Microwave satellite remote sensing is an important source of information about the polar ice sheets. Radiation in the microwave frequency range interacts with the firn volume, and firn properties such as density and microstructure have been found to strongly influence the scattered or emitted signal. Hence, a correct interpretation of the data is only possible if the interaction of microwave radiation with polar firn is understood sufficiently. In this context, ground truth measurements are indispensable. However, a direct comparison between pointwise field measurements and satellite data is often difficult, since the variability within the satellite footprint needs to be considered. In the case of dry polar firn, the variability of density and grain size over the signal penetration depth also needs to be take into account. In order to investigate the latter effect, we conduct a statistical analysis of the layering properties in a number of polar firn cores. We show that there is a correlation between density variability and grain size variability and introduce a procedure which enables us to use measured densities as a proxy for grain size values. Scattering and emission models often use mean profiles of grain size and density to describe the snow volume. However, these profiles ignore the strongly pronounced layering of the snow and firn, which impacts the dielectric contrast and hence potentially introduces a bias into the model results. Using our improved representation of microstructure variability, we conduct a sensitivity study where we examine the impact of firn layering on the microwave signal.