Surface heat flux, which can be defined as the heat flowing out of the interior of a planetary body, provides important constraints on the present-day thermal state of the lunar interior. Measurements, performed in situ during the Apollo program, and from orbit by Chang’E 1, Chang’E 2, and by the Diviner radiometer instrument, indicate important lateral variations in surface heat flux on the Moon (~5-180 mW/m²). The differences between Apollo 15 and Apollo 17 measurements have been explained by the presence of an anomalous region, enriched in uranium (U), thorium (Th), and KREEP (Potassium, Rare Earth Elements, and Phosphorus) elements, located on the lunar nearside. Previous modeling efforts also identified crustal thickness and thermal conductivity variations as secondary causes for surface heat flux differences. However, detailed explanations for the remaining two estimates and their implications for the evolution of the Moon remain highly debated. Additionally, the structure and properties of this putative KREEP-rich layer have remained uncertain.

Therefore, this study proposes a new global geodynamic model of lunar thermal evolution that includes lateral variability in the distribution of radiogenics, crustal thickness, and thermal conductivity. The present setup is capable of simultaneously explaining the variability between the Apollo 15, Apollo 17, and Region 5 heat flux values, while also providing further constraints on the KREEP layer structure and lunar crustal properties. The research question addressed in this work is: What is the effect of crustal structure (radiogenics, thickness, and thermal conductivity distribution) on 3D thermal evolution models of the Moon?

Here, we model the interior dynamics and thermal evolution of the Moon after magma ocean solidification using the fluid solver GAIA. We investigate the abundance and distribution of radiogenics on the Moon, and how they shape the interior temperature distribution through time. Additionally, we account for a spatially variable crustal thickness, derived from gravity and topography data. We also include a laterally variable thermal conductivity model, derived from porosity data, which we constrain using the nearside-farside differential thermal state of lunar basins. We model and vary the extent of a putative KREEP layer underlying the PKT (Procellarum KREEP Terrane) region. We enrich the KREEP layer and crust in heat-producing elements compared to the mantle, simulating an asymmetrical distribution of radiogenics as an initial condition.

We find that measurably lower heat flux values at the lunar south pole compared to Apollo 15 and 17 require KREEP material to extend at least partly beneath Mare Serenitatis. In this case, the Apollo 15 measurement would be representative of the KREEP region average heat flux, while Apollo 17 would lie on its edge. On the other hand, a smaller KREEP region (<1200 km in diameter) would make the Apollo 17 location representative of non-KREEP terrane, and show heat flux comparable to south pole values. This is incompatible with estimates based on the Diviner Lunar Radiometer Experiment onboard Lunar Reconnaissance Orbiter, although uncertainties associated with these estimates are unclear. Heat flux measurements that will be performed by the NASA CLPS-CP12 mission at Schrödinger crater will provide key information to exclude one of these two scenarios, and thus potentially constrain the extent of the KREEP layer underneath the PKT region.

Our results also show that a laterally variable thermal conductivity helps reduce the interior temperatures, while maintaining surface heat flux distribution unchanged. We find an effective farside crustal conductivity of ~2 W/(mK) (comparable to that of compact anorthosite) to best match the differential basin relaxation constraints, implying negligible effect of a porous megaregolith layer on the thermal conductivity. In the KREEP region, we favour models with effective crustal conductivity below 2 W/(mK), suggesting that lunar volcanic basalts may have an even lower bulk conductivity than the 2.6-2.7 W/(mK) values considered here. This could be due to the porosity of lunar volcanic material, a more complex layering of basaltic eruptions, or the effect of the temperature and pressure dependence of thermal conductivity. Although our model setup is simplified, it provides novel insights on the distribution of radiogenics and crustal properties on the Moon, and shows the potential to further constrain the asymmetrical character of lunar evolution.