Abstract. The thickness of supraglacial debris affects the surface energy balance and retreat patterns of mountain glaciers. Therefore, knowing a debris layer’s thickness is crucial for understanding the magnitude and timeframe of glacier melt. Field- based ground-penetrating radar (GPR) has recently gained attention as a possible method for measuring debris thickness. Airborne assessments achieve extensive coverage and characterization, but the use of GPR for such platforms remains relatively unexplored. We investigated the performance of 960 MHz and 2.6 GHz GPR signals through dry laboratory rock debris, and of 960 MHz over ∼ 2 km of transects on the debris cover of Changri Nup Glacier, Nepal Himalaya. On the glacier, 960 MHz profiles were characterized by no clear reflection from the ice interface and volumetric backscatter from within ∼ 10–40 cm, a depth that corresponds to approximate ground-truth debris thicknesses on all transects. The laboratory results show that the lack of an ice-debris interface return in field data was likely caused by a weak dielectric contrast between solid ice and porous dry debris and that surface scatter is coherent but weak. This suggests that the debris-ice interface reflection was also likely coherent, supporting our conclusion of a weak dielectric contrast. The laboratory 2.6 GHz results show significant penetration for only smaller clast sizes up to 4 cm. We used a statistical approach to estimate ice depth from volumetric scatter, which gave reasonable agreement with ground-truth depth measurements. We conclude that a remote system operating near 1 GHz could successfully estimate dry debris cover thicknesses based on depth of volumetric backscatter.