- A leaked version of the newest science report from the U.N. Intergovernmental Panel on Climate Change warns of looming, potentially catastrophic tipping points for Arctic sea ice melt, tundra thaw, savannification of the Amazon rainforest, and other planetary environmental thresholds beyond which recovery may be impossible.
- But what are tipping points, and how does one pinpoint what causes them, or when they will occur? When studying a vast region, like the Arctic, answering these questions becomes dauntingly difficult, as complex positive feedback loops (amplifying climate warming impacts) and negative feedback loops (retarding them) collide with each other.
- In the Arctic, one working definition of a climatic tipping point is when nearly all sea ice disappears in summer, causing a Blue Ocean Event. But attempts to model when a Blue Ocean Event will occur have run up against chaotic and complex feedback loop interactions.
- Among these are behaviors of ocean currents, winds, waves, clouds, snow cover, sea ice shape, permafrost melt, subarctic wildfires, aerosols and more, with many interactions still poorly understood. Some scientists say too much focus is going to tipping points, and research should be going to the “radical uncertainty” of escalating extreme local events.
For Arctic scientists, the summer of 2007 changed everything. That’s when, for the first time in history, record warmth melted the Northwest Passage, nearly opening it to shipping; turned a portion of the East Siberian Sea the size of Mexico into open ocean; and shrank the polar ice cap to a size never before reached so early in the March-to-September melt season, as documented by satellite since 1979.
Walt Meier, a senior researcher at the U.S. National Snow and Ice Data Center (NSIDC) in Colorado, was following the satellite images. “You’re watching and it’s neck and neck. Is it going to make it? Is it not going to make it? It’s like, we’re going to break this [annual low sea ice extent] record in August if it keeps up like this.”
By mid-September, when the melting had halted and the Arctic Ocean had begun its annual refreeze, the ice cap was 22% smaller than it had historically ever been. “We had set a record in 2002, 2005 before that,” says Meier. “Every couple of years we’ve kind of been dropping just a little bit lower than anywhere we had been before. But 2007 just smashed things.
He adds: “There was this thought that we have reached a tipping point.”
Tipping point. The expression has become a foundational concept in climate change science discussions and a mainstay of media headlines. Just last month, a leaked version of the newest upcoming science report from the United Nations Intergovernmental Panel on Climate Change warned of looming “tipping points” for Arctic, Greenland and Antarctic ice; subarctic permafrost melt, savannification of the Amazon Rainforest, and other planetary environmental thresholds beyond which recovery may be impossible. “Life on Earth can recover from a drastic climate shift by evolving into new species and creating new ecosystems … humans cannot,” warns the draft.
But what is a tipping point, and how do you know when you’ve reached one? And if you haven’t yet, how do you know you will? The term was popularized around 2000 by author Malcolm Gladwell, and has been applied since to everything from cultural studies to criminology to Arctic ice. The meaning shifts depending on source, but is often explained with the “ball in a basin” analogy: if you nudge the ball, it will tend to return to the bottom of the basin, a steady state. But nudge it hard enough, and it falls over the basin lip, down into another very different basin, or onto the floor, in a precipitous, irreversible change of state.
The challenge of defining the Arctic tipping point
While the ball-in-a-basin metaphor is neat and simple, things get messy and chaotic when assigning the concept to an entire Earth biome or region. In the Arctic Ocean, for example, the theory goes that, once global temperatures reach a certain high level, or sea ice extent reaches a certain minimum, or some other threshold is reached, a complete loss of sea ice is inevitable, bringing transformative repercussions to the Arctic, and possibly to climate around the globe.
But many polar scientists say the tipping point concept is at best an oversimplification of how sea ice works. At worst, it is a distraction from metaphors or concepts that are not only more descriptive of the science, but more powerful in preparing policymakers and the world’s people for a hotter future.
“I prefer to avoid the idea,” says David Barber, director of the Centre for Earth Observation Science at the University of Manitoba. “My view on it is that it’s more of a public thing. It’s more for explaining complex relationships to the public.” Those complex relationships are worth examining far more deeply, as they are key to understanding how and why the tipping point idea may, or may not, fit the case of Arctic ice — and, by extension, other biomes.
Feedback loops, within feedback loops
To explain why the tipping point concept has been so appealing, it’s maybe best to begin with an event that helped define it. The Arctic summer sea ice minimum that Walt Meier watched so assiduously back in 2007 was a key indicator of global warming — and a ringing wake-up call.
