Understanding the radar backscattering from flooded and nonflooded Amazonian forests: Results from canopy backscatter modeling

To understand the potential of using multiwavelength imaging radars to detect flooding in Amazonian floodplain forests, we simulated the radar backscatter from a flood plain forest with a flooded or nonflooded ground condition at C-, L-, and P-bands. Field measurements of forest structure in the Anavilhanas archipelago of the Negro River, Brazil, were used as inputs to the model. Given the same wavelength or incidence angle, the ratio of backscatter from the flooded forest to that from the nonflooded forest was higher at HH polarization than at VV polarization. Given the same wavelength or polarization, the ratio was larger at small incidence angles than at large incidence angles. Given the same polarization or incidence angle, the ratio was larger at a long wavelength than at a short wavelength. As the surface soil moisture underneath the nonflooded forest increased from 10% to 50% of volumetric moisture, the flooded/nonflooded backscatter ratio decreased; the decreases were small at C- and L-band but large at P-band. When the leaf size was comparable to or larger than the wavelength of C-band, the leaf area index (LAI) had a large effect on the simulated C-band (not L-band or P-band) backscatter from the flooded and nonflooded forests.