Vegetation affects the possibility of providing global surface mapping based on Synthetic Aperture Radar Interferometry (InSAR). The estimation of ground topography in vegetated areas is an important task both, for scientific and commercial applications. Therefore, the need of quantifying and removing the vegetation bias arises. In this paper, these two points are investigated with reference to the penetration capabilities of L-band over forests. More in detail, an already developed model-based estimation technique is applied to experimental, fully polarimetric data. This provides a first experimental validation of such a technique and a way to quantify the bias introduced at L-band.