FROM the ground, all you see is empty flat whiteness stretching to the
horizon. Even from the air, the most you can spot is an occasional jumble of
crevasses scoring the snow. Only from space does the whole picture finally
spring into view: mighty rivers of ice flooding down from the Antarctic heights
to spill into the Southern Ocean.
These ice-bounded ice streams are vast—fifty kilometres wide, hundreds
of kilometres long and more than a kilometre deep. More importantly, they
are fast. They move at more than a metre a day, which for ice is a breakneck
speed. And it’s their speed that attracts researchers, because for decades they
have feared that it could hasten global disaster.
Ice streams such as these exist only on the West Antarctic Ice Sheet, the
smaller of the continent’s two enormous ice sheets. Slighter it may be, but it
still holds a staggering 3.2 million cubic kilometres of ice—enough to
raise global sea level by 6 metres if all of it slid into the ocean. “Half the
world lives by the coast,” says glaciologist Bob Bindschadler of NASA’s Goddard
Space Flight Center in Maryland. If the western ice sheet were lost, cities and
towns from East Anglia to Florida would drown, and entire islands in the Pacific
would disappear beneath the waves. But how likely is such a catastrophe? After
all, ice sheets tend to change slowly, over thousands of years.
Advertisement
That’s where the ice streams come in. The danger is that because of their
speed they could respond very quickly to, say, a rise in temperature caused by
global warming. “With ice streams, changes can happen on a human timescale,”
says Richard Alley of Pennsylvania State University. The race is now on to
discover exactly how ice streams work—why they form where they do and why
they move so fast. To find out, researchers have had to work in some of the
continent’s most inaccessible spots. But in the past few years, their efforts
have started to pay off.
For the most part, Antarctica’s ice sheets maintain themselves in a steady
cycle, growing when snow falls in the high interior and shrinking as the ice
creeps downhill at a pace of about 10 metres a year, until it breaks off into
the ocean as icebergs. But for the western ice sheet, this stately progress is
disrupted by the ice streams, which can move at 60 times this speed and drain
two-thirds of its ice. The western sheet has a handful of streams on the Siple
Coast, which pour into the huge Ross Ice Shelf—a mass of floating ice the
size of Texas—and two on the opposite side of the continent, which spill
onto the equally large Filchner-Ronne Ice Shelf (see Map).
The first big puzzle facing researchers is why the streams run so fast while
the rest of the ice moves like molasses. One clue came from mountain glaciers,
which can sometimes surge forwards, sliding on a lubricating bed of thick mud.
In 1986, Alley, Don Blankenship of the University of Texas at Austin and others
obtained seismic data which suggested that ice stream B on the Siple Coast might
work the same way. From echoes of surface explosions reflected off the base of
the stream, the researchers deduced that a layer of soft sediment several metres
thick lay beneath the ice.
Other researchers, such as Barclay Kamb from Caltech in Pasadena, were
sceptical. “They had to make so many assumptions that I felt the whole thing was
like a house of cards,” he says. Nonetheless, in 1989, Kamb set off for ice
stream B with a hot-water drill. He planned to drill right through the
ice—which is more than a kilometre thick—and see what lay beneath.
To his astonishment, when his drill reached the bottom it hit a layer of sticky
grey mud, just as Blankenship and Alley had predicted.
So do all the ice streams flow so fast because they slide on mud? Kamb thinks
so. In the past three years, he and his coworkers have found the same mud at the
base of two other ice streams. And on the other side of the continent, seismic
studies by David Vaughan and Andy Smith from the British Antarctic Survey have
turned up evidence that the Rutford ice stream sits on the same soft
sediment.
But mud isn’t all that’s needed to make an ice stream. Only when the mud is
saturated with water does it turn into a high-pressure lubricant. This much was
clear when Kamb first drilled into ice stream B. Though the temperature close to
the surface was around –20 °C, at the bed it had risen to zero. What’s
more, the mud Kamb retrieved was crammed with water.
Theorists believe they understand how water can form beneath the ice. It
works like this. The ice traps some of the geothermal heat that flows from
the Earth’s interior. In some places, this could be enough to allow the base to
melt slightly and start to slip. This creates friction, which warms the bed
further, melting the ice and generating more water to lubricate the mud beneath
the ice. More lubrication means the ice can flow faster, which generates more
friction and more heat until the system runs away to create a speeding ice
stream. If this is correct, outside the ice stream there should be no liquid
water beneath the sheet. Sure enough, when Kamb went back in 1992 and drilled
through a region where the ice flowed only slowly, next to an ice stream, the
sheet was frozen to the bed and there was no liquid water to be seen.
