Climate Models: learning from the Past Long-Term Climate Variability

Projections of future climate change rely on global climate models. Such models simulate the dynamics and physics of the coupled atmosphere-ocean system and move towards models of the full Earth system. These models are evaluated on the present day mean climate as well as on climate variability over the last century. Although climate models agree on the overall patterns of future climate change, discrepancies are still displayed between models, both in amplitude and at the regional scale (e.g., IPCC AR5 (2007)). Past climates offer an unique opportunity to evaluate climate models on climate conditions different from the present-day conditions thanks to the reconstruction of past climates from paleodata (past vegetation, sea sediments and ice cores). They also allow to investigate the mechanisms of climate change and the importance of feedbacks in the climate system. The Paleoclimate Modelling Intercomparison Project, endorsed by both the international programs WCRP and IGBP, focuses on these issues (http://pmip2.lsce.ipsl.fr/). Main focus is on the last glacial maximum (LGM), ca 20 000 years before present (BP), and the mid-Holocene, ca 6000 years BP. The LGM simulation allows to examine the climate response to the presence of large ice sheets, cold oceans and lowered greenhouse gas concentrations. The mid-Holocene simulation allows to examine the climate response to a change in the seasonal and latitudinal distribution of incoming solar radiation caused by known changes in orbital forcing. Results show that the models used for future climate change projections are able to simulate the broad-scale features of both the LGM and mid-Holocene climate but that they tend to understand some of these features such as the polar amplification at the LGM and the northward extent of the African monsoon at the mid-Holocene. Debate however still prevails with regards to the change of the thermohaline circulation at the LGM. Paleoclimate modelling studies have also emphasized the importance of vegetation feedbacks at the regional scale, not yet included in future climate change studies. Other challenging issues with regards to future climate change are offered by long-term paleoclimate modelling studies such as ice-sheet melting at the last interglacial and during the last deglaciation, and understanding the past carbon cycle changes as inferred from ice cores.