THE BIG THAW
Frozen soil is thawing all over the arctic, with consequences that are potentially destructive and difficult to predict. But predictability is key to any plan of action, and Los Alamos is taking up that charge.
There are few places left on Earth where human footprints are not evident. Whether such footprints take the form of clustering skyscrapers, power lines across the desert, or congested highways, few places remain untouched by modern civilization. The Arctic is one of these places—or at least, it was, until the world's warming climate started to alter it. Now, lake-sized sections of tundra are collapsing into sinkholes and sliding from riverbanks to be washed away. These changes are not the same as parking lots and shopping malls popping up all over the Arctic, but they are consistent with our use of fossil fuels.
Global climate models show us that as the Earth warms, the Arctic warms faster than everywhere else. That's partially because the Arctic climate is influenced by the rest of the northern hemisphere and warms when the land warms (and most of the world's land is in the northern hemisphere), but mostly it's due to the Arctic seas. When white, sunlight-reflecting sea ice melts, it is replaced by darker seawater, which accelerates the melting because it absorbs more heat from the Sun. By 2030, when the global average temperature is predicted to rise by approximately one degree Celsius, the Arctic temperature is predicted to rise by approximately two degrees. And by the end of the 21st century, the Arctic is projected to be about six degrees warmer than it was when the century began.
Moreover, the Arctic is more vulnerable to warming-induced change because the vast majority of its land is permafrost. Temperate soil is naturally damp, and Arctic soil is naturally frozen, containing enormous chunks of ice in some places and small bits of ice intermixed with soil particles and pebbles in others. Either way, the frozen soil is ice-hard.
Most of the Arctic's continuous permafrost (blue) is likely to stay frozen as the climate warms throughout this century. Other areas are much more susceptible to thawing. CREDIT: National Snow and Ice Data Center, 2001
The term "permafrost" is reserved for any ground that has been continuously frozen, summer and winter, for the last two years (see map above). The longevity of permafrost is strongly influenced by not only its latitude, but also local factors, such as insulating vegetative cover or a north-facing hillside. Some tracts of permafrost have thawed and refrozen one or more times since a previous ice age, while other tracts remained continuously frozen for more than 100,000 years.
When permafrost thaws, there are serious impacts. Where large pockets of ice melt away, the ground collapses. On steep slopes, the surface soil slides away. Erosion patterns and water flows can change with the thaw, potentially accelerating the thaw rate. But the greatest impact of the thaw is underground, where bacteria and other microbes begin to multiply and migrate into the newly unfrozen soil. Bits of dead plant matter that have been safely stored in the frozen soil are now available for these bacteria to metabolize. The byproducts of that metabolism include carbon dioxide and methane—both potent greenhouse gases—and there's virtually nothing we can do to prevent these gases from reaching the atmosphere.
Estimates peg the amount of carbon stored in Arctic permafrost at a staggering 1.6 trillion metric tons. That's about twice the amount of carbon currently stored in the planet's atmosphere. Depending on how fast the permafrost thaws, and how efficiently its carbon is released into the atmosphere, the rate of Arctic carbon released by bacteria could rival that from all human-made sources. And unless some counterbalancing phenomenon arises to save the day, such as plants and other organisms removing the excess carbon from the atmosphere, we are all but guaranteed a worst-case scenario from among the range of global warming projections. Thus, climate scientists want to know two things about the thawing Arctic permafrost: how much and how fast?
Under the Surface
Permafrost soils reside just underneath the top layer of soil, known as the active layer. The active layer thaws every summer and refreezes every winter. In the northernmost reaches of the Arctic, the permafrost is hundreds of meters deep, underlying an active layer that's generally less than half a meter thick. Farther south, in the subarctic zone, 20–100 meters of permafrost typically support 1–2 meters of active layer. In some places, the full depth of permafrost has already thawed.
As the atmosphere gets warmer, the active layer warms up to the surrounding air temperature. This causes the permafrost underneath to thaw faster, partly because of conduction—the form of heat transfer that results when a warmer object (the active layer) is in direct contact with a cooler object (the permafrost just below). However, conduction is not the only way to transfer heat. There is also advection, which occurs when a fluid flows over an object. For example, a wind chill occurs when cold air flows over warm skin, causing heat to flow out of the skin and into the air. (Anyone who has ever experienced a wind chill knows how significant advection can be.) If a warmer fluid flows over a cooler object, the effect is reversed: heat flows from the fluid to the object.