The notion of ice sliding on a sloppy lubricant seems straightforward enough.
But what does it imply for the fate of the ice sheet? That depends on where the
sediment lies. If it stretches all the way to the heart of the ice sheet, for
instance, there may be nothing to stop the ice streams eating their way
backwards. “If the head can move inland, it could turn slow ice into fast ice
and dump the whole ice sheet in a hurry,” says Alley.
Unfortunately, nobody knows where the sediment lies—it’s not easy to
map the underbelly of an entire ice sheet. But using an ingenious combination of
techniques, researchers have managed to peer beneath one important area. Last
year, Sridhar Anandakrishnan, now at the University of Alabama at Tuscaloosa,
reported the results of his investigation of the upper region of ice stream C.
Working with Alley and Blankenship, he built up a seismic profile across the
edge of the ice stream. The team found virtually no sediment under the sluggish
ice, but beneath the ice stream they located a basin packed with soft mud.
No surprises there then. But Blankenship and Robin Bell from the
Lamont-Doherty Earth Observatory in New York had also taken airborne
measurements of the gravitational and magnetic properties of the bed beneath the
area. By themselves, these data aren’t enough to identify clearly whether the
bed is made of hard rock or soft mud. But by combining them with
Anandakrishnan’s local seismic survey, the researchers managed to distinguish
the two, and mapped the area. The upshot? It looks as though the sediment
disappears just where the ice stream starts.
Slow retreat
However, there are a few problems to be ironed out before the researchers can
be sure of this. For one thing, they had to judge where the stream starts moving
from satellite images, rather than making more accurate velocity measurements on
the ground. Blankenship and Anandakrishnan plan to go back to stream C to
pinpoint the source precisely and also to take more seismic measurements. This
time, rather than going to the side of the ice stream as before, they will
conduct their survey right at its head, where the sediment seems to pinch out.
If the ground measurements confirm the airborne results, this will suggest that
the head of ice stream C cannot retreat any farther.
Even so, nobody knows if the other ice streams are similarly stymied. If not,
they could yet eat their way into the heart of the ice sheet and cause it to
collapse. The next step, says Anandakrishnan, will be to repeat the whole
process at the head of ice stream D. Three years ago, Bindschadler and his team
measured the velocities of the ice there, and they have already established
where the ice stream starts. In two years’ time, Anandakrishnan and Blankenship
hope to take their seismic equipment there to discover whether stream D sits on
sediment, and—more importantly— if the sediment continues back into
the slow-moving ice sheet.
Another quirk of ice streams is that they sometimes come to a complete and
inexplicable standstill. This seems to have happened to stream C. At the head of
the stream—where Anandakrishnan and Blankenship did their seismic
work—the ice is still moving pretty quickly. But downstream; all signs of
motion have ceased. There are no crevasses at the surface, and the ice is now
moving at a paltry 20 metres a year or less. In 1993, by using radar to peer
through the surface snow, Charlie Bentley from the University of Wisconsin in
Madison found crevasses that, he calculated, must have been buried by snow
around 130 years ago. Presumably that’s when the ice stream stopped moving. The
question is, why?
One possibility is that the same process that accelerates ice streams can
work in reverse. Perhaps the stream began to slow a little, which reduced the
friction at its base, reduced the amount of meltwater and so on until it finally
froze to the bed, reverting to the slow, molasses-like flow of the rest of the
ice sheet. Two years ago, Kamb went to find out. Once again, he was astonished
by what he found. The bed of stream C looked just like all the others: sediment,
water, the works. It certainly wasn’t frozen fast. So why on Earth wasn’t it
moving as fast as the rest?
Alley and Anandakrishnan have an answer. They think that the head of stream C
had been creeping backwards into the ice sheet until, about 130 years ago, it
reached a point where the slope of the surface and bed conspired to divert much
of its liquid water. “The water that should have gone down C is now taking a
left turn and going down B instead,” says Alley.
This “water piracy” would not in itself be enough to bring stream C to a
halt: it would take something like a million years to dry out the sediment
completely. But the team found something else. Throughout the sluggish lower
part of the ice stream, Anandakrishnan detected the continuous rumble of
microearthquakes, which occur when the ice jolts over sticky parts of the bed.
The researchers suggest that the rock beneath all the ice streams is dotted with
similar “sticky spots”, and that it is only when the bed is fully saturated with
water that they can slide over them comfortably.
This theory holds two implications for the fate of the ice sheet. First, at
least for the past few centuries, it seems that something made the head of
stream C eat backwards into the ice sheet. Second, whatever caused this
behaviour might be doing the same for the other ice streams. If Alley and
Anandakrishnan are right, it was a serendipitous combination of geometries that
brought stream C to a halt. With the other streams, we might not be so
lucky.
