Quantifying the minerals abundances on planetary surfaces using VIS–NIR spectroscopy, what uncertainties should we expect? General results and application to the case of phyllosilicates and carbonates on Mars

Abstract

Over the last few decades, visible and near-infrared spectroscopy has proven to be an efficient technique to characterize planetary surface mineralogy, in particular thanks to the presence of diagnostic features appropriate for the identification of most minerals of interest. A more quantitative analysis of the VIS–NIR reflectance spectra constitutes the next major step in understanding the planetary bodies’ history as the retrieval of the mineral assemblages and their relative abundances enables to constrain the chemical and physical conditions of their formation and, thus, the past and present geologic and climatic processes.Here, we evaluate the capability to retrieve quantitative properties (abundance, grain size) of intimately mixed materials (the most common mineral mixture among planetary surfaces) from typical space VIS–NIR reflectance spectroscopic data. Such results are key to correctly assess the accuracy and relevance of the retrieved mineral information. For that purpose, we developed an inversion model based on a Monte Carlo Markov Chains (MCMC) scheme with a Bayesian approach to invert VIS–NIR spectra. This approach allows to properly propagate the uncertainties from the data to the retrieved properties, and finally assess what such uncertainties imply for the interpretation. Different binary and ternary mixtures with minerals of interest in planetary sciences and displaying a large variety of albedos and spectral features were tested. Typical uncertainties, both for the abundance and the grain size, were derived and sensitivities on specific parameters/trends were identified. In particular, the role of absorption features in the spectra is quantified. Tests were performed using both the Hapke and the Shkuratov radiative transfer models. The case of unidentified endmembers in the mixture is also discussed. In particular, results show that if the unidentified phase does not display any significant spectral feature, the lack of knowledge about its optical properties does not significantly impact the inversion. These different results will be key in the quantitative analyses of VIS–NIR spectra from planetary bodies.Finally, we analyze more specifically the case of phyllosilicates and carbonates, two families of minerals of high importance in understanding the Mars geologic and climate history. Typical uncertainties on their relative abundances and grain sizes are derived in various cases, providing a critical supporting dataset for the characterization of the martian mineralogy and the associated geological processes.