Simulated Greenland Surface Mass Balance in the GISS ModelE2 GCM: Role of the Ice Sheet Surface

The rate of growth or retreat of the Greenland and Antarctic ice sheets remains a highly uncertain component of future sea level change. Here we examine the simulation of Greenland ice sheet surface mass balance (GrIS SMB) in the NASA Goddard Institute for Space Studies (GISS) ModelE2 General Circul...

Full description

Bibliographic Details
Main Authors: Tedesco, M., Nowicki, S. M. H., Alexander, P. M., Fischer, E., Kelley, M., Schmidt, G. A., Fettweis, X., LeGrande, A. N.
Format: Other/Unknown Material
Language:unknown
Published: 2019
Subjects:
Online Access:http://hdl.handle.net/2060/20190000703
Description
Summary:The rate of growth or retreat of the Greenland and Antarctic ice sheets remains a highly uncertain component of future sea level change. Here we examine the simulation of Greenland ice sheet surface mass balance (GrIS SMB) in the NASA Goddard Institute for Space Studies (GISS) ModelE2 General Circulation Model (GCM). GCMs are often limited in their ability to represent SMB compared with polarregion Regional Climate Models (RCMs). We compare ModelE2 simulated GrIS SMB for presentday (19962005) simulations with fixed ocean conditions, at a spatial resolution of 2 latitude by 2.5 longitude (~200 km), with SMB simulated by the Modle Atmosphrique Rgionale (MAR) RCM (19962005 at a 25 km resolution). ModelE2 SMB agrees well with MAR SMB on the whole, but there are distinct spatial patterns of differences and large differences in some SMB components. The impact of changes to the ModelE2 surface are tested, including a subgridscale representation of SMB with surface elevation classes. This has a minimal effect on ice sheetwide SMB, but corrects local biases. Replacing fixed surface albedo with satellitederived values and an agedependent scheme has a larger impact, increasing simulated melt by 60100%. We also find that lower surface albedo can enhance the effects of elevation classes. Reducing ModelE2 surface roughness length to values closer to MAR reduces sublimation by ~50%. Further work is required to account for meltwater refreezing in ModelE2, and to understand how differences in atmospheric processes and model resolution influence simulated SMB.