In polythermal glaciers, specific climatic, topographic, and exposure conditions may lead to the formation of englacial lakes that can produce catastrophic effects downstream in the event of abrupt natural drainage. We have determined how a combination of ground-penetrating radar (GPR) and surface nuclear magnetic resonance (SNMR) surveys helped to locate and visualize the evolution of a water-filled cavity within the Tête Rousse glacier (French Alps). We have used GPR results to delineate the roof of the cavity and monitor the cavity deformation caused by artificial drainage. Because the glacier bed and cavity have complex 3D geometries, we needed dense acquisition lines and 3D GPR views to qualitatively identify out-of-plane reflections. This 3D approach made it possible to establish a precise map of the glacier bed topography, the accuracy of which was verified against borehole observations. Then, repetitive GPR measurements were used to obtain a quantitative estimate of the vertical deflection of the cavity’s roof and changes in crevasse geometry observed in response to the decrease in the water pressure when of water was drained by pumping. We have used 3D SNMR imaging to locate water accumulation zones within the glacier and to estimate the volume of accumulated water. The SNMR monitoring revealed that in two years, the cavity lost approximately 73% of its initial volume, with 65% lost after the first drainage. Knowledge of the water contained in the ice provided a better understanding of GPR images and thus a more reliable interpretation of GPR data. However, SNMR imaging had a much lower resolution in comparison with GPR, and consequently GPR allowed a more accurate study of the evolution of cavity geometry caused by consecutive drainage and refilling. This study demonstrated the value of combining GPR data with SNMR data for the study of polythermal glaciers.