Snowballs, Ice Ages and CO2

Earth's climate history includes numerous incidents of rapid warming and cooling. While Pleistocene ice-age glacial terminations are arguably the most dramatic recent examples of sudden climate change, during the last glacial period the climate of the Northern Hemisphere experienced several other significant episodes when the climate rapidly warmed. Scientists call these episodes Dansgaard-Oeschger (DO) events after the Danish and Swiss researchers who documented them using ice-core studies. These rapid oscillations are marked by rapid warming, followed by slower cooling. The most prominent coolings are associated with massive iceberg discharge into the North Atlantic Ocean known as Heinrich events (HE). The melting icebergs add large volumes of cold fresh water to the ocean, disrupting circulation patterns and causing further climate changes. Scientists look to past events like these to help us understand how Earth's climate system functions—what causes our planet to cool or suddenly warm. Recently, new data on past climate changes have led one commentator to predict the end of winter skiing in the American Southwest.

In an attempt to better understand the impact of such events on the American Southwest, Asmerom et al. examined oxygen isotopic data from a well-dated stalagmite recovered from central New Mexico. It seems that speleothems are providing scientists a wealth of new paleoclimate data these days. The data presented in this report, “Variable winter moisture in the southwestern United States linked to rapid glacial climate shifts,” utilizes a new analysis technique, different from the one used to identify ancient sea-level changes I reported on earlier (see “Ancient Sea-levels Rewrite Ice Age Transitions”). Here researchers were trying to fill in a gap in our knowledge regarding precipitation changes in days gone by. As they explain in the article:

Climate in the Northern Pacific basin is partly modulated by the El Niño–Southern Oscillation on interannual timescales and the Pacific Decadal Oscillation on decadal timescales. On centennial to millennia timescales, it has been shown that the East Asian monsoon (EAM) responds to Northern Hemisphere climate-forcing through modulation of the intertropical convergence zone (ITCZ). Over longer, orbital timescales, it has been suggested that the high stands of lakes in western North America during the Last Glacial Maximum were associated with the southward shift of the polar jet stream. The effect of centennial- to millennial-scale Northern Hemisphere climate modulation is less clear, in part owing to the lack of high-resolution proxies that can be absolutely dated.

This lack of data has been particularly true for arid regions, but the authors used speleothems and new multicollector inductively coupled plasma mass spectrometric techniques to analyze U-series isotopes, providing a new look at the climate history of the American Southwest. The ratio of oxygen isotopes in the calcite that makes up the speleothems from the sample site is a proxy for the relative amounts of summer and winter precipitation in the past. The researchers obtained 68 high-precision uranium-series dates using mass spectrometry with typical age uncertainties of less than 1%, providing a way to precisely date the precipitation proxy data from the oxygen isotopes.

The cave, typical of other parts of the southwestern United States, has two rainy seasons, consisting of summer North American monsoon rainfall, derived from the Gulf of Mexico and Pacific-derived winter precipitation. Asmerom et al.

These techniques showed that the stalagmite they collected from approximately 1 km into the Fort Stanton Cave in central New Mexico had grown continuously from 55.9 to 11.4 thousand years ago. Along with the date measurements, they took 1209 δ18O measurements. The authors found a close match between precipitation changes and the record of DO events, which they interpret as a result of a shift of the polar jet stream and intertropical convergence zone to the north during warm periods. This change in turn causes a reduction in winter precipitation.

The theory is that during warming periods, the pole-to-Equator temperature gradient decreases, shifting the polar jet stream and the Northern Hemisphere summer ITCZ further north. The opposite changes would be expected during cold swings. Analysis of later Holocene climate and historical data show that years of severe droughts are primarily the result of deficit in winter precipitation. Because of all of these factors the authors state “global warming may result in profound changes in precipitation,” which prompted Science writer H. Jesse Smith to quip “Ski While You Can,” a reference to expected temperature increases due to global warming. But that is not the full story.

According to the authors: “The poleward shift during DO events occurred at a time when the earth was in a glacial state. An example of ‘extra’ warming during an interglacial, expressed as a +4 to +6 m rise in sea level relative to today, at the end of the last interglacial 125,000 yr ago has been reported.” As was seen from the Mallorcan sea-level study reported earlier on this blog, such higher sea-levels did occur during the Eemian interglacial and the DO events following it. This implies that the bouts of “extra” global warming can and do occur naturally. And when this paper's authors menacingly pronounce, “rapid DO-like warming due to greenhouse-gas forcing during the present interglacial stage could push SWNA into an even more arid phase, unseen since the early Holocene, or even go beyond this, into conditions not represented since 125,000 yr ago,” it is pure speculation.