Ancient trigger
But what could have caused stream C and the others to migrate backwards? One
possibility is that a rise in temperature increased melting at the head of the
stream, allowing it to move inland. The one thing all the researchers agree on
is that such a temperature rise could not have come from recent global warming.
The ice streams are more than a kilometre thick, and ice transmits heat so
slowly that it would take thousands of years for a change at the surface to
reach the bed, where all the action is. But there was a previous, natural global
warming around 10 000 years ago at the end of the last ice age. That temperature
rise may still be working its way down to the base of the ice streams, causing
more melting and allowing them to glide over their sticky spots more easily, and
move more quickly. “Perhaps the trigger has already been pulled,” says
Alley.
There is no doubt that the Antarctic ice has receded over recent millennia.
At the height of the last ice age, the West Antarctic Ice Sheet was three times
its present size. Just how the ice retreated—how it responded to the
warming at the end of the ice age—is a question being asked by John
Anderson of Rice University in Houston, Texas. But whereas other ice stream
researchers drill through the ice to see what lies beneath, he goes where the
ice used to be—to the seas around the continent. Although much of this ice
has now melted, it left traces of its passing in the seafloor as it scraped by.
Anderson and his team cruise around Antarctica collecting sonar images of the
seafloor, from which they deduce the ice sheet’s history. They can also see
troughs where the ice streams used to run.
Three weeks ago, Anderson completed a cruise investigating how the ice
streams on the Siple Coast have retreated over the past 10 000 years, and he is
now poring over his findings. Already he has found something intriguing. He knew
from a cruise last year that when the stable eastern ice sheet retreated it did
so steadily, keeping close contact with its bed. But for the western sheet,
where the ice streams are, Anderson has found a stark difference. Instead of
retreating steadily, the ice sheet seems to have lifted up and lurched
backwards, its feet not touching the floor. If this interpretation is correct,
though the ice streams are now pinned to their beds, there is always the chance
that they are preparing for another dramatic lurch backwards.
But there is one last twist to this tale. While Anderson and the rest are
struggling to decide whether the ice streams are the villains of the piece, a
few theorists are suggesting that they could be the heroes. Rather than causing
the ice sheet to become so small that it disappears, the ice streams could be
preventing it from growing too big and collapsing under its own weight.
This notion comes from modelling work carried out in 1992 by Doug MacAyeal at
the University of Chicago. MacAyeal calculated that if an ice sheet sitting on a
layer of sediment grew large enough, the pressure could cause it to start
melting at its base. Then, just as for the ice streams, it would begin to move,
generate friction, pick up speed and finally slide wholesale into the sea.
Modellers Tony Payne, from the University of Southampton, and Richard Hindmarsh,
from the British Antarctic Survey in Cambridge, believe that the West Antarctic
Ice Sheet would be in danger of suffering just this fate if it weren’t for the
ice streams.
Last year, Payne published a model of the West Antarctic Ice Sheet in which
he allows ice to flow quickly wherever there is melting at the base. He started
with the topography under the real ice sheet, covered it with ice and set it
running. To his delight, ice streams formed in all the right places. What’s
more, the streams shifted and changed just as they do on the real ice sheet.
Stream C even stopped in its tracks. But—and here’s the key—the
overall volume of the ice sheet scarcely changed, no matter how dramatic the
changes in the ice streams. “The streams buffer changes locally,” says Payne.
And this removes the possibility that the whole ice sheet will surge forward and
slide into the sea.
Hindmarsh is also modelling the unstable behaviour of the ice streams,
concentrating on water pressure changes under the streams. Right now, he is
investigating how big the ice stream instabilities can grow. But he too believes
that the ice streams may be helping to avert catastrophe rather than cause it.
“They can start up and slow down so quickly that they never grow big enough, or
fast enough, to eat up the whole ice sheet,” he says. “I think they could be the
˛ő˛ą±ąľ±´ÇłÜ°ů˛ő.”
So who’s right? Are the ice streams heroes or villains? “I haven’t decided
yet,” says Alley. “There are still fundamental questions that we just don’t have
the answers for.” The modellers too need more data to be certain. Still, the
researchers on the ground have plenty of plans for providing them with better
information over the next few years. If they succeed, we will be in a better
position to know how vulnerable the ice sheet really is, and whether it’s likely
to collapse. “I think that if we humans know what’s coming we can respond
wisely,” says Alley. “And I believe that eventually we’re going to know what’s
ł¦´Çłľľ±˛Ô˛µ.”