A Neoproterozoic Snowjob: Testing the Limits of the Snowball Earth Hypothesis
Nicholas Christie-Blick (Columbia University, New York, New York, USA)
Ohio State University’s Geology Department colloquium (Columbus, Ohio, USA)
25 April 2002
Paul Hoffman’s Snowball Earth Hypothesis (SEH) doesn’t work very well.
A classic locality for seeing Neoproterozoic tillite is in northern Namibia.
Marinoan Ice Age (= Varanger Ice Age) - ~600 m.y. & the Sturtian Ice Age - ~750-700 m.y. These appear to be the most severe Ice Ages ever on Earth.
Showed a photograph of a large dropstone deforming iron formation (photo from Paul Hoffman).
Evidence for a cold climate at sea level ~600 m.y. ago: striated pavements, paleo-permafrost (sandstone frost wedges), adjacent to tidal bundles in sandstone.
The Elatina Formation in the central Flinders Ranges, South Australia has the best evidence for low-latitude glaciation at sea level - paleomag. shows this locality was at 7.5˚ North latitude (results are close to a primary position). See lots of magnetic reversals in the Neoproterozoic glacial interval of the Elatina Formation. This area really made the case for low-lat. glaciation.
The typical cap carbonate facies - occurs above Snowball Earth glacial sediments. The cap carbonates are peculiar. In the late Precambrian, can see carbonates above the glacial interval. The cap carbonate sediments (deep water and a continuous interval) are event beds - turbidites (sole marks, flute casts) of finely-laminated, homogeneous dolomicrite. They directly overlie glacial rocks everywhere. They are present even in terrigenous sections. Most sections have a few meters only of cap carbonates, though Hoffman focuses on the few sections that have 100s of meters of cap carbonates. See Kennedy, Christie-Blick, & Prave (2001).
Low isotopic values in cap carbonates are perceived to be peculiar.
-5.5 to -6 d13C (PDB) is the mantle carbon isotopic value. The modern world has Corg about -28.5 and Cinorg.CO3 of 0. We see mantle carbon values in Neoproterozoic cap carbonates - implies zero Corg burial - the source of the suggestion of completely ice-covered oceans in the Hoffman Snowball Earth Hypothesis (SEH). d13C curves go crazy during the Neoproterozoic, compare with pre-NP and post-NP.
Snowball 101: the basic ideas of SEH.
1) Freezing phase - the entire ocean surface is frozen (even in the tropics) from a runaway albedo feedback. When sea ice reaches 30˚ latitude, the rest rapidly freezes over. Then, primary productivity in the oceans ceases, which accounts for mantle carbon values in cap carbonates. Then, atmospheric CO2 increases to ~120,000 ppm due to the shutdown of hydrologic cycle and the shutdown of silicate weathering (both sinks for CO2). The increasing CO2 is coming from the mantle (continued volcanism from continuing plate tectonic processes).
2) Melting phase - catastrophic melting phase from greenhouse effect, on a scale of 100s of years - very rapid and very warm. This renews silicate weathering, resulting in drawdown of atmospheric CO2, which delivers alkalinity and base cations to the ocean. The cap carbonates record the transfer of excess atmospheric CO2 to the ocean. The trend of decreasing carbon isotope depletion upward in cap carbonates is due to: 1) protracted shutdown of marine autotrophic activity; 2) high fractional burial of carbonate carbon; and 3) Rayleigh distillation.
This is the freeze-fry cycle of the SEH (Hoffman, 2000). This freeze-fry cycle is on the order of 10 million years. The SEH is advocated by Hoffman and others because it plausibly explains many paradoxes in the record.
How to test SEH?
Global climate model simulations - the catastrophic freeze over starting at 30˚ latitude is actually the result of an artifact in a previous climate model.
Look at carbon isotopes in synglacial carbonates (look at non-eroded carbonates only, though).
Sr isotopic information - a measure of weathering influx.
Implications to evolution - what are the evolutionary responses and what about bottlenecks?
Continental locations - equatorially-located continents are “surprised” in the SEH model - the ice creeps toward them from the poles, then reaching 30˚ latitude, and quickly covers the remainder of the planet. Are the continents really all equatorially-located, though, at the required times?
Well, lots of modeling is going on now - each has different purposes, and it is difficult to compare them all. But, can you freeze over the whole ocean at all in climate models? Maybe, but very hard to do. Not a chance, say lots of people.
Models start off with CO2 at 315 ppm (corresponding with pre-industrial levels). In the Marinoan world, many continents are at low latitudes (including Australia). The models can’t get an ice sheet remotely close to Australia. What about changing solar luminosity? In the Neoproterozoic, sunlight intensity was less than now, estimated to ~94% of modern intensity. This value is based on astrophysicists’ estimates of a 30% fainter Sun at 4.5 by ago, and scaled forward to the late Neoproterozoic. No astrophysicist has yet argued for changes in the rate of solar luminosity increase. So, change models to 94% of modern luminosity. Still can’t get ice at low latitude Australia in the Neoproterozoic. Add a lower CO2 value of 40 ppm to the model, can get oceanic ice sheets close to the tropics, but still don’t have an all frozen ocean - not close.