In the case of thawing permafrost, advection comes into play whenever the permafrost makes contact with flowing water, such as groundwater. Even meltwater, with a temperature just barely above freezing, can transfer heat to the permafrost and increase the rate of thawing. So the fate of Arctic permafrost is not just a question of temperature, it's also a question of hydrology, the movement of water.
Consider, for example, any one of the many lakes found in the Arctic lowlands that occupy much of northern Canada, northern Siberia, and central Alaska. Only the upper two meters of the lake actually freeze in winter; the cold air can't freeze the deeper water. But the lake is nestled into the surrounding permafrost, which has been frozen solid for hundreds of years, if not longer. It is perhaps somewhat counterintuitive that this arrangement—without moving water—is normally stable: The permafrost is unable to freeze the lake, because drawing heat from the lake would raise the temperature of the permafrost, rendering it too warm to draw more heat from the lake. Similarly, the lake is unable to thaw the permafrost, because transferring heat to the permafrost renders the lake too cold to heat up the permafrost any further. It is a delicate equilibrium that the warming climate threatens to disrupt. A very small increase in temperature could tip the balance, because even a trickle of meltwater is enough to bring advection into play.
As a result of the climate warming, perennially warmer lake water starts to thaw little pockets in the surrounding permafrost. Pores develop in the frozen soil. Over time, the pores link together into channels flowing with water. These channels, small at first, grow and multiply and potentially mingle with the flow of groundwater below. The slightly warmer lake water heats up the cooler groundwater, affecting areas downstream. In this way, changes to the region's hydrology allow advection to degrade the permafrost much more rapidly than conduction alone.
In turn, damaged permafrost can lead to the formation of additional lakes. Whenever a large pocket of ice-rich permafrost collapses, it creates a depression in the ground known as a thermokarst. If that depression subsequently fills with water from surface flows, precipitation, or thawing permafrost, the result is a new lake, which could damage the surrounding permafrost further.
Thermokarst lakes form in the depressions that arise when large chunks of permafrost thaw, collapsing the ground surface. Past and present thawing episodes have left such lakes covering much of the Arctic landscape. Lakes and other surface water features can accelerate the damage to the underlying permafrost by creating pores and channels that could potentially connect to the groundwater below.
But warming lakes are not the only landscape effects of concern. For example, different depths of snow in winter and different types of vegetation throughout the rest of the year alter the insulating properties of the ground surface. As the Arctic landscape becomes warmer, the surface vegetation tends to shift from mosses to shrubs, and the full effect of this shift is not yet known. Because shrubs are darker than mosses, they absorb more heat from the Sun in summer, but they also shade and cool the ground. In winter, shrubs anchor a thicker and more insulating (warming) snow pack. But their dark, bare branches warm quickly and hasten the snowmelt, leading to less insulating (colder) ground conditions in the spring. Vegetation changes like this can begin with changing patterns of weather, erosion, and wildfire, all of which increase with a warming climate. In other words, a small increase in temperature could trigger a variety of processes—thermokarsts, erosion, changes in vegetation, lake-related advection, and others—to arise and contribute to the thaw.
At Los Alamos National Laboratory, a team of scientists is working to understand how local changes in hydrology might bring about major changes to the Arctic landscape, including the possibility of a large-scale carbon release from thawing permafrost. First, they must understand the local changes—from tiny, thawed pores in the permafrost to entire watersheds—before they can assess how the widespread occurrence of these local changes might become a planetary concern.
Bryan Travis, an expert in fluid dynamics, and Joel Rowland, an expert in hydrogeology and permafrost-dominated river systems, are performing Laboratory-directed research and development to understand Arctic hydrology. Travis is author of the Mars global hydrology numerical computer model, or MAGHNUM, used for calculating heat and fluid transport phenomena. (MAGHNUM was previously used to model hydrological phenomena under freezing conditions on other planets, including Mars.) Travis advanced the MAGHNUM software with a variety of improvements and additional components into a new program, called ARCHY, a comprehensive Arctic hydrology model. Travis and Rowland work with Cathy Wilson, a Los Alamos scientist who specializes in predictive hydrology and geomorphology and heads up several Laboratory efforts on Arctic warming and climate. Together, the team's goal is to make ARCHY capable of accurately modeling Arctic topography, thawing, and erosion. Because it includes advective heat transport, ARCHY will help to predict how quickly and how extensively the Arctic permafrost will thaw.
Travis has already used ARCHY to demonstrate the relative importance of advection. The figure below shows the predicted temperature gradients beneath an Arctic lake under two modeling scenarios. Without advection, 90 years of conduction leads to only minor thawing of the permafrost immediately supporting the lake. But when advective effects are included, the same 90 years is enough time for lake water to tunnel downward and connect with the groundwater flow, warming it by about one degree Celsius. The warmer groundwater then accelerates the demise of the permafrost downstream.