As Earth heated up, polar ice had been melting back progressively since before 2005. Climate models predicted this decline and observations in 2007 confirmed it. But sea ice minimums are noisy events, fluctuating wildly year to year. Still, the trend line was clear. The steep drop in 2007, followed by a further 20% drop in 2012, sparked a reappraisal of the Arctic’s future.
What once looked like a linear curve downward in sea ice extent now plunged in a way that augured the imminent loss of all or most Arctic sea ice, a so-called Blue Ocean Event. There seemed to be a shift in the physics of the system. That hypothesis prompted prominent Arctic researcher Peter Wadhams to predict that Arctic summers could be entirely ice free by 2017. He was wrong, it turned out, but some Arctic watchers continue to predict an imminent Blue Ocean Event or some equivalent major state change.
The physics behind the deep dip in ice extent, which is prompting such uncertainty, is marked by a series of complex positive feedback loops — processes that amplify change, potentially turning it from linear to exponential. Positive feedback loops are everywhere in nature. Why do apples all seem to ripen together, for instance? A single ripening apple in a basket releases the hormone ethylene, stimulating the ripening of nearby apples, which release more ethylene, which stimulates other apples, until all rush to ripeness.
In the apparent simplicity of the Arctic Ocean (a region dominated only by ice and water, transformable by just a nudge either side of -1.8° Celsius, or 28.8° Fahrenheit, the freezing point of seawater), feedback loops can have an outsize impact. To model those effects, all that should be needed is an understanding of those loops.
The best-known and largest is the albedo effect, albedo being a measure of reflectivity. White ice has a high albedo, reflecting more solar radiation back out into space, and keeping the Arctic cool. But exposed darker ocean absorbs it, storing heat that needs to be dissipated in the fall before the water can refreeze again. (A recent paper estimates that the open Chukchi Sea off Alaska alone accumulates enough heat in summer to warm the Arctic troposphere by 1°C, or 1.8°F.)
It’s a simple rule: as more and more ice melts and more Arctic Ocean is exposed, the warmer the water becomes, in a self-reinforcing cycle, until at some point a Blue Ocean Event happens.
But that simplicity is deceptive. In fact, Arctic feedback loops are like wheels within wheels within wheels, sometimes amplifying each other, in the form of positive feedback, or diminishing each other, as negative feedback.
The majority of Arctic melt feedback loops identified to date seem to be positive. For example, it’s theorized that warmer Arctic air temperatures alter the path of the polar circulating atmospheric jet stream, potentially drawing more storms to the region, which break up sea ice. Likewise, warmer North Atlantic Ocean waters are intruding into the Arctic, an effect called Atlantification.
Other Arctic positive feedback loops: receding ice exposes the sea to more wind, which can churn warmer, heavier Atlantic water to the surface, melting yet more ice. In turn, thin first-year ice disappears every summer and now rarely has a chance to become more stable multiyear ice, resistant to wind and melt. Meanwhile, once-strong ice dams, known as ice arches, forming annually at the north and south ends of the narrow Nares Strait between Greenland and Canada are collapsing about a week earlier every year on average, allowing multiyear floes that build up behind them to escape south, where they melt, creating more open water, and more warming.
But that’s just a start: There are other feedback loops whose effects haven’t been well quantified. For example, as global warming thaws Arctic permafrost on land, the frozen vegetation it contains rots, releasing CO2 and methane, greenhouse gases that could warm the Arctic further.
Likewise, as the polar region warms, snow cover disappears earlier, decreasing albedo and warming the Arctic further. Increased warming brings more subpolar vegetation, along with greater risk of fires, which can deposit soot on the remaining sea ice, reducing albedo further. Studies have also found that climate change has raised moisture and energy levels in the Arctic atmosphere, causing an increase in Arctic lightning strikes, which helped ignite the record wildfires in Siberia, Scandinavia, Alaska and northern Canada in the last four years.
Still, there’s much we don’t know. Recent research is turning up ever more esoteric feedback loops. Barber’s research, for example, specializes in a phenomenon known as frost flowers, delicate ice crystals that spring up from the briny surface of newly formed sea ice.
