Thawing a planet-sized snowball

November 5, 2010 By Michael Schirber,
The timing of Snowball Earths may be tied to changes in atmospheric oxygen. Credit:

These days the climate news is all about global warming, but global freezing was the biggest climate worry in Earth's distant past.

Long periods of severe cold – like Ice Ages on steroids – brought glaciers down to the equator and froze much if not all of the oceans. Scientists still debate what triggered these so-called Snowball Earths, but equally uncertain is how the Earth unfroze itself. One research group is studying the hyper-greenhouse warming that would be needed to end a million-year-long winter.

The evidence for Snowball Earths comes from paleomagnetic data taken from ancient glacial deposits. From their magnetic properties, geologists can tell that some of these ice-induced rocks originated from low latitudes. This surprising result implies a cold planet that only would get colder as the ice reflected away more heat.

"Once you have ice cover in the tropics, it all 'Snowballs' from there," says Alexander Pavlov of the University of Arizona.

Ice reflects more than twice the sunlight of bare ground and more than five times that of water, so as the ice spread, less heat was retained at the surface. The average global temperature is estimated to have dropped to minus 50 degrees Celsius (minus 58 degrees Fahrenheit).

"This is why it is so hard to escape a Snowball," Pavlov says. "There is much less radiation being absorbed to warm the planet."

But escape the Earth did. Pavlov and his colleagues are modeling the necessary conditions that would push the icy Snowball temperatures back to a more comfortable level. As part of NASA's Astrobiology: Exobiology and Evolutionary Biology program, they also will be considering the implications in the search for habitable planets. If breaking out of a Snowball event turns out to be very difficult, then other worlds that we would expect to have liquid water instead may be permanently frozen.

Deep freeze

The geologic record bears the signatures of three Snowball Earths, although there may have been more. The first event occurred around 2.3 billion years ago and seems to roughly coincide with the rise of atmospheric oxygen. The other two happened more recently at 710 and 640 million years ago.

Scientists have proposed several theories for what caused the Snowball Earths. The most well-known of these involves a decrease in heat-trapping when too many continents drifted near to the equator.

It might seem strange that atmospheric levels of CO2 might be controlled by continental position, but the connection is through the chemical weathering of rocks. Carbon dioxide in the air dissolves into rainwater, creating carbonic acid that dissolves rock minerals like silicate. Through the reaction, the carbon dioxide is transformed into carbonate, thus reducing the amount of this greenhouse gas in the atmosphere.

Typically, this weathering is faster where the climate is warm, so having most of the continents along the equator should increase weathering and lead to a significant drop in CO2 that will chill the planet.

In the geochemical carbon cycle, chemical weathering transforms carbon dioxide into carbonates. These carbonates enter the sediment, only to later be freed by volcanoes. Credit:

There is some debate among scientists over just how much of the Earth was covered in ice during Snowball episodes. Many biologists do not think life on Earth could have survived being locked under a kilometer-thick sheet of ice, so they argue that there was open water somewhere in the tropical seas.

However, Pavlov believes that it is very hard to prevent ice from going everywhere once it reaches around 35° latitude. Moreover, there is geologic evidence that suggests the ocean was effectively cut-off from the atmosphere during the extended glacial periods.

Breaking the ice

Assuming that the planet did freeze over completely, how did the temperature trend ever reverse?

Previously, researchers suggested that the could re-warm itself by turning back the dial on carbon dioxide. Volcanoes would be constantly releasing CO2, and there would be virtually no weathering on an ice-covered planet to consume this greenhouse gas. Calculations have shown that once the atmosphere has accumulated 0.2 bars-worth of CO2 (over 600 times what we have now), the greenhouse warming would be enough to start melting the ice.

This set of Hubble images shows seasonal changes in the northern ice cap on Mars. Much of the frozen carbon dioxide sublimates during the martian summer, only to return in the winter. This same CO2 cycle may have occurred on Earth's poles during a Snowball event. Credit: NASA/JPL-Caltech

However, there are some holes in this warming model. Carbon dioxide will condense into "dry ice" at around minus 80 degrees Celsius (minus 112 F). The wintertime polar temperature should drop below this limit during a Snowball episode, so a large fraction of CO2 could end up being trapped in seasonal ice caps at the poles (similar to what happens on Mars).

Pavlov and his colleagues are currently redoing the models to account for CO2 condensation and evaporation. They also will be looking at whether clouds of CO2 might form that could block some of the sunlight from reaching the ground.

It may turn out that carbon dioxide from volcanoes won't be enough to thaw out the planet. His group will therefore consider the effect of other greenhouse gases, such as methane released from ice deposits, or sulfur dioxide emitted from volcanoes.

The team also will be readdressing the reflection, or "albedo" of the ice. Often this is treated as a single parameter, but Paul Hoffman from Harvard University says that the buildup of dust or salt on the surface, as well as the daily melt cycles, can have a big effect on just how much of the Sun's heat gets reflected away rather than absorbed.

