New Constraints on Thermal and Dielectric Properties of Lunar Regolith from LRO Diviner and CE‐2 Microwave Radiometer

Abstract

We derive a new constraint on the thermal and dielectric properties of the lunar regolith layer by reconciling data from the Lunar Reconnaissance Orbiter (LRO) Diviner infrared radiometer and Chang'E-2 (CE-2) microwave radiometer (MRM). The bolometric Bond albedo of the lunar surface, which characterizes the ability of the lunar surface to reflect visible radiation, is a function of incidence angle. We determined the Bond albedo by using the Lunar Orbiter Laser Altimeter 1,064-nm normal albedo and the surface temperature at noon as a function of latitude. The results suggest a modification to existing regolith thermal conductivity models based on a fit to the diurnal variation of Diviner data. Based on the thermal model, a 1-D radiative transfer and dielectric properties model is developed to fit MRM data for the global Moon. With a new dielectric loss tangent equation for highland regolith applied, our model matches MRM data well at 19.35 and 37 GHz, which are generally accepted to be well calibrated. A global map of loss tangent of the Moon at these frequencies is also obtained by fitting the diurnal amplitude of microwave brightness temperature (T B ) of each location on the Moon. We find that the loss tangent of highlands regolith has a slight frequency dependence and is larger than previous studies. We also identify a large discrepancy between our theoretical model and T B obtained by CE-2 MRM at low frequencies, which is attributed to issues caused by contamination on calibration horn. Plain Language SummaryWe derive the physical properties of the dusty layer on the surface of the Moon by using the data from Lunar Reconnaissance Orbiter Diviner infrared radiometer, Lunar Orbiter Laser Altimeter, and Chang'E-2 microwave radiometer. Our results show that visible radiation reflected by the Moon surface is a function of incidence angle. The dusty layer is much less transparent at microwave band as we expected before. The dielectric properties of the layer vary a lot from bright highlands to dark Maria, which can be used to distinguish different compositions. We also recognize some calibration errors within the microwave radiometer data set.

Lunar Titanium and Frequency‐Dependent Microwave Loss Tangent as Constrained by the Chang'E‐2 MRM and LRO Diviner Lunar Radiometers

“…The map of b′ broadly resembles visible albedo (the global "bolometric bond albedo" plotted here is derived as 0.39*LOLA reflectance, from Lucey et al, 2014 as discussed in Feng et al, 2020) shown in Figure 11c. This should be expected, as albedo is essentially also a measure of composition, such as the ferrous iron content of the silicate minerals and the state of maturity (nanophase iron abundance, Lucey et al, 2014).…”

“…Assuming the density dependence of the Figure 5. Difference between measured and modeled equatorial brightness temperature data versus local time for the Chang'E MRM using loss tangent derived by Feng et al (2020). Note the systematic variation in 3.0-and 7.8-GHz data with local time.…”

“…dashed line represents the results of our standard thermal model (Feng et al, 2020;Hayne et al, 2017) with H ¼ 7 cm density increase using the fit values of b and c for Mare Tranquillitatis. Here, we found the density prefactor of 0.312 from Carrier et al (1991) returned a good fit, so we leave this term unchanged.…”

“…Here, we use a uniform value of ρ d = 1,800 kg m −3 and ρ s = 1,100 kg m −3 , H = 7 cm, which is broadly consistent with Apollo drill cores and penetrometer measurements (Carrier et al, 1991). Other thermal model quantities are detailed in Feng et al (2020).…”

“…We now would like to correct for the “terminator issue” affecting data near 6 AM/6 PM in the 3.0‐ and 7.8‐GHz maps, using an updated highlands loss tangent form from Feng et al (2020) where f is the frequency with a unit of GHz and ρ is density; we can characterize the offset quite well. This loss tangent equation was fit to 19‐ and 37‐GHz CE‐2 highlands data, and as shown in Figure 5, it provides a good fit to the data for all local times.…”