Integrating GPR and ice-thickness models for improved bedrock detection: The case study of Rutor temperate glacier

Abstract

Estimating the ice volume contained in glaciers is a topic of increasing interest, because the cryosphere has rapidly evolved during the last decades of global warming. Many disciplines collaborate to study the global warming impacts on glaciers. From the perspective of a geophysical research team, we examine the advantages of integrating glaciological ice-thickness models into the workflow for geophysical data processing. In temperate glaciers, the widespread englacial water content often challenges the analysis of Ground Penetrating Radar (GPR) data, the most commonly used geophysical technique for measuring ice thickness. In fact, recognizing the ice-bedrock interface is hard where the englacial water content generate high levels of scattering in the GPR data. A past GPR survey estimated that the Rutor glacier (European Alps) stored about 150 million m3 of ice in 2008. However, this estimate proved unrealistic after analyzing the geodetic mass balance of the following decade. Therefore, we analyzed new GPR measurements on the same glacier, which highlight the difficulty in identifying the true bed reflection. On these data, we tested the idea that ice-thickness models can help the GPR data analyst to better recognize the ice-bedrock interface in the radargrams. We selected four models, OGGM, GlabTop2-Py, Original-GlabTop2 and GlaTE, that estimate bedrock topography starting from surface topography, following principles of ice flow theory, ice dynamics and mass conservation. Combined visualization of the GPR and model data in 2D and 3D helped the analyst to manually select the ice-bedrock interface. This proved useful especially where the GPR data was scattered and interpretation was uncertain. The GPR data were then used to constrain one of the models, GlaTE, to produce an ice-thickness map that is the result of both model estimates and GPR information. Without this modelling context, the GPR data could be misinterpreted, and the resulting GPR-derived ice-thickness estimates might then be used to constrain a subsequent model in a way that would introduce significant errors. According to this methodology, the Rutor glacier stored about 450 million m3 of ice in 2021, about three times the previous estimate. The workflow is openly available in the section "Code and data availability"and may improve future GPR surveys of temperate glaciers, especially when facing scattered data due to englacial water content or other sparse reflectors such as debris. More accurate ice-thickness estimates will improve local studies and provide better calibration data for regional studies.