Arctic Sea Ice: What, Why, and What Next

Ramez Naam
September 21, 2012

On September 19th, NSIDC, the National Snow and Ice Data Center, announced that Arctic sea ice has shrunk as far as it will shrink this summer, and that the ice is beginning to reform, expanding the floating ice cap that covers the North Pole and the seas around it. The Arctic Sea Ice extent this September was far smaller than the previous record set in 2007. At 3.4 million square kilometers of ice coverage, this year s Arctic minimum was 800,000 square kilometers smaller than the 2007 record. That difference between the previous record and this year s is larger than the entire state of Texas. An ice-free summer in the Arctic, once projected to be more than a century away, now looks possible decades from now. Some say that it looks likely in just the next few years.

What s happening in the Arctic? Why is it happening? And does it matter for the bulk of us who live thousands of miles away from it?

Figure 1 - Ice in the Arctic is increasingly melting, exposing dark waters below. Photo credit: National Snow and Ice Data Center

Faster and Faster

Conditions in the Arctic change dramatically through the seasons. In the depths of winter, the Earth s tilt puts the Arctic in 24 hour-a-day darkness. Temperatures, cold year round, plunge even lower. The sea surface freezes over. At the height of summer, the opposite tilt puts the Arctic in 24 hour-a-day sunlight. While it s a cold cold place even at these times, the constant sunshine, warmer air, and influx of warm waters from further south serve to melt the ice. The ice cap usually starts shrinking in March, and then reaches its smallest area in mid-September, before cooling temperatures and shorter days start the water freezing and the ice cap growing once again.

When scientists and reporters talk about an ice-free Arctic, they re usually speaking of the Arctic in summer, and especially in September, when ice coverage reaches its minimum.

The amount of ice left at that minimum has indeed been plunging. In 1980, the ice shrank down to just under 8 million square kilometers before rebounding in the fall. This year s minimum extent of 3.4 million kilometers is less than half of what we saw in 1980. Strikingly, two thirds of the loss of ice has happened in the 12 years since 2000. The ice is receding, and the process, if anything, appears to be accelerating.

Figure 2 - Arctic sea ice coverage in September has dropped in half since 1980, and the drop appears to be accelerating.

As recently as a few years ago, most models of the Arctic ice anticipated that summers would remain icy until the end of the 21st century, and well into the 22nd century. But the trend line above makes that look unlikely. The amount of ice remaining, this year, is about the same as the ice lost between the mid-1990s and today. If ice loss continued at that pace, we d see an ice free summer sometime around 2030, give or take several years.

Is that plausible? Opinions differ substantially, even among climate scientists.

At one end of the spectrum are those who see the ice lasting in summer for another 20 or 30 years, or perhaps even a bit longer.

For example, Lars-Otto Reierson, who leads the Arctic Monitoring and Assessment Programme told Reuters that most models predict the summer ice disappearing by 2030 or 2040.

Similarly, a paper published this year in Geophysical Research Letters by multiple scientists, including several from the National Snow and Ice Data Center, found that an ice-free summer in the Arctic in the next few decades was a distinct possibility.

A recent assessment from Muyin Wang at the University of Washington and James Overland at the National Oceanographic and Atmospheric Administration, using the most up to date Arctic ice models and data, projected a nearly ice free Arctic around 2030.

And Ceclia Bitz, a professor of Atmospheric Sciences at the University of Washington at part of the Polar Science Center sees a 50/50 chance that the Arctic will be ice free in summer in the next few decades.

On the other end of the spectrum are those who think the melt could happen much sooner. Peter Wadhams, who leads the Polar Ocean Physics Group at the University of Cambridge, has predicted since 2008 that the Arctic ice could be gone in summer by 2015. He now believes there s a chance that it could happen even sooner.

Similarly, Mark Drinkwater, the European Space Agency s senior advisor on polar regions and a mission scientist for the CryoStat satellite that measures arctic ice, believes that the Arctic could be ice free in September by the end of this decade.

When will the ice melt? While the range of possibilities is wide today, it s shrunk dramatically from just a few years ago, when most climate scientists expected the ice to survive through the 21st century. Now the question is whether it will be gone in decades or in mere years.

Why is the Ice Melting?


Why is the Arctic ice cap growing so much smaller in summer? And why this summer in particular?

Some of the effect is seasonal variation. Like anything in climate, there is a fair bit of noise. This year, in particular, an immensely powerful high latitude storm in August stirred up waters, bringing warmer water from further south up to the Arctic, and accelerating the melt of the ice.

