1Department of Geoscience, University of Wisconsin-Madison, Madison, WI, USA
2NASA Goddard Institute for Space Studies and Columbia University, New York, NY, USA
3College of Earth, Ocean, and Atmospheric Sciences, Oregon State University, Corvallis, OR, USA
4Pacific Climate Impact Consortium, University of Victoria, Victoria, BC, Canada
5Department of Earth Sciences, University of New Hampshire, Durham, NH, USA
Abstract. Simulations of past climates require altered boundary conditions to account for known shifts in the Earth system. For the Last Glacial Maximum (LGM) and subsequent deglaciation, the existence of large Northern Hemisphere ice sheets provides a profound change in surface topography and albedo. While ice-sheet extent is fairly well known, numerous conflicting reconstructions of ice-sheet topography suggest that precision in this boundary condition is lacking. Here we use a high-resolution and oxygen-isotope-enabled fully-coupled global circulation model (GCM) (GISS ModelE2-R), along with two different reconstructions of the Laurentide Ice Sheet (LIS) that provide maximum and minimum estimates of LIS elevation, to assess the range of climate variability in response to uncertainty in this boundary condition. We present this comparison at two equilibrium time slices: the LGM, where differences in ice sheet topography are maximized, and 14 ka, where differences in maximum ice sheet height are smaller but still exist. Overall, we find significant differences in the climate response to LIS topography, with the larger LIS resulting in enhanced Atlantic meridional overturning circulation and warmer surface air temperatures, particularly over Northeast Asia and the North Pacific. These up and downstream effects are associated with differences in the development of planetary waves in the upper atmosphere, with the larger LIS resulting in a weaker trough over Northeast Asia that leads to the warmer temperatures and decreased albedo from snow and sea-ice cover. Differences between the 14 ka simulations are similar in spatial extent but smaller in magnitude, suggesting that climate is responding primarily to the larger difference in maximum LIS elevation in the LGM simulations. These results suggest that such uncertainty in ice-sheet boundary conditions alone may greatly impact the results of paleoclimate simulations and their ability to successfully simulate past climates, with implications for estimating climate sensitivity to greenhouse gas forcing utilizing past climate states.