New York
Ice skating in Central Park would have been a breeze 20 000 years ago. Or
anywhere else in Manhattan, for that matter. The last ice age was at its peak,
and giant sheets of ice stretched as far south as London and New York. Around
the world, average temperatures were about 6 °C colder than today, the sea
lapped coastlines more than 100 metres lower and ocean currents ran more slowly.
The atmosphere was dry and dusty and the whole climate was unstable, prone to
sudden swings between mild and freezing spells. Yet a hundred thousand years
before that, things were very similar to today. And 20 000 years before that,
the Earth was in the grip of another big freeze.
This lurching from cold to warm and back to cold again has been going on for
the past million years. But despite decades of research, no one agrees why they
happen. And yet, if we can understand these dramatic shifts in the natural
climate system, we might also learn more about how the Earth will respond to our
own climate-changing efforts.
Advertisement
Until recently, the main contender was an explanation combining the effects
of changing sunlight patterns and the action of currents in the North Atlantic.
Then, in 1996, two scientists from California suggested that these climate
changes could all be down to dust pouring onto the Earth from space. Though this
idea appears to have fallen by the wayside, the challenge it posed to the
traditional model triggered a frenzy of research activity. In its wake a new
explanation is rapidly gaining ground.
The conventional explanation of the ice ages dates back to the 1920s, when a
Yugoslavian astronomer called Milutin Milankovitch suggested that small wobbles
in the Earth’s orbit might be to blame. Milankovitch noticed that the Earth’s
orbit around the Sun is distorted by the gravitational pull of the Moon and the
other planets. This changes three things about the way the Earth orbits: the
shape of the orbit—more oval or more round—how much the Earth’s spin
axis tilts away from vertical, and the time of year when the Earth is closest to
the Sun
(see p 30).FIG-21215201.jpg
Though the distortions make only a small difference to the total amount of
sunlight reaching the Earth, they have a big effect on the amount of sunlight
arriving at different parts of the globe, and at different times of year. The
upshot is that regular cycles in the sunlight pattern have periods of 100 000
years in the ovalness, 41 000 years in the tilt, and 23 000 years in the timing
of closeness to the Sun. Milankovitch thought that the Earth’s climate is
controlled by these changes in the pattern of sunlight.
Sure enough, scientists looking at how climate changed in the past have found
signals with all three of these frequencies. For the 23 000 and 41 000-year
cycles, the pattern and relative sizes of climate changes around the Earth match
well with the pattern of sunlight changes. But with the crucial 100 000-year
cycle—the timescale on which ice ages occur— Milankovitch’s theory
hits a problem. The climate changes going into and out of ice ages are ten times
the size of the ones on the other two cycles. Changes because of the ovalness
have the right timescale to explain the ice ages, but they are far too
small.
Chilling out
To get round this, Milankovitch suggested that something else is amplifying
the effect of changes in sunlight patterns. His favourite candidates were the
large ice sheets that form on the continents of the northern hemisphere during
the ice ages. When there is relatively little sunlight falling on these areas
during the summer, ice sheets don’t get a chance to melt, so they keep growing
year after year. A big, growing ice sheet cools the surrounding region both by
directly chilling the air that blows over it, and by reflecting more of the
Sun’s warmth back into space. So a small initial change in temperature could
trigger growth of the ice sheets, which would, in turn, cause more cooling.
But how could changes in this one region affect the rest of the planet? In
the 1980s, Arnold Gordon and Wally Broecker, both at the Lamont-Doherty Earth
Observatory of Columbia University in New York, discovered that the North
Atlantic comes equipped with its own system for transmitting local climate
changes around the world: it is the starting point for the deep-ocean currents
that drive a global “conveyor belt” of ocean water, carrying heat along with it.
Water sinks in the North Atlantic, flows southward, around Antarctica, and then
northward into the Pacific before returning to the surface. Changing the local
climate in the North Atlantic could easily change the rate at which the water
sinks and hence change the whole circulation system.
So a general consensus began to emerge. Changes to sunlight patterns in the
northern hemisphere changed ice sheets around the North Atlantic, which
influenced the ocean conveyor belt, which cools the planet by changing the way
heat is transported around the Earth.
