by Christina Hulbe
Babylon had its Hanging Gardens, Rhodes had its Colossus, Alexandria had its Lighthouse, and the Arctic has its sea ice. The floating ice pack that defines the Arctic Ocean as we know it is a major player in Earth's climate system. It is an integral part of the Arctic Ocean ecosystem. It has been both a blessing and a curse to northern travelers through the ages. And it appears now to be diminishing very quickly.
A new review of the observed change in Arctic sea ice was published this month in the journal Science as part of a special issue timed to coincide with the start of the International Polar Year. Mark Serreze and his colleagues (Science, 2007, vol. 315; abstract) examined trends in monthly ice extent from 1979 through 2006 and found that for every month of the year there is a statistically significant downward trend in sea ice cover. Substantial interannual variability is superimposed on the downward trend but the trend is real. The decline has been most rapid for September, the end of the melt season, for which it is -8.6 +/- 2.9% per decade. Is this a global warming signal? And if so, what does the future hold in store?
what is sea ice?
Sea ice is frozen ocean water. The Arctic sea ice pack is often confused with ice from terrestrial sources (such as floating ice shelves or ice caps & ice sheets) but it is a distinct part of the cryosphere, with its own important role in climate and ecosystems.
Sea ice forms slowly. Dissolved salt lowers the freezing point of ocean water below that of fresh water, for example, -1.9 degrees Celcius for a salinity of 33 parts per thousand. This means that more cooling is needed to freeze salt water than to freeze fresh water. Second, salinity affects the temperature-density relationship of water. Fresh water has the unusual property that it reaches its maximum density above the temperature at which it freezes (changes phase from liquid to solid). This is why lakes overturn in the fall and spring and why ice floats. The salt content of ocean water complicates things because it increases the density of the water, such that cooling water may sink before ice crystals begin to grow. When sea ice does begin to form at the ocean surface, it insulates the underlying water from additional cooling.
As the sea surface begins to freeze, small (3 to 4 millimeter) frazil ice crystals form. Where we go from here depends on the sea state. In calm conditions, the frazil crystals accumulate in a soupy film called grease ice (photo), and over time knit together into a continuous expanse called nilas (photo). Once a continuous cover forms, the only way to grow more ice is freezing onto the underside, a slow process that produces congelation ice. Despite its slow growth rate, congelation ice usually makes up the bulk of a sea ice floe.
In rough seas, frazil crystals jostle together into circular disks of pancake ice (photos: young, well-developed, old). The pancakes have relatively thick rims due to collisions with other pancakes. If the wave action is strong enough, relatively thin floes will raft over each other (photo), thickening the ice cover. Where the ice is thick, collisions cause the sea ice to bend, fracture, and pile up into pressure ridges. Over time, individual pancakes cement together into large floes.
Sea ice grows through the winter, drifting according to the ocean and wind circulation. The return of warmer air temperatures in the spring starts the melt season (photo in a neat zoom sequence). Any new, first-year, ice that persists through the summer and into fall thickens again the following winter and is classified as multiyear ice. Sea ice in the Arctic basin (map) is quite different from ice around the Antarctic in that multiyear ice makes up a large proportion of the Arctic ice pack. This year-to-year persistence allows the old ice to grow quite thick, up to 4 or 5 meters (thicknesses around the Antarctic are 1 to 2 meters at most).
The seasonal cycle of new ice production and melting in the Arctic basin yields a maximum ice extent in March and a minimum extent in September. The typical range in ocean surface area covered is from a maximum of about 16 million square kilometers to a minimum of about 7 million square kilometers. This annual range also contrasts with the Antarctic, where the annual range is from about 18 to 3 million square kilometers.
why does sea ice matter?
Arctic sea ice is important to Earth's climate in part for its direct effect on the global energy balance. Its bright surface reflects most of the solar radiation it receives. This reflectivity is called the albedo of the surface. As sea ice cover is reduced, a very reflective material gives way to a very absorptive one, the dark ocean surface. The result is an increase in the amount of incoming solar radiation absorbed by Earth's surface and thus available to warm the atmosphere.
Sea ice also affects global climate through is its role in defining the characteristics of ocean water masses. As water freezes at the ocean surface, salt is left behind in the liquid, increasing the density of the remaining liquid and causing it to sink. Conversely, when sea ice melts, the shallow water is freshened. In the Arctic, much of that relatively fresh, cool water accumulates in the Beaufort Gyre, a wind-driven circulation north of Alaska that embraces an area larger than the Gulf of Mexico. The Beaufort Gyre is an essential part of the Arctic climate system. The low-density, cool water insulates sea ice from underlying warmer, saltier (and thus denser) water from the Atlantic. Freshwater flux from the Arctic could also play a role in the formation of North Atlantic Deep Water.
Sea ice also plays a fundamental role in Arctic ecology and in the livelihood of native peoples. The underside of the ice is a habitat for photosynthetic algae and a nursery for both invertebrates and fish. As the ice melts in spring it releases organisms and nutrients into the shallow ocean, promoting large ice-edge blooms that are a key component of the Arctic Ocean food web (examples of coastal and pelagic webs).
