The global, systematic observation of land surfaces and continental biosphere characteristics is crucial to determine what fraction of carbon dioxide is absorbed by the vegetation. This is needed to analyze ecosystem response to climatic fluctuations and to derive the parameters of surface/atmosphere exchanges.

It is possible to determine vegetation cover indicators from the multispectral data acquired by the optical imaging sensors. The annual and inter-annual variation of these indicators may then be analyzed. It is also possible to derive parameters which describe vegetation cover (such as leaf area index) and primary production through statistical relationships.

Nevertheless, two major problems have to be overcome when using measurements from space : atmospheric effects (mainly due to aerosols), for which no global method of correction is currently available; and surface directional effects, not usually taken into account in vegetation index computations.

POLDER, whose spectral channels are well suited to land surface observation, also has measurement specificities that should lead to some improvement with respect to the abovementioned problems.

Since the atmospheric molecules and aerosols polarize the scattered radiation, whereas the surface does not, the polarized reflectance measured by POLDER is essentially an atmospheric contribution. It is then possible to fit an aerosol model to the observations, derive the atmospheric optical thickness and thus satisfactorily correct for the atmospheric effect.

Hot-Spot directional signature measured from the spaceborne POLDER instrument at 443 (blue), 670 (green) and 865 (red) nanometers. The reflectance is shown as a function of the phase angle, that is the angle between the sun and view directions. See details in Bréon et al., 2002, J. Geophys. Res., 107, 4282.

POLDER's multidirectional measurements may also be used to fit a directional model to the surface signature and derive a reflectance corrected for directional effects and normalized to a standard viewing geometry as well as a precise albedo, which is a crucial parameter in determining the radiation budget. Once normalized, the surface spectral signature may be used to derive various vegetation indices.

The bidirectional surface reflectance measurements constitute a specific signature that it should be possible to relate to certain structural parameters for vegetation cover (particularly the vegetation's interception and absorption of radiation), which should improve the quality of the surface cover classification on a global scale.