SOME guys have all the luck. Two years ago Dan Schrag was in Indonesia diving
for fun when he made his discovery. 鈥淲e had a glorious day,鈥� he says. 鈥淎fter a
morning dive when we saw a huge school of barracuda, we stopped for lunch. I
took a walk down the beach, behind the mangrove swamp, and saw a massive coral
head, incredibly well preserved.鈥� He chiselled a piece out and headed for home.
That fossilised coral, later dated at 125 000 years old, now looks as if it
could transform our understanding of El Ni帽o鈥攖he Pacific Ocean
phenomenon that is the crucible of much of the world鈥檚 climate.
The find wasn鈥檛 entirely serendipitous. Schrag, who is based at Harvard
University鈥檚 Department of Earth and Planetary Sciences, had been looking for
well-preserved fossil coral on four previous trips to Indonesia. The idea
was to use growth rings in the coral to look for signs of El Ni帽o in the
distant past. Schrag鈥檚 earlier expeditions had yielded plenty of samples, but
what he really needed was one with enough rings to record a series of El
Ni帽os. He had been about to return home empty-handed yet again when he
took his postprandial walk down the beach of Bunaken Island, a speck of old
atoll off the Indonesian island of Sulawesi. 鈥淚 guess I got really, really
lucky,鈥� he says now. The coral he found was the first鈥攁nd so far the
only鈥攑iece located by researchers that is large enough and well enough
preserved to give a good snapshot of ancient El Ni帽os. What鈥檚 more, says
Schrag, Sulawesi is the 鈥渂ull鈥檚-eye鈥� of El Ni帽o. A record from there
could tell us whether El Ni帽o was different in the past, and how it might
change in the future.
It is an important issue. Until recently, climatologists looked on El
Ni帽o as an aberration in the tropical Pacific, of only passing interest
to the outside world. But in the past two decades it has become the fifth
horseman of the Apocalypse, a bringer of devastating floods, fires and famine
from Ethiopia to Indonesia to Ecuador, and a sender of weird weather round the
world. It has been appearing more frequently, with effects that last longer than
ever. Its activity is unparalleled in the historical record. And yet nobody
could be sure if this is a perfectly normal blip, or an alarming consequence of
human-induced climate change.
Advertisement
Schrag鈥檚 coral could change all that. Along with other evidence now pouring
in, it makes a strong case that the climate system is changing beyond all
recognition.
For climatologists, El Ni帽o is the flywheel of the world鈥檚 climate, a
redistributor of heat and energy that kicks in when the regular circulation
systems cannot cope. In normal times, the winds and waters flow across the
tropical Pacific from the Americas in the east to Indonesia in the west, driven
by the Earth鈥檚 rotation. In the tropical heat, the water warms as it goes. The
result is the gradual accumulation of a pool of warm water around Indonesia that
can be 40 centimetres higher and several degrees warmer than water on the other
side of the ocean. This cannot last and, typically every three to seven years,
this warm water breaks out and flows back across the surface of the ocean. As
the pattern of ocean currents shifts, so do the wind and air pressure systems
associated with it, and with them the weather. So the wet rainforest climate of
Indonesia drenches the normally arid Pacific islands, and often reaches the
coastal deserts of the Americas. Meanwhile, Indonesia and much of Australia dry
out (see Map).
But scientists have been uncertain about how far back El Ni帽o goes.
Reliable climate and ocean records cover only a century or so; delving further
requires an alternative source of information. To this end, some researchers
have been digging up the beds of old Andean lakes, in the expectation that they
will bear the scars of the occasional El Ni帽o-inspired floods that hit
the normally arid region. In January, Donald Rodbell of Union College in
Schenectady, New York, reported findings from a lake in southwest Ecuador, in
which he dated sediments associated with occasional heavy flooding going back 15
000 years (Science, vol 283, p 516). For the first half of the period,
the floods seemed to come only once every 15 years or less, he says. Then they
speeded up quite abruptly to settle some 5000 years ago at an average return
period of between two and eight years鈥攖he classic El Ni帽o pattern
that has broadly held to the present.
