A test of the Hapke photometric model

Abstract

We conducted a test of the Hapke (1981, 1984, 1986, 2002) photometric model to determine what physical parameters could be confidently linked to a surface's photometric behavior. The first author selected and measured the bidirectional reflectance distribution functions of 14 different samples in up to three different wavelengths using the Bloomsburg University Goniometer Laboratory. A total of 29 data files were obtained, each file containing more than 700 measurements from different viewing geometries; phase angles varied from 3° to 130°. The 29 files were initially sent “in‐the‐blind” to the second author, who was charged with inverting the data files and extracting best fit model parameters. Our baseline model used the most recent Hapke (2002) formulation with a two‐term Henyey‐Greenstein particle phase function and shadow‐hiding backscatter opposition effect (SHOE) model. We also inverted the data with three other variations that included three‐term Henyey‐Greenstein phase functions and both SHOE and the coherent backscatter opposition effect (CBOE) models. Our results were compared with the known physical properties of our samples. We found no compelling evidence that individual photometric parameters could be uniquely interpreted to reveal the physical state of our samples, either in an absolute or relative sense. Rather, combinations of physical properties such as albedo, roughness, and porosity were convolved within each retrieved photometric parameter. On the basis of our empirical evidence, we speculate that the fault lies with the inability of radiative transfer models to adequately account for discrete media and the effects of porosity, and its deficient assumption that individual particles are the primary scattering units.