Synergistic use of multitemporal RAMP, ICESat and GPS to construct an accurate DEM of the Larsemann Hills region, Antarctica Original Research Article

An enhanced digital elevation model (DEM) of the Larsemann Hills region, east Antarctica, is constructed synergistically by using highly accurate ground-based GPS measurements, satellite-derived laser altimetry (GLAS/ICESat) and Radarsat Antarctic Mapping Project (RAMPv2) DEM-based point elevation dataset. Our DEM has a vertical accuracy of about 1.5 times better than RAMPv2 DEM and seven times better than GLAS/ICESAT-based DEM. The accuracy is improved by validating the RAMPv2 DEM elevation by supplementing with GLAS/ICESat and DGPS survey data, when compared to that of DEM constructed by using GLAS/ICESat or RAMPv2 alone. With the use of accurate GPS data as ground control points reference elevations, the DEM extracted is much more accurate with least mean RMSE of 34.5 m than that constructed by using a combination of GLAS/ICESat and RAMPv2 as true reference. The newly constructed DEM 7 achieves highest accuracy with the least average elevation difference of 0.27 m calculated using 46 ground reference points. Available DEMs of Antarctic region generated by using radar altimetry and the Antarctic Digital Database indicate elevation variations in the range of 50–100 m, which necessitates the generation of local DEM and its validation by using ground truth. This is our first attempt of fusing multi-temporal, multi-sensor and multi-source elevation data to generate a DEM of any part of Antarctica, in order to address the ice elevation change to infer the ice mass balance. Our approach focuses on the strengths of each elevation data source to produce an accurate DEM.