Modeling the warm last interglacial period with CESM2.1

The modeling of paleoclimate has long been used to understand and explain past environments and as out-of-sample tests of models being used for projections of future climate. Although the last interglacial period ( ~ 129 to 116 thousand years ago ) was discussed in the First Assessment Report of the IPCC, it gained more prominence in the IPCC Fourth and Fifth Assessments. Proxy records indicate a much different climate in this period. Arctic summers were warmer and boreal forests spread northward, replacing tundra to the Arctic Ocean in some locations. Reconstructions indicate a global mean sea level of more than five meters higher than present, with these high sea levels lasting for several thousand years. The last interglacial period is thus an important period for testing our knowledge of interactions of the atmosphere, ocean, sea ice, ice sheets, and biosphere, during a warm period, with implications for the stability of the Greenland and Antarctic ice sheets.

Forced with the strong astronomically induced seasonal forcing of insolation as compared to present, the Last interglacial is one of two interglacial epochs included in the suite of Paleoclimate Modeling Intercomparison Project (PMIP4) simulations in the Coupled Model Intercomparison Project (CMIP6). NCAR scientists Bette Otto-Bliesner, Esther Brady, and Bob Tomas from the Paleo and Polar Climate section, completed the CMIP6 Tier 1 simulation for 127 ka with CESM2.1 at the same resolution as the CMIP6 DECK and future projection simulations.

Arctic summer sea ice thickness and annual surface mass balance (SMB) as simulated by CESM2.1 for 127 ka during the Last Interglacial
Figure: Arctic summer sea ice thickness and annual surface mass balance (SMB) as simulated by CESM2.1 for 127 ka during the Last Interglacial.

The results indicate a much warmer climate at high northern latitudes. Simulated Arctic sea ice remains through the summer but is much thinner than modern. This agrees with proxy indicators for LIG sea ice, such as the relatively new biomarker IP25 whose presence in sediments is direct proof for presence of past Arctic sea ice. As compared to modern, the simulated LIG surface mass balance (SMB) of the Greenland ice sheet indicates a more positive SMB than modern in the southeastern quadrant, associated with increased snowfall associated with the winter storm tracks in this warm interglacial. As well, the positive SMB at high elevations in central and northern Greenland agrees with Greenland ice core data. Regions of negative SMB, indicating that annual ablation exceeds accumulation, are expanded in the LIG simulation as compared to modern along the western and northern margins of the ice sheet. The next step in this project is to couple the Community Ice Sheet Model (CISM) to CESM to allow the Greenland ice sheet to respond to these SMB changes and assess its contribution to the LIG highstand.


  • Otto-Bliesner, B.L. et al., 2017: The PMIP4 contribution to CMIP6 – Part 2: Two interglacials, Scientific objectives and experimental design of the PMIP4-CMIP6 Holocene and Last Interglacial simulations. Geoscientific Model Development, 10, 3979-4003,