Glacial cycles of the late Quaternary are shaped by the asymmetrically varying mass balance of continental ice sheets in the Northern Hemisphere. The surface mass balance is mostly positive during about four precssional periods and turns strongly negative at glacial terminations. The surface mass balance is governed by processes of ablation and accumulation. Here two ablation schemes, namely the positive-degree-day (PDD) method and the surface energy balance (SEB) approach, are compared in

more »

... ient simulations of the last glacial cycle with an Earth system model of intermediate complexity. The standard version of the CLIMBER-2 model simulates ice volume variations reasonably close to reconstructions. It uses the SEB approach which comprises fluxes of short-wave and long-wave radiation and of sensible and latent heat and accounts explicitly for snow albedo changes from dust deposition and snow aging. The PDD-driven ablation is computed offline in ensemble simulations to study the sensitivity with respect to short-term temperature variability and to melt factors for snow and ice. With standard literature values, the anomaly between the 130 ka-long ablation series from the two schemes is minimized but, more suitable are smaller values for inception than for termination and larger values for ice sheets in America than in Europe. Accordingly, PDD-online simulations require smaller values for inception than for termination to reproduce global ice volume variations. However, a reproduction at inception involves afterward excessive ice volume growth up to twice as large as reconstructed at LGM while a reproduction at termination implies ice volume growth about half as reconstructed at LGM. The PDD-online simulation with standard values generates at LGM a huge sea level drop of 250 m and a global cooling of 8 °C. The PDD-online simulation reproducing the LGM ice volume produces insufficient ablation at the turning point from glacial to interglacial climate, hence termination is delayed. According to our simulations, the SEB approach including effects of changing snow albedo, in particular at the American ice sheet margins, proves superior for simulating glacial cycles.