Everyone who studies the ice melt agrees that natural weather variation from year to year has played a role. Both 2007 and 2012 were unusually bad years for the ice, and would have been even if the planet s climate hadn t changed. Looking over the data from 1979 to 2011 (before this year s new record low), a team of scientists led by Julienne Stroeve at the National Snow and Ice Data Center found that those random variations accounted for around 40 percent of the change in ice cover to date, and that human activity accounted for around 60 percent of the change.

Other research points to an even greater human component. In 2011, a team led by Chilean scientist Christophe Kinnard published a paper in Nature that used data from 69 sites around the Arctic to reconstruct the extent of the ice over the last 1450 years all the way back to the 6th century AD. What they found was that late summer ice coverage over that entire fourteen and a half century period stayed between 9 million and 11 million square kilometers, a little higher than it was before satellite observation started in 1979, or roughly three times the minimum that we hit this September.

Kinnard was kind enough to send me the team s underlying data. Combining it with satellite based observations from 1979 onward, the last few decades pop out. Ice coverage fluctuates for centuries, but stays in a narrow band, until suddenly, in the last few decades, the amount of ice left in late summer plunges.

Figure 3 - Arctic sea ice coverage, as reconstructed from 69 sites around the area, fluctuated in a fairly narrow range for more than 1400 years. In the last few decades, it has plunged abruptly.

All the sea ice loss, including that before satellite observation occurred, has happened since the start of the industrial revolution, and the beginning of human emissions of CO2 and other greenhouse gases on a massive scale.

In this context, random variations look rather small. In fact, there are at least three distinctly non-random factors leading to the disintegration of the ice.

1. A Warming Planet.

The first factor is the most obvious. Ice melts faster in warm air than in cold air, and faster in warm water than in cold water. And the Arctic has warmed more rapidly than any place on Earth. In the last 40 years, the world as a whole has warmed by around 0.8 degrees Celsius. That alone would accelerate the melt of ice. But the warming in the Arctic has been twice at fast, at roughly 1.9 degrees Celsius, or 3.5 degrees Fahrenheit. That may not sound like much, but when ambient conditions are so close to freezing already, that additional heating can make a tremendous difference in the rate of melting.

2. Positive Feedback

The second factor is positive feedback. Not the good job kind of positive feedback, mind you. This type of feedback is more similar to what happens when a microphone comes too close to a speaker, and a random piece of noise gets amplified out of control. Any time a process can amplify itself, it s a positive feedback loop. The melting ice is, in fact, amplifying its own destruction by helping to fuel accelerated arctic warming and thus more rapid disappearance of the ice that remains.

Here s how: Ice and snow reflect light. Thick ice covered with snow will reflect the large majority of the sun s energy back into space, absorbing only 10 to 20% of the sunlight as heat. Ocean water, on the other hand, reflects very little of the sun s energy back into space, absorbing more than 90% of it as heat.

So the less ice there is, the more the sun warms the waters that remain. And 90% of the ice cap is under the surface, bathing in that water. Less ice means faster melting of the ice that remains.

Figure 4 - Arctic sea ice reflects most solar energy into space. As ice melts it reveals dark waters below, which absorb more than 90% of the solar energy that hits them, leading to faster warming both locally and globally. Image courtesy of NASA.

The melting ice and warming atmosphere cause a second, less appreciated feedback loop. Water vapor is an important greenhouse gas. It captures heat radiating from the planet and traps it in the atmosphere, warming the planet by tens of degrees Celsius.

Cold air doesn t hold much water, though. The air above the Arctic is dry a supercold desert. As a result, Arctic air doesn t trap much of the heat that radiates away from the water and ice. But as the air warms, it can hold more water – about 7% more for every degree Celsius the temperature rises. And that added humidity traps more heat in the atmosphere above the Arctic, raising local temperatures and further accelerating the melt of the ice.

All in all, there s good reason to believe: the less ice remains, the faster the ice that s left will go.

3. Thin Ice.

There s one more important factor to bear in mind the increasingly thin state of the ice. Over the last few decades, the average thickness of the ice that covers the Arctic in summer has dropped in half, from an average of more than 2 meters thick (7 feet) to roughly 1 meter thick (3 feet).

That drop in thickness has happened, in large part, because the ice that remains in the Arctic in summer is now predominantly very young ice, most of it formed just in the past winter. In the mid 1980s, only around 25% of the summer ice was this new, thin, extremely vulnerable stage, and almost two thirds of the ice was older than two years old old enough to have accumulated more water and grown thicker and sturdier. Now the numbers have nearly reversed. Very little of the thick multi-year ice remains. The ice that s left, being thinner, takes less heat to melt.

Figure 5 - Average ice thickness in the Arctic in summer has dropped in half, leaving behind ice that is increasingly vulnerable and prone to melting. Photo courtesy of NOAA.

