A revised algorithm for calculating TiO₂ from Clementine UVVIS data: A synthesis of rock, soil, and remotely sensed TiO₂ concentrations

[1] Investigating mare basalt compositions, at both the sample and remote-sensing level for the Apollo and Luna mare sites, reveals the need for a more complex regression procedure than previously proposed in order to extract accurate TiO₂ concentrations from Clementine spectral reflectance (CSR) data. The TiO₂ algorithm of Lucey and coworkers is modified by using two different sets of regression parameters to relate measured regolith compositions from sampling locations to the CSR properties of these sites. One regression trend fits the majority of Apollo data, and the second regression is a fit to the Apollo 11, Luna 16, and Luna 24 data, which were considered to be anomalous in previous TiO₂ calibrations. These three sites have unusually low albedo compared to other mare landing sites, and some 32% of nearside mare regions appear to share this characteristic. Possible reasons for these differences related to proximity of the other sites to mare-highland boundaries are discussed. Using the dual-regression method, we find (1) that TiO₂ concentrations calculated for the basaltic landing sites faithfully reproduce a bimodal distribution as seen in the sample data, (2) that when coupled with the effects of other thermal neutron absorbers, Ti concentrations are more consistent with observed epithermal-to-thermal neutron-flux ratios than are previous Clementine-based derivations of TiO₂ for basaltic regions, and (3) that basalts of intermediate-TiO₂ concentrations occur most frequently in the Oceanus Procellarum region and that these intermediate concentrations appear to be inherent to the flows underlying the regolith and presumably to the basalt source regions.