"What conduction alone would take a century to do happens in a few decades when advection is included as well," Travis says. Wilson adds, "It's clear that these hydrological processes need to be represented in the next generation of climate models if we hope to predict the impact of climate change on the rates of permafrost thaw and carbon release in the Arctic."
ARCHY must also be able to apply its thawing predictions to the bacterial processing of carbon in the permafrost. This, too, adds interesting and important complexity to the problem. Some bacteria metabolize carbon-rich matter in the soil aerobically—making use of oxygen from the air. They convert organic forms of carbon in the soil to carbon dioxide (CO2), which can then enter the atmosphere. Other bacteria are capable of converting organic carbon into methane (CH4) under anaerobic conditions, such as when the soil is so damp as to limit the available oxygen. Both CO2 and CH4 are major greenhouse gases, but CH4 has a substantially greater warming effect; therefore, it is important to know the ratio of the two emission rates.
Because the emission of CH4 results from anaerobic processes in wet, oxygen-deprived soils, it is enhanced by rain (which is predicted to increase with global warming) and watery surface conditions. If, in addition, thawing creates more bogs and swamps, the greenhouse gas ratio will include more CH4. The ratio will also vary seasonally, because some anaerobic metabolism continues even under partially frozen conditions, while aerobic metabolism does not. All of these processes must be accounted for if ARCHY's output is to accurately capture the Arctic's effect on global climate change.
Computer modeling reveals the importance of the advective—that is, fluid-flow-based—mode of heat transfer. Each frame shows an underground vertical slice through the same Arctic lake embedded in the permafrost after 90 years of simulated time evolution. The simulation on the left allowed only conductive heat transfer—no moving water. At right, water was allowed to flow through thawing soil pores, inducing advective heating. With advection, the lake water reached the groundwater, flowing right to left about 50 meters below the surface, and more rapidly thawed the surrounding permafrost. The numbers on the axes are distances in meters, and the color scale spans temperatures from -2 (blue) to +4 degrees Celsius (red). The liquid-solid (unfrozen-frozen) boundary in both frames is the solid, zero-degree contour line. Dashed contour lines indicate below-freezing temperatures.
Evidence for the big thaw is already written large throughout the Arctic landscape. In addition to thermokarsts and other varieties of collapsing land, much of the Arctic is characterized by patterned ground (see below photograph), where large subsurface ice wedges form the boundaries between segments of soil shaped like polygons. These features form in regions where the active layer and permafrost shrink and crack under cold winter temperatures. In the spring, surface water seeps into the cracks and freezes in the still-frozen permafrost. Because water expands as it freezes, the original crack must widen each time new water is frozen into it. When climate conditions warm, the temperature of the shallow permafrost is too high to quickly freeze the new water, and the ice wedges below start to melt. Eventually the soil over the melted ice wedge collapses, forming warm ponds that cause even more permafrost thaw and ice wedge melt.
Patterned ground appears where large ice wedges form beneath cracks in the surface soil. These cracks widen whenever surface water enters the cracks and then freezes. But with climate warming, the water takes longer to freeze, allowing enough time for the ice wedges below the cracks to melt. This leaves a void into which the surface soil collapses. Water gathers in the collapsed depression and accelerates the thaw.
Perhaps the most striking landscape change occurs when enormous chunks of land disappear entirely. On the moderately steep banks of the Selawik River in Alaska, for example, Rowland studies a phenomenon known as a retrogressive thaw slump (see below photograph). The thawing permafrost destabilizes the ground, causing it to break off and slide downhill into the river, whose current carries the soil downstream. It is retrogressive because it progresses backwards: permafrost left behind after a slide is suddenly exposed, rendering it vulnerable to faster thawing and destabilization. The thawing land continues to break off and slide away, year after year. How far it progresses and what causes it to eventually cease are not yet known. The Selawik slump, for example, started in 2004 and has receded about 15 to 20 meters every year since.
Retrogressive thaw slumps, like this one along Alaska's Selawik River, form when thawing permafrost weakens the land, causing it to slump as the former terrain runs off into the river. The exposed soil at the top of the slump is thawing now, and will probably break away again. At other sites, this cycle has been observed to repeat for up to 50 years, widening up to a kilometer from the riverbank.
