Inverting a radiative transfer model for sediment density retrieval from hyperspectral BRDF data

Abstract

Approximate radiative transfer model solutions have been developed by Hapke [1][2]. Important geophysical parameters such as density and grain size distribution play an important role in parameterizing these solutions, however the radiative transfer solutions are typically nonlinear functions of these important parameters. We have developed a method which isolates single scattering properties of the radiative transfer solution, and by requiring these properties to be constant under different illumination conditions for hemispherical observations of the angular dependence of the spectral reflectance, we can arrive at an estimate of the surface fill factor [10]. In this paper, we consider details of the calculation with laboratory measurements of hemispherical conical reflectance factor (HCRF) data from two scans at differing illumination geometries, showing that details of the interpolation scheme required play an important role in the quality of the estimate.