C. T. Russell, G. Schubert and P. J. Coleman, Jr.

  Institute of Geophysics and Planetary Physics, University of California, Los Angeles, Calif., USA

Originally published in:
Proceedings of the Fifth Lunar Conference, Vol. 3, pp. 2747-2760, 1974.


Returned samples, surface observations and orbital surveys reveal extensive magnetization at the lunar surface. The most probable reason for this is that the lunar crust was magnetized by an ancient lunar dynamo. The magnetization caused by an internal source will cause no external field except at non-uniformities such as holes in the magnetized crust.

The direction of the present magnetic field is not predominantly north-south in the lunar equatorial regions as would be expected if the ancient lunar magnetic moment were aligned along the present rotation axis. Rather it is mainly radial and east-west. Possible explanations of this include a dynamo mechanism not dominated by Coriolis forces, and movement of the rotation axis relative to the crust after the magnetization of the crust.



All available evidence indicates that the lunar crust everywhere possesses significant remanent magnetization. The returned lunar samples contain stable natural magnetization [1]; the surface magnetometers measured permanent fields at every site with magnitudes of up to over 300 [2]; and orbital surveys at altitudes near 100 km with the Apollo subsatellites reveal large regions of coherent magnetization [3]. In fact, the low altitude coverage (< 50 km) obtained with the Apollo 16 subsatellite just prior to crashing into the lunar surface showed a continuously magnetized surface in the limited region surveyed [4].

On the other hand, there is apparently no global magnetic field. Fourier analysis of the tangential and radial components of the Apollo subsatellite magnetic field observations in the relatively steady field and near vacuum conditions of the geomagnetic tail provides a precise measure of the projection of the lunar magnetic moment in the subsatellite orbit plane. Analysis of the seven-month sample of Apollo 15 results in an upper limit of 1.3 x 1018 gauss cm3 in the Apollo 15 subsatellite orbit plane [5]. If we assume that the moment is parallel to the lunar rotation axis this limits the total moment to < 3 x 1018 cm3. Analysis of the smaller 2-month sample of Apollo 16 data gives a moment of 6.5 + 5.4 cm3 in the Apollo 16 orbit plane [6]. Combining the two results we obtain a moment of (2, - 6, -9) x 1018 cm3 along the three selenographic axes. However since both the Apollo 15 and 16 measures do not differ significantly from zero the combined result is also consistent with zero.

Analysis of the magnetization of the returned lunar samples requires an ancient magnetizing field of over 1 oersted [7]. It has been hypothesized that this field was of external origin, either the terrestrial field [8] or a field in the solar nebula at the time the moon condensed [9], or that this field was of internal origin, due to a dynamo [10]. The absence of any detectable present day dipole moment places constraints on these models. It is the purpose of this paper to discuss these constraints, together with the implications derived from the altitude dependence and direction of the observed remanent fields.


An Ancient Uniform Magnetization?

There are basically two hypotheses regarding the magnetization of lunar samples by an external field: first, that the lunar interior was magnetized at the time of formation and the resulting dipolar field magnetized the surface material at a later time, and second, that the lunar surface was directly magnetized by the external field. Had the moon ever acquired a uniform magnetization, present measurements of the moment indicate that all traces of this uniform magnetization have disappeared. If the ancient lunar magnetization were sufficient to produce a 1 oersted field at one time and if it were distributed uniformly throughout the moon, but was erased from the inside out as the moon warmed up due to radioactive decay, then only 11 meters of such material could be present today [5]. Furthermore, it is improbable that the moon could have been magnetized sufficiently to give a 1 oersted surface field strength [11]. The typical saturation remanent magnetization of lunar basalts is 3 x 10-3 gauss, whereas a magnetization of 3 X 10-1 gauss is necessary for a 1 oersted field. Thus, we may rule out the possibility that the moon became uniformly magnetized and the resultant planetary field in turn magnetized surface material as it cooled through its Curie point.

The next possibility, that the observed magnetization in the returned lunar samples was directly acquired in an external field, is consistent with only the terrestrial field as a possible source. The observation of magnetization in rocks with ages from 3.2 to 4.0 aeons then constrains the moon to be very close to the earth for about 109 years. Further, the magnetization imposed by the terrestrial field would be roughly uniform, although possibly to a variable depth. If we take a moderately high value, 10-4 emu g-1, for the remanent magnetization of the lunar crust [12], the limits on the dipole moment restricts the uniform magnetization to a layer less than 300 m thick [5]. This latter restriction is not a serious one for the terrestrial field could have reversed many times during this period. The 300 in limit would then be on the net uncancelled thickness of magnetized crust. On the other hand, the restriction on the evolution of the lunar orbit is not so easily overcome.


