Fossil raindrops and ancient air

An analysis of fossil imprints of ancient raindrops suggests that the density of the atmosphere 2.7 billion years ago was much the same as that today. This result casts fresh light on a long-standing palaeoclimate paradox. See Letter p.359

When a main-sequence star such as our Sun ages, its inner core becomes denser and the temperature at which its hydrogen is fused to helium increases. As a result, the Sun is currently more luminous, and delivers more energy to Earth's surface, than in the past — its energy output around 2billion years ago is inferred1 to have been less than 85% of that today. Such a faint Sun should not have been able to warm Earth's surface above the freezing point of water2, yet the geological record indicates that liquid water abounded at that time. The apparent contradiction between theory (sub-freezing Earth surface temperatures) and observation (liquid water) is known as the 'faint young Sun' paradox. On page 359 of this issue, Som et al.3 address this paradox using seemingly unlikely evidence: fossilized imprints left by ancient raindrops.Footnote 1

There are ample indications of liquid water under the faint Sun. In South Africa, for example, exposed rocks more than 3billion years old contain features associated with flowing or standing water, including sedimentary deposits that preserve ripple marks and mud cracks4, glassy 'pillow' lavas that were rapidly quenched3 and alga-like microfossils5. Indirect evidence, meanwhile, extends the record of terrestrial oceans to as early as 4.4billion years ago6. So how can this be explained?

Solutions to the paradox fall into two general categories: those contending that Earth's atmosphere retained heat more efficiently in the past than it does now, for example because of increased concentrations of greenhouse gases, and those arguing that the albedo, or reflectance, of Earth was lower in the past, perhaps because there were fewer clouds and/or less ice. Most models used to explain the paradox are purely theoretical, and were designed to highlight the conditions necessary to solve it. Unfortunately, few observational constraints to support or refute these models have been identified, and those that have been proposed7,8,9,10 tend to be controversial.

Som et al.3 have implemented a clever approach — first attempted11 in 1851 by a pioneer of geology, Charles Lyell — to determine the density of the atmosphere early in Earth's history. This information is crucial for assessing whether greater concentrations of greenhouse gases (such as carbon dioxide12), or of other gases (such as nitrogen13) that amplify the effects of greenhouse gases, could explain the faint young Sun paradox.

Specifically, the authors inferred the velocity of falling ancient raindrops from the geometry of fossilized raindrop-impact marks preserved in a 2.7-billion-year-old sedimentary rock from South Africa (Fig. 1). The atmosphere exerts a drag on raindrops such that they typically fall at a terminal velocity that is inversely proportional to the density of the atmosphere. Any difference between the inferred velocity of ancient raindrops and that of those that fall today may therefore reflect a change in atmospheric density. Of course, the imprint generated by a raindrop falling at a given velocity depends both on the size of the drop and on the nature of the substrate onto which it falls. To constrain these variables, the authors observed the size distributions of naturally occurring raindrops, and coupled this information with data from experiments in which they let water droplets fall onto volcanic ash — mimicking the conditions in which the fossil raindrops formed.

Figure 1: Solid evidence.


Som et al.3 have analysed the fossilized imprints of raindrops, such as those shown here, to determine the atmospheric density 2.7 billion years ago. Rule, 5 cm.

Som et al. conclude that the atmospheric density 2.7billion years ago was probably 50 to 105% of that today. This finding immediately calls into question solutions to the faint young Sun paradox that invoke elevated concentrations of greenhouse gases, unless small increases of greenhouse-gas concentration were able to exert a large warming effect. It is also unlikely that higher concentrations of greenhouse-enhancing nitrogen could have caused the paradox, because concentrations of twice or more the present atmospheric abundance would have been required to provide sufficient warming in the presence of a modest increase in carbon dioxide13. Under such conditions, the atmospheric density would have been greater than that predicted by the authors. It therefore seems that elevated concentrations of highly effective greenhouse gases, such as methane14, ethane15 and/or carbonyl sulphide16, may be required to explain the paradox, possibly in combination with moderately higher concentrations of less-effective greenhouse gases such as carbon dioxide14. A lower planetary albedo, caused by the reduction or absence of continental ice sheets, could also have contributed to warming.

Although raindrop size distributions associated with typical storms are well known, it is possible — albeit unlikely — that the ancient raindrops responsible for the fossilized imprints were unusually large. Small errors in the inferred size of the raindrops would result in significant errors in the atmospheric pressure predicted by Som and colleagues' method3. The accuracy of the method is further limited by lack of information about factors (such as moisture content) that would have affected the cohesiveness of the ash in which the fossil imprints were made; the cohesiveness can affect the morphology of impact craters. The atmospheric density 2.7billion years ago could therefore have been more than twice that of the modern atmosphere if the circumstances under which the fossil imprints formed were unusual.

It is to be hoped that Som and colleagues' work will stimulate further studies of fossil raindrop imprints, including perhaps those originally observed by Lyell. In particular, it will be interesting to see whether coherent temporal trends in atmospheric pressure can be inferred from imprints in deposits of varying ages. With increasing recognition and analysis of such features in the geological record, it may be possible to establish a chronological record of atmospheric pressure on Earth throughout the past 3.5billion years. Such a record would shed light on otherwise poorly constrained aspects of climate change deep in Earth's history.


  1. 1.

    *This article and the paper under discussion3 were published online on 28 March 2012.


  1. 1

    Ribas, I. in Solar and Stellar Variability: Impact on Earth and Planets (eds Kosovichev, A. G., Andrei, A. H. & Rozelot, J.-P.) Ch. 1 (Cambridge Univ. Press, 2010).

  2. 2

    Sagan, C. & Mullen, G. Science 177, 52–56 (1972).

  3. 3

    Som, S. M., Catling, D. C., Harnmeijer, J. P., Polivka, P. M. & Buick, R. Nature 484, 359–362 (2012).

  4. 4

    Ramsay, J. G. Trans. Geol. Soc. S. Afr. 66, 353–401 (1963).

  5. 5

    Schopf, J. W. & Barghoorn, E. S. Science 156, 508–512 (1967).

  6. 6

    Harrison, T. M. Annu. Rev. Earth Planet. Sci. 37, 479–505 (2009).

  7. 7

    Reinhard, C. T. & Planavsky, N. J. Nature 474, (2011).

  8. 8

    Dauphas, N. & Kasting, J. F. Nature 474, (2011).

  9. 9

    Goldblatt, C. & Zahnle, K. J. Nature 474, (2011).

  10. 10

    Rosing, M. T., Bird, D. K., Sleep, N. H. & Bjerrum, C. J. Nature 474, (2011).

  11. 11

    Lyell, C. Edinb. N. Phil. J. 51, 70–74 (1851).

  12. 12

    Kasting, J. F. Precambr. Res. 34, 205–229 (1987).

  13. 13

    Goldblatt, C. et al. Nature Geosci. 2, 891–896 (2009).

  14. 14

    Sheldon, N. D. Precambr. Res. 147, 148–155 (2006).

  15. 15

    Haqq-Misra, J. D., Domagal-Goldman, S. D., Kasting, P. J. & Kasting, J. F. Astrobiology 8, 1127–1137 (2008).

  16. 16

    Ueno, Y. et al. Proc. Natl Acad. Sci. USA 106, 14784–14789 (2009).

Download references

Author information

Correspondence to William S. Cassata or Paul R. Renne.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Cassata, W., Renne, P. Fossil raindrops and ancient air. Nature 484, 322–324 (2012).

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.