Monday, March 15, 2010

Snowball Earth Proven?

Snowball Earth is a popular, controversial idea in which Earth is thought to have been completely covered at least once in ice during the Sturtian glaciation of roughly 718-700 Ma. This glaciation episode, bookended by two others, the Kaigas (~785-728 Ma) and the Marinoan (~670-624 Ma), comprise the period of the Late Precambrian known as the Cryogenian Period and are a precursor to the Cambrian Explosion of roughly 530 Ma.

This is proposed as a hypothesis to explain some sediments seen worldwide that some geologists are having trouble characterizing any other way. Glacially-associated sediments are distinct from the more common wind- and water-driven deposits. Because of the viscosity of ice, it is capable of carrying sediments of varying grain size from dust to megaboulder. Some sediments are sorted out during melting into eskers or loess, but other sediments are dropped where they are when the glacier starts to melt. Advance/retreat periods during melting can push up ridges of unsorted sediments at the glacier head known as moraines.

Skeptics of the hypothesis argue that there are problems with the idea of worldwide glaciation, and their arguments are very well-founded. Those of us who see snow every year are aware of how bright the world can get when snow is on the ground and the sun is shining. This is because snow and ice have a high albedo, or ability to reflect light. If light is reflected rather than absorbed, not much energy is available at surface level to melt snow or ice. It's possible that runaway glaciation would be VERY hard to reverse once established simply because of the albedo argument. I believe some models back this up. Others don't. It's controversial, which is what makes it a compelling story of Earth's past.

There are other arguments, but I strongly suggest reading through the Wikipedia article linked at the top of this post. It's fascinating.

Sediments such as these have been observed in paleoequatorial regions. Using paleomagnetic data embedded in rocks (that hasn't been too far altered by subsequent metamorphic processes), it is possible to extrapolate a paleolatitude of a particular rock formation. This is done by isolating the remanent magnetic signature (thought of as a vector) in the native Fe2O3 (hematite), Fe3O4 (magnetite), and FeTiO4 (ilmenite) crystals present in a rock. Alignment may occur in the internal structure of a crystal during cooling and crystallization of an igneous rock (if it has enough time), or during deposition of iron-bearing sedimentary rocks (if the magnetic field has enough influence to align enough of the grains in situ). If it is possible to establish an age for these same rocks, we know when that piece of continent was at that particular latitude (but NOT the longitude, because assuming a perfect magnetic dipole, the magnetic vector is the same at all longitudes at that latitude).

Better yet if you find glacial sediments hanging out in close association with rock types typically found in tropical environments. Uniformitarianism for the win.

That's what scientists have managed to do with a suite of rocks in northwestern Canada. Tropical sediments and presumably paleomagnetic data (not directly mentioned in the article) were used to identify what could be a smoking gun for Snowball Earth.

As a corollary, some glacial sediments are now found at tropical latitudes, but that plus paleomag provides a smoking gun in favor of plate tectonics. Typical glaciations throughout Earth's history were not at the level of the Cryogenian glaciations.

1 comment:

Dog Momster said...

Just a note for non-geology readers... I had to check a few terms, because they *look* so close to "everyday" terms... they are not misspellings! All the words you see in the post truly exist in the world of Geology!