杏吧原创

High noon. 25 February 2000. For the band of researchers
that has descended on northern Oklahoma, noon is the crucial time of the day.
The plane loaded with instruments is airborne, the
antenna on the tarmac tracking its every move. 杏吧原创s at computer screens
scrutinise the stream of incoming data.

Everyone is rushing to get the job done while the Sun is still high in the
sky. But first they have to wait for their quarry, and so far today, it鈥檚 a
no-show. Clouds can be that way.

The US Department of Energy (DoE) sent the team to this rural airstrip in
Oklahoma in February to settle a debate that has been raging for almost six
years. In 1995, three research groups said they had spotted something impossible
in the sky: clouds absorbing far more solar radiation than could be explained by
the laws of physics. To say there is no consensus on these findings is a massive
understatement. More than a few climate researchers feel strongly that the
phenomenon simply cannot exist. Others merely refer to the strange effect as
鈥渧exing鈥�. Some refuse to express an opinion, so charged has the debate
become.

But this year the Oklahoma team might just come up with an answer to the
mystery. The team of over seventy is made up of both sceptics and leading
proponents of enhanced cloud absorption鈥攖hey鈥檒l be using state-of-the-art
instruments and their considerable expertise to tease out what happens to the
Sun鈥檚 energy when it strikes a thick layer of cloud. Right now, no one fully
understands what could be causing the extra absorption, so finding an answer
would be a major coup for atmospheric science. And it would change our
understanding of how solar energy is distributed between the atmosphere and the
surface of the Earth, giving us more reliable computer simulations of future
climate change.

Computer simulations that calculate the scattering and absorption of solar
energy rely on 鈥渞adiative transfer鈥� equations鈥攅quations that are based on
fundamental physical principles dating back to the 19th century. In other words,
new discoveries are simply not expected. 鈥淚n the atmosphere, we鈥檙e dealing with
classical physics,鈥� says Thomas Ackerman, chief scientist at the DoE鈥檚
Atmospheric Radiation Measurement (ARM) programme, the sponsors of the Oklahoma
experiment. 鈥淲e鈥檙e not out here doing high-energy physics where collisions
generate new particles we don鈥檛 understand. We鈥檙e talking about solving
Maxwell鈥檚 equations, which have been known since the 1880s.鈥�

It鈥檚 not that researchers don鈥檛 know that their simulations tend to
underestimate the amount of sunlight absorbed by real clouds. In fact they鈥檝e
know this since the 1950s. They simply thought that the discrepancy was too
small to bother about: models predict that on average a cloudy atmosphere should
soak up about the same amount as a completely clear sky, perhaps a few per cent
more. And in many cases, the discrepancy measured in the field isn鈥檛 far off
this value.

For this reason, these models don鈥檛 attempt to compute the real interaction
between clouds and incoming sunlight. In the simulated atmosphere, a photon
meets a cloud as a point striking a one-dimensional plane, either passing
through it or bouncing off. A photon bumping into a real cloud, however, takes a
much more complicated path. It ricochets from one cloud droplet or ice crystal
to the next like a pinball, it bounces between and within cloud layers, and it
seldom leaves a cloud directly below the point where it entered. 鈥淢ost people
would agree that there is an anomaly, that there鈥檚 more absorption in reality
than in theory, because we don鈥檛 include all the constituents of the atmosphere
in our models,鈥� says Graeme Stephens of Colorado State University. But on the
whole, climate researchers felt the anomaly was insignificant.

This issue exploded in 1995 when three studies published in Science
reported 50 per cent more solar absorption in cloudy skies than in clear. The
lead paper was written by veteran climate researcher Robert Cess of the State
University of New York at Stony Brook. Cess used satellite records of solar
radiation at the top of the atmosphere and ground-based measurements of solar
radiation to see how absorption in the atmosphere varied between clear and
cloudy skies. In the five locations he investigated, he found 50 per cent more
absorption in cloudy skies but 鈥渘o obvious explanation鈥�. A second paper by V.
Ramanathan of Scripps Institution of Oceanography at La Jolla, California,
arrived at about the same figure by modelling heat transport in the Pacific
Ocean. The title of his paper raised the provocative question of a 鈥渕issing
辫丑测蝉颈肠蝉鈥�.

But Francisco Valero of Scripps and Peter Pilewskie of NASA鈥檚 Ames Research
Center near San Francisco took the most direct approach. Flying two aircraft in
stacked formation above and below clouds during an experiment in the South
Pacific, they measured the solar radiation entering and leaving the column of
atmosphere between the planes. The values they calculated were even higher than
those reported by Cess.

