Sources of Spread in Multimodel Projections of the Greenland Ice Sheet Surface Mass Balance

Abstract

he many studies investigating the future change of the Greenland Ice Sheet surface mass balance from climate model output exhibit a wide range of projections. This study makes projections from the Coupled Model Intercomparison Project phase 3 models used in the Intergovernmental Panel on Climate Change Fourth Assessment Report and explores the underlying physical processes behind their spread. The projections are made for three Special Report on Emissions Scenarios, B1, A1B, and A2, with a focused analysis on the A1B scenario. The estimate in the study suggests that about 60% of the intermodel difference in the twenty-first-century ablation rate change under the A1B scenario is accounted for by the global annual mean temperature change. In the current study, other processes responsible for the spread in model projections are investigated after excluding this global effect. A negative correlation (?0.60) was found between the simulated summer temperature bias over the Greenland Ice Sheet under present-day conditions and the ablation rate increase during the twenty-first century, partly because maximum warming over ice is approximately limited to the melting temperature. Models with relatively larger ablation rate increases during the twenty-first century exhibit greater warming with a greater reduction in sea ice cover. The authors found that these models also simulate relatively cooler summer conditions in high latitudes with more sea ice cover in the late twentieth century, suggesting the importance of sea ice feedbacks. Also, an anticorrelation (?0.75) is found between weakening of the Atlantic meridional overturning circulation and the ablation rate increase during the twenty-first century. The relation in the model spread between the twenty-first-century ablation change and the late twentieth-century climate conditions is then used to investigate the impact of model bias on the multimodel ensemble of projections. The result suggests that the models? underestimation of present-day sea ice concentration near the coast of Greenland may cause an underestimation of future Greenland Ice Sheet ablation rate increase in the ensemble projection. These results emphasize the importance of correctly simulating present-day conditions and understanding the underlying multiple physical processes behind the intermodel difference to reduce the uncertainty in future projections.