Characterizing landslides through geophysical data fusion: Example of the La Valette landslide (France)
Résumé
Many studies show that certain geophysical methods, such as seismic and electrical-resistivity imaging, appear to be well adapted for investigating the internal structures of landslides and understanding the related hydro-mechanical mechanisms. These are methods that allow the direct and non-intrusive measurement of acoustic (P) and shear (S) wave velocities and electrical resistivity (ρ), which are three physical parameters considered as essential for estimating the mechanical properties of moving reworked material. We applied these techniques to the La Valette landslide (Southern French Alps), a typical example of an intra-material landslide, carrying out measurements simultaneously along two profiles, 400 m and 300 m long and respectively perpendicular to and along the slide direction. We then used suitable inversion algorithms to estimate both the P- and S-wave velocity fields and the electrical resistivity field from the recorded data. The results, aided by field surface observations, show that a correlation exists between the state of the material and the seismic-velocity and/or electrical-resistivity data, thus confirming that the simultaneous use of the two methods provides complementary information on the geomechanical behavior of the landslide. More particularly, the seismic data provide information on fissure density variations and the presence of shear-bent material, whereas the electrical resistivity data provide information on the groundwater content. To enable a more integrated petrophysical interpretation, we applied a data-fusion strategy based on fuzzy subsets to the geophysical datasets. Through combining the tomograms we identified a surface layer of soft material along the two profiles; the bottom of this layer was also recognized in a borehole. From a methodological point of view, the results show the applicability of adopting geomechanical hypotheses as inputs of geophysical data fusion for identifying areas where sediment mobilization could occur.