Modeling the 100-kyr glacial-interglacial cycles

Investigations during the last twenty-!ve years have demonstrated that the astronomically related 19, 23 and 41-kyr quasi-periodicities actually occur in long records of the Quaternary climate. But the same investigations identi!ed also the largest climatic cycle as being about 100-kyr long. This cycle, the most striking feature of the Quaternary paleoclimate records, is characterized by long glacial periods followed by a short interglacial (!10 to 15 kyr long). Different sources for this so-called 100-kyr cycle have been found in the astronomical parameters and in the insolation itself. The most popular astronomical one is certainly the eccentricity with the largest spectral components around 100-kyr being 94 945, 123 297, 99 590 and 131 248 years (Berger, 1978). For insolation, it is known that there is only a very weak signal around 100-kyr coming from eccentricity itself. Moreover, the 100-kyr signal in eccentricity is fading away in the upper Pleistocene, at the same time that it appears to be stronger and stronger in paleoclimate records. Therefore, eccentricity cannot be related to either the orbital forcing or to the climate response by any simple linear mechanism. Actually, the variance components centered near the 100-kyr cycle seem to be in phase with the eccentricity cycle, but its exceptional strength in the climate record demands a non-linear ampli!cation. It was already suggested that this can be done by the ice sheet (Imbrie and Imbrie, 1980), the carbon cycle (Shackleton, 2000) and/or the ocean circulation (Imbrie et al., 1993), all arguments which imply that climate model must be used to test the origin of this 100-kyr cycle in paleoclimate records. Such a model has been developed in Louvain-la-Neuve for the Northern Hemisphere and used to perform sensitivity analyses to the astronomically-driven insolation changes and to the atmospheric CO2 concentration over the Quaternary. Assuming a CO2 concentration decreasing linearly from 320 ppmv at 3 Myr BP (late Pliocene) to 200 ppmv at the Last Glacial Maximum, the model simulates the intensi!cation of glaciation around 2.75 Myr BP, the late Pliocene–early Pleistocene 41-kyr cycle, the emergence of the 100-kyr cycle around 900 kyr BP, and the glacial–interglacial cycles of the last 600 kyr (Berger and Loutre, 2004). Simulations with different CO2 reconstructions over the last 1 Myr have con!rmed that the model can sustain the glacial–interglacial cycles of the late Pleistocene (Berger et al., 2004). Although the model results agree pretty well with the reconstruction in phase and amplitude over the last 400 kyr, before MIS-11 it neither keeps enough ice during the interglacials nor produces the reduced amplitude of the glacial–interglacial cycles as shown in deep-sea (Imbrie et al., 1984) and ice cores (EPICA, 2004)