These studies startled their fellow researchers. What could be behind this
strange effect? They couldn鈥檛 incorporate enhanced cloud absorption into the
radiation transfer codes that drove their climate models until they could find a
clear physical explanation.

Initially, best guesses said that the answer lay somewhere in the limitations
of the one-dimensional cloud-radiation models. So several groups tried
three-dimensional cloud models to see if the real-world effects of complex cloud
geometry could account for the discrepancy. In 1998 Catherine Gautier and
William O鈥橦irok of the University of California at Santa Barbara simulated a
towering tropical cumulonimbus cloud. They found that with the Sun high
overhead, photons tended to 鈥渓eak鈥� from cloud edges to the lower atmosphere
where they were absorbed by water vapour in the air. However, this effect could
only account for 25 per cent of the extra absorption.

Other clues began to emerge. Using satellite and ground-based measurements on
a global scale, Zhanqing Li of the Canada Centre for Remote Sensing in Ottawa
discovered that clouds in the tropics seemed to absorb as much solar radiation
as Cess claimed, but much less at other latitudes. He wondered if aerosols from
burning forests and grasslands in the tropics could be the cause. However, when
other researchers added aerosols to their models, they couldn鈥檛 match these
patterns of absorption. 鈥淲e had clouds doing everything except sing and dance
for us, and we still couldn鈥檛 explain it,鈥� says Ramanathan, whose models
included 3D effects and heat-absorbing aerosols.

Without a plausible explanation some scientists decided that the observations
must be flawed. 鈥淚t鈥檚 possible there are things out there we don鈥檛 know anything
about,鈥� says Ackerman. 鈥淏ut it鈥檚 hard for me to accept that there鈥檚 a physical
mechanism out there that we have missed completely that is that big.鈥�

After more than two weeks of waiting, the prospects still don鈥檛 look good.
The sky over the airstrip is streaked with strands of thin cirrus cloud. A bad
sign. Inside a squat building at the end of the gravel airport road, half a
dozen researchers are already at their computers. Field manager Will Bolton
checks his watch. At 8 o鈥檆lock sharp the call he鈥檚 been waiting for comes in.
It鈥檚 mission scientist Bob Ellingson. The latest satellite images are
discouraging, he says: there is no slow-moving, solid cloud layer headed their
way. Bolton tells everyone that today鈥檚 mission is scrubbed; according to
Ellingson鈥檚 forecast they may have to wait three days to get another shot.

This is ARM鈥檚 second Oklahoma airborne experiment dedicated to settling the
cloud absorption debate. In the autumn of 1995, Cess mounted an airborne
experiment with Valero to try and confirm their initial findings. Three aircraft
flew in stacked formation over ARM鈥檚 research compound in northern Oklahoma.
After working through their data, Cess and Valero believed they had not only
confirmed their earlier work but had recorded even more absorption than before.
鈥淯nder heavy overcast conditions, clouds absorbed twice as much as clear skies,鈥�
recalls Cess. 鈥淚t was phenomenal.鈥�

But when the scientists reported these results at the ARM meeting in San
Antonio in 1997, they got a stormy reception. Other researchers vigorously
attacked the data, pointing out inconsistencies in the results from different
instruments. Worse, they maintained, with only one day of heavily overcast
skies, there simply wasn鈥檛 enough data to confirm the findings.

Three years on and defences and critiques of this experiment are still being
published. The single case of extremely high absorption by heavily overcast
skies on 30 October 1995 particularly puzzled some researchers. 鈥淭he
[reflectivity] of this very thick cloud only reaches something on the order of
55 per cent. This really set alarm bells ringing for me,鈥� says Peter Francis,
manager of airborne radiation research at the Meteorological Office at Reading
in Berkshire. He says that they often see much higher reflectivity than that, as
high as 70 or even 80 per cent. 鈥淲hy are our clouds different than the ones they
were seeing?鈥�

One possible explanation is if the sensors flying above the clouds
underestimate the amount of sunlight reflected from the clouds, those uncounted
photons will show up as being absorbed by the cloud. Francis says that there is
little support for Cess鈥檚 findings, at least within the European atmospheric
radiation research community. The results from the first Oklahoma experiment,
and allegations of instrument or data analysis error, further polarised the
research community. 鈥淭he field is pretty divided,鈥� Cess acknowledges, 鈥渁nd it鈥檚
been pretty emotional.鈥�

Independent confirmation of the enhanced cloud absorption findings by other
research groups has yet to arrive. An analysis of cloud observations collected
over several years around the world by the Met Office鈥檚 airborne research
programme agreed well with current radiation transfer models for a negligible
cloud absorption anomaly. The Japanese Cloud-Climate Study in 1998 also found no
excess absorption in clouds over the East China Sea.