Some researchers interpret this as showing that El Ni帽o started 6000
years ago. Others say that the 10 000 years before that were merely a quiet
phase, caused perhaps by abnormal seasonal patterns which were in turn triggered
by wobbles in the Earth鈥檚 orbit. Others again say that the lake record may not
be reliable, because local glaciers could have interfered with the El
Ni帽o 鈥渟ignal鈥�.
Schrag鈥檚 chunk of coral sidelines that debate by putting the date of the
first recorded El Ni帽o back by more than 100 000 years, to before the
last ice age. In a paper due for publication shortly in Geophysical Research
Letters, Schrag and his colleague Konrad Hughen will reveal their analysis
of the isotopic signature of the annual growth layers inside the Sulawesi coral,
and use it to plot the pattern of the ancient El Ni帽os.
When water evaporates, molecules containing the lighter isotope of
oxygen鈥攐xygen-16鈥攖end to evaporate slightly faster, leaving behind
seawater that is enriched with oxygen-18. So in the Indonesian islands during El
Ni帽os, when rainfall ceases and drought ravages the islands, both the
seawater and the coral growing in those years contain more oxygen-18. By
measuring the relative amount of oxygen-18 in his coral, Schrag has come up with
a year-by-year El Ni帽o record over the 65 years covered in its annual
rings.
According to Schrag, the pattern of El Ni帽o events revealed in his 125
000-year-old coral looks exactly like the modern period before 1976, but nothing
like the post-1976 period. He has examined in detail the 鈥渞eturn period鈥� for El
Ni帽os, both in the ancient coral and modern meteorological and coral
records, and found that in the modern record prior to 1976 the dominant return
period for El Ni帽o was around six years. That was also the case in the
65-year time slice in his ancient coral. But the post-1976 record shows a peak
return period at 3.5 years. The implication is that the cycling of El
Ni帽o was highly stable over hundreds of thousands of years, but has
changed fundamentally in the past quarter-century.
The crucial question is what lies behind this change. Has El Ni帽o been
disturbed by some external factor, such as global warming, or is it simply on a
short-lived, exuberant joyride? Many oceanographers support the joyride theory.
They point out that El Ni帽o has always had decades when it is unusually
quiet or busy or just plain weird. Mark Cane of the Lamont-Doherty Earth
Observatory at Columbia University in New York has compiled one of the most
respected models of El Ni帽o, one that has successfully predicted the
onset of El Ni帽os. That model, he says, generates such fluctuations as
part of the natural variability of El Ni帽o, without introducing any
outside element.
Kevin Trenberth, head of climate analysis at the National Center for
Atmospheric Research in Boulder, Colorado, thinks differently. Trenberth was one
of the first researchers to spot the unusual state of the tropical Pacific after
1976, and he believes that the recent El Ni帽o shenanigans could well be
down to global warming. Global warming up to 1976 may have been modest enough
for the 鈥渘ormal鈥� climate system to cope with quite happily, he suggests. Only
after the magnitude of the warming hit a threshold did it begin to trigger
unusual effects in the El Ni帽os.
One way to check this, says Schrag, is to look for signs of recent warming in
the ocean. Together with Tom Guilderson from Lawrence Livermore National
Laboratory in California, he has recently pointed out that the unique signature
of the post-1976 El Ni帽os is down to a very specific warming of surface
waters in the eastern Pacific during the cold season. Maximum sea temperatures
in the area changed very little, but minima jumped from a typical 23.5 掳C
before 1976 to above 24.5 掳C thereafter.
This area of ocean is a constant battleground between warm waters at the
surface and cold waters that well up from the deep. Most of the time the
upwelling is dominant. But during El Ni帽os, when warm waters wash across
the Pacific from the west, the upwelling is shut off. What seems to have
happened is that this shut-off has become near permanent.