The reduced thickness of the ice not only bodes poorly for its survival, it tells a dramatically different story of how far from an ice free summer we might be. Most models of arctic ice coverage predict that some ice will remain even at the September low point – for decades to come. Looking at the trend of ice coverage and doing the very simplest extrapolation, we could see the first ice free Arctic days by 2030, a little sooner than many experts predict, but still nearly a generation in the future.

But coverage is just the area the ice covers. Volume the area covered multiplied by the thickness is how much ice there actually is. Imagine a sheet of ice covering a frozen lake in early spring. The ice may cover the entire lake, but a two inch thick layer is much sturdier and will take longer to melt than a one inch thick layer.

When we look at volume instead of area, we don t see that half of all the ice has disappeared since 1980. Instead, we see that almost 80% of the September ice has disappeared in that time. And most of that loss has been in the last 12 years. 70% of the ice volume we saw in 2000 has disappeared. Less than one third of that ice volume from just 12 years ago is what we see today.

Figure 6 - Artic sea ice volume has dropped far faster than coverage. In September the ice that remains is down by nearly 80% from its volume in 1980. At current pace of volume loss, the first ice-free Arctic day could come in just a few years.

If the Arctic sea ice loses volume at the same rate that it has over the last 12 years, then the first ice-free Arctic day in September could happen in the next 5 years. If the rate of ice volume loss continues to accelerate, as it has been, then that day could be even sooner. Ice volume tells a story much more like that of Peter Wadhams, the leader of the Polar Ocean Physics Group at Cambridge, who for years has been predicting an ice free Arctic September as early as 2015.

An Ice Free Solstice?

Here s what the future holds. One September, perhaps as early as 2015, perhaps decades later, we ll learn that the Arctic is ice-free. There will still be ice bergs and occasional floes of ice in the Arctic, but they ll be scattered, surrounded by far more water than ice. Nowhere will ice constitute the majority of the Arctic.

You ll probably hear this referred to as the first ice-free Arctic summer. But, in reality that first ice-free Arctic will most likely be a period of a few days or a week or two in mid-to-late September. August will still have an ice cap, if a small one. July will have a larger one than that. June larger still. And March, the month of the year when the Arctic ice cap is at its maximum, will still have seen plenty of ice.

Yet from that year on, the ice-free period will likely grow, expanding in duration to start earlier and end later year over year. Most likely, it ll follow the same jagged, two-steps-forward one-step-back progression we see in climate in general. The first year after the first ice-free Arctic year, we may see some ice cap persist all the way through the summer again. For that matter, next year, 2013, it s quite possible that we ll see more ice than this year. Climate is bumpy that way. But, if recent history is any lesson, bit by bit, step by jagged step, the ice free period will lengthen, and the ice coverage in other months of the year will shrink.

This is particularly important because September is not a very sunny month in the Arctic. The sun never rises high above the horizon, and so its heating power is muted. Those dark waters are absorbing more of the solar energy that strikes them than ice would, but there s simply less solar energy striking the arctic in September than there is for the spring and summer months leading up to it.

The sunniest time of year in the northern hemisphere is the summer solstice, in late June, and the weeks preceding and following it. For several weeks the sun s rays are at their most intense and the Arctic receives 24/7 sunlight, giving it a double whammy of heating. In fact, in June, July, and the latter half of May, the Arctic receives more total solar energy per day than regions at the equator do at any time of year. The sun s rays are never as powerful in the Arctic as they are at the equator, but the 24/7 availability of sun more than makes up for that. (If you doubt this, see NASA s Earth Observatory page on the topic or use NASA s monthly insolation-by-latitude calculator.)

Thus, every patch of dark ocean water revealed by melting ice in June, or May, or July, has a warming effect much larger as much as five or six times larger – than the same change in ice coverage in September, when the ice hits its minimum today. The loss of ice in September is one thing. Loss of ice in June would have a far bigger impact on the region and the planet.

So far, the ice extent change in June has been modest. Between 1979 and 2012, NSIDC reports that the Arctic ice coverage in June shrank by around 10%, compared to the roughly 50% shrinkage that the September ice coverage has seen. And at first blush, the trend looks more or less linear. At that pace, we wouldn t expect to see an ice-free June in the Arctic for another 300 years.

Figure 7 - Arctic sea ice coverage in June - the sunniest month of the year - has dropped by roughly 10 percent since 1980, far less than the drop in September coverage. Image courtesy of NSIDC.

