Snowball or slushball Earth

In 1964, Brian Harland noticed that Proterozoic glacial deposits occur on nearly all continents and proposed a great Neoproterozoic ice age. At the time Harland wrote, the position of Neoproterozoic continents was very uncertain, and it was a possibility that the continents experienced glaciation at different times as they drifted closer to the poles.

140 135 130 125 120 115 110 105 100 150 -i-

Time (thousands of years from present)

Time (thousands of years from present)

Figure 4.3 Difference in the strength of the solar beam at mid-month dates of Berger (1978) paired by equal geometry of the incoming solar beam. (a) 24,000 years ago to 12,000 years in the future. (b) 140,000 to 103,000 years ago. Notice the switches from autumn to spring modes. Source: Reprinted from Global and Planetary Change 40(1-2), G. Kukla and J. Gavin, Milankovitch climate reinforcements, 27-48. Copyright © 2004, with permission from Elsevier.

However, he did find glacial deposits associated within marine sedimentary rocks characteristic of low latitudes. The intriguing possibility of tropical glaciation at sea-level begged many questions, including the crucial one of whether it was feasible to freeze the entire planet. Interestingly, experiments with a simple two-dimensional energy-balance climate model showed that if the Earth's climate were to cool, and ice were to form at progressively lower latitudes, the planetary albedo would rise at an accelerating rate because there is more surface area per degree of latitude on approaching the Equator. The model pre dicted that once ice formed beyond a critical latitude (around 30 degrees north or south, which is equivalent to half the Earth's surface area), the positive feedback was so strong that surface temperatures nose-dived, freezing the entire planet (Budyko 1969). The ocean bottoms stayed unfrozen, owing to heat leaking from the Earth's interior. However, 1-km-thick layer sea ice formed, which was thickest at the poles and thinnest at the Equator.

At the time these results appeared, the idea that the Earth had experienced a global glaciation was regarded at interesting but academic. Such a catastrophe would have presumably killed all life and caused temperatures to plummet so low that the ice-cover would last forever, so few researchers believed that it had actually happened. However, the discovery in the late 1970s of organisms living in deep-sea hydrothermal (hot water) vents, and later in the extremely cold and dry mountain valleys of East Antarctica, raised the possibility that some of organisms might be able to survive a global glaciation. Later, it became evident that plate tectonic processes might be capable of reversing the ice-albedo feedback mechanism that sustained the frigid global climate.

Joe Kirschvink (1992) coined the term 'snowball Earth' to describe a global glaciation and showed that, even under full glacial conditions, volcanism associated with plate tectonic processes would still supply carbon dioxide to the atmosphere and oceans. However, if the Earth were so cold that there was no liquid water on the continents, weathering reactions would effectively cease, allowing carbon dioxide to build up to incredibly high levels. Eventually, the carbon dioxide-induced warming would offset the ice albedo, and the glaciation would end. Given that solar luminosity 600-700 million years ago was about 6 per cent lower than today, a carbon dioxide concentration about 350 times the present concentration would have sufficed to overcome the albedo of a snowball Earth (Caldeira and Kasting 1992). Using current rates of volcanic carbon dioxide emissions as a guide, the Neoproterozoic snowball Earth would have lasted for millions to tens of million of years before the sea ice would begin to melt at the Equator. It is now known that at least two Proterozoic glaciations probably took the form of 'snowballs' - the Sturtian (about 710 million years ago) and the Marinoan (about 635 million years ago) - each of them global and each enduring for at least several million years.

Paul Hoffman and his co-workers were instrumental in firming up the snowball Earth hypothesis (e.g. Hoffman et al. 1998a, 1998b; Hoffman and Maloof 1999; Hoffman and Schrag 2000, 2002). Figure 4.4 summarizes an interpretation of the basic sequence of events going into, through, and out of a snowball climate. Two emerging lines of evidence strongly supported the snowball Earth hypothesis. The first was a dolostone cap, suggestive of high surface temperatures, sitting on Neoproterozoic glacial deposits across Australia (and later found to occur world-wide). The transition from glacial deposits to capping dolostone is sharp and shows no significant interruption, the indication being that Neoproterozoic glacial epochs ended with abrupt climatic warmings. A second line of evidence was an unusual pattern of variation in the ratio of carbon-13 and carbon-12 in the cap carbonates. The carbonate rocks beneath the glacial deposits in northern Namibia record a remarkable carbon isotope excursion, showing carbon-13 enrichment by as much as 1.5 per cent relative to volcanoes, and much more than in modern carbonate sediments. These enriched sediments represent at least 10 million years. The implication is that buried organic carbon in the Neoproterozoic accounted for nearly half the total carbon removed from the ocean. However, just before the deposition of the glacial deposits, the amount of carbon-13 falls steeply to levels equivalent to the volcanic source, stays at those levels through the deposition of the cap carbonates atop the glacial deposits, and then slowly rebounds to higher levels several hundred metres above. Such a rapid excursion in the

