Polarimetry - Forestry - Polarimetric Parameters Co-polarization Signatures

Polarimetric data may provide unique information on forest canopies related in particular to canopy architecture and the consequent scattering mechanisms. Backscatter mechanisms include direct backscatter from branches (single bounce/volume scattering), backscatter from trunks (single bounce), scattering from branch-ground interaction (double bounce), scattering from trunk-ground interaction (double bounce), and direct backscatter from the ground (surface scattering). The relative contribution of each of these depend on the nature of the canopy and imaging parameters such as incidence angle and frequency. Components within the canopy (leaves and twigs) may play a significant role in the scattering and attenuation interactions depending on frequency. It is expected that use of polarimetric data analysis will help understand the nature of scattering within a canopy.

An example modelled co-polarization signature representative of a hardwood forest at L-band shows that backscatter from a heavy forest cover at HH and VV polarizations is similar with that at VV being slightly lower, suggesting the dominance of multiple branch scattering combined with a weak double bounce component (Figure 9-28a). The pedestal height indicates a large unpolarized component in the backscatter return indicative of multiple scattering Durden et al.

Modelling of the same canopy with a branch density that is an order of magnitude lower results in a polarization signature that is characteristic of a dihedral corner reflector (double bounce) (Figure 9-28b), suggesting the dominance of trunk-ground scattering mechanisms for this forest with fewer branches. The lower branch density also results in a significantly lower pedestal height due to a lower unpolarized component, indicating a smaller amount of volume scattering

Figure 9-28
P = Normalized Power

Figure 9-28. Co-polarization signature at L-band of, a) a hardwood forest and b) a hardwood forest with a branch density that is of an order of magnitude less than (a) (from Durden et al).

Figure 9-29 shows the modelled co-polarization signatures for a forested area and a clear cut area. The clear cut area has a lower pedestal height indicative of a lower unpolarized component due to greater direct return from the ground surface and a smaller amount of volume scattering. The stronger return in the VV case compared to that at HH is indicative of surface scattering.

Figure 9-29
P = Normalized Power

Figure 9-29. Co-polarization signature for, a) forested area and b) clear cut (from Durden et al). Polarization Phase Difference

The Co-polarization Phase Difference can be useful in understanding the scattering mechanisms for a particular target. Single bounce scatterers generally result in a phase difference close to 0o whereas ideal double bounce scatterers have a phase difference of plus_minus180°. In the example from the work of LeToan et al. Le Toan et al 1992 for a forest canopy the Co-polarization Phase Difference has been correlated with stand age (Figure 9-30), and to a lesser extent with forest stand height, and trunk biomass In this case, the mean Co-polarization Phase Difference obtained for the clear cut area ( = 6.8°) indicates that the backscatter is largely a function of surface scattering from the ground. The mean value for the mature forest stand is much higher (approximately 66°) and has a larger standard deviation (90°) than that obtained for the clear cut area (70°). This is indicative of volume scattering.

Figure 9-30
F = Frequency of occurence (%), D = Polarization phase difference (Deg)

Figure 9-30. Histograms of P-band Copolarization Phase Differences between HH and VV polarizations () for (I) a clear cut area and (II) a 46 year old pine forest plantation (Le Toan et al 1992).