In the past, global maps of major oxides and Mg # of the lunar surface had been derived from spectral data with “ground truth” geochemical information from Apollo and Luna samples. These compositional maps provide insights into the chemical variations of different geologic units, thus the regional and global geologic evolution. In this study, we produced new global maps of major oxides (Al2O3, CaO, FeO, MgO, and TiO2) and Mg # with imaging spectral data of KAGUYA multiband imager (MI) with the one dimensional-convolutional neural network（1D-CNN）algorithm, taking advantage of recently acquired geochemical information of China’s Chang’E-5 (CE-5) samples. The coefficients of determination (R2) and Root Mean Squared Error (RMSE) were selected as the model evaluation indicators, and compared with the models used by Wang et al. (2021) and Xia et al. (2019), the results showed that the 1D-CNN algorithm model used in this study had a higher degree of fit and smaller dispersion between the ground true value and the predicted value. The 1D-CNN algorithm generally performs better in describing the complex nonlinear relationship between spectra and chemical components. In addition, we present regions of mare domes in Mairan Dome (43.76°N, 49.90°W), and irregular mare patches (IMPs) in Sosigenes (8.34°N, 19.07°E) to demonstrate the geologic implications of these new maps. With the highest spatial resolution (~ 59 m / pixel), these new maps of major oxides and Mg # will serve as an important guide in the future study of lunar geology.

The overturn of mantle cumulates following the crystallization of the lunar magma ocean – particularly the sinking of ilmenite-bearing cumulates (IBC) – provides an explanation for several aspects of the Moon’s evolution. However, the growth of a stagnant lid due to the temperature dependence of the viscosity tends to prevent IBC from sinking. Here, we investigate the dynamics of the overturn based on a composition-dependent rheology coupling diffusion and dislocation creep of major lunar mantle minerals via three alternative mixing models: isostrain, isostress and Minimized Power Geometric (MPG). The pre-overturn structure is obtained from a fractional crystallization model of the lunar magma ocean, which predicts the formation of a 36-km-thick IBC layer. The possibility of overturning this layer strongly depends on the choice of the rheological mixing model. The isostress model allows for a rapid and complete overturn, while the isostrain and the experiment-based MPG models do not allow IBC to sink. If IBC started sinking and mixing with the underlying mantle during magma ocean solidification, IBC could be initially distributed across a layer with a thickness of up to 150 km, whose partial overturn is always possible, independent of the rheological model. These results highlight the importance of improving rheological models for relevant lunar materials. In all cases, the overturn occurs via small-scale instabilities.