Onset of Quaternary glaciation

The cause for the onset of widespread Northern Hemisphere glaciation about 2.7 million years ago remains disputable. Previous hypothesis related this event with a general decline of the atmospheric CO2 concentration or change in the North Atlantic circulation associated with the closure of the Panama seaway. In collaboration with an international group of scientists in 2005 we proposed in Nature a new hypothesis which relates the onset of the Northern Hemisphere glaciation with the changes in the hydrological regime of the Northern Pacific (Haug et al. 2005). Paleoclimate data indicate a strong and abrupt change in biogeochemical tracers in the Northern Pacific just prior to the onset of the Northern Hemisphere glaciation. This change was interpreted as the onset of permanent stratification in the Northern Pacific. In turn, using the Earth system model of intermediate complexity (EMIC) CLIMBER-2 model, we have shown that the establishing of permanent stratification in the Northern Pacific strongly affects seasonal variations of temperature and precipitation over Northern America and sufficient to trigger fast growth of continental ice sheets under a favourable orbital configuration.


Stability analysis

Although the Quaternary climate had phases of high transiency and was probably never in full equilibrium, an investigation of stability branches gives insight in the mechanism underlying the system. In particular, thresholds and bifurcations are best to be inspected in the phase space of the system. Further, a stability analysis shows the bounds in which a system operates. We found that the maximum summer insolation (the maximum of insolation during a season) at 65° N is a good proxy of orbital forcing. Of course, our model treats the full seasonal and latitudinal dependence of orbital forcing. But the response of the model, e.g. if one considers the mass balance of the ice sheets, correlates relatively well with the maximum boreal summer insolation. This is one of the reasons why our stability analysis is carried out is the phase space of maximum summer insolation. By slowly varying the Earth’s orbital parameters and, alternatively, by performing several equilibrium simulations with different temporal constant orbital parameters, we obtained a stability diagram of ice volume versus maximum summer insolation. The analysis showed that the climate-cryosphere system exhibits hysteresis behaviour in the phase space of maximum summer insolation (Calov and Ganopolski 2005). For the first time, this endeavour has been undertaken with an Earth system model of intermediate complexity. Because of that, the simulated positions of the glacial to interglacial transition and vice versa are based on a far more realistic approach than others.


Hysteresis diagram of the climate-cryosphere system. Click on image to enlarge.

New conceptual model of glacial cycles

Understanding of how variations in orbital parameters with periodicities of 20 and 40 kyr drive strongly asymmetric 100 kyr cycles represents a long-standing challenge for the scientific community. A number of conceptual models which successfully reproduce some important features of the glacial cycles have been proposed in recent decade. However, all these models were based on ad hoc assumptions which are difficult to verify. Results of the stability analysis of the climate-cryosphere system in the phase space of orbital forcing performed with the CLIMBER-2 model (Calov and Ganopolski 2005) provide an opportunity to design a new, physically based conceptual model of the glacial cycles. This model successfully reproduces the reconstructed from palaeodata ice volume variations, with the dominant 100-kyr cyclicity over the past million years and obliquity dominated cyclicity prior to 1 million years BP. The model explains glacial cycles as a direct and a strongly nonlinear response of the climate-cryosphere system to the Milankovitch forcing, whilst 100 kyr cyclicity results from the phase locking of the glacial cycles to the eccentricity cycle.

Abrupt climate changes and climate-cryosphere interaction

Palaeoclimate records indicate that abrupt climate changes were common during glacial age. A better understanding of interaction between climate and ice sheets is required to explain mechanism of abrupt climate changes. In the framework of ICE-QUEST project we continued the study of abrupt climate changes with the special emphasis on the bi-directional interaction between climate and cryospheric components of the Earth system. In particular, it has been shown in Arz et al. (2007) that the reorganizations of the Atlantic meridional circulations might affect considerably the mass balance of the Northern Hemisphere ice sheets, hence providing a strong feedback between oceanic and cryosphere processes. New simulations performed with the fully interactively and bi-directionally coupled climate and cryopshere components of the Earth system model CLIMBER-2 model reveal a rather good agreement between simulated and observed millennial scale variability during the glacial cycles.

Large-scale instabilities of ice sheets

Undoubtedly, the most exciting outcome of the inland-ice group at PIK is the first successful simulation of Heinrich-type surges of the Laurentide ice sheet during glacial time. Here, we again use the climate-system model CLIMBER-2 which is coupled with the thermomechanical ice-sheet model SICOPOLIS. In our model, Heinrich events are quasi-periodic self-sustained oscillations of the Laurentide ice sheet. The recurrence time of HEs is about 7000 years, while the duration of a Heinrich surge is some hundred years. During a Heinrich event, the ice thickness over Hudson Bay drops by about 1.2 km. The resulting rise in sea level agrees with records from palaeo research. A detailed description of the mechanism can be found in Calov et al. (2002). The animation (link to mpg file, 1.4 Mbyte) shows the Laurentide ice sheet in several surge events. A further contribution to ICE-QUEST is the model intercomparison initiative HEINO (Heinrich Event INtercOmparison) in which results from different ice-sheet models in a schematic model setup are inspected. HEINO is under the umbrella of the external project ISMIP. The goal is to identify the status quo of contemporary ice sheet model in simulation of self-sustained oscillations and try to propose improvements of model physics.


Laurentide ice sheet during a Heinrich event. Click on image to enlarge.

Glacial inception

One of the most enigmatic phenomena is the cyclic glaciation and deglaciation of the Earth’s surface during the Quaternary. As a first step for understanding, we performed simulations of last glacial inception with the Earth system model of intermediate complexity CLIMBER-2. We found that glacial inception can be understood as a bifurcation in the climate system (mainly) caused by snow albedo feedback when the summer insolation (incoming solar radiation) in the model drops below a certain value. Surprisingly high was the speed of increase of ice cover in northern North America at about 118 kyr BP. This is because for the increase in ice area during the first phase of glaciation the relatively fast components of the climate system as the atmosphere and the upper ocean are important players. The slower grows in ice volume follows after the fast increase in ice area. The different phases of glacial inception are shown in the figure and in the animation (link to mpg file, 0.8 Mbyte). We further investigated the roles of the ocean, terrestrial vegetation and atmospheric CO2 content for glacial inception. It shows that in our model ice cover develops although less expanded, if ocean (sea surface temperature) and vegetation are kept at its present-day value. The same happens when the atmospheric CO2 content is held constant on the pre-industrial 280 ppm. Only if constant present orbital forcing applies to the model while CO2 may drop glacial inception ceases completely. The conclusion from these findings are very clear: The orbital forcing and the snow albedo feedback are a very crucial factors to initiate glacial inception while the other factors only amplifying the increase in ice volume to the its value as found in the palaeo records.


Surface elevation of the Laurentide ice sheet during different phases of glacial inception. Click on image to enlarge.

Glacial cycles

As a further step towards an understanding of the Quaternary climate change, we simulated the last four glacial cycles with the EMIC CLIMBER-2. In addition to the (natural) orbital forcing, the atmospheric CO2 content is prescribed here, although we plan simulations without any artificial forcing. The modelled time series of ice volume agree well with the correspondent sea level change in the palaeodata. An even better agreement is achieved if the modelled δ18O is compared with that from marine records. The glaciation and deglaciation and, in particular, the termination phases are displayed properly by the model. The geographical distribution of ice cover is broadly due to the geographical findings. The animation (link to mpg file, 29 MByte) shows the last four glacial cycles from a model simulation.


Modelled δ18O in comparison measured with one. Click on image to enlarge.