Runcorn's Theorem

Although the moon does not presently generate a planetary magnetic field by means of an internal dynamo, it might have at one time when its spin rate and/or its internal conditions were different. One might expect that this dipolar field would have left sonic trace in the lunar crust which cooled and became magnetized while the dipole field was still present. However, Runcorn has shown that internally generated fields do produce magnetization patterns in a spherically symmetric crust whose magnetic field lines are confined to within the crust [13, 14]. This is illustrated in Fig. 1 [15]. Thus, a crust magnetized in strength and orientation in a dipolar fashion has no external field. We note that holes in this crust or other irregularities will lead to local surface-correlated fields such as the magnetic anomalies observed with the subsatellites [3].

Fig. 1. The magnetization (left panel) and magnetic induction (right panel) for a dipolar shell of magnetization, illustrating that there is no magnetic field exterior to the shell.

The simple model, then, of an original internal dynamo, the subsequent magnetization of the crust as it cools through the Curie point, and then the destruction and/or enhancement of the magnetization in certain regions of the moon explains both the absence of a global field, the presence of magnetic anomalies and the general magnetization of lunar material.


Second Order Corrections

While no field "leaks" out of the magnetized shell in the spherically symmetric moon considered by Runcorn, the real moon is not spherically symmetric, the fractional difference between principal axes being a few times 10-3. Further, we expect a difference between the magnetization of the nearside basalts and the highland anorthosite. These will lead to dipolar and higher multipole terms dependent on the magnitude of the asymmetries and the depth of the magnetized crust. To measure these effects we require coverage of the entire lunar surface to determine the spherical harmonic coefficients, and thereby separate effects of different order and degree [14].

A simpler effect is the fact that the lunar crust could not have cooled through the Curie point simultaneously at all depths. Presumably, the outer crust cooled first. The cooled outer crust then produced a uniform field internal to itself in which the lower crust became magnetized when it cooled. This would lead to a dipole moment that was proportional to the square of the efficiency of magnetization K, where K is generally from 10-3 to 10-4.

In a two-layer cooling model in which a is the radius of the moon, b is the inner radius of the outer shell and c is the inner radius of the inner shell, it can be shown that the ratio of the dipole moment due to the magnetized material ms to the moment of the internal magnetizing field m, is given by

For the case in which both shells are 200 km thick ms/mi = 0.023 K2. Taking ms to be 5.2 x 1024 cm3 corresponding to an ancient 1 oersted surface field, and K to be as large as could reasonably be assumed (10-3), mi is 1.2 X 1017 cm3, which is over an order of magnitude less than our observed present day upper limit. Thus, we cannot make use of this second order term to test the hypothesis of an ancient lunar dynamo with the present subsatellite observations.


The Altitude Dependence

Fig. 2. The attitude dependence of the magnitude of the fine scale (< 600 km) lunar magnetic field for tile four lunar selenographic quadrants centered oil the 0o, 90o, 180o and 270o meridian using only the Apollo 15 data. The quadrants are labeled according to whether they are predominantly mare, M, highland, H, highland with some mare H(M), etc.

The altitude dependence of the lunar remanent magnetic field depends both on its scale size and on its depth. Fig. 2 shows the altitude dependence of the fine scale lunar magnetic field (scale sizes < 600 km) as measured with the Apollo 15 subsatellite over the four lunar quadrants centered on the 0o, 90o, 180o and 270o selenographic meridians. We have drawn three curves corresponding to dipoles buried at depths of 50, 100 and 200 km. The curves have been drawn to cross each other and the trend in the data at 100 km altitude. Deviation of the observations from the models occurs at high altitudes as the observations approach the digitization noise level of the magnetometer All four quadrants are consistent with source depths of 100-200 km and perhaps even greater from 225o to 315o. A 100-200 km source depth is, of course, consistent with a depth of magnetization of from 200 to 400 km and is consistent with the ideas discussed in the previous two sections. On the other hand, this apparent source depth could also be due to a surface scale size of from 200 to 400 km which is not unlike that observed.

Fig. 3. The altitude dependence of the Apollo 16 subsatellite measurements of the fine scale magnetic field for the four lunar selenographic quadrants centered on the 45o, 135o, 225o and 315o meridians. Apollo 15 data have been included where this satellite covered a similar ground track. Comments in caption of Fig. 2 apply.

