A study of the relationship between radar backscatter and regenerating tropical forest biomass for spaceborne SAR instruments

Abstract

The relationship between mean backscattering coefficient, as measured by spaeebome SAR instruments, and the areal density of above-ground biomass in regenerating tropical forest is investigated for a study area in the central Amazon basin. Measurements of tree height, diameter, and species composition were made in 1994 in the Tapajós region of Pard State in Brazil in order to estimate the biomass density in 15 widely distributed sample plots. These plots were chosen so as to characterize homogeneous forest areas representing a range of ages of regeneration from new regrowth to mature forest. The mean backscattering coefficients of these forest areas, as measured by SAR instruments on the ERS-1 and JERS-1 satellites and by SIR-C on the Space Shuttle, was determined so that its dependence on the biomass density of regenerating forest at different radar wavelengths and polarizations could be quantified. Results confirm the findings of similar studies of different types of forest, and of radar backscatter models, that the longer wavelength (L-band) SAR imagery may be used to discriminate between different levels of forest biomass up to a certain threshold and that cross polarized baekscatter is more sensitive to changes in biomass density. The shorter wavelength (C-band) SAR imagery is limited to differentiating between vegetation and bare soil when it is dry. The biomass density limit of around 60 tonnes per hectare, above which the L-band baekscatter shows no further sensitivity to increased biomass density, suggests that spaceborne SAR imagery is suitable for estimating the biomass density and determining the extent of cleared and regenerating forest in tropical regions.