| Summary: | International audience Knowledge on the thermal state and ice distribution of steep alpine rock faces and rock glaciers is crucial to assess potential geohazards associated with the degradation of permafrost. Temperature measurements at the rock surface or in boreholes are however expensive, invasive, and provide spatially-limited information. Electrical conductivity tomography has been broadly used to detect and monitor mountain permafrost since thirty years on rock walls and rock glaciers. Recent studies show the interest of coupling, at several scales, methods of electrical conductivity tomography, induced polarization tomography, and self-potential measurement in the field in parallel with petrophysical characterization of core samples in order to develop petrophysical transforms between geophysical observables and parameters of interest such as temperature. The advantages of these geophysical methods are their low cost, their non-invasive character, and the fact that they provide 2D, 3D, and 4D tomograms/images of the subsurface.For the steep rock walls, geophysical datasets acquired at the lower Cosmiques ridge, Aiguille du Midi, and Aiguille des Grands Montets (Mont Blanc massif) show the permafrost distribution. We tested a recently-developed petrophysical model based on the use of an exponential freezing curve applied to both electrical conductivity and normalized chargeability to infer the distribution of temperature below the freezing point. We then applied this approach to obtain the temperature distribution from electrical conductivity and induced polarization field data obtained across these profiles. The comparison of this result with 1D diffusive transient thermal model from surface temperature time series confirms the presence of permafrost in a way that is consistent with that obtained independently from the geophysical datasets. On the rock glaciers, geophysical datasets collected at the Pierre Brune, Thorens and Col de Vés rock glaciers (Vanoise massif) show the distribution of ice content ...
|
|---|