Abstract
This paper compares the use of biophysical indices and coefficients of variation derived from airborne synthetic aperture radar for discriminating and characterising land cover classes in the Tucurui reservoir. It assumes that the structural and dielectric properties of the land cover classes are sufficiently distinct to permit their discrimination using airborne C-band SAR data at different polarizations. Analyses of the results show that any of the proposed indices can be used effectively to discriminate among the land cover classes in the Tucurui reservoir. Certain combinations of polarizations and biophysical indices can, however, improve the saparability of the land cover classes.