Résumé
Normal faults parallel to the strike of the mountain ranges have always been though as an evidence of gravitational collapse. The Pyrenees and the Alps (western Europe) provide an invaluable laboratory to check for gravitational collapse. Indeed, earlier GPS studies show extremely low deformation if any within the Eurasia-Iberia plate, new GPS results from 10 to 20 years of surveys show no significant horizontal motion to the level of 0+ or -0.2 mm/yr, leveling and continuous GPS depict uplift rate up to 2 mm/yr, and focal mechanisms suggest extension normal to the range axis. The extremely low horizontal deformation associated with significant uplift rate and extension in the inner range is rather puzzling, and one can wonder if this could be related to erosion of the range rather than gravitational collapse. We use a 2D finite element model to study the effects of the erosion on the deformation of mountain ranges accommodating low convergence rate if any. We first use a conceptual model with a symmetric topography, and a crustal root isostatically compensated. We test average erosion rate from 0 to 1.2 mm/yr. Several erosion laws are used, as well as different geotherms and an alternate model with topography corresponding to the Pyrenean asymmetric average topography and the crustal root geometry based on seismic imaging. We also extend the rate of convergence up to 9 mm/yr in order to check if extension related to erosion could occur within active ranges. We obtain a robust solution since the variations on the parameters induce no significant differences. Our study shows that gravitational collapse cannot explain the observations for this western end of the Alpine-Himalayan range. This is only when we introduce erosion that we are able to fit the observations. Hence, we show that erosion associated to isostatic rebound explain not only the extension within the range, but also the uplift rates. We further suggest that depending on the erosion rate, extension could be observed within the inner part of a range even if this range is currently accommodating convergence at the level a several mm/yr. Therefore this new model could further explain why extension occurs within the upper part of the Albanian mountain range and maybe in the Andes.