But there have always been some nagging worries. For instance, although the
average effect of changing the ovalness of the Earth’s orbit is a sunlight cycle
of 100 000 years, a more careful look shows that some of these cycles are 95 000
years long while others are 125 000. Climate records, however, show only 100
000-year cycles. It may be that the records just aren’t detailed enough to pick
out both cycles, or it might be that the ovalness of the orbit is not what
causes the 100 000-year cycle at all.
In 1996, Richard Muller of the University of California at Berkeley and
Gordon MacDonald of the University of California at San Diego, suggested a
controversial new approach. They pointed out that the plane of the Earth’s orbit
wobbles relative to the average plane in which the whole Solar System rotates.
This wobble gives rise to just one cycle exactly 100 000 years long. This new
“inclination” model no longer suffers from the ovalness problem of two different
cycles. But as the wobble makes no difference to the sunlight reaching the
Earth, how could it affect climate?
Muller and Macdonald suggested that climate might be affected by cosmic dust
entering the atmosphere. Perhaps, they say, when the Earth orbits in the same
plane as the rest of the Solar System, more dust enters the atmosphere, blocking
sunlight and triggering the ice ages. Then, when it tilts out of the plane, less
dust arrives and the planet warms up again.
But was there enough cosmic dust to have an effect? One way to work out how
much has arrived over the years is to look at ocean sediments. Franco
Marcantonio of Tulane University of Louisiana in New Orleans and Ken Farley of
the California Institute of Technology in Pasadena have each been working on
this question (żěè¶ĚĘÓƵ, Science, 9 December 1995, p 22). The jury
is still out, but it looks as if there are no significant changes in the cosmic
dust flux during past climate cycles—bad news for the inclination
hypothesis.
Strange changes
The lack of any mechanism to link inclination changes with climate has meant
that most researchers have now discounted Muller and Macdonald’s model. But the
controversy it raised prompted many scientists to look back at the existing
data, to collect new information and to question the traditional explanations
for ice ages. Just how well does the theory about the northern hemisphere’s ice
sheets hold up to scrutiny?
One indication that all might not be well comes from research done by Chris
Charles from Scripps Institution of Oceanography in California. In 1996, Charles
and three of his colleagues made a puzzling discovery: changes in the North
Atlantic and its conveyor belt seem to lag behind changes elsewhere. The
researchers studied sediment that had been laid down in the South Atlantic over
the past 80 000 years. They extracted the shells of tiny sea creatures called
foraminifera from different depths in the sediment. As the foraminifera grow,
their shells act as a record of the local chemical conditions—conditions
that depend both on the temperature of the surface of the ocean where the
creatures live, and on the amount of water passing through from the North
Atlantic on the great ocean conveyor belt.
The shells revealed that whenever the temperature in the South Atlantic ocean
changed, the amount of water passing through also changed. But the surprise was
that the circulation change occurred slightly after the temperature change. In
other words, the South Atlantic was changing before the North Atlantic. How
could this be, if temperatures in the south were controlled by changing
conditions in the north?
Charles suggested that the North Atlantic was itself responding to changes
elsewhere. It was lagging behind, he believed, because the ice on the northern
continents slowed down the currents’ response—a bit like a storage cooler.
As the sediments that Charles analysed only went back 80 000 years, he was not
studying the full ice-age cycles but shorter climate oscillations occurring
within the last ice age. But they demonstrate that the North Atlantic is not the
sole driver of global climate.
Core of the problem
At Christmas, Charles joined around thirty scientists sailing for two months
aboard the Ocean Drilling Program ship, the JOIDES Resolution, to the same part
of the South Atlantic. This time, the researchers plan to drill longer cores
into the sediment. When analysed, the hope is that they will back Charles’s
original results and show whether the same effect is seen over more than just
the past 80 000 years.
There is other evidence that suggests Charles is on the right track. In the
past few years, Mike Bender and Todd Sowers of the University of Rhode Island
compared climate records in ice cores taken from the north and south poles to
try to see which changed first. When you are dealing with two cores, rather than
one, it is much more difficult to be sure about the relative timing of climate
changes. Bender and Sowers solved this problem by measuring oxygen that they
found in small air bubbles trapped in the ice cores. Oxygen is well mixed in the
atmosphere, so any changes should show up simultaneously all over the world. By
lining up the changes in oxygen, the researchers managed to compare the timing
of the temperature records.
