This work aims to address two main scientific objectives. First, it seeks to rigorously compare ice thickness estimates from GPR datasets with those derived from various modelling approaches. Second, it examines warm and cold ice areas identified by GPR in relation to 2D thermal modelling performed along selected profiles. The analyses focus on two nearby glaciers in Greenland, surveyed in different years (2014 and 2024) and seasons (August and February) and with different GPR antennas, namely 50 MHz unshielded and 100 MHz shielded. We found that global-scale ice thickness models provide relatively accurate volume estimates at regional scale, while they have limitations in local accuracy, as well as the ice thickness models, especially when the bedrock topography derived from GPR data is complex. 2D thermal modelling results were only partially consistent with warm and cold ice occurrence derived from GPR data, indicating the unique and complex thermal structures of polythermal glaciers with irregular shape and geometry. Due to the differences between the two surveys, we believe that the results are relevant not only to the specific test site, but also to a wider range of geographical and climatic conditions and may provide useful guidance for similar applications.