The Lunar Prospector Electron Reflectometer has obtained the first global map of lunar crustal magnetic fields, revealing that the effects of basin-forming impacts dominate the large-scale distribution of remanent magnetic fields on the Moon. The weakest surface magnetic fields (<0.2 nT) are found within two of the largest and most recent impact basins, Orientale and Imbrium. Conversely, the largest concentrations of strong surface fields (>40 nT) are diametrically opposite to these same basins. This pattern is present though less pronounced for several other post-Nectarian impact basins larger than 500 km in diameter. The reduced strength and clarity of the pattern for older basins may be attributed to: (1) demagnetization from many smaller impacts, which erases antipodal magnetic signatures over time, (2) superposition effects from other large impacts, and (3) variation in the strength of the ambient magnetizing field. The absence of fringing fields stronger than 1 nT around the perimeter of the Imbrium basin or associated with craters within the basin implies that any uniform magnetization of the impact melt must be weaker than ∼10−6 G cm3 g−1. This limits the strength of any steady ambient magnetic field to no more than ∼0.1 Oe at the lunar surface while the basin cooled for tens of millions of years following the Imbrium impact 3.8 billion years ago.

Introduction: Lunar Prospector neutron spec- trometer (NS) observations have previously been com- bined with spectral reflectance estimates of FeO and TiO 2 (1, 2, 3, 4) to infer the distribution of the rare earth elements (REEs) gadolinium and samarium on the lunar surface (5). This is possible because Fe, Ti, Gd and Sm are the most important thermal neutron absorbers in lunar materials. Here the Fe contributions are removed using preliminary gamma ray spectrometer data. Analysis of Clementine spectral reflectance (CSR) data provides the nominal Ti-contributions to the NS data, and GRS thorium can be used as a proxy for the Gd and Sm. However, some deviations from a good Th-REE correlation are apparently related to overestimates and underestimates of the CSR-derived FeO and TiO2 abundance values. Here we use Prospector NS and GRS data to help constrain FeO and TiO2. We find evidence suggesting that CSR TiO2 values are too high in several of the nearside maria. Approach: In previous work we have relied on CSR FeO and TiO2 estimates to derive REE abun- dances from the NS data. Here we will use preliminary estimates of FeO abundance derived from the Prospec- tor GRS to determine the degree of neutron absorption due to Fe, and we will use GRS estimates of thorium to calculate the REE contribution to absorption. The latter comes from a well-documented correlation of Th and REEs in returned samples of KREEP. We use the GRS FeO and CSR TiO2 abundance estimates to cal- culate Σeff, the macroscopic absorption cross section. In turn, Σeff is directly related to the ratio of the epither- mal neutron flux to the thermal neutron flux. We can then relate part of the deviations from the ideal flux ratio relationship with Σeff to the presence of REEs, and remove this effect by using GRS Th as a proxy. Whatever deviations remain must then be due primarily to errors in the assumed TiO2 abundance. Results: Figure 1 is a map of ΔΣeff, the deviation of the calculated Σeff from the ideal flux ratio relation- ship. This map delineates high positive values (yel- low-red) that are associated with concentrations of REEs in KREEPy terrains. Negative values (purple- magenta) that reflect the apparent overestimate in major element neutron absorber, namely FeO, TiO2 or both. Figure 2 is a scatter plot of GRS Th vs ΔΣeff. Blue points correspond to low-Ti, where errors in es- timated TiO2 have a small effect on Σeff. This illus- trates the trend between the REEs gadolinium and sa- marium, and thorium. Red points correspond to high- Ti locations; many of these lie to the left of the Th- REE trend, indicating that ΔΣeff is pulled to lower values by overestimates of FeO, TiO 2 or both.