Ackerman, who became chief scientist for the ARM programme in the midst of
the cloud absorption controversy, says they鈥檝e reached an impasse. 鈥淲e have data
sets which purport to show anomalous absorption, but we have no viable theory
that explains it. And each of these data sets has internal inconsistencies.鈥�

That鈥檚 why the DoE decided to fund a second experiment, which began this
February in Oklahoma. To resolve questions about instrument performance, seven
different radiation-measuring instruments would fly side by side to track the
total solar radiation and particular wavelength regions. After some
arm-twisting, all the instrument teams agreed to participate in the experiment,
collecting more extensive measurements of individual wavelength regions in the
solar radiation than anyone has ever done before. By seeing which wavelengths
are absorbed, scientists will get clues to what might be absorbing the light.
Many of the researchers feel this will be the key. 鈥淲e really don鈥檛 understand
the wavelength at which this phenomenon is occurring,鈥� says Cess. 鈥淭hat鈥檚 been
the big bugaboo.鈥�

To simplify data analysis the target this time would be unbroken, thick cloud
layers鈥攖he closest nature can get to the simplified cloud sheets of the
computer models. Only one aircraft would fly above the cloud tops, looping back
and forth in a tight daisy-wheel pattern centred directly over the world鈥檚 most
sophisticated cloud-and-radiation observatory.

Now the Oklahoma cloud trackers are beginning to get nervous. After almost
three weeks in the field, they have bagged their prey only once. Four days of
observations is deemed the absolute minimum for a successful experiment, and as
clear spring skies approach, time is running out.

Finally the weather starts to 鈥渋mprove鈥�. On 17 March satellite images show
the entire state blanketed with a thick cloud layer. The aircraft takes off in
the late morning and flies above the clouds, collecting data for an hour and a
half while the instruments below record the sunlight reaching them.

By the time the experiment ends in early April, the aircraft and instruments
have gathered three more days of observations over stratiform cloud layers,
giving the mission its minimum quota of cloudy sky data. But is this enough to
settle the debate?

Ackerman hopes so. And many researchers are anxiously waiting for a look at
the new data. But Stephens, one of enhanced absorption鈥檚 sternest critics, is
not optimistic. 鈥淲hen you have a debate that鈥檚 developed like this, there鈥檚
never a time when that debate ends, when there鈥檚 an absolute right or wrong.鈥� He
suspects the issue will gradually fade away as the discrepancy between models
and observations gets smaller.

Ramanathan believes the problem may be too complex to solve with a modest
experiment focused on tracking photons. 鈥淭his system is so highly turbulent, we
will never agree on how to sample it. You鈥檙e not going to capture all the
photons to settle the issue.鈥� The solution, he feels, may not come from the
atmospheric radiation community alone. His hunch is that there may be some
complex chemistry on the surface of cloud droplets that lets aerosol particles
stick to their surface, which would make them absorb more. 鈥淚f there is
absorption, I think this is the only thing that can explain it.鈥�

Susan Solomon of the National Oceanic and Atmospheric Administration鈥檚
Aeronomy Laboratory in Boulder, Colorado, is taking up the challenge 鈥渙ne
molecule at a time鈥�. Solomon thinks that nitrogen dioxide generated by lightning
inside storm clouds could account for some of the cloud absorption seen in
Cess鈥檚 research. But the high levels of the gas needed to cause significant
absorption are tied to local phenomenon like thunderstorms and air pollution
sources, and couldn鈥檛 account for the enhanced absorption reported globally.
Solomon and colleagues at the University of Bremen are now trying to get a
global picture of nitrogen dioxide distribution with observations from a remote
sensing satellite.

The scientists collaborating on this year鈥檚 experiment plan to gather in a
couple of months to discuss their new data. Ackerman expects this meeting will
iron out questions of instrument performance and data analysis before
atmospheric scientists at large get their hands on the results. This way, he
hopes, the heat of the debate will gradually dissipate, and some light can
finally shine through.