Biomass estimation from satelliteborne SAR data

The total backscatter of a forest is a combination of the backscatter from ground and vegetation. Previous studies have shown that the seasonal effects, such as soil freezing and thawing, snow cover and soil wetness, drastically change the level of the total backscatter in the boreal forest. Thus, the correlation between the biomass and the backscattering coefficient is dependent on the above mentioned factors. Nevertheless, these seasonal effects may also benefit the inversion if suitable models are used with a multitemporal data set. Moreover, the direct application of the backscattering coefficient for forest biomass inventory is limited by saturation at L- and C-band. The retrieval of boreal forest biomass from multitemporal ERS-1, JERS-1 and RADARSAT images is investigated by employing radar observations obtained for a 73000 hectares large test site located in southern Finland. Our new adaptive inversion method for the stem volume estimation is based on a semi empirical backscattering model. The model presumes that the backscattering from a forest canopy is affected by the stem volume (biomass), soil moisture and vegetation moisture