Look at synglacial carbonates - we now have data about this from 4 continents. Look at marine cements, peloids, oolites (in-situ carbonates in glacial interval). The values are scattered, but typically they are positive, +2 or +3, not -5 as SEH says (supposed to be close to mantle carbon values, since all Earth is frozen, and only carbon input is from mantle, through plate tectonic-driven volcanism).
Strontium ratios are ~invariant throughout the glacial to post-glacial interval. No evidence from Sr for a 1000 fold variation in weathering rates (as expected in Hoffman’s SEH). The maximum weathering rate for an atmosphere with 120,000 ppm of CO2 is <50x present.
Secondary hypotheses have been proposed to circumvent the Sr problem. They assert a substantially longer time scale, and they call upon weathering of carbonates, so not weathering silicates. There are difficulties with these. The longer timescale argument is inconsistent with high fractional burial of carbonate carbon, plus the cap carbonates likely represent ~104 years, plus carbonate weathering cannot draw down CO2. Silicate weathering can, but carbonate weathering can’t. So, cap carbonates are not a product of CO2 drawdown, which is a key Hoffman idea. Carbonate weathering drives cap carbonate d13C values into a positive upward trend.
The Pleistocene glacial maximum had sea level at 120 meters lower than now. There is still 73 meters worth of sea level trapped in ice.
Australian glacial facies and magnetic reversals indicate a glacial retreat over 105 to 106 years (much longer than Hoffman’s SEH says).
Glacial-eustatic rise continued after glacial-isostatic rebound at low latitudes (cap carbonates are late glacial - deposited when there was still ice somewhere on Earth).
Australia has 30 meter thick cap carbonates (event beds - turbidites).
Alternative Hypotheses for the Origin of Cap Carbonates
1) Post-glacial upwelling (Kaufman et al., 1991). But, it is difficult to stratify a glacial ocean, and upwelling of nutrient-rich water leads to enhanced productivity.
2) Gas Hydrate Hypothesis (GHH) - accounts for the source of light carbon seen in cap carbonates. Gas hydrates are ice that have methane in the lattice. They are common in continental shelves - buried frozen permafrost. Cap carbonates represent destabilization of permafrost methane hydrates during post glacial flooding of continental margins and interior basins. Temperature increases, and methane is delivered into the water column. Need to have a methane source from a reservoir in shallower areas (permafrost reservoir, for example). Deeper marine reservoirs of methane aren’t suitable sources for this because a sea level rise will increase pressure and will stabilize the deep ocean methane reservoirs.
Evidence for GHH? Widespread features in cap carbonates (CC) are consistent with cold seep facies. See bedding expansion and cementation in deformed beds below CC in Australia. See sheet cracks in CCs - cement is growing into empty spaces. See tubes in CCs - gas escape tubes? They are real tubes - can see sediment falling into them. See roll-up structures in sub-photic zone biohermal communities in CCs. See barite and aragonite fans in CCs - cold seep facies features.
So, GHH calls upon a pulse addition of light carbon from permafrost methane followed by steady state recovery. CCs represent ~104 years (rapid accumulation).
Mass balance calculations show the plausibility of GHH, using modern ocean constraints. Carbon introduced by methane release that accounts for a -5‰ shift - need about 3 x 1017 moles of CH4. Carbon buried in cap carbonates - estimated 8 x 1017 moles of CaCO3. Same order of magnitude.
How big were the ice sheets in the Neoproterozoic? There is good evidence in Utah - see incised valley systems in a series of sections - rivers cut into marine sediments (~160 meters of cutting) - they are the same age as the glacial interval.
Apparent sea level change of 193 meters for the Pleistocene Ice Age (120 m + 73 m) - equivalent to a eustatic change of 130 m (need to multiply 193 m by a factor of 1.4 - this accounts for difference of sea level change values seen from islands, which experience water loading with sea level rise, compared with sea level changes that will be seen on the continents).
How many Neoproterozoic ice ages? All sections show only 2. Multiple events or not? Multiple events seem to be based on miscorrelation of the 2 ice ages, especially in southern to northern Namibia. Two main events, apparently. One is at 600 m.y., and one is at ~750 m.y.. All workers agree that they were the most severe ice ages in history. Not >2 events.
SEH is not consistent with C and Sr data together, or with eucaryotic evolution (some call on a soft snowball, versus Hoffman’s hard snowball, or a slushball hypothesis), or with the scale of ice-volume changes, or with the duration of glacial retreat (>105 years, not 102 years).
GHH is consistent with climate models, outcrop evidence (including strange features in CCs that have been unexplained for a long time), isotopic data, biological evidence. Permafrost and associated hydrates should have been unusually widespread, because lots of glaciation occurred in the Neoproterozoic, which is accepted by all.
The upwelling hypothesis doesn’t work. SEH doesn’t work. Any other hypotheses out there? They are welcome.
Call any of these ideas “Snowball Earth” if you want - the name doesn’t matter.
Methane is coming from marine sedimentary basins.
There was lots of Precambrian organic carbon, including Precambrian oils and Corg isotopic values that hover around zero back to the Archean.