“When I started my career in the ’80s, frost flowers were not a big thing in the Arctic because it was mostly multiyear sea ice. But now that it’s covered by this first-year ice, there’s big, huge areas where these things start to grow in the winter,” says Barber. “You can have literally miles and miles and miles of these things that form in these young ice covers.” They trap snow, which insulates ice against spring warming, but also possibly slowing autumn cooling — potentially acting as both a positive and negative feedback loop.
Clouds have proven to be one of the most confounding of feedback loops, sometimes causing, or sometimes inhibiting, warming, depending on altitude and other factors. Like ice, clouds can cool the Arctic by reflecting the sun’s radiation back into space. But they can also act as a blanket, trapping long wave radiation emitted upward by the ocean, a positive feedback that warms the region.
Which effect prevails depends on when and where the clouds form, and even how. Clouds condense around tiny airborne particles called aerosols. Until recently aerosols were thought to originate from outside the Arctic, as dust and soot blown in from the south. But new research by a U.S.-Swedish team has found that iodic acids from sea spray kick-start reactions that produce aerosols. The reactions are subtle, and still poorly understood, but crucial to understanding Arctic climate change. “You only have a few [aerosol] particles per cubic centimeter during certain days of the year,” says Paul Zieger, a physicist at Stockholm University and one of the study contributors. “It’s such a pristine area that you can easily change this delicate balance of particles, forming clouds.” Adding loops within loops, it appears that frost flowers may also contribute to aerosol formation.
‘Brakes on the system’
With so many positive feedback processes compounding each other, it may come as no surprise to learn that the Arctic is warming at twice the global average, a phenomenon known as Arctic or Polar Amplification. You might even expect the rate of sea ice melt to be greater, exponential in fact. Back in 2007, that seemed a reasonable expectation.
But despite all the feedbacks amplifying warming, the rate of melting isn’t accelerating; if anything it has slowed. The U.S. National Oceanic and Atmospheric Association (NOAA) identifies three periods in sea ice minimums: From 1979 to 1992, sea ice loss declined about 6% per decade. From 1993 to 2006, the decline accelerated to 13.3%. But from 2007 to 2020, the slope softened to just 4%.
Extrapolating future decline rates is notoriously difficult, since the range of weather variation and other chaotic events is almost as strong as the overall trend. But in a 2017 report, the Arctic Monitoring and Assessment Program (AMAP) of the Arctic Council felt confident enough to declare: “There is no evidence for an internally caused tipping point (a run-away condition based on internal climate physics) of rapid change.”
“If we’re slowing down on the [rate of Arctic sea ice loss and] warming, there must be some negative, some brakes on the system,” explains John Walsh, chief scientist at the University of Fairbanks Climate Adaptation Science Center and an AMAP report co-author. “From the Arctic’s perspective, there may be some reason to think that the warming is not going to just take off and get completely out of control.”
Just as there are positive Arctic feedback loops that reinforce warming, there are also negative loops that retard it. One of these, says Walsh, is the Planck effect, a phenomenon by which warmer regions radiate disproportionately more heat to space than cooler regions. So as Arctic temperatures climb, the amount of heat lost to space increases faster. Aerosols, as already mentioned, can also increase solar heat reflection to space, having an Arctic cooling effect.
Another potential de-amplifying loop: the difference in temperature between polar regions and southern latitudes drive currents in the North Atlantic, which bring warm water from the tropics north. There it cools in the colder air, and returns at depth. As the north-south temperature difference declines, however, the currents slow, bringing less warm water poleward. (Lest this be mistaken for good news, the slowing of these currents, knowns as the Atlantic Meridional Overturning Circulation, could eventually be disastrous on a global scale, as widely covered in the media this year, maybe eventually causing Europe and the U.S. Northeast to cool drastically.)
Melting may also have slowed simply because the thickest multiyear ice, four years and older, that used to dominate the Arctic is now gone, accounting for a mere 1% of ice today, most of it in the highest latitudes. “We’ve gotten rid of a lot of the quote-unquote easy ice to melt away,” says Walt Meier. The vast majority today is fragile younger ice that melts out each summer, but at least returns each year with the winter freeze.
Many words for ‘ice’
Tipping point or no, Arctic ice is definitely in trouble. The rate of melt may be slower, but the direction points downward. Sea ice minimums over the last 12 years are the 12 lowest on record. 2020 saw the second-lowest minimum ever. And predictions for what is left are dire — if frustratingly vague. Models suggest the Arctic could be ice-free in the summer any time between the 2030s and the end of the century.