(a) Incipient glaciation

(a) Incipient glaciation

see co, ~i.o

(b) Runaway ice-albedo

(b) Runaway ice-albedo

(c) End of snowball

(c) End of snowball

(d) Hothouse aftermath

(d) Hothouse aftermath

Figure 4.4 Cartoon of a complete 'snowball' episode, showing variation in planetary albedo, atmospheric carbon dioxide levels, surface temperature, tropospheric depth, precipitation, glacial extent, and sea-ice thickness. Source: Reprinted by permission from P. F. Hoffman and D. P. Schrag (2002) The snowball Earth hypothesis: testing the limits of global change. Terra Nova 14, 129-55.

Figure 4.4 Cartoon of a complete 'snowball' episode, showing variation in planetary albedo, atmospheric carbon dioxide levels, surface temperature, tropospheric depth, precipitation, glacial extent, and sea-ice thickness. Source: Reprinted by permission from P. F. Hoffman and D. P. Schrag (2002) The snowball Earth hypothesis: testing the limits of global change. Terra Nova 14, 129-55.

carbon-13 content of calcium carbonate must represent an abrupt dive in the fraction of carbon leaving the ocean as organic matter that lasts long enough to change the composition of the entire reservoir of dissolved inorganic carbon in seawater. It probably represents a drop in biological productivity as ice formed over the oceans at high latitudes, and the Earth 'teetered on the edge of a runaway ice-albedo feedback' (Hoffman and Schrag 2000).

The snowball Earth hypothesis poses two key questions: How could Earth enter the snowball state and, once there, how could it escape it? Raymond T. Pierrehumbert (2002) thoughtfully reviewed efforts to answer these posers. He saw answers hinging upon the behaviour of the pole-to-equator temperature gradient (or equivalently, the meridional heat transport carried by the atmosphere and oceans). Conditions for entering a snowball state appear to be anomalously low carbon dioxide concentrations, coupled with Sun shining 6 per cent less brightly than today in the Neoproterozoic. Several modellers used general circulation models coupled to a mixed-layer ocean (but without ocean dynamics) in an attempt to trigger a snowball climate (Jenkins and Smith 1999; Chandler and Sohl 2000; Hyde et al. 2000). Some simulations produced a global glaciation when carbon dioxide concentrations were down to 100 ppm, but some did not. The results painted a somewhat messy picture of the controlling factors. This may be owing to the use of specified ocean heat transports estimated from the present ocean in the climate models. To be sure, 'if ocean heat transports are eliminated altogether, the Earth falls easily into a snowball state at low CO2' (Pierrehumbert 2002, 193). One study found that dynamic ocean heat transport was an effective inhibitor of glaciation (Poulsen et al. 2001), but this does not rule out the possibility of global glaciation, in view of 'the relatively short simulations used, the highly idealized sea-ice model, and myriad other uncertainties plaguing ocean simulation' (Pierrehumbert 2002, 193).

Fewer researchers have studied the mechanisms for exiting the snowball state, which would involve a truly spectacular deglaciation. The fate of the snowball hypothesis plainly depends upon the explanation of deglaciation, since, if the Earth did indeed enter a snowball state, then it indubitably escaped it. The most obvious exit strategy relies on greatly elevated carbon dioxide concentrations, and the cap dolostones seem to support this explanation. The thickness of cap carbonates, which gives an indication of the amount of carbon dioxide degassed during the glaciated period, would be consistent with deglaciation at carbon dioxide partial pressures of about 0.2 bar. Such an ultra-high carbon dioxide atmosphere would have raised temperatures to the melting point at the Equator. The scenario envisaged runs as follows (Hoffman and Schrag 2000). Once melting had begun, the ice-albedo feedback would have reversed and combined with the extreme greenhouse atmosphere to drive surface temperatures upwards. The warming would have proceeded rapidly because the change in albedo would have begun in the tropics, where insolation and surface area are maximal. Evaporation would have restarted, adding water vapour to the atmosphere and contributing powerfully to the greenhouse effect. Tropical sea-surface temperatures would have become very warm in the aftermath of a snowball Earth, powering an intense hydrologic cycle. Sea ice hundreds of metres thick globally would have disappeared within a few centuries. Intense chemical weathering of silicate rocks and dissolution of carbonate rocks would have resulted from the strong hydrologic cycle, the low pH of carbonic acid rain, and the large surface area of frost-shattered rock and rock flour produced by the grinding action of glaciers. Rivers would have carried the products of chemical weathering reactions - cations and bicarbonate - to the ocean, where they would have neutralized the acidity of the surface waters and driven massive precipitation of inorganic carbonate sediment in the rapidly warming surface ocean. Characteristically, cap dolostones pass upwards into much thicker, deeper-water clays or limestones, perhaps reflecting a rise in sea level as continental ice sheets melted, and suggesting that the cap carbonates precipitated extremely rapidly, perhaps in only a few hundred years. Textures in the dolostones and limestones, such as gas-escape tubes and crystal fans consistent with precipitation from seawater highly supersaturated in calcium carbonate, support this idea. Cap dolostones are no paradox; they are the expected consequence of the ultra-greenhouse conditions unique to the transient aftermath of a snowball Earth.