By contrast, Heinrich events, named after German oceanographer Hartmut Heinrich, are relatively brief and occur on average every seven-thousand years during glacial periods. Abrupt shifts to warmer climate follow immediately after Heinrich events, the last event (HE1) marked the onset of the termination which ended the last glacial. The events seem to occur only when the ice volume is relatively large, the temperatures are relatively cold, and the sea level is 130-250 feet (40-60 meters) below present values. The cause of Heinrich events—internal ice dynamics or external climate change—is not fully understood but it is pretty safe to say we are not threatened by them during the current interglacial.

Atmospheric CO2 composition and climate during the last glacial period. Red numbers denote DO events. Ahn et al.

The data presented by Asmerom et al. in no way show a causal connection between CO2 or anthropogenic global warming and DO events. In fact, a connection to the latter could not exist, since the humans that were around back then had no factories or SUVs. The gratuitous comments at the end of this paper and the fatuous remark from Mr. Smith show how pervasive the AGW mindset has become among climate scientists. As I reported in “Modeling Ice Age's End Lessens Climate Change Worries,” it may not be possible to induce an abrupt onset warming under a gradual forcing.

Sizing Up The Snowball

The time before the Cambrian explosion of life, the Neoproterozoic, was an era of great environmental and biological change. Unfortunately, direct and precise knowledge about the age of strata from this time has prevented accurately linking the early development of complex life with changes in the Precambrian environment. In a study of rocks in northwestern Canada, Francis A. Macdonald et al. have linked large perturbations in the carbon cycle, a major diversification and depletion in the microfossil record, and the onset of the Sturtian glaciation. The researchers set the scene and motivation for their work at the begining of their paper “Calibrating the Cryogenian”:

Middle Neoproterozoic or Cryogenian strata [850 to 635 million years ago (Ma)] contain evidence for the breakup of the supercontinent Rodinia, widespread glaciation, high-amplitude fluctuations in geochemical proxy records, and the radiation of early eukaryotes; however, both relative and absolute age uncertainties have precluded a better understanding of the nature and interrelationships of these events. Several first-order questions remain: How many Neoproterozoic glaciations were there? How were they triggered? What was their duration and extent? How did the biosphere respond?Answers to all of these questions hinge on our ability to precisely correlate and calibrate data from disparate stratigraphic records around the world.

Some of the most dramatic climate events of this period are the so called “Snowball Earth” super ice ages. Snowball Earth refers to the hypothesis that Earth's surface became nearly or entirely frozen over at least once during three ice age periods between 650 and 750 million years ago. The snowball Earth hypothesis, originally put forth by W. Brian Harland, was developed in response to strong paleomagnetic evidence for low-latitude glaciation from the Elatina Formation in Australia.

In 1964, Harland published a paper in which he presented data showing that glacial tillites in Svalbard and Greenland were deposited at tropical latitudes. Since continents drift over time, it is difficult to figure out exactly where they were positioned on the globe at a given point in Earth's past. When sedimentary rocks form, magnetic minerals within them tend to align themselves with the Earth's magnetic field. Through the precise measurement of this palaeomagnetism, it is possible to estimate the latitude (but not the longitude) where the rock was deposited.

One of the things that may have contributed to the several periods of Neoproterozoic ice-house conditions was the configuration of the continents. During the time period in question the ancient supercontinent Rodinia was in the process of breaking up into a number of smaller land masses. It has been suggested that when there is a large land mass over either of the polar regions conditions become favorable to an ice age. One reconstruction of the geography of the late Precambrian is shown below.

Possible geography of late Precambrian Earth. Hyde et al.

The Sturtian snowball earth is linked to sedimentary deposits found on virtually every continent some evidence suggests that it lasted for millions of years. It is commonly called the “Sturtian,” after glacial sediments in South Australia described in 1908 by the geologist Walter Howchin. It is uncertain whether the Sturtian glacial epoch consisted of one discrete glaciation that lasted tens of millions of years, or multiple glacial episodes including the low-latitude glaciation at ~716.5 Ma. “Prior to this study, minimum and maximum age constraints on the Sturtian glaciation were provided by a sample from South China dated at 662.9 ± 4.3 Ma and the Leger Granite in Oman dated at 726 ± 1 Ma, respectively,” state the authors. They suggest that 717.43 ± 0.14 Ma is the maximum age constraint on the low-latitude glaciation.

Models suggest extremely rapid ice advance once ice extends below 30° latitude, implying that such glaciation of equatorial latitudes should be synchronous around the globe. According to the authors “we conclude that the Sturtian glaciation at ~716.5 Ma was global in nature.” Even so, there remain a number of unanswered questions, including whether the Earth was a full snowball or a “slushball” with a thin equatorial band of open water during these episodes.