But again, while coverage matters, it s volume of ice that really tells us how healthy the ice is, and how much there is left of it. And June ice volume in the Arctic has sunk fast by more than half since 1980. What s worse is that the trend is accelerating. In the decade from 1980 to 1990, June ice volume dropped by around 3.8 thousand cubic kilometers. In the last decade, between 2002 and 2012, it dropped by almost 9.5 thousand cubic kilometers, two and a half times as fast. If the pace of the last decade were to continue on this recent pace and didn t keep on accelerating June ice would be gone by 2026, exposing the dark waters of the Arctic Ocean to the year s most intense influx of solar energy, which would then be captured as heat.

Figure 8 - Arctic sea ice volume in June tells a much worse story than coverage alone. June ice volume in 2012 was less than half of the ice volume seen in 1980. June ice volume loss appears to be accelerating.

Of course, if we see one thing in the loss of Arctic sea ice, it s that the process isn t continuing at a steady pace. It s accelerating. And as more dark water gets exposed to the warm June sun, the lifetime of the remaining ice is likely to shrink even faster than it currently is. An ice free June could be quite a bit closer than we imagine.

Just a few years ago it would have seemed incredibly alarmist to predict even a single ice free Arctic day this decade. Increasingly, it seems foolish not to accept that ice free Arctic summer days, weeks, or months this decade are, if not a certainty, then at least a very real possibility.

The Good News

So is this the end of the world? Or is it just an interesting but ultimately unimportant phenomenon we get to observe? Could it even be an opportunity?

First, some good news. The melting Arctic ice will not cause sea levels to rise to any noticeable degree. The Statue of Liberty isn t about to be reduced to a head and single upraised arm, forlornly holding her torch just above water. The Arctic ice cap is sea ice. It floats already. And just as a melting ice cube in your drink doesn t raise the overall level of fluid in the glass, the melting of floating ice in the Arctic won t directly raise sea levels. (This is quite different from the effect of land-based ice, such as that on Greenland or Antarctica. The melting of ice that is currently sitting on land does raise sea levels. But such ice is also far harder to melt.)

More good news, or at least the absence of terrible news: The melting Arctic ice is unlikely to suddenly stop the deep ocean conveyor , the current that brings warm water to Europe and keeps the continent much of which is at the same latitude as Canada fairly warm and temperate. While the deep ocean conveyor belt, also known as the thermohaline circulation, does appear to be slowing a bit, calculations show that the amount of fresh water needed to stop it is far greater than the amount of water currently trapped in Arctic ice. (A breakdown of the thermohaline circulation, by the way, is the vague explanation given for the rather jumbled science of the movie The Day After Tomorrow. So, among other good news, take note that you won t need Dennis Quaid to snowshoe across a frozen landscape to come to your rescue.)

Finally, good economic and natural resource news: The receding ice will open up new trade routes, making it easier, cheaper, faster, and more efficient to ship goods between northern Europe, Canada, Russia, and the United States. Cargo that once had to be placed on ships that passed through either the Panama Canal or the Suez Canal will now, in many cases, have a shorter route one that saves on both time and fuel.

The opening up of the Arctic will also open up exploration for minerals and for fossil fuels. The US Geological Survey estimates that the region has the world s largest remaining untapped reservoirs of oil and natural gas as much as 90 billion barrels of oil and almost 1.7 trillion cubic feet of natural gas. Those numbers would make the Arctic home to 13% of the world s remaining oil and 30% of the world s remaining natural gas. And that is a very rich prize, spurring investment in exploration, and increasing jockeying between the world s powers including some countries, like China, that don t even have a physical presence in the Arctic to gain access to those resources.

Tapping those minerals and oil and natural gas would be an economic boon to communities in the Arctic. It would also be a win for the global economy as a whole, helping to keep energy prices lower by bringing new supply to market. But of course, burning that oil and natural gas would also accelerate climate change the very process that has set the Arctic on the path to melting.

The Bad News

So the seas will not rise, the ocean currents won t suddenly end, and there will be some economic benefits. What is it about the melting ice that worries us?

By now many of us have seen pictures of lonely polar bears, seemingly stranded on a patch of ice surrounded by water. There are indeed threats to polar species. But I want to focus on wider threats that extend beyond the Arctic and into the rest of the world. There are three in particular that should concern us.

1. More Extreme Weather

The most palpable impact of climate change for those of us who live far from the poles is the increase in extreme weather. Around the world, record highs are occurring at more than twice the rate of record lows. The US drought of 2012 was the worst since the Dust Bowl of the 1930s. A drought nearly as bad struck Texas and the American south in 2011, and even more destructive heat waves hit China and Russia in 2010, and Europe as a whole in 2003.