It is possible that much higher carbon dioxide concentrations set deglaciation in motion, but the radiative properties of water vapour and clouds enter crucially in both the glaciation and deglaciation problems, and the failure to consider them may present a severe impediment to solving the snowball problem (Pierrehumbert 2002). In an attempt to evaluate the role of clouds and water vapour, Pierrehumbert (2002) used an idealized model, the results of which were 'regime diagrams' that show the effects of cloud cover and clear-sky relative humidity on the initiation and termination of a snowball state (Figure 4.5). These diagrams show where the snowball state starts (at 100 ppm carbon dioxide) and where the state ends (at a carbon dioxide partial pressure of 0.2 bar) as a function of cloud fraction and clear-sky relative humidity. When the outgoing long-wave radiation from clouds (OLRcloud) is 120 W/m2, the snowball state initiates with sufficient cloud cover, with less cloud cover needed at low relative humidity (note that, in these calculations, relative humidity refers to the humidity of the clear-sky areas between clouds). However, no combination of cloud cover and humidity allows deglaciation. When OLRcloud is 90 W/m2, the clouds have greater warming effect and deglaciation can occur. However, under these conditions, the Earth will not enter the snowball state unless humidity and cloud cover are very low. When OLRcloud is 100 W/m2, a tiny sliver of parameter space permits both the initiation and termination of a snowball state, but initiation requires dry, clear-sky conditions. The results are consistent with general circulation model studies that show (at least in the absence of dynamic ocean heat fluxes) it is possible to start up a snowball state even with cloud feedbacks, and that cloud feedback can in fact assist the initiation. Pierrehumbert's work perhaps helps to explain why general circulation model predictions differ so much in their initiation behaviour, given that changing cloud radiative properties by less than 30 W/m2 can completely alter the climate regime outcome (Pierrehumbert 2002). Later work by Pierrehumbert (2004, 2005) used a general circulation model with elevated carbon dioxide levels to estimate the deglaciation threshold. His simulations included several supposedly significant features of a snowball Earth not considered in previous studies with less sophisticated models - a reduction of vertical temperature gradients in winter, a reduction in summer tropopause height, the effect of snow cover and a reduction in cloud greenhouse effects. The model was 30°C short of deglaciation with atmospheric carbon dioxide concentrations as high as 550 times the present levels (0.2 bar of carbon dioxide). In consequence, deglaciation of a totally frozen world (hard snowball state) seems unlikely, even at very high carbon dioxide levels, unless unknown feedback cycles not included in the model come into play.

Another putative cause of Proterozoic glaciations is high obliquity. The Earth's axial tilt may have stood at more than 60 degrees, making the tropics colder than the poles, for 4 billion years following the lunar-forming impact (e.g. Williams 1975b). However, the high-obliquity hypothesis fails to account for several major features of the Neoproterozoic glacial record. It cannot explain the abrupt beginnings and ends of discrete glacial events and their close association with large negative shifts in carbon-isotope rations, nor with the dep

(a) OLRdoud = 90 W/m2 (fa) OLRdoud = 100 W/m2 (c) OLRcl0lld = 120 W/m2

(a) OLRdoud = 90 W/m2 (fa) OLRdoud = 100 W/m2 (c) OLRcl0lld = 120 W/m2

Figure 4.5 Regime diagrams showing the effects of cloud cover and clear-sky relative humidity on the initiation and termination of the snowball Earth state. All calculations used Neoproterozoic insolation and an ice albedo of 0.7. Source: Reprinted by permission from Macmillan Publishers Ltd: Nature 419 (Pierrehumbert 2002), copyright © 2002.