In the 1960s Martin J. S. Rudwick, working with Brian Harland, proposed that the climate recovery following a huge Neoproterozoic glaciation paved the way for the explosive radiation of multi-cellular animal life during the following Cambrian period. In a 2000 article in Scientific American, Paul F. Hoffman and Daniel P. Schrag conjectured: “It has always been a mystery why it took so long for these primitive organisms to diversify into the 11 animal body plans that show up suddenly in the fossil record during the Cambrian explosion. A series of global freeze-fry events would have imposed an environmental filter on the evolution of life. All extant eukaryotes would thus stem from the survivors of the Neoproterozoic calamity.” Indeed, Hoffman (no relation) and colleagues had previously reported on evidence of this freeze then bake cycle:

Negative carbon isotope anomalies in carbonate rocks bracketing Neoproterozoic glacial deposits in Namibia, combined with estimates of thermal subsidence history, suggest that biological productivity in the surface ocean collapsed for millions of years. This collapse can be explained by a global glaciation (that is, a snowball Earth), which ended abruptly when subaerial volcanic outgassing raised atmospheric carbon dioxide to about 350 times the modern level. The rapid termination would have resulted in a warming of the snowball Earth to extreme greenhouse conditions. The transfer of atmospheric carbon dioxide to the ocean would result in the rapid precipitation of calcium carbonate in warm surface waters, producing the cap carbonate rocks observed globally.

Could complex life on Earth owe its existence to the evolutionary pressure exerted by these global freeze overs? According to Macdonald et al., “It is clear that a diverse biosphere persisted through the Neoproterozoic glaciations, but the impact of global glaciation on eukaryotic evolution remains unresolved.” Though Earth has not frozen solid since the advent of truly complex life, ice ages have occurred throughout the Phanerozoic.

A modern Snowball Earth. Image by Neethis.

A geographic configuration similar to the Middle Neoproterozoic existed during the Permian, when the continents once again had clustered together to form another supercontinent—Pangaea. During this period the great Karoo Ice Age occurred, ending around 260 mya. This was followed by the worst of all the mass extinction events, the Permian-Triassic at 251 mya. Following that catastrophe life almost had to begin anew and Earth soon entered the age of the dinosaurs. And of course, H. sapiens bounced back from the edge of extinction and exploded across the globe following the last deglaciation only 14,000 years ago.

What impact these new findings will have on thinking regarding our current ice age remains to be seen. Considering the geographic dispersal of the continental land masses today it is unlikely that Earth will be cast into another snowball earth episode any time soon. There as been, however, another new report in Nature Geoscience that casts doubt on at least one theory about CO2 and the end of the glacial period just prior to the current Holocene warm period.

Goin' Through Them Changes

Over the past 500,000 years, each of the five identified glacial periods ended abruptly. Rapid warming caused continental ice-sheets retreated in roughly one-tenth of the time it took for the Earth's climate to cool and for ice sheets to reach to their maximum extent. Ice-core records from Greenland and Antarctica show that atmospheric CO2 levels rise by ~80 ppm during such deglaciations, leading scientists to seek a source for the carbon dioxide. According to current thinking the Southern Ocean as the main area of exchange with reservoirs of deep old water, which then makes its way to lower-latitude waters. Ricardo De Pol-Holz et al. investigated the hypotheses that an injection of carbon dioxide with low radiocarbon activity from an oceanic abyssal reservoir was the source of the increase (see “No signature of abyssal carbon in intermediate waters off Chile during deglaciation”).

“The fundamentals behind the atmospheric CO2 increase and Δ14C decrease during the so-called mystery interval (17.5–14.5 kyr BP) and the contemporaneous deglaciation have remained elusive,” state De Pol-Holz et al. Ice-core records show that the overall increase in atmospheric CO2 was ~100 ppm (parts per million) during the deglaciation. At the same time, the concentration of radioactive carbon 14 (14C) decreased without any record of a decrease in cosmic ray activity that would explain the drop. The proposed source for the upsurge in depleted CO2 is an upwelling of old water from the deep.

Supposedly, the deep ocean water had remained isolated from the atmosphere for several thousand years during the last ice age, becoming progressively depleted in 14C because of radioactive decay. Previous work by Marchitto et al. had documented two pulses of extremely depleted water coinciding with the mystery interval and the Younger Dryas (11,500–12,900 years ago). This resulted in changes to the radiocarbon content of Antarctic Intermediate Water (AAIW) because of mixing with old abyssal water south of the subantarctic frontal zone. To test for the presence of such water a number of cores were taken from ocean bottom sediment. What they found was unexpected—the upwelling of deep, carbon 14 deplete water could not have come from the depths off Chile.