You might not think that what happens in the Artic has much bearing on what happens in Texas or Moscow or southern provinces of China, but a study published in 2012 in Geophysical Research Letters has drawn a convincing connection. Blowing around the periphery of the Arctic is the polar jet stream a region of high speed wind that blows west to east, and helps drive wind circulation around much of the northern hemisphere. The jet stream is powered by the temperature difference in fall and winter between the Arctic and the more temperate areas just to its south. But as the Arctic ice has receded, the Arctic Ocean waters have absorbed more heat in late summer and early fall. In late fall and early winter, they ve given that heat up, back into the atmosphere. That, in turn, has led to warmer Arctic autumns and winters, which has reduced the temperature difference that fuels the jet stream.

The result is that the jet stream is now weaker than it once was about 14% weaker than it was in 1980.

Why does this matter? Because a slower jet stream makes it easier for blocking weather patterns to develop. Blocking weather patterns are the ones that hover over a region rather than moving on like the drought that basted Texas in 2011 and decimated its forests and hay and wheat crops to the tune of more than $7 billion in damage, and like the heat wave that enveloped Moscow and much of the rest of Russia for most of the summer of 2010, killing an estimated 55,000 people in July and August of that year.

Figure 9 - The polar jet stream drives weather circulation around the northern hemisphere. As Arctic warming weakens it, 'blocking' weather patterns that hover over regions become more common, amplifying the effect of extreme weather. Image courtesy of NASA.

Climate change is driving more extreme weather – by heating up the atmosphere, pumping more energy into storms, and heating the air to the point that it can more easily suck away moisture or concentrate it in one point. As the planet continues to warm, all of those factors will increase, leading to more heat waves, more droughts, and more floods.

And the changes to the Arctic, it seems, will exacerbate this, by slowing down the jet stream, and making it more likely that the extreme weather conditions that develop get locked in place, hammering the same regions for protracted lengths of time.

2. Accelerated Warming

The second thing to fear about loss of Arctic sea ice is the potential to accelerate climate change on a global basis.

A black object gets hotter in the sun than a white object. That much is common sense. Earlier, in describing how melting ice accelerated the melt of more ice, I talked about the fact that dark sea waters absorb up to 90% of the sun s energy that strikes them, while snow-covered ice absorbs only 10 to 20% of that same energy. The exposure of darker waters speeds up heating of the Arctic, and thus the loss of more ice.

But the impact is larger than that. And indeed, it s large enough to make a difference on a global scale.

In June, the Arctic ice cap covers around 2% of the Earth s surface around 11 million square kilometers of Arctic ice cap out of a total of 510 million square kilometers of Earth s land and oceans. And that 2% of the Earth s surface, for a period of roughly two months, receives more solar energy per day than even the sunniest areas on the equator.

Analyzing this, Peter Wadhams of the Global Oceans Physics Program at Cambridge calculates that the loss of the Arctic ice throughout the summer would have a warming effect roughly equivalent to all human activity to date. That is to say, with the ice gone in summer, the planet would have an additional heating effect just as large as the heating effect of all human CO2 and other greenhouse gasses to date.

In other words, the complete meltdown of the Arctic could roughly double the rate of warming of the planet as a whole.

Figure 10 - The Arctic ice cap in June covers roughly 2% of the globe, and is the sunniest place on Earth for two months. The replacement of white ice and snow with darker ocean and land would capture roughly as much additional heat per year as current human-caused warming does today.

There are important caveats and uncertainties to that analysis. First, Wadhams doesn t take into account the effect of clouds. Darker waters absorb more energy from sunlight only if the sunlight reaches them. Cloud cover in the Arctic something which may increase as rising temperatures enable Arctic air to carry more moisture may reflect sunlight back into space before it ever touches the water, thus reducing the warming effect of ice melt.

Conversely, Wadhams calculations only take into account the loss of sea ice in the Arctic Ocean itself. If the accelerating warming of the region melts permafrost in Siberia, Canada, and Alaska something it seems to be doing it could well change the reflectivity of the land in those areas, and thus how much sunlight they absorb. How large a change could this be? And in what direction? Climate scientist Judith Curry points out that warming in the Arctic is also associated with more winter snowfall in Canada and Europe, which could make the land more reflective and exert a cooling influence. That change, however, would have an impact mostly in winter (when sunlight is most scarce) and in early spring. In late spring and summer the effect of a warmer Arctic on tundra is likely to be towards less light-colored permafrost and more dark-colored thaw lakes (the bodies of water that form as permafrost melts), dirt, and growing plants.

Climate, as always, is incredibly difficult to model. But the main takeaway here is that the replacement of white snow and ice in the Arctic with dark colored water is in the ballpark where it could rival and add to all human greenhouse gas emissions to date.

Global climate models, it should be said, don t anticipate the ice being gone until well after 2100, and so they don t include this added heating effect when they calculate future temperatures. Thus, they may underestimate the temperature changes to come.

