Lunar regolith properties derived from LRO/Diviner data and thermophysical modelling

Introduction
The Moon as our nearest celestial object is one of the most important bodies for space resource exploration and planetary science. However, knowledge of the physical properties of the lunar regolith is required for the exploitation of lunar resources and for understanding the Moon's geologic history. This knowledge comes mainly from Apollo in-situ experiments and returned samples (Heiken et al., 1991), but the global distribution of these properties is still poorly understood. Remote sensing measurements offer the opportunity to derive properties of unsampled areas with the help of models.

In our study, a microphysical thermal model for the lunar regolith was developed and the simulated surface temperatures were compared with thermal emission measurements from the Diviner radiometer (Paige et al., 2010) on board the Lunar Reconnaissance Orbiter (LRO) to derive regolith properties. This work expands upon previous investigations of lunar regolith properties using Diviner data (e.g., Hayne et al., 2017), by more directly simulating physical properties such as particle size and porosity.

Thermophysical model
The thermophysical model is based on the heat transfer equation and takes as input the physical and thermophysical properties of the modelled material, namely the bulk density, thermal conductivity and heat capacity of the lunar regolith. In order to describe these properties as continuous functions of e.g., depth or temperature, parameterised models of these properties are again needed. 

The bulk density profile is described by the stratification model developed by Schräpler et al. (2015), in which the volume filling factor is a function of the grain radius and depth. The steepness of the transition between loose packing at the surface and dense packing at large depths is described by a single parameter. In addition, the density at the deepest model layer, defining the highest volume filling factor, is also a free parameter. The thermal conductivity at all depths is modelled as a function of grain radius, local volume filling factor and temperature (Gundlach & Blum, 2012) and the heat capacity is temperature dependent (Hayne et al., 2017). In our study, the highlands and the maria are modelled separately, taking into account their difference in albedo (Feng et al., 2020), mass density (Kiefer et al., 2012) and thermal conductivity (e.g., Cremers & Birkebak, 1971; Cremers & Hsia, 1974). Fig. 1 shows an example of modelled surface temperatures at the lunar equator.

Derivation of lunar regolith properties
Comparison of simulated surface temperatures with regolith temperatures measured by Diviner (Bandfield et al., 2011) allows us to constrain the free parameters in our thermophysical model (Fig. 2). We selected lunar regolith temperature data from a narrow band around the lunar equator for the highlands and the maria separately and determined the best fitting parameter set to investigate differences between these two types of terrains. For this comparison, only night-time temperatures were used, because they are most sensitive to subsurface thermophysical properties, such as thermal conductivity and bulk density. Fig. 2 shows the Diviner measurements together with the best-fitting thermophysical model runs. The best fits suggest smaller grain sizes and greater bulk densities in the maria compared to the highlands. By comparing the simulated and measured regolith temperatures for individual pixels on the lunar surface, maps of the investigated regolith properties can be produced. The final results along with the derived properties will be presented at the conference.

Conclusion
In this work, we derived lunar regolith properties by matching synthetic surface temperatures from a microphysical thermal model with LRO/Diviner regolith temperature measurements. The investigated parameters are regolith grain size, the transition width from low to high volume filling factor and the deep layer density. The surface temperatures in the highlands and the maria were analysed and modelled separately. During the EPSC, we will report on our findings and present the models and data in more detail.

References
Bandfield et al. (2011), J. Geophys. Res., 116, E00H02.
Cremers & Birkebak (1971), Lunar Planet. Sci. Conf. Proc. 2, 2311.
Cremers & Hsia (1974), Lunar Planet. Sci. Conf. Proc. 3, 2703-2708.
Feng et al. (2020), J. Geophys. Res., 125, E06130.
Gundlach & Blum (2012), Icarus, 219, 618-629.
Gundlach & Blum (2013), Icarus, 223, 479-492.
Hayne et al. (2017), J. Geophys. Res., 122, 2371-2400.
Heiken et al. (1991), Lunar Source Book, Cambridge University Press.
Kiefer et al. (2012), Geophys. Res. Lett., 39, L07201. 
Paige et al. (2010), Space Sci. Rev., 150, 125-160. 
Schräpler et al. (2015), Icarus, 257, 33-46.