Planetary science

Magnetic impact craters

Aerial surveys of the Vredefort impact crater in South Africa suggest that it is only weakly magnetic. The rocks themselves tell a different story, but does this apply to giant impact basins on Mars?

Evidence about the history of Mars can be gleaned from magnetic surveys — hence the importance of data sent back by the Mars Global Surveyor in the 1990s1. The magnetometer on the satellite measured much stronger magnetic fields over some parts of the southern highlands of the planet than fields at similar altitudes over Earth. But the fields are notably weak over the giant impact basins Hellas and Argyre (Fig. 1), suggesting that the dynamo in the martian core was inactive during the era of major meteorite impacts 3.5–4 billion years ago2,3,4. This conclusion has had a considerable influence on ideas about how Mars cooled, and when its mantle and inner and outer cores differentiated.

Figure 1: Magnetic lows over the Hellas (left) and Argyre impact basins on Mars.

Superimposed on the topography are the peak shock pressures, in gigapascals, estimated to have been produced by the impact. The light and heavy magnetic-field contour lines are for 20 and 40 nanotesla, respectively, and were measured by the Mars Global Surveyor at about 400 km altitude. (Reproduced from ref. 3 with permission from the authors and the American Geophysical Union.)

The hypothesis of demagnetized impact basins and all it implies about Mars' evolution is called into question by Carporzen et al.5 on page 198 of this issue. They measure unusually strong magnetizations of bedrock samples in the giant Vredefort impact crater of South Africa, yet aerial measurements of magnetic fields over the crater are lower than over surrounding areas. This paradox is explained by variations in the directions of sample magnetization vectors over distances of 10 cm or less. The strong, but spatially incoherent, magnetic signal of the bedrock is essentially randomized when viewed from larger distances. Meteorite craters can then seem to be magnetic or non-magnetic, depending on how close the magnetometer is to the source. Viewed from satellite altitudes of 100–400 km, martian impact basins would appear magnetically featureless if the magnetic vectors of their source rocks vary in direction over distances of a few kilometres or less.

The Vredefort crater is 2 billion years old and was originally 300 km in diameter. Vredefort is Earth's oldest and largest impact structure, and Carporzen et al.5 argue that, as such, it is our best analogue for the giant martian impact basins. Moreover, other terrestrial impact structures, the Charlevoix and Slate Islands impact craters in Canada among them, also have strong but spatially dispersed sample magnetizations and low aeromagnetic signatures.

But what causes this unusual behaviour? Rocks typically become magnetized when they are cooled from a melt or other high-temperature conditions. This thermal remanent magnetization, like virtually all other mechanisms of magnetization, forms parallel to the magnetic field that acts during cooling of the magnetic minerals in the rock6. This parallelism is central to the success of the palaeomagnetic method: a rock's remanent magnetism is a fossil compass that tracks the ancient magnetic field, recording movements of continents, sea floor and tectonic plates. To account for the magnetization of the Vredefort crater, some other magnetizing mechanism must have been at work during the meteorite impact.

Carporzen et al.5 suggest that highly charged plasmas produced in the impact created magnetic fields that lasted only a matter of minutes, but which were as much as 1,000 times stronger than Earth's magnetic field. The plasma currents that created the transitory fields would have varied on a scale of centimetres and are proposed to be the cause of the strong but randomly directed bedrock magnetizations. The magnetite that carries the bedrock magnetization is confined to planar deformation features in shocked quartz grains. The magnetic carrier certainly formed as a direct result of the shock: unshocked but otherwise similar rocks lack this magnetite. The orientation of the magnetite grains in the quartz cannot explain the random directions of magnetization — any external field in a constant direction would produce a distribution of magnetic vectors over a half-sphere, not over 360° as observed. The field itself must have had a random orientation, and remanent magnetization6 of the magnetite that crystallized during the brief existence of that field is responsible for the bedrock magnetization.

Supporting evidence comes from rocks that melted during the impact, which were also analysed by Carporzen et al.5. Unlike the shocked but unmelted bedrocks, these ‘dyke’ and ‘vein’ samples cooled for days before their magnetization became permanently fixed. By this time, the plasma fields had long vanished and the melt rocks acquired thermal remanent magnetizations as they cooled in Earth's field. Their magnetizations are much weaker than those of the surrounding bedrocks, but they are all in the same direction.

In drawing analogies between the Vredefort crater and giant impact basins on Mars, we must bear in mind the differences in scale, and in the types of rocks involved. Earth's continental crust, where most impact craters are preserved, tends to be granitic, rich in aluminium and silicon but poorer in magnesium and iron. The crust of the martian highlands is more basaltic7, similar to the composition of Hawaii, Iceland and the ocean floor. These rocks contain much more iron, and abundant iron-bearing magnetic minerals can form during cooling without the intervention of shock. The meteorite that formed the Vredefort crater was probably 10–20 km in diameter, but to produce a cavity 1,000 km wide, as Argyre initially was, requires an impacting body closer to 100 km in size. Shock-wave pressures are correspondingly larger and their effects are felt at greater depths.

The strongest magnetic fields measured by the Mars Global Surveyor require at least the upper 30 km of the crust to be magnetized8,9. At these depths, shock and thermal blanketing would cause demagnetization of rocks that are already coherently magnetized — in addition to the new but randomly directed magnetization that plasma fields would create in crystallizing minerals. At the surface, the spatial coherence of bedrock magnetizations could be tested using an outcrop magnetometer on a future rover mission to Hellas or Argyre. It is noteworthy that some lunar rocks returned by the Apollo missions had strong and stable magnetizations whose directions varied by 60–120° between adjacent subsamples10 — these are a possible extraterrestrial example of the suggested magnetization mechanism at Vredefort.


  1. 1

    Acuña, M. H. et al. Science 284, 790–793 (1999).

    Article  ADS  PubMed  PubMed Central  Google Scholar 

  2. 2

    Hood, L. L., Richmond, N. C., Pierazzo, E. & Rochette, P. Geophys. Res. Lett. 30, 1281 (2003).

    Article  ADS  Google Scholar 

  3. 3

    Rochette, P., Hood, L. L., Fillion, G., Ballou, R. & Ouladdiaf, B. Eos Trans. Am. Geophys. Un. 84, 561, 567 (2003).

    Article  ADS  Google Scholar 

  4. 4

    Mohit, P. S. & Arkani-Hamed, J. Icarus 168, 305–317 (2004).

    Article  ADS  Google Scholar 

  5. 5

    Carporzen, L., Gilder, S. A. & Hart, R. J. Nature 435, 199–201 (2005).

    Article  ADS  Google Scholar 

  6. 6

    Dunlop, D. J. & Özdemir, Ö. Rock Magnetism: Fundamentals and Frontiers (Cambridge Univ. Press, 1997).

    Google Scholar 

  7. 7

    McSween, H. Y., Grove, T. L. & Wyatt, M. B. J. Geophys. Res. 108, 5135 (2003).

    Article  Google Scholar 

  8. 8

    Connerney, J. E. P. et al. Science 284, 794–798 (1999).

    CAS  Article  ADS  PubMed  Google Scholar 

  9. 9

    Dunlop, D. J. & Arkani-Hamed, J.

  10. 10

    Fuller, M. Rev. Geophys. Space Phys. 12, 23–70 (1974).

    Article  ADS  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Dunlop, D. Magnetic impact craters. Nature 435, 156–157 (2005).

Download citation


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.