That is also troubling because current proposals for tackling climate change involve dramatically ratcheting down human greenhouse gas emissions, and allowing greenhouse gas concentrations in the atmosphere to gradually drop. That plan assumes that those greenhouse gas emissions are the major source of human caused warming. The addition of a new source of warming may dash any hopes of arresting climate change by reducing greenhouse gas emissions. If we want to keep global temperature increase below two degrees Celsius we may need to either stop the melting of the Arctic something which may no longer be possible or find ways to combat climate change that go beyond simply reducing our greenhouse gas emissions.

3. The Arctic Methane Bomb

The final risk is the largest in the very long term, though the extent to which it will affect us in the coming years and decades is still a matter of great uncertainty.

The Arctic and the region immediately surrounding it are home to immense amounts of buried carbon. The permafrost of Siberia, Canada, and Alaska is estimated to hold around 1.7 trillion tons of carbon mostly in the form of dead plant matter. The sea bed beneath the shallower parts of the Arctic Ocean holds anywhere up to another 10 trillion tons of carbon trapped in a semi-frozen state called methane hydrates.

By contrast, all human CO2 emissions over the last century amount to only 1.1 trillion tons of carbon. The permafrost carbon, alone, could exceed the effect of all human burning of fossil fuels. The Arctic Sea bed deposits are close to ten times all carbon humans have released. What s worse is that much of that carbon will end up released as methane (CH4) instead of carbon dioxide (CO2).

A molecule of methane absorbs and traps roughly a hundred times as much heat as a molecule of carbon dioxide. Fortunately, methane degrades quickly in the atmosphere, lasting on average for around 10 years before being converted into CO2, which can last for a hundred years or more. Even so, over the course of a century, a molecule of methane released today will have 25 to 30 times the heating impact of a molecule of CO2 released today.

If even 10% of the northern permafrost s buried carbon were released as methane, it would have a heating effect over the next decade equivalent to ten times all human greenhouse emissions to date, and over the next century equivalent to roughly four times all human greenhouse emissions to date.

And the permafrost is melting. In Fairbanks, Alaska, ground that s been frozen solid for 10,000 years is melting, opening up sink holes. In the town of Newtok, Alaska, the permafrost melt has been so bad that the residents recently voted to move the entire town rather than stay and watch it sink into the once frozen land.

Figure 11 - Once frozen solid, permafrost near the Arctic is melting, creating conditions for decomposition of organic matter and the release of carbon as CO2 and methane. Image courtesy of NASA.

Historically, climate modelers haven t expected the bulk of the carbon in northern permafrost to be released any time soon. Instead, as recently as 2006, climate scientists expected only around 100 billion tons of permafrost carbon to make its way into the atmosphere this century. How much warming effect that carbon release will have will depend on how much of it emerges as CO2 and how much emerges as methane. When plant matter decays in the presence of oxygen, the carbon will be produced as CO2. When plant matter decays anaerobically, without oxygen (for example, in a pool of slushy soil that was once permafrost) then much of the carbon will emerge as methane. If one third of that 100 billion tons expected this century were released as methane, then the heating effect each year would be roughly equivalent to that of the amount of CO2 and other greenhouse gases human civilization releases each year.

That is to say, we could end all our burning of fossil fuels take them all the way to zero – and still see greenhouse gas levels rising just as rapidly as they are today.

Unfortunately, climate models today don t take into account this release of carbon. As a paper in 2011 on permafrost melt noted, none of the climate projections in the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report, none of the recent permafrost projections, and none of the projections of the terrestrial carbon cycle account for the release of carbon from melting permafrost.

Worse, even the predictions of 100 billion tons of carbon entering the atmosphere from melting permafrost may be out of date and too conservative. In 2008, a study from the National Center for Atmospheric Research (NCAR) and the National Snow and Ice Data Center (NSIDC), published in Geophysical Research Letters, found that melting sea ice threatened permafrost as much as 900 miles inland, and that in past periods of rapid sea ice loss, Arctic land warmed three and a half times as fast as the warming that models predict for the 21st century. Permafrost melt models haven t taken into account the rapid feedback cycles driving the warming of the Arctic. They haven t even taken into account the effect that greenhouse gases released from the permafrost itself will have.

The projections of the last few paragraphs deal with the on-land deposits of carbon in the permafrost alone. I ve ignored the even vaster methane deposits in methane hydrates frozen below shallow waters of the Arctic Ocean s continental shelves. That store of carbon is enough that, if all of it were to go, it would have a warming effect equivalent to hundreds of times the total human carbon emissions to date.