Map of ocean salinity at 1 km depth showing the location of cores. In the figure above, intermediate-depth low-salinity signatures of AAIW and North Pacific Intermediate Water are shown in purple–pink. De Pol-Holz et al.

“Increasing evidence from the Atlantic, both in observations and in models, and from the West Pacific indicates a re-invigoration of AAIW production during the deglaciation,” the researchers concluded, “We therefore believe that the idea of a deglacial AAIW being formed elsewhere and routed in a completely different way so as to not affect the intermediate waters off Chile is unlikely.” What does this mean for science's understanding of ocean water movement during a deglaciation? It means that we do not really understand the conditions that prevailed during the “mystery interval” more than 14 thousand years ago. Unless an alternative formation mechanism is found for the young intermediate waters observed in the record, the upwelling intensity proxy for the Southern Ocean may need further revision. The paper concludes this way:

The increasing number of sedimentary records showing anomalous low radiocarbon at intermediate depths around the world demands further investigation on its causes. Our work shows that their connection with the Southern Ocean overturning strength is not entirely straightforward and alternative explanations for their presence should be explored. Understanding glacial–interglacial CO2 cycles still remains an elusive test for our proficiency in the study of the Earth as a system.

Science continues to refine the knowledge we have of Earth's climate, whether it be from 14,000 years ago or half a billion, new discoveries are constantly being made. Despite centuries of study, we still do not know what makes conditions right for Ice Ages to come and go: some special combination of the size and shapes of the continents; the circulation patterns of ocean waters and the atmosphere; the wobbling, subtly changing path of Earth around its star; and changes in the Sun's output may all have to come into convergence to trigger radical climate change. Meanwhile, the link between radical climate change and CO2 remains fuzzy at best. What we do know is that to rebound from a Snowball Earth took CO2 levels 350 times those of today. Those who claim that human activity will trigger a sudden warming or a sudden cooling are just blowing hot air.

As for the threat to skiing in the American Southwest, consider what has happened this year. One of the things that happens during an El Niño is an increase in snowfall across the southern Rockies and Sierra Nevada mountain ranges. Many have speculated that a warming climate could enter a phase of nearly continuous El Niño. Though this sounds like good news for snow skiers, not bad, I wouldn't count on it either way. After all, the climate change prognosticators haven't been right yet.

Be safe, enjoy the interglacial and stay skeptical.

Not likely to go away anytime soon.

Cooling first

The most immediate impact of volcanic eruptions has been observed to be cooling due to the ejection of particulates and sulfur compounds into the upper atmosphere. The 1980 eruption of Mt. St. Helens lowered global temperatures by 0.1°C, however, the much smaller eruption of El Chichon in Mexico lowered global temperatures three to five times as much. Although Mt. St. Helens emitted a greater amount of ash into the stratosphere, the El Chichon eruption emitted a much greater volume of sulfur-rich gases. The amount of sulfur-rich gases appears to be more important than ash when it comes to cooling. It can take several years for the ash and sulfur compounds to be removed from the atmosphere.

This initial cooling effect can be followed by enhanced warming caused by the emission of CO2. However, a far greater amount of CO2 is released into the atmosphere by human activities each year. Here I'm talking about the familiar explosive type of volcanic eruption that spews billowing clouds of ash into the air and forms cone shaped mountains. In the past, much larger eruptions of a different kind have occurred: the Siberian Traps, the Deccan Traps, and the CAMP among them. These are basaltic flow type eruptions where lava oozes out of fissures, covering vast areas often to depths of a kilometer or more. These types of eruptions are light on the ash and heavy on the CO2, meaning they are apt to cause enhanced warming with no significant cooling prelude.

I see no suggestion in the literature that volcanic ash coating Earth's surface causes a significant change in the planet's overall albedo. Ash falling on the ground is fairly quickly subsumed by returning plant life and ash falling on the ocean would sink, having little effect. The only place it might cause noticeable change is if it fell on glacial ice, where I would expect the cycling of the seasons and precipitation to remove or cover it in a year or so. Any affect would be short lived and localized. If you can find references that indicate otherwise please let us know.

Continents at the Poles Build Up Glaciers

The Snowball periods (4 of them at least) are easily explained by Continental Drift. When a significant fraction of the continents are locked together over one of the poles, the glaciers continue to build up until most of the landsurface is covered in 5 km to 0.5 kim high glaciers.