Fig. 3 shows a similar set of curves for the Apollo 16 subsatellite including Apollo 15 data where the Apollo 15 and 16 suborbit tracks cross. Again, the data are best fitted by altitude dependences due to dipoles buried at from 100 to 200 km. We note that the one exception in the 180o-270o quadrant, where the altitude dependence is much more rapid from 40 to 60 km altitude than even the 50 km curve, may be due to the fact that all these observations were made with the Apollo 16 subsatellite over a slightly different ground track than that covered by the Apollo 15 data which dominate at higher altitudes. Again we must conclude that the data are consistent with a crust which is magnetized to depths of up to several hundred kilometers.


The Direction of Magnetization

If the moon were magnetized by a dynamo similar to that of the earth's we might expect the original dipole moment to be roughly parallel to the lunar rotation axis. We would then expect the magnetization of the crust to be mainly north-south in the equatorial regions and radial near the poles. However, inspection of the contour maps of the lunar magnetic field in the equatorial regions [3, 4] reveals the north-south fields to be the weakest of the three components. Furthermore, detailed examination of the altitude dependence of the vector field [15] and modeling of the Van de Graaff region [16] shows principally radial and east-west fields in the near-equatorial regions. Possibly the dynamo did not act in a manner similar to the terrestrial dynamo. For example, if the importance of the Coriolis force were much less in the lunar dynamo than in the terrestrial dynamo, we might expect the magnetic pole to assume some other orientation, perhaps along the moon-earth line, or perhaps wandering with time. Alternatively, the lunar rotation axis may have shifted relative to the present lunar crust. We note that if the moon were magnetized by an external field, we would be forced to conclude that the lunar rotation axis had shifted, for an external field will only magnetize a rotating body parallel to its rotation axis.



The absence of a global field in the presence of wide spread magnetism on the lunar surface can be most easily explained in terms of an ancient lunar dynamo which magnetized the lunar crust and then disappeared. The only external field consistent with the measurements is the terrestrial field. However, the moon would have had to remain close to the earth for close to 109 years if this were the source of the magnetizing field.

The upper limit on the present day magnetic dipole moment, and the altitude dependence are both consistent with a deep magnetized crust but do not prove its existence. The direction of magnetization that is consistent with the observed fields is not predominantly north-south, but is radial and east- west. This implies that the ancient lunar magnetic dipole moment was not along the present rotation axis.

These results from suborbit tracks covering only a limited portion of the lunar surface are very intriguing. Hopefully, future missions will provide data over a large portion of the lunar surface to confirm and extend these results and to explore other questions such as the asymmetries in the magnetized lunar crust.



We gratefully acknowledge many fruitful discussions of these observations with S. K. Runcorn. This work was supported by the National Aeronautics and Space Administration under NASA grant NGR 05-007-351.



[1] M. Fuller Rev. Geophys. 12, 23 (1974).

[2] P. Dyal et al., Rev. Geophys. 12, 586 (1974).

[3] C. T. Russell et al., Color contour maps of Lunar magnetic field, Proc. 5th Lunar Science Conf. frontispiece, Pergamon Press, Oxford and New York 1974.

[4] C. T. Russell et al, Color contour maps of Lunar magnetic field, Proc. 6th Lunar Science Conf., frontispiece, Pergamon Press, Oxford and New York 1975 (in press).

[5] C. T. Russell et a]., Proc. 5th Lunar Science Conf., Pergamon Press, Oxford and New York 1974 (p. 2747).

[6] C. T. Russell et al., Space Research XV, 621 (1975).

[7] A. Stephenson et al., Proc. 5th Lunar Science Conf., Pergamon Press, Oxford and Now York 1974 (p. 2859).

[8] H. Alfven and L. Lindberg, The Moon 10, 322 (1974).

[9] S. K. Runcorn and H. C. Urey, Science 180, 636 (1973).

[10] S. K. Runcorn et al., Proc. Apollo 11 Lunar Science Conf., MIT Press, Cambridge, Mass. 1970 (p. 2369).

[11] A. Stevenson et al., Proc. 6th Lunar Science Conf., Pergamon Press, Oxford and New York 1975 (p. 3049).

[12] G. W. Pearce et al.. Proc. 4th Lunar Science Conf., Pergamon Press, Oxford and New York 1973 (p. 3045).

[13] S. K. Runcorn Nature 253, 701 (1975).

[14] S. K. Runcorn Phys. Earth Planet. Interiors 10, 327 (1975).

[15] C. T. Russell et al., Proc. 6th Lunar Science Conf., Pergamon Press, Oxford and New York 1975 (p. 2955).

[16] D. W. Strangeway et al., Proc. 6th Lunar Science Conf., Pergamon Press, Oxford and New York 1975 (p. 2975).

Back to CT Russell's page More On-line Resources
Back to the SSC Home Page