Thermochemical Convection Models of Lunar Evolution

Dave Stegman, Mark Richards, Mark Jellinek, John Baumgardner (LANL)


We consider the early thermo-chemical history of the Moon and specifically address the question of how the Moon could have had an internally generated magnetic field suddenly 'switch-on' somewhat late in its evolution and then just as quickly 'switch-off'. It is commonly assumed the Moon never underwent mantle convection, given the majority of the surface geology is likely the original crust formed nearly simultaneous with the Moon. This can partially be explained by the fact that thermal evolution may well occur under the regime of stagnant-lid convection. Furthermore, there are a few tantalizing clues that perhaps the Moon possessed a brief, internally generated magnetic field (Cisowski et al., 1983) and by implication, that the interior of the Moon was once convecting. Lunar samples returned from the Apollo missions provide a few which may contain a remnant thermal magnetism, possibly acquired during the Moon's 'magnetic era' (Cisowski et al., 1983). These samples also reveal the near side of the Moon contains large areas flooded with volcanic material, the lunar mare. These mare (Latin for sea), erupted during a pulse of magmatism beginning 0.5 billion years after the Moon had formed and mostly ended after 1 billion years of activity. We have recently shown (Stegman et al., 2003) that chemical overturn models suggested to explain the eruption of the Maria basalts may also account for the hitherto unexplained existence of a lunar magnetic field (core dynamo) at about the same time ( 3-4 billion years ago). We have chosen the approach of parallel computing to solve governing equations using the 3-D spherical finite element model (for which a considerable amount of effort was spent implementing a Lagrangian tracer algorithm). Our convection models bring together the main features of early lunar post-magma-ocean history, and carry an important testable prediction - that further analysis of lunar samples may yield a definite onset time at  4 Ga for the lunar core dynamo.

Generating an Early Lunar Dynamo

The Moon presently has no internally-generated magnetic field (i.e. core dynamo). However, paleomagnetic data combined with radiometric ages of Apollo samples record the existence of a magnetic field from approximately 3.9 to 3.6 Ga ('magnetic era') possibly due to an ancient lunar dynamo (Cisowski et al., 1983; Collinson, 1993). A dynamo during this time period is difficult to explain(Collinson, 1993; Stevenson, 1983), because current thermal evolution models for the Moon (Konrad and Spohn, 1997) yield insufficient core heat flux to power a dynamo after  4.2 Ga. In Figure 38.1, we show that a transient increase in core heat flux following an overturn of an initially stratified lunar mantle may explain the existence and timing of an early lunar dynamo. Using a 3-D spherical convection model (Baumgardner, 1985), we show that a dense layer, enriched in radioactive elements ("thermal blanket"), at the base of the lunar mantle initially prevents core cooling, thereby inhibiting core convection and magnetic field generation. Subsequent radioactive heating progressively increases the buoyancy of the thermal blanket, ultimately causing it to rise back into the mantle. The removal of the thermal blanket, proposed to explain the eruption of thorium and titanium-rich lunar Mare basalts Hess and Parmentier, 1995), plausibly results in a core heat flux sufficient to power a short-lived lunar dynamo.

Figure 38.1: Thermochemical evolution models in stable (a-c) and unstable (d-h) thermal blanket regimes as seen in temperature (a,d,g), composition (b,e,h) and core heat flux (c,f) compared with paleomagnetic data (i). The equatorial cross-sections of (a) temperature and (b) composition for model TB-1 at 400 million years show thermal blanket material is too dense to to become buoyant, but some entrainment occurs. However, equatorial cross-sections of temperature and composition for model TB-2 show a marginally stable thermal blanket interacting with mantle convection at 100 million years (d,e) and that by 400 million years (g,h) has sufficient thermal buoyancy to rise back towards lunar surface. Core thermal history (c) for reference model TB-$\emptyset$ shows heat flux values (blue line) well below adiabatic core heat flux (shaded region) while model TB-1 (red line) has nearly zero heat flux (black line). Such core heat flux values ranging between the black line and shaded region indicates a thermally stratified core, in which all core heat loss is by conduction and no dynamo is supported. A core heat flux equal to or above the shaded region indicates core convection and likely occurrence of a dynamo, as seen in models TB-2,3 (f). Paleointensity measurements (i) from Apollo samples (modified from Cisowski et al. 1983) where dots indicate absolute paleointensity measurements (Thellier-Thellier method in red, other techniques in blue) and crosses indicate scaled normalized relative paleointensities. In our models, asymmetric thermal blanket removal leads to a localized distribution of partially-molten thermal blanket material at relatively shallow mantle depths, confirming a plausible explanation for the eruption of high-Th, high-Ti mare basalts, similar to the models of Zhong et al., 2000. Our models make no attempt to evaluate melt transport to the surface.


We thank B. Buffett, C. Johnson, R. Jeanloz, M. Manga, and H-P. Bunge for helpful discussions. This work was supported by IGPP LANL, NASA CT project, NSF, Miller Institute for Basic Research. We dedicate this work to the memory of the late Stephen Zatman.


Collinson, D. W., Magnetism of the Moon - A lunar core dynamo or impact magnetization? Surv. Geophys., 14, 89-118, 1993.

Stevenson, D.J., Planetary magnetic fields, Rep. Prog. Phys., 46, 555-557, 1983.

Hess, P.C. and E.M. Parmentier, A model for the thermal and chemical evolution of the Moon's interior: implications for the onset of mare volcanism, Earth Planet. Sci. Lett., 134, 501-514, 1995.

Konrad, W. and T. Spohn, Thermal history of the moon - Implications for an early core dynamo and post-accretional magmatism, Adv. Space Res., 19, 1511-1521, 1997.

Baumgardner, J.R., Three dimensional treatment of convective flow in the Earth's mantle, J. Stat. Phys., 39, 501-511, 1985.

Stegman, D.R., A.M. Jellinek, S.A. Zatman, J.R. Baumgardner, and M.A. Richards, An early lunar core dynamo driven by thermochemical mantle convection, Nature, 421, 143-146, 2003.

Zhong, S., E.M. Parmentier, and M.T. Zuber, A dynamic origin for the global asymmetry of lunar mare basalts, Earth Planet. Sci. Lett., 177, 131-140, 2000.

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail:
© 2004, The Regents of the University of California