This study examines the recent evolution of the Greenland ice sheet and its six major drainage basins. Based on laser altimetry data acquired by the Ice, Cloud and Land Elevation Satellite (ICESat), covering the period September–November 2003 to February–March 2008, ice surface height changes and their temporal variations were inferred. Our refined repeat track analysis is solely based on ICESat data and is independent of external elevation models, since it accounts for both ice height changes and the local topography. From the high resolution ice height change pattern we infer an overall mean surface height trend of −0.12 ± 0.006 m yr−1. Furthermore, the largest changes could be identified at coastal margins of the ice sheet, exhibiting rates of more than −2 m yr−1. The total ice volume change of the entire ice sheet amounts to −205.4 ± 10.6 km3 yr−1. In addition, we assessed mass changes from 78 monthly Gravity Recovery and Climate Experiment (GRACE) solutions. The Release-04 gravity field solutions of GeoForschungsZentrum Potsdam cover the period between August 2002 and June 2009. We applied an adjusted regional integration approach in order to minimize the leakage effects. Attention was paid to an optimized filtering which reduces error effects from different sources. The overall error assessment accounts for GRACE errors as well as for errors due to imperfect model reductions. In particular, errors caused by uncertainties in the glacial isostatic adjustment models could be identified as the largest source of errors. Finally, we determined both seasonal and long-term mass change rates. The latter amounts to an overall ice mass change of −191.2 ± 20.9 Gt yr−1 corresponding to 0.53 ± 0.06 mm yr−1 equivalent eustatic sea level rise. From the combination of the volume and mass change estimates we determined a mean density of the lost mass to be 930 ± 11 kg m−3. This value supports our applied density assumption 900 ± 30 kg m−3 which was used to perform the volume–mass-conversion of our ICESat results. Hence, mass change estimates from two independent observation techniques were inferred and are generally in good agreement.