Why the uncertainty, when 15 years ago some scientists confidently predicted ice-free summers within the decade?
One possible reason: Arctic ice is not nearly as featureless as all that whiteness would make it appear, says NSIDC researcher Matthew Druckenmiller. In fact, it is actually highly variable: “If you fly over the Arctic — and any month of the year, you’re flying over sea ice — you’ll see that sea ice is a very heterogeneous material. It’s broken, it’s ridged, snow drifts in certain places, and doesn’t drift in others.” Those fine-grained details can affect how something as simple as albedo is calculated. “You can have a process modelled perfectly, but how you actually assign those parameters can lead to huge changes in the model output,” he explains.
Northern Indigenous people have long understood these nuances. The language of the Iñupiat, who live along the Alaskan side of the Bering Strait, have words for ice attached to the shore, ice floating freely, for thin ice skims forming atop open water, for soft melting ice, for ice ridges rising up to 12 meters (39 feet) high where floes collide, and for ice crystals floating in the air, among many others. More than linguistic curiosities, each of these features plays a complicated role in the Arctic climate — from capturing carbon, to reflecting solar radiation, to helping generate clouds.
But the satellite imagery that most climate researchers rely on shows none of this. Our current knowledge depends on models barely beginning to capture the region’s complexity, leaving major gaps in our forecasting ability. As cloud researcher Paul Zieger notes, the possible error range associated with cloud and aerosol impacts on Arctic temperatures is larger than the estimated impact itself.
Why don’t we know more? “It’s difficult to get there,” he laments. “It’s quite expensive.”
Living in an age of ‘radical uncertainty’
More data and better climate models may one day help predict precisely when the plotted sea ice curve will cross the x-axis, creating a Blue Ocean Event, or maybe not. But that may be missing the weather for the warming, says James Overland, of NOAA’s Pacific Marine Environmental Laboratory. Just as the Arctic is warming faster than the rest of the globe, weather extremes — the random ups and downs natural to any chaotic system — are getting larger too. Where scientists previously discarded extremes as noise in the system, Overland says they are the system.
“We weren’t really concerned about [local weather extremes] even two years ago,” he says. “And they’re very hard to pin down. They can accelerate overall, and they’re not very well handled by the climate models that are used to project out 50 or 80 years.”
As warming rises, tundra collapses, vegetation is left uncovered by snow, shorelines are exposed to waves, and as open ocean is exposed to extreme wind, extreme weather takes a greater toll through droughts, wildfires, storms and flooding. Unlike long-term trends, however, these extreme effects are local, striking different places harshly at different times — and doing terrible damage. In January 2021, Alaska and northern Canada were unseasonably warm and wet, while Siberia was much colder than normal. This summer, the Siberian town of Verkhojansk, one of the coldest places in Russia, hit a record high of 48°C (118.4°F). In the last decade, all these regions have suffered devastating wildfires, though in different years.
However, because these local weather extremes are so geographically restricted, they can fall through the cracks of large-scale observational data and predictive models. Moreover, the complex interactions between these fluctuating local extremes are far from understood.
For Overland, these interactions are what he calls “radical uncertainty,” a whole new physics. It is as if the ball in the bowl were to roll half way up to the edge of the basin, when the basin suddenly turns into a shag carpet — extremes making accurate forecasts fantastically difficult.
For this reason, he says, scientists, policymakers, Arctic residents, and the rest of the world should focus less on long-term trends — obsess less about approaching tipping points — and start planning for a multitude of extreme scenarios, any of which could strike anywhere tomorrow. Tipping points, it turns out, may be more of a human construct, with little connection to nature’s utterly unexpected, chaotic and unpredictable creations.
Echoing a sentiment from 2007, though for different reasons, Overland says: “We’re in a different ball game from where we thought we were a couple of years ago.”
Banner image: The Swedish icebreaker Oden close to the North Pole during the U.S.-Swedish Arctic Ocean 2018 expedition. The icebreaker was moored to an ice floe for about five weeks, where an ice camp was established to perform various observations. Tethered balloons were used to record profiles of atmospheric variables and to sample aerosol particles and clouds. Image by Paul Zieger.
FEEDBACK: Use this form to send a message to the author of this post. If you want to post a public comment, you can do that at the bottom of the page.