Estimating the modal mineralogy of eucrite and diogenite meteorites using visible-near infrared reflectance spectroscopy

Reliable quantitative mapping of minerals exposed on Vesta's surface is crucial for understanding the crustal composition, petrologic evolution, and surface modification of the howardite, eucrite, and diogenite (HED) parent body. However, mineral abundance estimates derived from visible-near infrared (VIS-NIR) reflectance spectra are complicated by multiple scattering, particle size, and nonlinear mixing effects. Radiative transfer models can be employed to accommodate these issues, and here we assess the utility of such models to accurately and efficiently determine modal mineralogy for a suite of eucrite and olivine-bearing (harzburgitic) diogenite meteorites. Hapke and Shkuratov radiative transfer models were implemented to simultaneously estimate mineral abundances and particle size from VIS-NIR reflectance spectra of these samples. The models were tested and compared for laboratory-made binary (pyroxene-plagioclase) and ternary mixtures (pyroxene-olivine-plagioclase) as well as eucrite and diogenite meteorite samples. Results for both models show that the derived mineral abundances are commonly within 5-10% of modal values and the estimated particle sizes are within the expected ranges. Results for the Hapke model suggest a lower detection limit for olivine in HEDs when compared with the Shkuratov model (5% versus 15%). Our current implementation yields lower uncertainties in mineral abundance (commonly <5%) for the Hapke model, though both models have an advantage over typically used parameters such as band depth, position, and shape in that they provide quantitative information on mineral abundance and particle size. These results indicate that both the Hapke and Shkuratov models may be applied to Dawn VIR data in a computationally efficient manner to quantify the spatial distribution of pyroxene, plagioclase, and olivine on the surface of Vesta.