A study of low frequency SAR clutter using model forests

Recent trials with an experimental, low frequency, airborne SAR operating in the frequency range 150-600 MHz have been complemented with simulation studies, designed to predict attenuation and clutter distributions from forests. Biologically accurate, tree architectural models have been combined with coherent wave, electromagnetic scattering calculations. By simulating the process of SAR image formation SAR clutter distributions are determined. Variation of mean clutter levels, and attenuation rates, with stand density and species have been investigated. Both mean backscatter levels and clutter distributions are seen to be sensitive to changes in stand density, and both co-polar channels indicate optimum frequencies for coupling of RCS to biomass. Intensity distributions are seen to become increasingly nonexponential as stand density is decreased. Resolution volume occupancy fluctuations are implicated as the strongest influence on texture at low frequency, in contrast to shadowing, which has been shown to control texture at more traditional frequencies. The potential exists to exploit the information content of low frequency SAR imagery texture, in addition to mean brightness, in order to retrieve forest biomass.