In a paper in Sciencelast year (vol 281, p 240), Guilderson and
Schrag showed that since 1976 the thermocline鈥攖he boundary zone between
surface and deep water that lies around 50 metres below the surface鈥攈as
deepened by 10 metres or more. This dramatic change is reflected in a variation
in the isotopic composition of the water. As the carbon-14 in seawater gradually
decays, surface water can replenish its stocks from the atmosphere, but deep
water cannot. So low levels of carbon-14 are an indicator of deep water. Since
1976, water at the surface in the eastern Pacific has been richer in carbon-14,
showing that deep water is not welling up as much as before. Upwelling normally
keeps the eastern Pacific cool, maintains the normal trade winds and so
suppresses the outbreak of El Ni帽os. Reduce the upwelling and the system
is permanently primed for an El Ni帽o. Schrag concludes that the post-1976
change in the thermocline may be responsible for the increase in the frequency
and intensity of El Ni帽o events since then.
Does any of this matter beyond the Pacific Ocean? As climatologists discover
more and more about the workings of the oceans and atmosphere, they realise how
central El Ni帽o is to the functioning of the entire climate system. The
Indian Ocean shows its own post-1976 shift. Analysis of weather statistics from
the remote Chagos Archipelago by Charles Sheppard of the University of Warwick
found that around the mid-1970s average air temperatures abruptly rose by a
degree, while cloud cover shrank by 50 per cent. The islands鈥� coral reefs, some
of the largest and most pristine in the world, have been wrecked as a
result.
Catastrophic climate
Is the shift in El Ni帽o the long-sought 鈥渟moking gun鈥� that will
convict greenhouse gases of causing climatic mayhem? Modelling studies provide a
hint that it may be so. Mojib Latif and colleagues at the Max Planck Institute
for Meteorology in Hamburg recently developed the first global climate model
detailed enough to reproduce El Ni帽o cycles. It successfully predicted
the 1997-98 El Ni帽o. And in April this year Latif reported in
Nature(vol 398, p 694) that when they simulated global warming in the model
it generated more frequent El Ni帽o-like conditions.
鈥淢odels are not proof of what will happen,鈥� Latif acknowledges. 鈥淏ut for the
past 50 years, our model shows well what has actually happened.鈥� The model
predicts that the average climate in the 21st century will become more like the
typical El Ni帽o conditions at the end of the 20th century. And Schrag鈥檚
coral results underline the growing feeling that some fundamental change is
afoot.
The suspicion is taking hold that sometime in the 1970s, a shift took place
that stacked the odds in favour of events such as major El Ni帽os, and
perhaps other catastrophic climate events too. Seen in isolation, the sea change
of 1976 and the El Ni帽os of 1982-3 and 1998-9 were all quite plausible
鈥渘atural鈥� events. But all three of them within such a short space of time? That
does sound as if greenhouse gases may have loaded nature鈥檚 dice.
Cane remains cautious. 鈥淚f you ask me, as a scientist, if the unusual
behaviour of El Ni帽o lately has anything to do with greenhouse gases then
the answer, at the level of confidence the customs of science demand, is a clear
no. But if you ask me at a cocktail party if there is a causal link, I would say
`probably鈥�.鈥� To some extent, time will tell, says Schrag. 鈥淚f pre-1976
conditions don鈥檛 return soon that will tend to support the global warming
丑测辫辞迟丑别蝉颈蝉.鈥�
Meanwhile, Schrag is unlikely to be joining Cane at many cocktail parties. He
is off in search of new fossilised coral on the shores of Sulawesi. There is, he
says, an urgent need to find corals from different times in the geological past,
to bridge the huge gap in the history of El Ni帽o between 125 000 years
ago and the past century. 鈥淚 am optimistic that over the next several years we
will find enough to put together a much more complete story,鈥� he says. Happy
diving, Dan.