Quantitative estimation of physical parameters of the Earth's surface is currently an active area of research. Multidimensional synthetic aperture radar (SAR) systems appear in this framework as a tool capable of performing this task on a global scale and independently of the weather conditions. In particular, polarimetric SAR data have been successfully employed, in combination with different electromagnetic inversion algorithms, to extract physical information from measured data. Nevertheless, SAR data are affected by speckle noise. The influence of speckle noise and its filtering in the quantitative estimation of physical parameters is analysed. To perform this study, a novel model for speckle noise in multidimensional SAR data is considered. Finally, the effect on the quantitative estimation of surface parameters is investigated.