"Tropical ice albedo is the 'elephant in the room' in Snowball modeling," says Hoffman, who is not involved in this project.

Snowballs elsewhere?

Our planet was able to escape its Snowball events, but would other planets be so lucky?

"The more landmass a planet has, the better it is protected from runaway Snowball," Pavlov says. "It is much harder to build a glacier inside a large continent if it is not at the pole."

So a watery planet with little or no dry land might get itself stuck in a and never escape. Astronomers eventually may have to consider that too much water can be a bad thing when it comes to the habitability of a planet.

"I would say that the main conclusion is that we should not be focused so much on the water-rich planets," Pavlov says. "Dry planets with some water can be habitable at farther distances from their stars."

Explore further: Scientists study 'Snowball Earth'

Related Stories

Scientists study 'Snowball Earth'

November 28, 2006

Canadian scientists have determined the factors involved in ending a severe ice age 750 million years ago that nearly completely froze Earth's oceans.

The greenhouse gas that saved the world

August 18, 2009

When Planet Earth was just cooling down from its fiery creation, the sun was faint and young. So faint that it should not have been able to keep the oceans of earth from freezing. But fortunately for the creation of life, ...

Recommended for you

Weathering of rocks a poor regulator of global temperatures

May 22, 2017

A new University of Washington study shows that the textbook understanding of global chemical weathering—in which rocks are dissolved, washed down rivers and eventually end up on the ocean floor to begin the process again—does ...

Smoke from wildfires can have lasting climate impact

May 22, 2017

The wildfire that has raged across more than 150,000 acres of the Okefenokee Swamp in Georgia and Florida has sent smoke billowing into the sky as far as the eye can see. Now, new research published by the Georgia Institute ...


Adjust slider to filter visible comments by rank

Display comments: newest first

2 / 5 (4) Nov 05, 2010
Please, make also calculations regarding a
huge asteroid/comet colliding with Earth. Seems possible it could break and stack the sea ice and blackening the planet and hence change the albedo. Then do the same regarding supervolcanoes.
3.4 / 5 (5) Nov 05, 2010
This looks to reveal some interesting things about the earth and also may have something to say about the claims of current climate science. Many Climate scientists would say that volcanoes never produce enough CO2 to have a significant effect on global warming compared to manmade CO2. This study could possibly have ramifications for modern climate models as well as for the claims of the portion of the scientists who adhere to the models.

Interesting study nonetheless.
2.5 / 5 (2) Nov 05, 2010
i would have to go for the huge asteroid impact and that also might have subsequently triggered intense global vulcanic activity as well, with lots of vulcanic dust covering the white albedo layer, also an asteroid on the poles would have evaporated the dry-ice CO2, its somewhat ironic that greatest deathbringers are lifegivers as well
4 / 5 (2) Nov 05, 2010
I'd go with solar oscillation considering how relatively uniform the cooling curves are leading to snowball earths, especially as each protracted cold period was on a shorter and shorter measurement of time.
3 / 5 (5) Nov 05, 2010
I'm inclined to agree with Skeptic_Heretic for the earlier periodicity.
1 / 5 (1) Nov 06, 2010
Was anyone around back then to "worry" about anything at all?
not rated yet Nov 07, 2010
Ok I'll put my penny in and say it was Polar Shifting as the poles move around, the magnetics of the earth change the ice caps moved with it
1 / 5 (2) Nov 07, 2010
What has ice formation got to do with magnetic pole shifts?
1 / 5 (1) Nov 08, 2010
"I'd go with solar oscillation considering how relatively uniform the cooling curves are leading to snowball earths"

That's really the question, isn't it? I'll play devil's advocate this time and politely and respectfully decline to accept any current theory due to lack of sufficient evidence and understanding. I think it's better to hold your bets and leave the door open for any of the proposed theories, or a combination of them. Perhaps you could assign a range of likelyhood for each scenario at best, but to say "I believe" is presumptuous. The leading experts are all still scratching their heads over this one, so I'm certainly not going to form an opinion yet. There is a finite but HUGE number of factors that could combine in currently unpredictable ways. Who says that each supposed snowball earth had the same causes, and who says that they came to an end by the same mechanism each time? The face of the earth was quite different in each of the proposed time periods I think.
not rated yet Nov 08, 2010
This looks to reveal some interesting things about the earth and also may have something to say about the claims of current climate science. Many Climate scientists would say that volcanoes never produce enough CO2 to have a significant effect on global warming compared to manmade CO2. This study could possibly have ramifications for modern climate models as well as for the claims of the portion of the scientists who adhere to the models.

Interesting study nonetheless.

Find a single climate scientist that ever said that volcano's do not emit enough CO2 to change the climate! In this case however, we are talking a million years of volcano's... not 100 years.
1 / 5 (2) Nov 08, 2010
Actually, several have said this but I would have to comb the entire literature to dig up the quotes... Not going to happen anytime soon. Sorry.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.