Greenland surface mass balance variability throughout the last deglaciation

Modeling the surface mass balance (SMB) from long-term climate simulations with Earth System Models (ESMs) remains a major challenge, as the spatial resolution of ESMs is often too coarse to represent processes governing accumulation and melt accurately. Hence, previous studies of the SMB are limited to the last century, due to the availability of observations and the computational limitations of regional climate models. By using transient simulations of the last 26,000 years with a comprehensive ESM in combination with a sophisticated energy balance model (EBM), we extend previous research and study changes in the Greenland SMB and equilibrium line altitude (ELA) for deglacial climate conditions.

For the simulation of the deglaciation, the ESM is forced with atmospheric greenhouse gases, insolation and ice sheets from reconstructions. The EBM is used to downscale the SMB from the coarse resolution ESM grid onto a higher spatial resolution, which allows to determine SMB variations caused by topographic gradients not resolved by the ESM. It accounts for changes in the snow albedo due to varying snow properties (age, depth, melting and refreezing of water) and cloud cover, as well as key physical processes like percolation and refreezing of meltwater. An evaluation of the setup for historical climate conditions shows that derived SMBs are in good agreement with SMBs from regional modeling.

Throughout the deglaciation, changes in insolation dominate the Greenland SMB: 1) The increase in insolation and associated warming early in the deglaciation result in an ELA and SMB increase. The SMB increase is largely driven by changes in the accumulation, as a warmer atmosphere precipitates more. Only after 13 ka before present (BP) melt begins to dominate and the SMB decreases. 2) The decline in insolation after 9 ka BP leads to an increasing SMB and decreasing ELA. Superimposed on these long-term changes is significant millennial-scale variability, associated with episodes of significant SMB/ELA decreases. Drivers of this variability are slowdowns of the Atlantic Meridional Overturning Circulation (AMOC) that lead to cooling over most of the Northern Hemisphere.