■ Initiates and I I Stable polar ice deglaciates I-1 or ice-free state

Figure 4.5 Regime diagrams showing the effects of cloud cover and clear-sky relative humidity on the initiation and termination of the snowball Earth state. All calculations used Neoproterozoic insolation and an ice albedo of 0.7. Source: Reprinted by permission from Macmillan Publishers Ltd: Nature 419 (Pierrehumbert 2002), copyright © 2002.

osition of strange, world-wide carbonate layers (cap carbonates) during postglacial sea-level rise, and the return of large sedimentary iron formations, after a 1.1 billion year hiatus, exclusively during glacial events (Hoffman and Schrag 2002). A snowball event, on the other hand, should start and finish rapidly, particularly at lower latitudes; it should last for millions of years, because outgassing must build up an intense greenhouse in order to overcome the ice albedo. While the ocean is largely ice-covered, it should become anoxic and reduced iron should form. The soluble iron would be transported everywhere and precipitated as iron formations wherever oxygenic photosynthesis occurred, or upon deglaciation. Moreover, the intense greenhouse would guarantee a short-lived postglacial regime of enhanced carbonate and silicate weathering. This regime would impel a flux of alkalinity capable of accounting for the world-wide incidence of cap carbonates (see also Shields 2005). The resulting high rates of carbonate sedimentation, coupled with the kinetic isotope effect of transferring the carbon dioxide burden to the ocean, should push down the carbon-isotope ratio of seawater, as observed in the sedimentary record (Hoffman and Schrag 2002).

The snowball hypothesis has its critics. A source of disagreement is the extremity of the freezing. Some claim that thick ice, including the oceans, covered the entire world, with life surviving in small warm spots where sunlight penetrates the icy cover - a so-called 'hard snowball' planet. Such an icy world would change rapidly in a greenhouse environment. Others favour a 'slushball hypothesis' that has large areas of thin ice or even open ocean, mainly in the tropics; this not-quite-so icy world would experience a slower deglaciation (Hyde et al. 2000; Crowley et al. 2001). The open waters would have acted as a refuge for multicellular animals.

Some observational and computer modelling work since 2000 supports the snowball hypothesis. On the observational front, the base of cap carbonates in cores from Zambia and the Democratic Republic of Congo have yielded iridium, a constituent of cosmic dust. The iridium occurs in such quantities at the end of the Sturtian and Marinoan glaciations that ice must have covered much of the planet for at least 3 million years and perhaps as long as 12 million years (Bodiselitsch et al. 2005). The cosmic dust would have accumulated on and in the global ice cover and then precipitated during the rapid melt associated with deglaciation. The iridium could have come from volcanoes, but the attendant elements in the dust are indicative of a meteoritic origin. On the modelling front, climate simulations models have helped to clarify the role of different factors in producing a snowball planet. A key ingredient in making a snowball glaciation appears to be uncharacteristically low carbon dioxide concentrations. GEOCLIM, a coupled climate-geochemical model, helped to assess the impact of change in palaeogeography preceding the Sturtian glaciation (Donnadieu et al. 2004). The simulation showed that the breakup of the supercontinent Rodinia (see p. 18) increased runoff and consequently the consumption of carbon dioxide through continental weathering. The result was a 1,320 parts per million drop of atmospheric carbon dioxide concentrations, which could have led to a progressive transition from a greenhouse climate to an icehouse climate during the Neoproterozoic. The model produced a snowball glaciation when the tectonic changes were combined with the effects of weathering the extensive basalt traps formed during the breakup of the supercontinent (see also Godderis et al. 2003). Another simulation study managed to produce a slushball glaciation, with a substantial area of ice-free ocean in the tropics (Figure 4.6) (Crowley et al. 2001).

An exciting line of enquiry considers the role that methane might have played in Earth's earliest climates and in aiding the onset of snowball states (Kasting and Siefert 2002; Pavlov et al. 2000, 2003). James F. Kasting (2004) explained that the young Earth managed to

Ice thickness

Ice thickness

' 4.6 An example of an open water solution to a near-snowball Earth model. (a) Ice sheets (thickness in meters). (b) Mean annual temperature (°C). Source: After Crowley et al. (2001).

avoid entering an icehouse state for the first 2.3 billion years of its history, despite the Sun's burning some 30 per cent less brightly than it does today, because the atmosphere contained relatively high levels of methane (a highly effective greenhouse gas) produced by methanogens. However, when atmospheric oxygen produced by cyanobacteria started to increase rapidly about 2.3 billion years ago, methane levels fell, perhaps making conditions favourable for the onset of the Huronian glaciation. It is possible that a second rise in oxygen levels in the Neoproterozoic encouraged the snowball glaciations by decreasing yet again atmospheric methane.

Was this article helpful?

0 0

Post a comment