Skip to content

Scientists have always thought that the vast majority of ice contained in the Antarctic ice cap was formed from frozen precipitation. Recent research has revealed that this is not totally correct. Over a large fraction of East Antarctica, the deepest part of the ice sheet contains ice that did not originate as surface snow but developed when subglacial meltwater was frozen onto the underside of the ice sheet. The amount of ice involved is much larger than the estimated volume of Antarctic subglacial lakes and may even exceed the volume of all glaciers on Earth outside of the two polar ice sheets. Current computer models predict that subglacial water escapes toward the ocean. These new findings indicate that water from areas of basal melting actually migrate to areas of basal freezing, something not accounted for by current ice sheet models. To scientists’ surprise, the East Antarctic Ice Sheet is getting thicker from the bottom up.

According to a Science report published online March 3, 2011, by Robin E. Bell et al., little is known about processes at the base of the ice sheets. In “Widespread Persistent Thickening of the East Antarctic Ice Sheet by Freezing from the Base,” researchers report that, although the surface accumulation of snow remains the primary mechanism for ice sheet growth, beneath the research site known as Dome A, 24% of the base by area is frozen-on ice. In some places, up to half of the ice thickness has been added from below. The unexpected thickness of these bottom layers of ice was detected using radar imaging during the International Polar Year 2007–2009.

The exact mechanisms for the formation of this refrozen ice depends on the location. According to Bell et al., the freeze-on at valley heads is primarily the result of conductive cooling over static bodies of water. Along the valley walls, ice is primarily the result of the hydrologic potential forcing water up steep valley slopes. But these processes are not mutually exclusive: individual packages of frozen-on ice could well have been produced by a combination of the two mechanisms. The authors speculate that the creation of bottom ice is widespread and has been going on since the beginning of persistent glaciation more than 30 million years ago.

In East Antarctica, basal freeze-on has continued in the same locations through the last glacial-interglacial transition and has probably been a persistent process since East Antarctica became encased in a large ice sheet 32 million years ago. The simple geometry of the subglacial topography and the stable ice flow in the Dome A region have enabled us to image this process for the first time. Although the surface accumulation, surface slope, and bed morphology vary distinctly on the northern and southern sides of Dome A, throughout the area almost a quarter of the ice sheet base consists of ice freeze-on from the bottom. Widespread freeze-on can change the rheology and modify the flow of the Antarctic and Greenland ice sheets. Inclusion of these basal processes is essential to produce robust predictions of future ice sheet change.

The addition of hundreds of meters of ice to the base of an ice sheet deforms the overlying ice, causing uplifted sections on the glacier’s surface. This changes ice sheet stratigraphy and may affect the surface accumulation by changing the slope of the surface. The thickest package of frozen-on ice the researchers found was 1110 m, located at the downflow end of a 20-km-long valley. The internal layers in the covering ice are deformed upward over 410 m at the valley head. The shape of the surface actually reflects the shape of the accreted ice body and not the underlying bedrock. Radar imaging and its interpretation are shown in the figure below.

AGAP radar and lidar data over the freeze-on ice package.

“The accretion sites in the Dome A region are typically coincident with 5- to 35-m mounds in the ice surface, indicating linkage between the basal processes and the ice surface morphology,” the report states. In fact, the freeze-on rates in the Dome A region may be locally greater than the surface accumulation rates from snowfall. Aside from changing accumulation rates and ice-flow dynamics, this discovery has implications for those trying to extend the ice core record farther back into time.

“the widespread melt required to support the freeze-on process may have destroyed the ice containing the ancient paleoclimate records,” Bell et al. state. “Without the inclusion of basal processes, simple models of ice sheet temperatures cannot accurately predict the location of the oldest ice.” Knowledge gathered from the deepest ice cores may need revision.

In an accompanying perspective article, “Antarctica’s Deep Frozen “Lakes”,” Slawek Tulaczyk and Saffia Hossainzadeh estimate the volume of water contained in the accumulated frozen-on ice. “If the average thickness of Antarctic accreted ice is in the range of 10 m to 100 m, its volume is around 100,000 to 1,000,000 km3, i.e., much larger than the estimated volume of Antarctic subglacial lakes, 10,000 km3,” they state, adding “it may even exceed the volume of all glaciers on Earth outside of the two polar ice sheets, 180,000 km3.”

Accreted basal ice layers were known from past Antarctic studies, Tulaczyk and Hossainzadeh note, but the newly discovered ones are considerably thicker and more widespread. West Antarctica boreholes had previously encountered accreted ice 5 to 15 m thick. In contrast, Bell et al. imaged accreted ice that is hundreds of meters thick. It is now thought that basal conditions favor freezing over approximately half of the area of the Antarctic ice sheet, as shown in the figure below. Melting is positive and shown in warm colors; freezing is negative and in cold colors.

Estimated Antarctic basal melting and freezing rates in mm/yr.

The implications of the research are clear. Bell and colleagues summed the situation up this way: “Widespread freeze-on can change the rheology and modify the flow of the Antarctic and Greenland ice sheets. Inclusion of these basal processes is essential to produce robust predictions of future ice sheet change.” Tulaczyk and Hossainzadeh are even more to the point:

The discovery of thick, widespread accreted ice layers changes in fundamental ways our understanding of the Antarctic ice sheet. Further mapping and modeling of these ice bodies is necessary to aid the ongoing search for the oldest ice on Earth. New models of subglacial water generation, flow, and freezing will have to be developed to account for this large internal mass and heat redistribution within the ice sheet

These scientists are saying that the new findings fundamentally change our understanding of ice sheets and their inclusion in new models is essential. Basically, this means all of the model predictions of how fast ice is melting and moving to the sea, in both Antarctica and Greenland, need to be revised. It turns out that all those model based predictions of an icy Armageddon, with glacial ice racing to the ocean at ever increasing rates, were not based on reality.

This discovery shows the danger of basing estimates of future conditions on model output, be it ice flow or climate change. As fodder for research models are fine, but as a foundation to base public policy on they are horrid. The public, media and politicians need to understand that models are not fact, they are at best imperfect representations of a poorly understood reality.

Be safe, enjoy the interglacial and stay skeptical.

The Resilient Earth, 24 April 2011