Overestimation of elevation-melt feedback in uncoupled projections of ice sheet mass loss

The elevation feedback on melt has been identified as a key process to explain (Gregoire et al, Nature, 2012) and project (Aschwanden et al., Sci Advances, 2019; Ridley et al, J Clim, 2005; Vizcaino et al, Clim Dyn, 2008) long-term deglaciation. It is also central to the theory of ice sheet evolution hysteresis, deglaciation thresholds/tipping points, and the problem of reversibility (Garbe et al, Nature, 2020; Gregory et al, TC, 2020; Gregory et al, Nature, 2004, Robinson et al, Nature Clim. Change, 2012). Ice-sheet-model-only estimates of this feedback rely on a largely unexplored parameter, the so-called “lapse rate”. This parameter is defined as the rate at which the near-surface atmospheric temperature changes strictly due to surface elevation change.

In this work, we use coupled an uncoupled ice sheet and climate simulations with two different General Circulation/Earth System Models (Vizcaino et al, GRL, 2015; Muntjewerf et al, JAMES, 2019) to estimate the temperature lapse rate over the Greenland ice sheet as it deglaciates. We find that this lapse rate is highly variable over seasons, with much reduced lapse rates during summer over melting surfaces. We propose that uncoupled  state-of-the-art projections are likely overestimating deglaciation rates due to too high summer lapse rates over the ablation area.