Once again, they were surprised by what they saw. The last time the planet
warmed suddenly, at the end of the last ice age, the warming started in
Antarctica and only later showed up in Greenland.
This approach shows that the southern hemisphere seems to warm first. But it
still only gives relative timing. If you could compare the temperature records
directly with the past sunlight pattern for different latitudes, it might be
possible to pinpoint the exact location of the ice-age amplifier. The
problem is that it is extremely difficult to obtain reliable dates for changes
in ocean sediments or in ice cores. So Isaac Winograd and Ken Ludwig and
colleagues at the United States Geological Survey in Denver and Virginia turned
to an unusual cave in Nevada where stalactites and stalagmites have been growing
continuously for the past few hundred thousand years. Just like the foraminifera
shells, these speleothems created a temperature record as they grew. But as well
as temperature they also absorb natural radioactive isotopes which can pin down
their age very precisely.
In 1992, Winograd and Ludwig produced a record, carefully dated, of the
timing of climate changes in these Nevada caves. They discovered that, when the
Earth’s climate makes the sudden shift out of an ice age, warming occurs before
changes in the northern hemisphere sunlight patterns but at about the same time
as changes in the southern hemisphere. Sceptics of this research have suggested
that the Nevada record reflects the local climate rather than the global one.
However, even if this is true, the record is certainly one more argument against
an ice age cycle driven by the traditional northern hemisphere ice sheet
model.
It seems that the evidence is pointing towards the southern hemisphere as the
culprit. But is there a mechanism for amplifying the weak change in sunlight
into a strong enough signal to cause the ice ages? Because there is less land in
the southern hemisphere, ice sheets in the south don’t change their size in the
drastic way that those in the north do so they are not likely to be the cause.
But the ocean surrounding Antarctica does contribute water to the conveyor belt
as it passes by. So some scientists believe that ocean circulation changes in
the south, rather than the north Atlantic, could be the key.
Another way in which the south could act as an amplifier is by changing the
amount of carbon dioxide in the atmosphere, and therefore changing the amount of
warming caused by the natural greenhouse effect. One mechanism for controlling
carbon dioxide in the south was first suggested during the 1980s by the late
John Martin who worked at Moss Landing Laboratories in California. He pointed
out that some areas of the oceans, especially the ocean around Antarctica, have
very low concentrations of iron—a key nutrient for the growth of marine
plant life. If these areas had more iron dissolved in the water during the ice
ages than they do now, there could be more biological activity in the water.
This in turn would consume more carbon dioxide from the atmosphere, as the
plants photosynthesised.
This idea was supported in an experiment in 1996 when a large team of
scientists lead by Kenneth Coale, one of Martin’s colleges from Moss Landing,
added bucket-loads of iron into one such area of the ocean
(This Week, 12 October 1996, p 4).
This produced a big increase in the amount of biological
activity, just as Martin had predicted, and a big increase in the carbon dioxide
taken up. During the ice ages, the climate is drier and the winds stronger, so
more dust blows into the ocean. So perhaps a small change in sunlight caused
drier, dustier conditions and this increased the amount of dust—which
contains plenty of iron—blowing onto the Southern Ocean, thereby sucking
carbon dioxide out of the atmosphere to cause the ice age.
Proponents of the northern-hemisphere model are not ready to go quietly,
however. Their model does explain a lot of what is known about the Earth’s past.
For instance, sea levels hit their highest values in the past just when the
northern-hemisphere model predicts the most melting of the ice sheets. So now
the race is on to reconcile the new data with the old model or to improve the
Southern Ocean models to make them as convincing as the northern hemisphere
model used to look.
To be really sure why London and New York were under ice 20 000 years ago, a
theory would need to unpick the intricate combinations of temperature, carbon
dioxide and ocean circulation—how the whole complex climate system fits
together. When we succeed, we may even learn more about how the climate system
will respond to the carbon dioxide released by mankind’s activities. Not just
where we’ve been, but where we’re going too.