The Arctic sea shelf is definitely giving off methane. In 2011, a Russian expedition found kilometer-wide plumes of methane bubbles rising up from the sea floor. Yet the total quantity given off remains quite low today at most a tiny fraction of overall greenhouse gas emissions. Indeed, it s possible that those methane plumes have been there for decades or centuries, and are only newly discovered rather than new. Climate realists point out other reasons to remain calm in the face of the methane at the bottom of the Arctic Ocean.

First, while most of the methane is believed to be buried roughly 200 meters below the sea bed, only the top 25 meters or so of sea-bed are currently thawed, and thawing seems to have only progressed by about one meter in the last 25 years a pace that suggests that the large bulk of the buried methane will stay in place for centuries to come.

Second, several thousand years ago, when orbital mechanics maximized Arctic warmth, the area around the North Pole is believed to have been roughly 4 degrees Celsius warmer than it is today and covered in less sea ice than today. Yet there s no evidence of a massive amount of methane release in this time. (Though it must be said that the Arctic is set to pass those temperatures sometime in the next few decades, and keep soaring beyond them, at which point we ll be in territory uncharted over the past few hundred thousand years.)

Third, the last time methane was released in vast quantities into the atmosphere during the Paleocene-Eocene Thermal Maximum 56 million years ago – the process didn t happen overnight. It took thousands of years.

Put those facts together, and we are probably not in danger of a methane time bomb going off any time soon.

However, even a slow, gradual release of just a tiny fraction of the methane buried beneath the Arctic Ocean could significantly add to the pace of climate change. If the Arctic sea floor methane deposits started to outgas at a rate that would empty them into the atmosphere in 10,000 years, that would still be an added annual warming effect roughly on par with the amount of carbon humans emit into the atmosphere each year. If the rate of Arctic sea floor methane release were faster more like a 1,000 year pace to empty those deposits then we d be looking at a warming effect each year from that methane outgassing that would be many times greater than the warming from the fossil fuels we burn.

The Triple Whammy, and the Perils of Prediction

So, in addition to the increased persistence of severe weather from a slowing jet stream, we face a triple whammy of raw warming effect.

  1. Warming from the greenhouse gases we emit already.
  2. Warming from the loss of ice and permafrost in the Arctic, and the exposure of dark water and dark land below.
  3. Warming from the release of more carbon into the atmosphere as the permafrost and the Arctic sea floor methane begin to go.

The first of these is certain.

The second, the darkening of the Arctic and the warming that will come with it, is high confidence, though we still have questions about the exact magnitude. That factor alone means our current climate change projections for the coming century may be out of date and overly conservative. It means that we may, in the next 10 to 20 years, reach a point where no matter what reduction we make in greenhouse gases, the planet will keep warming.

The third whammy, the risk of more carbon entering the atmosphere, is the most speculative. There s a range of possibilities. At the low end, published work suggests that, at the very minimum, CO2 release from melting permafrost will add 10-15% to human emissions. In the middle of the range is the possibility that permafrost will melt more rapidly than expected, or that at least a few percent of the carbon it gives off will be in the form of methane. In that range, permafrost melt alone could add a greenhouse effect equivalent to all human greenhouse gas emissions (in addition, of course, to the heating effect from a darker Arctic and the heating effect from human-released greenhouse gases). And at the high end there s the small but non-zero chance of much more rapid methane release from the permafrost or from the oceans, with the release even a small fraction of the methane trapped there leading to a warming effect that exceeds human contributions significantly.

It seems pretty likely then, that if the ice cap continues on its way towards rapid disappearance, we re on path to a rate of warming faster than current climate models. And at worst, far beyond that.

It s tough to make predictions, Yogi Berra once said, especially about the future. Climate is an incredibly complex system, with feedback loops in every direction, with variables that are tough to model, and with random noise in the year-to-year data that can make trends look slower or faster than they really are. It s quite possible that random fluctuations are making the trends in the ice look worse than they truly are. It s possible that Arctic ice will settle into a new and stable state at this reduced level. It s possible that the permafrost melt will be on the slow end of projections, or even well below any projections we have today. Perhaps increased snowfall from higher humidity in the Arctic will more than offset the higher temperatures, or perhaps cloud cover increase will more than cancel out the darkening of the Arctic.

But our recent track record in predicting what happens to the ice has not been good. The reality of changes to the Arctic has far outstripped most predictions. Only a few years ago, in the 2007 Intergovernmental Panel on Climate Change report, the bulk of models showed the Arctic ice cap surviving in summer until well past 2100. Now it s not clear that the ice will survive in summer past 2020. The level of sea ice we saw this September, in 2012, wasn t expected by the mean of IPCC models until 2065. The melting Arctic has outpaced the predictions of almost everyone everyone except the few who were called alarmists.