Satellite remote sensing provides the best data with which to establish trends for the entire Arctic basin. Satellite data are used to track the ice edge, estimate ice concentration (anything above 15% is classified as "ice covered"), and classify sea ice types. The high quality observation era began in late 1978 with the launch of NASA's Scanning Multichannel Microwave Radiometer. The Defense Meteorological Satellite Program launched the first Special Sensor Microwave/Imager in 1987. A very complete overview of sea ice and terrestrial snow observations is available here.
There is a statistically significant downward trend in Arctic sea ice cover from 1979 through 2006 in every month of the year. The decline has been most rapid for September, the end of the melt season, for which it is -8.6 +/- 2.9% per decade. Substantial interannual variability is superimposed on the downward trend. That variability offers insight into the processes involved in sea ice change over time.
Observations of sea ice thickness have also been made, via upward-looking sonar on submarines. Here too, the story is one of decline over recent decades, although the data are limited. Corroborative evidence, from native knowledge and changes in other components of the arctic cryosphere, also points to a significant downward trend in Arctic sea ice.
why the change?
The extent and thickness of Arctic sea ice change for both dynamic and thermodynamic reasons. Dynamic processes are those associated with the motion of the sea ice itself, due to wind and ocean currents. For example, if the rate of ice export out of the Arctic basin (almost entirely through Fram Strait; map) changes, so too does the Arctic ice pack.
Thermodynamic processes involve changes in air and ocean temperature, along with changes in the energy radiated to and from the ocean or sea ice surface (due, for example, to changes in cloud cover and surface type). These processes may be forced by global warming but may also have significant variability of their own. In order to really understand the sea ice signal, each component of the dynamic-thermodynamic system must be examined.
Interannual variability in Arctic climate is dominated by a strong atmospheric variability called the northern annular mode (NAM). The NAM is an oscillation of atmospheric mass between high and middle latitudes that results in distinct patterns of change in sea level pressure, temperature, storm tracks, and other weather phenomena.
It is possible that recent changes in Arctic sea ice are due to dynamic and thermodynamic consequences of the NAM. This possibility has been explored by a number of researchers and their results are summarized in the the new paper by Serreze and colleagues noted above (Science, 2007, vol. 315; abstract). In short, some, but not all, of the observed change in Arctic sea ice is associated with the NAM. A combination of greenhouse warming and atmospheric variability is needed to fully explain the observed changes from 1979 to the present.
Modeling studies of Arctic sea ice change were conducted for the Intergovernmental Panel on Climate Change's Fourth Assessment Report (AR4). As a whole, the models do a fair job of reproducing the observational record, with respect to both annual and seasonal variability. That said, there are significant differences among the models due to differences in how ocean circulation, clouds, and other processes are treated. This is a strength, as inspection of those differences allows climate modelers to gain insight into these complicated systems and improve model performance.
Overall, sea ice continues to decline through the 21st century and about half the AR4 models project September ice-free conditions by 2100. The sea ice does not vanish entirely though, ice still grows in the wintertime, although the thickness of the late winter ice pack declines.
News coverage of the sea ice story has emphasized the idea of tipping points beyond which positive feedbacks lead to abrupt change. This is a topic of some concern because the paleoclimate record tells us that abrupt change is possible. The guys at RealClimate have written a nice overview of tipping point hypotheses. Here, I'll focus on the "ice-albedo feedback" and a tipping point related to the NAM.
The ice-albedo feedback: As the atmosphere warms and the summer melt season intensifies, sea ice retreats, replacing its bright reflective surface with the dark ocean surface. The ocean surface absorbs relatively more incoming solar radiation and is warmed by that energy. The warmer ocean surface radiates that energy back to the air above, amplifying the initial warming effect. This delays the return to ice growth in the autumn, yielding a relatively thinner, and more quickly melted, ice pack the following spring. It's not quite that simple, though, because exposing the ocean surface can enhance evaporation and perhaps cloud formation, something that would boost the polar albedo. A nice summary of the two effects can be found here.
A tipping point: One possibility is that an Arctic sea ice pack thinned by global warming could be vulnerable to anomalous ocean heat transport associated with the natural variability of the northern annual mode. Such variability is observed and affects the layer of cool, low density water that separates the Arctic sea ice from relatively warm north Atlantic waters (as discussed above). Abrupt decline to September ice-free conditions by 2040 is found in one coupled climate system model (technical discussion; popular, with movie). Does this mean that such a transition is likely? The short answer is that nobody knows, but it is a possibility. But really, extreme events may be a bit of a red herring. We don't need extreme events to be concerned about change in the Arctic sea ice.
The current decline in Arctic sea ice is in part a product of global warming, a process that will continue into the foreseeable future. All of the climate models tasked to study this phenomenon project future declines in Arctic sea ice extent and most of them indicate that the rate of loss will increase in the decades ahead. This change will have far-reaching consequences for Arctic ecosystems and northern hemisphere climate (technical abstract).