Lunar H 2 O / Oh - Distributions : Revised Infrared Spectra from Improved Thermal Corrections

Introduction: There has been considerable interest in the spectral response of the lunar surface near 2.8 and 2.95 μm, where OHand H2O have strong absorptions. Previous studies have identified these absorptions with variable strengths across the lunar surface [1–7]. In particular, the absorptions have been found to be most prominent at higher latitudes and at early and late local times (e.g., [1–2]), leading investigators to consider a dynamic environment where much of the OH/H2O migrates around the Moon on diurnal timescales (e.g., [1,4,8–10]). The accuracy of these results are highly dependent on the accounting and removal of emitted radiance from the measured spectra beyond ~2 μm. Current emitted radiance corrections for lunar spectral data assume an isothermal surface and/or a predictive spectral continuum (e.g., [11–12]). These corrections have an advantage of being relatively simple to implement, without additional models or datasets. However, the assumptions inherent in these corrections can lead to a significant mis-estimation of the emitted radiance. For example, surface temperatures and emitted radiance are systematically underestimated in the thermally corrected Chandrayaan-1 Moon Mineralogy Mapper (M3) Level 2 radiance data. To understand and accurately model thermal emission from the lunar surface, we must understand the thermophysical properties of the surface. The spatial, angular, and temporal variations of temperature reflect regolith properties, such as particle size distribution and density. These properties represent a delicate balance of lunar surface processes that produce a highly structured regolith over time (e.g., [13–16]). In particular, the effects of surface roughness on thermal emission have been recognized as a dominant factor in the thermal emission of airless bodies, including the Moon [17–18]. This understanding has come with the collection of more precise and systematic datasets, and the development of increasingly sophisticated models. The current thermal corrections applied to these data are insufficient in part because they do not account for these effects. As a result of surface roughness, lunar surface temperatures derived from NIR measurements will commonly show brightness temperatures that are much higher than the average surface temperature in the measurement field of view (Fig. 1). In addition, the brightness temperature will not be constant with respect to wavelength, even within limited spectral ranges. Properly accounting for anisothermality and correcting for this emitted radiance is crucial for the characterization of the 3 μm OHand H2O absorptions. Methods and Data: For this work, we use a simple radiative equilibrium model to predict the temperature of each surface facet [19]. To model roughness, we use a Gaussian distribution of slopes similar to that of [20]. This reduces the surface slopes/roughness to a single parameter (RMS slope distribution), while maintaining reasonable fidelity to natural surfaces. Using the modeled temperatures for each slope/azimuth combination and slope distribution, the mixture of Planck radiances are calculated in proportion to their contribution to the measurement field of view. The resulting modeled spectral radiance has been compared with Lunar Reconnaissance Orbiter (LRO) Diviner Radiometer measurements. These results indicate that typical lunar regolith surfaces are consistent with a mm to cm-scale RMS slope distribution of 20–25° [19]. The correction of M3 data using the output of the roughness thermophysical model (radiance as a function of wavelength) is relatively straightforward. We assume a Lambertian surface and that Kirchoff's Law applies (ε = 1 – R). The examples shown here assume a surface slope distribution of 20° RMS, similar to typical lunar regolith [19]. Results: The initial application of the roughness model for the removal of thermal contributions from M3 data has a dramatic effect on the resulting spectra (Fig. 2). At wavelengths greater than ~2.5 μm, the