A theoretical analysis of the effect of forest structure on synthetic aperture radar backscatter and the remote sensing of biomass

Stand level forest canopy structure as measured by the size, density, and distribution of the stems, branches, and leaves may have a strong effect on SAR backscatter. This study used the Michigan Microwave Canopy Scattering model (MIMICS) and forest canopy biometric data from tropical and subtropical broadleaf forests to simulate a series of forest stands having equivalent above ground biomass while still exhibiting substantial structural differences. The model stands were made to represent a wide range of basic structural differences found in Earth’s broadleaf forests. The radar response to these structural differences is shown for the NASNJPL AIRSAR P-, L-, and C-band quadpol. configuration at incidence angles of 20°, 40°, and 60°. Results for three sets of equal-biomass-foreststands were generated: five structural types having a total above ground biomass of 5 kg/m
²
(50 tonha), three structural types at 15 kg/m
²
(150 tonsha), and three structural types at 30 kg/m
²
(300 tonsha). Simplified model input were used to reduce the number of geometric variables tested across the range of structural types. The potential effect that extreme structural differences might have on the biomass signal and saturation threshold was assessed. Results indicate that structure can have a considerable effect on the SAR return for forests with equivalent above ground biomass. Differences in backscatter of up to 18 dB were predicted for some bands and polarizations. A forest canopy structural descriptor derived from the vegetation surface area to volume ratio (SAN), which is a measure of stuctural consolidation, appears to explain the differences in backscatter. In many vegetation stands structural consolidation is directly related to an increase in biomass during the "thinning phase" of forest succession. This structural effect may explain the good relationship between SAR backscatter and biomass in these cases.