Forest canopies are semitransparent to microwaves at both L- and P-bands. Thus, a number of scattering sources and different types of scattering mechanisms may contribute to a single range cell of a synthetic aperture radar (SAR) image. By appropriately combining the SAR data of multiple parallel flight paths, a large 2-D aperture is synthesized, which allows for tomographic imaging of the 3-D structure of such semitransparent media and the underlying ground. A separate paper deals with the actual tomographic imaging part that leads to the 3-D data cube. In particular, three focusing techniques are described and analyzed: multilook beamforming, robust Capon beamforming, and multiple signal classification beamforming. In this paper, the resulting data products obtained by tomographically focusing two airborne multibaseline SAR data sets of a partially forested area, one at L-band and another at P-band, are subject to a detailed analysis with respect to the location and the type of backscattering sources. In particular, the following aspects are investigated: 1) The forest structure, as obtained from the vertical profiles of intensities at sample plot locations within the forest, is compared to the height distribution of the top of the forest canopy, as derived from airborne laser scanning data, and profiles are presented for all polarimetric channels and focusing techniques, as well as at both frequencies; 2) the type and location of scattering mechanisms are analyzed as functions of height for the two frequencies, namely, L- and P-bands, and using the polarimetric channels, as well as the Pauli and CloudePottier decompositions thereof; and 3) the accuracy of the ground elevation estimation obtained from the different focusing techniques and the two frequencies is assessed with the help of a lidar-derived digital elevation model.