Radiation physics and modelling for off-nadir satellite-sensing of non-Lambertian surfaces

Abstract

The primary objective of this paper is to provide a deeper understanding of the physics of satellite remote-sensing when off-nadir observations are considered. Emphasis is placed on the analysis and modelling of atmospheric effects and the radiative transfer of non-Lambertian surface reflectance characteristics from ground-level to satellite locations. We evaluate the relative importance of spectral, spatial, angular, and temporal reflectance characteristics for satellite-sensed identification of vegetation types in the visible and near-infrared wavelength regions. Highest identification value is attributed to angular reflectance signatures. Using radiative transfer calculations to evaluate the atmospheric effects on angular reflectance distributions of vegetation surfaces, we identify atmosphere-invariant angular reflectance features such as the “hot spot” and the “persistent valley”. A new atmospheric correction formalism for complete angular reflectance distributions is described. A sample calculation demonstrates that a highly non-Lambertian measured surface reflectance distribution can be retrieved from simulated satellite data in the visible and near infrared to within about 20% accuracy for almost all view directions up to 60° off-nadir. Thus the high value of angular surface reflectance characteristics (the “angular signature”) for satellite-sensed feature identification is confirmed, which provides a scientific basis for future off-nadir satellite observations.