Detection and characterization of boreal coniferous forests from remote sensing data

Abstract

Advanced radiation transfer models capable of representing the reflectance of the coupled surface and atmosphere system have been used to generate look-up tables of simulated remote sensing measurements at the top of the atmosphere for typical conditions of forest cover and atmospheric composition found in northern Europe. These simulations were evaluated against actual observations under identical viewing and illumination geometries, available for the blue, red, and near-infrared spectral bands of the VEGETATION instrument to retrieve the most likely of a set of predefined solutions to the inverse problem. The accumulation of results over multiple days in the summer of 1999 permitted the establishment of maps showing the likelihood of identifying the predefined forest types, their corresponding structural characteristics, as well as the associated atmospheric optical depth on the day of retrieval. The proposed methodology is completely generic and thus can easily be prototyped for different biomes and instruments. Copyright 2001 by the American Geophysical Union.