New insights into lunar crustal magnetism from joint inversions

After the initial differentiation of the Moon, subsequent thermal, magmatic, and impact-related processes modified the crust over time, leaving a long-lived geophysical record. Through these modifications, a portion of the lunar crust acquired remanent magnetization, leading to pronounced magnetic anomalies (some reaching tens of nT at the surface) detectable from orbit [1–2]. Unraveling the structure and origin of magnetized sources that generate the observed strong magnetic anomalies is key to reconstructing the lunar magnetic history and understanding the thermal and magnetic evolution of the lunar crust.

The formation mechanism responsible for the lunar magnetic anomalies remains unknown. Many anomalies are associated with impact structures, supporting hypotheses involving the emplacement of impact-melt-bearing materials [3–4] and the acquisition of shock remanent magnetization [5]. However, only a small fraction of lunar craters exhibits such magnetic anomalies [6], indicating that impact-related mechanisms alone cannot explain the full range of observed features. Non-impact origins, such as magmatic processes, are less understood but may be significant for lunar crustal magnetism.

One set of geological features that may help bridge this gap are lunar swirls, high-albedo sinuous surface features that lack associated topographic variations but are consistently co-located with crustal magnetic anomalies [7]. Swirls are thought to be shaped by interactions between the solar wind and local magnetic fields, providing a direct connection between crustal magnetism and surface processes. Therefore, analysis of these features can provide insights into the geologic processes that shaped the crust. The best studied example, Reiner Gamma, has been linked to a subsurface magnetized structure associated with the Marius Hills volcanic complex [3,8], suggesting a possible magmatic origin.

If a magnetic anomaly originates from a geologic process that also generates a density anomaly (e.g., magmatic process), a correlation between observed gravity and magnetic data is expected. While dominant structures such as the large-scale linear gravity anomalies do not show such correlation [9], regional correlations of gravity and magnetic field data exist [10]. Such regions are well-suited for combined analysis of the data sets, but none have been conducted yet.

We apply an innovative 3D joint inversion of gravity and magnetic field data [11–12] to characterize the density and magnetization structure beneath the Dewar swirl. The gravity data are derived from GRAIL by removing the signals of surface and Moho relief, and the magnetic data are predicted using a crustal field model from [1]. The inversion begins with a strong coupling between density and magnetization anomalies, allowing for spatial collocation, and then we iteratively relax this coupling to improve the data fit. We focus on the result that fits the observations well while preserving a high degree of correlation between density and magnetization.

The resulting model reveals several prominent magnetization features, some of which are associated with density anomaly, where the Dewar swirl overlies the most prominent density and magnetization anomaly (Fig. 1). Magnetized bodies with positive density anomalies uniformly exhibit horizontal magnetization. Amongst these, the strongest anomaly is correlated with the Dewar swirl. Two key factors might explain the presence of the swirl: sufficiently strong horizontal magnetization and the presence of FeO-rich surface materials (Fig. 2). This is consistent with the local magnetic field shielding the surface from solar wind and thus reducing the effect of space weathering. The Dewar swirl lies within a geochemically anomalous region rich in FeO, whereas none of the other magnetized bodies with horizontal magnetization shares this surface geochemistry. Because space weathering of FeO generates opaque metallic iron, the regions that resist space weathering can form an albedo high in contrast to their surroundings.

Fig 1. Crustal structure in the Dewar region at 1 km depth: (a) density, (b) total, (c) vertical, and (d) horizontal magnetization. The swirl from [7] is highlighted in green. The solid and dashed boxes noted the magnetized bodies with a positive density anomaly and with no density anomaly, respectively. Additional magnetized bodies with negative density anomalies are noted by the arrow.

Furthermore, our results suggest a magmatic origin of the Dewar magnetic anomaly. The positive density anomaly can be explained by intrusive basaltic materials, where magnetization is the result of the enrichment of metallic iron generated from the reduction of ilmenite during its cooling [13] in the presence of a lunar dynamo. This interpretation aligns with previously identified Dewar cryptomare, the buried mare with elevated FeO and mantle-derived materials at the surface [14–15]. Other magnetized bodies with positive density anomalies lack association with geochemical anomalies and are not located near the known cryptomare. These might instead represent intrusive dikes or local emplacements of impact melt.

Fig 2. FeO abundance map in the Dewar region. The map was derived from the Clementine UVVIS datasets using the technique of [16]. The black dashed outline notes the cryptomare region identified by [15]. Other notations are the same as in Figure 1.

Finally, magnetized bodies with no density anomaly exhibit only negative vertical magnetization components of similar strength (dashed box in Fig. 1), suggesting a shared formation period and mechanism. These features could potentially represent the distal ejected materials from large impact events such as the South Pole-Aitken [4]. In contrast, weakly magnetized bodies accompanied by negative density anomalies may have formed through impact shock. Such a process could increase crustal porosity, which manifests in a negative gravity anomaly, while simultaneously generating shock remanent magnetization.

 

Reference

[1] Tsunakawa, H. et al. JGR-Planets (2015). [2] Ravat, D. et al. JGR-Planets (2020). [3] Oliveira, J. et al. JGR-Planets (2017). [4] Wieczorek, M. et al. Science (2012). [5] Hood, L. & Artemieva, N. Icarus (2008) [6] Yang, X. & Wieczorek, M. Icarus (2024). [7] Denevi, B. et al. Icarus (2016). [8] Oliveira, J. et al. JGR-Planets (2024). [9] Andrews-Hanna, J. et al. Science (2013). [10] Gong, S. & Wieczorek, M. JGR-Planets (2020). [11] Moorkamp, M. GRL (2022). [12] Mittelholz, A. et al. JGR-Planets (2025). [13] Liang, Y. et al. JGR-Planets (2024). [14] Lawrence, S. et al. JGR (2008). [15] Whitten, J. & Head, J. Icarus (2015). [16] Gillis, J. et al. GCA (2004).