Land cover type and forest biomass assessment in the Colombian Amazon

Results of an analysis of AIRSAR data of the Colombian Amazon (province Guaviare) acquired during the 1993 AIRSAR South American deployment of NASA are discussed. Research is focused on the potential of AIRSAR radar images to classify land cover types, such as primary and secondary forest, recently cut forest, pasture, savanna and bush, taking into account the influence of incidence angle and speckle level. Also the relation between backscatter and biomass level is studied. Very good classification results were obtained using combinations of two channels. High correlation between biomass and backscatter were found for many channels. The use of (partial) polarimetry and circular polarizations furthers results.