The Earth's Albedo will increase from 0.3 today to something around 0.45 in a continental configuration like the above. Temperature will decline by up 25C as a result of all that sunlight being reflected back into space.

Snowball ends when continental drift breaks up the polar supercontinent and the continents move off the poles toward 50S or so and the ice melts. Albedo declines and the Earth warms up again. This is the exact scenario that the continental drift reconstructions show happened.

The last snowball ended about 635 million years ago. It took another 50 to 55 million years before Cambrian explosion of complex life happened so the warm-up from the snowball was slow.

Look at Gondwana's continental drift to explain the Ordovician/Silurian and the Carboniferous ice ages.

Not the whole answer

The first supercontinent cycle may have taken place as much as 2.8 billion years ago (Ga). Since then there have been a number of supercontinents from the amalgamation and dispersal of a possible Neoarchaean supercontinent to the formation of the supercontinent Nuna, 1.9–1.8 Ga. This supercontinent cycle, having started so long ago and progressing so slowly in human terms, is currently lacking in historical detail. More recent supercontinents include Rodina, which began forming ~1.2 Ga with breakup starting ~750 Ma. Rodinia dominated the Earth for some 350 million years, a long time by any measure. It is interesting that there were several Snowball Earth episodes during the lifespan of Rodinia, the last occurring ~650 Ma when the supercontinent was well on its way to being broken up (though a large portion still covered the south poll).

By the Permian Period ~300 million years ago, the supercontinent cycle had come full circle again when the most recent supercontinent, Pangaea, was fully formed. The breakup of Pangaea started around 250 Ma but that supercontinent persisted over much of the Permian and Triassic periods, a span of some 100 million years. Matching the known ice ages with the comings and goings of supercontinents shows that ice ages do often occur when most of Earth's continental land mass is gathered together in one large continent. But the ice ages do not persist for the entire time span that a supercontinent exists.

Further more, ice ages have happened before and after supercontinent formation and breakup. It is tempting to say that the existence of a supercontinent is a necessary, though not sufficient condition for an ice age but even that is not a hard and fast rule. It seems that having a sufficiently large continent covering one of the polls is more important than having a single large continent—witness the current Pleistocene Ice Age. As Popper said, “science may be described as the art of systematic over-simplification.” While the configuration of Earth's continents undoubtedly plays a role in climate, I'm afraid that explaining ice ages as a result of the existence of an earthly supercontinent is an oversimplification. There is, however, a possible stronger link between the supercontinent cycle and increases in atmospheric oxygen (see “Formation of supercontinents linked to increases in atmospheric oxygen”).


gbaikie is correct that, many times, when the ice builds up to a certain extent, it can depress the continent by enough that the ocean floods in and this effectively breaks up the glaciers. It can take a long time for a continental area to recover (witness Hudson Bay, the Baltic, the Barents and the Kara seas - these are all on the continental shelves and were covered by 1 to 3 km high glaciers at the height of the last ice age).

And this is exactly what happened with Gondwana when it started moving across the south pole 450 million years ago. By 430 million years, the northern third of Africa went below sea level and the Ordovician/Silurian Ice Age ended. Southern South America suffered the same fate as it drifted across the south pole some time later.

In the snowballs however, enough ice is built up that it completely covers the continental shelves (and sea levels were probably 250 metres lower) so the situation is a little different.

Pangea is part of the supercontinent cycle, but it was centred over the equator and only a small amount of Antarctica was near the poles. This would produce very warm conditions compared to a Pangea over the south pole.

Glaciers build up on land. They can sometimes push out into the ocean over continental shelves but this takes a lot of snow accumulation in a central spreading region. And this only occurs when a continent is very near or on top of the poles (where there has always been 6 months of darkness and little solar insolation in the summer).

Otherwise, when there is only ocean at the poles, some sea ice can build up but it will not form glaciers. If it has deep ocean and surface ocean access to the rest of the oceans, the currents will be constantly moving warmth from the equator to the poles and moving the sea ice out to be melted at lower latitudes. Think of the Pangea example when there was unfettered surface ocean access to both poles (70% to 75% of the planet was one big ocean). So, sea ice in the winter only, brutishly hot conditions in the centre of Pangea at the equator. Mid-latitude deserts on Pangea would be twice as big as the Sahara. Shallow enclosed oceans like the Paleo Tethys sea could reach 40C (year-round).

Headed for another crunch

Earth's geology remains active and the continents continue to drift. Christopher R. Scotese, of the University of Texas at Arlington, has predicted the formation of one final supercontinent before Earth becomes geologically dormant—Pangaea Ultima. This final gathering of the continents should happen in about 250 million years. A map of Pangaea Ultima from the PALEOMAP Project can be seen here.