Eemian Greenland Surface Mass Balance strongly sensitive to SMB model choice

Abstract

Abstract. Understanding the behavior of the Greenland ice sheet in a warmer climate, and particularly its surface mass balance (SMB), is important for assessing Greenland's potential contribution to future sea level rise. The Eemian interglacial, the most recent warmer-than-present period in Earth's history approximately 125,000 years ago, provides an analogue for a warm summer climate over Greenland. The Eemian is characterized by a positive Northern Hemisphere summer insolation anomaly, which introduces uncertainties in Eemian SMB when using positive degree day estimates. In this study, we use Eemian global 5 and regional climate simulations in combination with three types of SMB models -a simple positive degree day, an intermediate complexity, and a full surface energy balance model -to evaluate the importance of regional climate and model complexity for estimates of Greenland SMB. We find that all SMB models perform well under the relatively cool pre-industrial and late Eemian. For the relatively warm early Eemian, the differences between SMB models are large which is associated with the representation of insolation in the respective models. For all simulated time slices there is a systematic difference between 10 globally and regionally forced SMB models, due to the different representation of the regional climate over Greenland. We conclude that both the resolution of the simulated climate as well as the method used to estimate the SMB, are important for an accurate simulation of Greenland's SMB. Whether model resolution or SMB method is most important depends on the climate state and in particular the prevailing insolation pattern. We suggest that future Eemian climate model inter-comparison studies are combined with different SMB models to quantify Eemian SMB uncertainty estimates.

Eemian Greenland ice sheet simulated with a higher-order model shows strong sensitivity to SMB forcing

“…6). MAR-SEB and MAR-BESSI are two Eemian SMBs from a wide range of simulations analyzed in Plach et al (2018). Note that the same global climate model (NorESM) is used as a boundary condition for the SMB models.…”

“…The SMB is calculated with (1) a full surface energy balance model implemented within MAR (MAR-SEB) and 2an intermediate complexity SMB model (MAR-BESSI; BErgen Snow SImulator; BESSI; Born et al, 2018). These two SMB estimates are the best guess Eemian SMB simulations from a wider range of simulations discussed in Plach et al (2018). MAR-SEB is used as a control because it has been extensively validated against observations in previous studies (Fettweis, 2007;Fettweis et al, 2013Fettweis et al, , 2017 and MAR-BESSI is used to test the sensitivity of our ice sheet simulations to the SMB forcing (c.f., discussion in Sec.…”

“…The higher-order approximation is equally well suited to simulate slow as well as fast ice flow and applying it to the entire domain avoids any model-inherent discontinuities of "hybrid models" (i.e., combining SIA and SSA; Pollard and DeConto, 2009;Bueler and Brown, 2009;Pollard and DeConto, 2012;Aschwanden et al, 2016) at the boundaries between these two approximations. Plach et al (2018) show that the simulation of the Eemian SMB is strongly dependent on the choice of SMB model.…”

“…Here, we test SMB forcing derived from dynamically downscaled Eemian climate simulations and two SMB models (a full surface energy balance model and an intermediate complexity SMB model) as described in Plach et al (2018). Furthermore, we perform sensitivity experiments varying basal friction for the entire GrIS, as well as localized changes below the outlet glaciers.…”