Modelling the glacier dynamics of the Greenland Ice Sheet is a central part of cli- mate research. To be able to model the processes that influence a glacier’s flow, the position of its ice-bed contact is a basal boundary condition. Using five different P-wave reflection seismic datasets, the subsurface structure of Russell Glacier, which is a land-terminating glacier in southwest Greenland, is inves- tigated. The data have been recorded by the glaciology group of the Alfred-Wegener Institute, in the ablation zone of Russell Glacier. To investigate the efficiency of the minivibrator source ELVIS III on polar glaciers having ice thicknesses up to several kilometers, a dataset using the ELVIS source is compared to an equivalent explosive seismic dataset. The question to answer is, if the approximately 600m deep glacier bed and additional subglacial structures are displayed by the vibroseis data. As the maximally 10-fold vibroseis data, in contrast to the maximally 2-fold ex- plosive seismic data, don’t reveal any reflection signals, it is concluded that the penetration depth of the ELVIS signal is too small to resolve subglacial structures on polar glaciers. The subsurface structure of the survey area is investigated with the help of two explosive seismic datasets. In the resulting depth profiles lying perpendicular to each other, a distinct topography of the glacier bed as well as a stratification of the subglacial material is visible. The propagation velocities of seismic P-waves in the subsurface layers are firstly determined by applying Dix’ method at two additional explosive datasets. As the subglacial material couldn’t be determined that way, for the reason that the ap- purtenant velocity value determined with Dix’ method are not reliable due to the topography of the covered reflector, the polarity of the reflection coefficient associ- ated with the first layer boundary are reconstructed using the material classification from Christianson et al. (2014). For future applications of Dix’ method in order to determine seismic velocity values, a preliminary processing of seismic data in the field would be productive to be able to record the additional common-midpoint datasets in preferably horizontal strati- fied areas. The resulting two-layer subsurface model finally consists of the top 500-600 m thick ice pack with P-wave velocities in the range of v1 = 3500−3700 m s−1 and below lying subglacial sediments with P-wave velocities in the range of v2 = 1700 − 1900 m s−1, corresponding to unconsolidated sediments or dilatant till. The orthogneiss forming the bedrock in the survey area (Klint et al., 2013) couldn’t be resolved by the used seismic datasets, from which it can be inferred that its depth amounts to more than 600 m.