A new study from a team of French researchers from the Institut de Physique du Globe de Paris (CNRS/IPGP/Université Paris Diderot) working on conjunction with scientists from Brazil and the United States has challenged the belief that Earth was completely covered in ice 635 million years ago, creating what is now colloquially known as the Snowball Earth hypothesis.
The work was published in the journal Nature on October 6, and specifically challenges the hypothesis that states the Snowball Earth retreated from its full glaciation as a result of an increase in volcanic sourced carbon dioxide.
There have been several extreme glaciations in Earth’s history, two of which took place during the Cryogenian period which occurred some 710 to 630 million years ago.
However, in 1992 and 1998 scientists hypothesised that approximately 635 million years ago our planet saw a massive glacial shift which ended up covering the entire planet in ice.
However there has always been one underlying problem: given that ice better reflects sunlight than rock, and if the whole planet was covered in this wonderful reflective covering, how did it come to an end?
The original theory was that enough carbon dioxide of volcanic origin had built up in the atmosphere to warm the surface enough to melt the ice away. This scenario requires carbon dioxide concentrations of as much as 120 000 ppmv(1) (i.e.12%), which is 300 times greater than CO2 concentrations today.
The French, Brazilian and US researchers set out to assess the atmospheric carbon dioxide concentrations of the period by studying carbonates deposited 635 million years ago during the Marinoan glaciation. The researchers found that the carbon dioxide levels at the time were actually closer to what they are today (less than 3,200 ppmv), which is far from what would be necessary to increase the temperature in the atmosphere enough to melt away a Snowball Earth.
Not only does this work challenge the Snowball Earth hypothesis, but also the general strength of the glacial episodes which, according to the data, were not as strong as had previously been suggested.