A quantitative model to estimate major oxide abundances on the Moon based on in situ reflectance spectral data of Chang'e missions

The oxide abundances of the lunar regolith are fundamental for understanding the geological evolution of the Moon both regionally and globally. Several models have been constructed to estimate the oxide contents of lunar regolith from the visible and near-infrared (VNIR) spectra based on Apollo and Luna samples. In recent years, China's successful lunar missions, Chang'e-3, Chang'e-4, and Chang'e-5 have yielded new ground-truth data for developing the models. In this study, we propose a novel approach to estimate the FeO, TiO2, MgO, CaO, and Al2O3 contents from VNIR spectra, based on spectral derivative, genetic algorithm, and partial least squares regression. We validate the model's robustness by comparing our estimated oxide abundances with those derived directly from in situ or laboratory measurements. Using our model, the chemical compositions along the Chang'e-4, Yutu-2 traverse are estimated, providing important constraints on the mineralogical and geological properties of the landing area.