A New Approach to Characterize Global Land Surfaces

Abstract

The upcoming generation of satellite sensors will provide more and much higher quality data (specifically much better spectral and angular sampling of the radiative fields emerging from terrestrial surfaces) than have been available so far. These data are of primary interest for providing an improved characterization of land surfaces which, for global scale investigations, have traditionally been based on the temporal analysis of a vegetation index (such as the Normalized Difference Vegetation Index) computed from AVHRR data. Indeed, the radiance fields measured by satellites do depend on the radiative properties of the surface (among other factors), which include a number of key variables such as the Leaf Area Index (LAI). Estimating the values of these surface variables on the basis of remote sensing data reduces to the solution of an inverse problem. An algorithm designed to identify the most probable solutions amongst a set of potential solutions that are predefined in a Look-Up Table (LUT) has been implemented and explored. These solutions have been derived from direct simulations achieved with a one-dimensional physical model of radiation transfer in vegetation canopies. In this approach, every pre-defined solution is associated to the set of values for all the physical variables required by the radiation transfer model used to generate the LUT.