Figure 12 - Arctic sea ice has receded dramatically faster than the mean of IPCC models projected, reaching levels not expected until 2065. Positive feedback loops such as those that have accelerated sea ice loss might accelerate the loss of permafrost and the release of additional CO2 and methane from the Arctic.

I bring up this case of climate change happening far faster than we expected here, at the end of this article, to convey a key point. The future is uncertain. Changes in climate can at times move far more slowly than we expect. They can also move far more rapidly. The most important thing for us to understand is that we don t know, for certain, what changes will come. We only know the range of possibilities. And at one end of that range, things may not be so terrible. At the other end of the range we have deep reason for concern.

Hope for the best, goes the English proverb, but prepare for the worst.

Let s hope the Arctic sea ice stabilizes, or reverses course. But let s not count on it. An ounce of prevention is worth a pound of cure. Every step we take to cut greenhouse gas emissions today is far easier than fighting the triple whammy we could be facing just a few years in the future.

References and Further Reading

Information about the state of the Arctic ice can be retrieved from the US National Snow and Ice Data Center (NSIDC) at In particular, for data on the coverage of Arctic ice, I ve relied on NSIDC s Sea Ice Index (

Sea ice volume data comes from the Pan-Arctic Ice Ocean Modeling and Assimilation System (PIOMAS) operated by the Polar Science Center at the University of Washington. You can access that data here:

The finding that human activity is responsible for roughly 60% of the Arctic ice melt from 1979 2011 is from: Stroeve, J. C., V. Kattsov, A. Barrett, M. Serreze, T. Pavlova, M. Holland, and W. N. Meier (2012), Trends in Arctic sea ice extent from CMIP5, CMIP3 and observations, Geophys. Res. Lett., 39, L16502, doi:10.1029/2012GL052676. You can find that paper here.

Long term reconstructions of Arctic ice coverage for the 1,450 years are from: Christophe Kinnard, Christian M. Zdanowicz, David A. Fisher, Elisabeth Isaksson, Anne de Vernal & Lonnie G. Thompson, Reconstructed changes in Arctic sea ice over the past 1,450 years, Nature 479, 509 512 (24 November 2011) doi:10.1038/nature10581. You can find that paper here.

Note that the graphic showing that trend is mine, and combines data from Kinnard s study with contemporary ice coverage data from NSIDC. Any error there is mine.

Peter Wadhams observations about the heating effect of a darkening Arctic have been repeated widely in the press. I know of no primary source in the literature, but one version of the math can be found in a blog post here. I find the math in the blog post generally right but wrong in specifics. An easier approach is to take the fraction of the Earth s surface covered by the Arctic ice cap in the height of summer (about 2% of the planet) and multiply that by the average insolation the region receives (170 180 watts / m^2) and then by the plausible change in albedo (perhaps 0.5). That gives a change in the amount of energy captured by the earth before taking into account clouds and such of about 1.75 watts / m^2 (averaged across the whole planet). That compares to 1.6 watts / m^2 of heating effect caused by humans via other changes to the earth system. The math is slightly different than Wadhams , but the answer is roughly the same a warming effect (a climate forcing in the parlance of the field) roughly as large as all current human-caused warming.

On the topic of permafrost, three important papers on the rate of permafrost thaw and the amount of carbon which could be released are Vulnerability of Permafrost Carbon to Climate Change: Implications for the Global Carbon Cycle by Edward Schuur and colleagues, Amount and timing of permafrost carbon release in response to climate warming by Kevin Schaefer and colleagues, and Accelerated Arctic land warming and permafrost degradation during rapid sea ice loss by David Lawrence and colleagues.

On the topic of methane hydrate deposits on the sea floor, a relevant don t panic yet paper is Siberian shelf methane emissions not tied to modern warming by Colin Schultz. A reminder that some periods of the last 10,000 years have been a bit warmer than our present, seemingly without triggering runaway explosive release of Arctic methane, can be found in Ice free Arctic Ocean, an Early Holocene analogue by Svend Hunder. (With the caveat that the planet seems to be well on its way past the high temperature marks of the early Holocene.)

Finally, the last figure in this article is adapted from a key paper by Julienne Stroeve and colleagues titled simply: Arctic sea ice decline: Faster than forecast. Because that paper dates from 2007, I ve updated the data for actual sea ice melt to bring it up to speed with the events of the subsequent 5 years. I ve left the IPCC projections alone, though for simplicity I m displaying only the mean of projections, and not the range. Any error in adapting and updating Stroeve s figure is mine.


Follow Scientific American on Twitter @SciAm and @SciamBlogs. Visit for the latest in science, health and technology news.
© 2012 All rights reserved.