Has Mars always been cold and dry or was it once warm and wet? Reanalysis of spacecraft data reveals a signature from the surface rocks that indicates their composition has been altered in the presence of water.
Two years ago, Bandfield et al.1 published interpretations of data from an instrument on the Mars Global Surveyor spacecraft that caused many planetary scientists to rethink their ideas about the geological evolution of the red planet. On page 263 of this issue, Wyatt and McSween2 describe how they have reinterpreted the same data and arrived at a fundamentally different view. So what's going on?
Bandfield et al. reported that the mineralogy of volcanic regions on Mars was best explained by the existence of two very different rock types — basalt in the southern highlands and andesite in the northern plains. The mineral make-up of a volcanic rock reflects the composition of its source region and whether it has undergone melting or crystallization. On Earth, basalt (which is dominated by the minerals plagioclase and high-calcium pyroxene) is the typical rock of the ocean floor; it is formed by partial melting of mantle rocks beneath the Earth's crust and extrusion of the melt to the surface. Andesite, a more silica-rich rock, is most commonly formed by partial melting of mantle in the presence of the water released as oceanic crust descends into the mantle along subduction zones.
Bandfield and colleagues' conclusion, that there are large areas of both andesite and basalt on Mars, precipitated a number of questions. Has recycling of crust occurred on Mars? Are there significant quantities of water in the mantle? Are the mineralogical divisions due to some other magmatic process?
Wyatt and McSween's reinterpretation2 of the data leads to a very different, but no less intriguing, conclusion. Their calculations also produce a mineralogy that is consistent with basalt for the southern highlands. But for the northern plains, their inferred composition consists of typical basalt minerals plus clays and sheet silicates that are characteristic of low-temperature aqueous alteration — in other words, basalt weathered in the presence of water. It happens that the distribution of this altered basalt corresponds well with the vast areas of the northern plains in which water, debouched from enormous outflow channels that emanate from the southern highlands, may once have been concentrated into an ocean (Fig. 1)3,4. The spatial coincidence is not perfect, and there are concentrations of altered basalt in the southern highlands that are not correlated with enclosed basins or water-related features. Yet the association is strong enough to take the implications seriously.
If Wyatt and McSween's solution is correct, a key question is when the basalt alteration took place. The timing would affect the specific mineralogy of the alteration, as well as the prospects for potential habitats for life. If the volcanic material was erupted into an ocean (or ice, if the ocean was frozen), then we might expect copious hydrothermal activity to have occurred, which would have provided abundant energy sources and habitats for life. Furthermore, the alteration could have been widespread in places and penetrated deep into the crust. If, on the other hand, the oceans were created well after the volcanic rocks were in place, then without heat sources the amount of chemical exchange would have been far less. In this case, the alteration would have been relatively weak and mainly confined to the surface, and the habitats for life would have been far less favourable.
A third alternative is that the altered basaltic rocks are sedimentary in origin, with the alteration occurring at some unspecified time and location upstream from the basin, and the altered rocks subsequently being transported to their present location. In this case, the association of the altered basalt with the ocean basin would be purely a consequence of sediment transport. Certain geomorphological features in the northern plains suggest that interaction between volcanoes and water or ice may have occurred. But these features are sparse relative to the extent of alteration mapped by Wyatt and McSween2. So the question of when and how the alteration occurred remains unresolved.
One might ask how such distinctly different conclusions1,2 can be derived from the same data set. The analyses were conducted with spectra acquired by the thermal emission spectrometer (TES) on Mars Global Surveyor, which is still in operation. Crystal-lattice vibrations in minerals modify the thermal emission spectra from surfaces in distinct ways that can be used to identify minerals and their mixtures. However, the signature from the surface of Mars is relatively weak and, compared to that from terrestrial rocks, bland. As Wyatt and McSween show2, there are several possible solutions for the TES observations that fit the data equally well. Which is the right one? In fact, the solutions differ only subtly. It turns out that high-silica glass and some clay minerals have very similar spectral properties in the TES wavelength range. Bandfield et al. used the glass spectra, Wyatt and McSween did not. But this made a big difference to the interpretations of rock type and thus of geological history.
One source of uncertainty in these new results is that the surface of Mars is quite commonly covered to a variable extent by mobile materials such as dust and sand. True exposures of the crust are rare, although abundant rocks of likely local origin were observed at the Viking landing sites, which are in the northern plains. Poleward of 30°, both north and south, there is a patchy and discontinuous layer of cemented dust and soil, 1–10 m thick, which was formerly ice-rich5. This layer becomes more continuous with latitude, until it covers most of the surface at latitudes above 60° (ref. 6). Water in this layer may be responsible for the alteration suggested by Wyatt and McSween, and poleward of 60° may obscure the signatures of the underlying bedrock.
To see beneath the mantle, we need to measure spectra from isolated, small exposures of bedrock, and that is not possible with the current generation of sensors, which offer only coarse spatial resolution. However, a new series of sensors, with increased spatial resolution and covering complementary wavelength regions, is scheduled to fly to Mars in 2003–2006. By combining all these sources of data, it should be possible to identify true exposures of the martian crust and determine their mineral compositions.
Bandfield, J. L., Hamilton, V. E. & Christensen, P. R. Science 287, 1626–1630 (2000).
Wyatt, M. B. & McSween, H. Y. Jr Nature 417, 263–266 (2002).
Parker, T. J. et al. J. Geophys. Res. E 98, 11061–11078 (1993).
Head, J. W. et al. Science 286, 2134–2137 (1999).
Mustard, J. F., Cooper, C. D. & Rifkin, M. L. Nature 412, 411–414 (2001).
Kreslavsky, M. A. & Head, J. W. J. Geophys. Res. E 105, 26695–26711 (2000).
About this article
Journal of Geophysical Research (2011)