Reducing Uncertainties in Greenland Surface Mass Balance Using IceBridge and ICESat Altimetry, GRACE Data and Regional Atmospheric Climate Model Outputs

Abstract:

The mass of the Greenland Ice Sheet is undergoing rapid changes due to increase in surface melt and ice discharge. Considerable progress has been made to reduce the overall uncertainty of mass balance assessments. Here, we address the uncertainty in runoff production, which is of the largest remaining uncertainty in ice sheet mass balance. Runoff model outputs are difficult to evaluate due to a lack of in-situ monitoring networks. We document the uncertainty in runoff and how it varies spatially by comparing runoff products from different regional climate models (RCM) with two other sets of observations: 1) GRACE regional time series calculated using a least-squares mascon approach and corrected for regional ice discharge - with an emphasis on regions where ice dynamics is less significant: southwest and north Greenland; 2) At the smaller scale, NASA IceBridge and ICESat surface elevation change products, dh/dt, in the ablation zone to compare the observed volume changes with those predicted by RCMs. These two comparisons help evaluate how well seasonal melt and runoff are modeled by RCMs. The results help constrain uncertainties in present-day surface mass balance and runoff, as well as identify sources of RCM error. We also examine the constraints applied to the models (re-analysis data, albedo, energy budget, scheme to implement water retention, etc.) to gain insights into the processes responsible for the difference between models. Overall, we find substantial differences between MAR and RACMO results, and the results vary by region in terms of magnitude, timing and duration of surface melt. For instance, melt-water production and refreeze in the southwest are greater in RACMO, but total runoff is greater in MAR, and RACMO agrees better with GRACE. In the northeast, MAR predicts less runoff than RACMO and agrees better with GRACE. In the southeast, we find that the current version of MAR over-predicts runoff production. This work was funded by NASA’s Cryospheric Science Program.