Ground truth constraints and remote sensing of lunar highland crust composition

Abstract

AbstractWe review constraints on the magnitude and possible causes of discrepancies, or at least major disparities, among global and near‐global data sets for lunar highland surface composition. When compared with data from other sources, reported mafic mineral abundance results from the Kaguya Spectral Profiler (Kaguya SP) spectral reflectance method for four Apollo 16 soils appear systematically low by a factor of 0.6, or an even more extreme factor (~1/3) if viewed in relation to the soils’ nonglass or CIPW mineralogy. Also, whether evaluated on a global median basis or on the basis of site‐by‐site comparison (for Apollo 16, Luna 20, and Apollo 17), the compositions found by the Kaguya SP technique show discrepancy, or at least disparity, versus other mafic abundance observations by that same factor of ~1/3. Spectral reflectance does not supply a simple bulk analysis of the target soil. The reflectance mineralogical signal is preponderantly determined by the nonglass fraction, and especially the masswise subordinate 10–20 µm grain size fraction. Literature data show that in anorthositic lunar soil, chemical composition is fractionated, more extremely anorthositic, for the nonglass component compared to the glass component. Also, the grain size fraction (10–20 μm) that most closely matches bulk reflectance has a significantly higher abundance of impact/agglutinitic glass than does the coarser material that dominates the soil mass. The Kaguya SP mafic abundance calibration needs adjustment by a factor of nearly 3 if results are to be interpreted as indicative of the mineralogy of the underlying crust. A claimed detection of several hundred lunar 500 m scale purest anorthosite (PAN; ≥98 vol% plagioclase) locales among millions of spectral reflectance observations is dubious, in part because with large data sets, compositional extremes are inevitably exaggerated as a byproduct of analytical uncertainty. Preponderance of PAN composition is rare among terrestrial layered intrusive anorthosites and is neither required nor expected for the flotation crust of a global magma ocean. Buoyant flotation and compaction would not suffice to yield pure plagioclase unless adcumulus growth was negligible, and trace element contents of ferroan anorthosites show that their mafic silicate components are for the most part of adcumulus, not “trapped melt,” derivation. A PAN‐dominated crust would imply a curiously fractionated (low) thorium/aluminum ratio for the crust, an implausibly high mantle/crust Th concentration ratio, and an oddly low Th/Al for the bulk Moon. Remote sensing techniques for planetary regolith composition are not easy to calibrate, particularly near the extremes of composition‐space and sensitivity.