"Beyond the obvious effect on the local hillside, the impact of a retrogressive thaw slump can be felt throughout the river system," Rowland points out. "When that much sediment is added to a river system all at once, there can be important consequences downstream." Indeed, the runoff from the Selawik slump flows downriver to a fishery. If too much sediment from the slump accumulates in the fishery, it affects the river's sheefish population. Like salmon, sheefish need clear, sediment-free water to oxygenate their eggs, so a murky river bottom threatens not only a fish population, but a local village's livelihood as well.
Additionally, river currents can deposit excess sediment as new sandbar islands in the middle of the river. The warming climate favors the rapid colonization of these islands by shrub vegetation, which tends to stabilize the soil and give the islands some permanence. The islands constrict and accelerate the river's flow. While earth scientists see no particular catastrophe as a result of this, they do expect some outcomes—affecting boat navigation, fish habitats, and nutrient loads, for example—even far downstream where the river meets the ocean.
Arctic rivers are experiencing other changes, too. Rivers naturally migrate over time, shifting locations as their riverbanks erode. (See below photograph, and satellite image below.) Rowland and others have found that river migration in the Arctic, unlike that in other regions, depends more on the temperature of the water than its flow rate because Arctic riverbanks are made of permafrost, which must partially thaw before they can be significantly eroded. So global warming is likely to increase the rates of Arctic river migration and floodplain erosion, and thereby provide additional mechanisms for releasing long-frozen carbon into the atmosphere and oceans. Thus, river migration and other sediment redistribution processes join the list of Arctic climate phenomena that Los Alamos scientists seek to understand and ultimately predict.
The rate of riverbank erosion in temperate and tropical climates is primarily determined by the water flow rate, with larger flows washing away more soil. However, the rate of erosion of an Arctic riverbank, like this one along Alaska's Yukon River, is primarily determined by the water temperature, with warmer water causing more damage to the permafrost underlying the active layer.
This 2009 satellite image of Alaska's largest river, the Yukon, has been overlaid with the same stretch of river as it appeared in 1974. By comparing the river's recent location (lighter blue) with its location 35 years earlier (darker blue), one can observe the river's migration and a new population of vegetated islands, spawned by the increased sediment deposition from the destabilization and erosion of permafrost upstream. Satellite image courtesy of GeoEye
Should it prove critical to slow down the thaw, Cathy Wilson, Joel Rowland, and Bryan Travis are holding on to a last-ditch option.
Computer models predict global climate change by breaking the world's land, sea, and atmosphere into manageably small blocks—but not so small that the computers can't juggle them all. Unfortunately, the scale of the Arctic hydrological changes in progress—to lakes, rivers, and underground flows—is often much smaller than a single block. How, then, can the Arctic's important hydrological effects be incorporated into global models? How can one aggregate a widespread pattern of collapsed and fractured ground? What thawing effect can be expected from a cluster of a thousand lakes? Arctic researchers at Los Alamos are studying various Arctic phenomena to address these questions. Success will mean characterizing all the small-scale changes in a way that can be accurately and conveniently represented on the global scale.
The Los Alamos researchers are deploying sensors and reviewing other data to address many pieces of the overall puzzle—pieces that will be assembled in the next few years. Will a thawing region tend to deform in a way that concentrates the surface water into streams, thereby drying out the rest of the landscape? Los Alamos is examining this with high-resolution topographical data obtained by lidar, a laser-reflection measurement system. Will lakes expand and merge with more meltwater, or will they drain through underground channels into the groundwater? Los Alamos is addressing these questions by training a sophisticated computer algorithm to recognize the changing surface water distribution in satellite imagery of the lake country (above). Such information, once aggregated, will be used to inform and improve scientists' understanding of the global climate.
The degree of surface damage the Arctic will ultimately suffer is not yet known. How much carbon its thawing permafrost will release into the atmosphere and how soon it will happen are not yet known. It's not even clear that the carbon release poses a problem; perhaps the Arctic ecosystem will surprise us with some way of consuming it. What is clear is that the Arctic temperature is rising, its permafrost is thawing, and its surface is crumbling. Scientists have observed a 200 percent increase in surface scars (thermokarsts, active layer erosion features, and thaw slumps) over the last 25 years. The Los Alamos work helps link these changes into a deeper understanding of the global climate system.
Data from instrumentation on the ground and in space are needed to constrain the range of climate and landscape changes expected above the Arctic Circle. Computer-based algorithms mine satellite images for subtle signatures of widespread change in the Arctic. Images like this one of the many lakes of the Lena River delta in Siberia help researchers to identify changes over time and their impacts on local hydrological patterns. Inset: This ground-based sensor system measures rainfall, solar radiation, air temperature, humidity, soil temperature, and soil moisture.