Replicating the Ice-Volume Signal of the Early Pleistocene with a Complex Earth System Model

Milankovitch theory proposes high-latitude summer insolation intensity paces the ice ages by controlling perennial snow cover amounts (Milankovitch, 1941). According to theory, the ~21 kyr cycle of precession should dominate the ice-volume records since it has the greatest influence on high-latitude summer insolation. Modeling experiments frequently support Milankovitch theory by attributing the majority of Northern Hemisphere high-latitude summer snowmelt to changes in the cycle of precession (e.g. Jackson and Broccoli, 2003). However, ice-volume proxy records, especially those of the Early Pleistocene (2.6-0.8 Ma), display variability with a period of ~41 kyr (Raymo and Lisiecki, 2005), indicative of insolation forcing from obliquity, which has a much smaller influence on summer insolation intensity than precession. Several hypotheses attempt to explain the discrepancies between Milkankovitch theory and the proxy records by invoking phenomena such as insolation gradients (Raymo and Nisancioglu, 2003), hemispheric offset (Raymo et al., 2006; Lee and Poulsen, 2009), and integrated summer energy (Huybers, 2006); however, all of these hypotheses contain caveats (Ruddiman, 2006) and have yet to be supported by modeling studies that use a complex GCM. To explore potential solutions to this '41 kyr problem,' we use an Earth system model composed of the GENESIS GCM and Land Surface model, the BIOME4 vegetation model, and the Pennsylvania State ice-sheet model. Using an asynchronous coupling technique, we run four idealized transient combinations of obliquity and precession, representing the orbital extremes of the Pleistocene (Berger and Loutre, 1991). Each experiment is run through several complete orbital cycles with a dynamic ice domain spanning North America and Greenland, and fixed preindustrial greenhouse-gas concentrations. For all orbital configurations, model results produce greater ice-volume spectral power at the frequency of obliquity despite significantly greater summer insolation variability from the cycle of precession. We find obliquity enhances the climate sensitivity to direct insolation forcing through positive high-latitude surface feedbacks between vegetation, sea-ice, and mean-annual insolation while the seasonal dichotomy of precessional forcing leads to climate counterbalancing that dampens the annual ice-volume response. Longer cycle duration further amplifies the ice-volume response to obliquity. Our results help remedy the discrepancies between Milankovitch theory and the ice-volume proxy records. However, summer insolation intensity remains the most important factor for determining ice-volume rate-of-change in our experiments. Consequently, we still find a significant ice-volume response to precession, which is inconsistent with the Early Pleistocene records. The disconnect is likely attributable to climate phenomena not accounted for in the model or our choice of initial conditions, which are poorly constrained for the Early Pleistocene and ice-sheet modeling in general. Future work will examine the importance of initial climate conditions on ice-volume response.