Current multidisciplinary studies confront the difficult problem of validation and verification of internal compatibility of geophysical and geological data of different nature, quality and origin. Our 'geodynamic' approach exploits the natural interdependence between the data related to different physical properties of geological structures (density, rheology, temperature, stress). Using the latter as input for a thermomechanical finite-element model, we can verify whether the inferred properties and geometries (i) are mechanically stable, (ii) require theologically consistent stresses, strains, realistic thermal conditions, etc., and (3) will evolve in a geologically realistic way in the future. This 'mechanically balanced' approach is tested on the Alpine orogen (ECORS and NFP20 profiles) in the framework of the GeoFrance 3D programme. The results suggest a number of corrections and adjustments that tan be made to existing seismically-and gravity-predicted geometries of structures such as the Ivrea body and those of the depleted subducted crust.