosphere–biosphere exchanges are taken into account with the SECHIBA land hydrology model (Ducoudré et al., 1993). The
Temperature and precipitation anomalies
The AGCM fields driving the ISM are the mean annual and summer 2-m air temperature and mean annual precipitation. To minimize the errors due to GCM model deficiencies, we used a perturbative method of the present-day climate: the anomaly of temperature deduced from the AGCM (i.e. the difference between the simulated control and a given past climate) is added to the present-day observed temperature (Legates and Willmott, 1990). This method is implicitly based upon the assumption that the errors
Results of the standard experiment
Our first experiment is designed to investigate whether the ice-sheet model driven by a climatology derived from AGCM simulations is able to account for the deglaciation process. The time-evolving climatology is provided by a linear interpolation through time of the anomalies of temperature and precipitation. Although it is well known that the deglaciation is a nonlinear process, this approach allows us to test the ISM response to a reasonable climatic evolution, and aims to evaluate the role
Impact of millennial-scale climate variability
One of the main reasons which could be at the origin of the lag of the simulated deglaciation process may be linked to the linear climate evolution we used to force the ice-sheet model. A linear climate forcing does not capture high frequency events observed in marine sedimentary records and ice core data. To evaluate the impact of the millennial-scale forcing, we performed an additional experiment using a climate evolution calibrated against the GRIP δ18O record (see Table 2, Experiment no.2).
Conclusions
The Last Deglaciation of the Northern Hemisphere continental ice sheets has been simulated with a 3-dimensional thermomechanical ice-sheet model driven by a time-evolving climatology provided by the time interpolation between AGCM snapshots simulated for different periods of the Last Deglaciation, from the LGM to the present. Although it is well known that the deglaciation is a nonlinear process, the so-called standard experiment, characterized by a linear interpolation between snapshots, and
Acknowledgements
We are very grateful to Masa Kageyama for helping us to improve the writing of the manuscript. Thanks are also expressed to William Hyde and to an anonymous reviewer for their fruitful comments which helped us to improve the clarity of the manuscript. This work is supported by the Programme National d’Etudes du Climat (PNEDC, CNRS). Computer time on Cray C90 was provided by the Institut du Développement et des Ressources en Informatique Scientifique (IDRIS/CNRS) for AGCM simulations. The
References (108)
Long-term variations of caloric insolation resulting from the Earth's orbital elements
Quaternary Research
(1978)- et al.
Modelling Northern hemisphere ice volume over the last 3 Ma
Quaternary Science Reviews
(1999) - et al.
The role of ocean-atmosphere reorganizations in glacial cycles
Geochimica et Cosmochimica Acta
(1989) - et al.
Geomorphological reconstruction of the Labrador sector of the Lautrentide Ice Sheet
Quaternary Science Reviews
(2000) - et al.
Effect of isostatic rebound on modelled ice volume variations during the last 200 kyr
Earth and Planetary Science Letters
(2001) - et al.
Simulations of continental ice sheet growth over the last glacial-interglacial cycleexperiments with a one-level seasonal energy balance model including seasonal ice albedo feedback
Palaeogeography, Palaeoclimatology, Palaeoecology (Global and Planeatry Change Section)
(1992) - et al.
Carbon stocks and isotopic budgets of the terrestrial biosphere at mid-Holocene and last glacial maximum times
Chemical Geology
(1999) - et al.
The Greenland ice sheet and greenhouse warming
Palaeogeography, Paleoclimatology, Palaeoecology (Global and Planetary Change section)
(1991) - et al.
Simulation of the climate of 18,000 yr BPresults for the North American/North Atlantic/European sector and comparison with the geologic record of North America
Quaternary Science Reviews
(1985) Glacial rebound and sea-level changean example of a relationship between mantle and surface processes
Tectonophysics
(1993)