HOW do you model the world鈥檚 oceans? This week, 300 scientists will gather at
the Netherlands Congress Centre in The Hague to answer that question. They鈥檙e
not interested in isolated readings of temperatures or tides, plankton or
pollution. Their goal is nothing less than to wrap the planet in a web of
sensors that will track the ever-changing physical, chemical, and biological
processes that occur in the oceans. They hope that this information will
ultimately help them to predict conditions above and below the surface of the
seas and oceans as readily and regularly as meteorologists anticipate conditions
in the atmosphere.
The Global Ocean Observing System (GOOS), as the project is known, will help
to revolutionise the world鈥檚 marine industry. 鈥淭he aim is to produce daily
forecasts and other forms of practical information for shipping, fisheries,
mineral exploration ventures, and other commercial operations,鈥� says John Woods,
one of the programme鈥檚 organisers and professor of oceanography at Imperial
College, London. Shipping will use ocean currents to cut fuel consumption and
journey times. Measuring the strength of deep-sea currents will allow oil
companies to design rigs that can withstand them. And governments will use the
data on levels of pollutants to monitor water quality.
Launched after the Earth Summit, held in Rio in 1992, GOOS links institutions
and government agencies from more than a hundred nations in a campaign to
develop oceanographic services for the new century. The scientists behind the
project are so sure that data from GOOS will have a huge commercial impact on
the marine industry and environmental monitoring that they expect governments
and industry to foot the bill, which may run to billions of dollars a year. The
money will not come from research budgets, which would be swamped by the cost of
the programme, but from other government agencies and the industries that expect
to benefit.
Advertisement
Moneymaker
The money will be well spent, contend the organisers of GOOS. The marine
economy is worth around $1000 billion a year. A better understanding of
the way the oceans work will both save money and generate new sources of income.
鈥淔rom the beginning, we said we would justify this investment on sound economic
grounds,鈥� says Woods.
With a few exceptions, information about oceans is now gathered haphazardly
and irregularly鈥攅ven barely at all in some, such as the Arctic Ocean.
There is no standard for storing or analysing information, so it can be
difficult to combine data from several projects.
For the relatively small regions such as ports and estuaries, where data is
taken regularly from many places, computer models can mimic a few marine
processes, such as three-dimensional tidal currents and coastal erosion. But
modelling entire oceans needs regular data from many points on and beneath the
surface. When this is not available, the computer must estimate the conditions
between two data points, often a hundred or more kilometres apart.
The first task for GOOS is to coordinate and build on the existing work.
鈥淔rom day one, it has been emphasised that GOOS will be built upon existing
activities,鈥� says Neil Anderson, a marine biologist at the University of
Maryland in College Park who is helping to shape research into coastal and
environmental issues. The project must standardise and automate the collection
and checking of data, so that it can be quickly distributed to those who need
it. This is no easy task: the data will be coming in from sources as diverse as
satellites making large-scale observations from space to thousands of automated
buoys measuring nutrients and salinity in locations all over the world.
The building blocks of GOOS will be the existing ocean measurement programmes
and a host of projects are already under way. Early next year, the World Ocean
Circulation Experiment, a seven-year study of oceanic currents sponsored by the
UN, will finish its task of data collection, and begin to analyse the haul. The
US, Japan, France and Australia are jointly funding a programme to monitor
conditions under the Pacific with a string of 60 buoys moored beneath the
surface, while the European Union is sponsoring a project to model the influence
and variability of the Asian summer monsoon and its link with the ocean. 鈥淎
patchwork of bits and pieces is already in place that works pretty well, and
gives GOOS a global skeleton,鈥� says Nicholas Flemming, director of EuroGOOS, the
enterprise鈥檚 European coordinating group, based at the Southampton Oceanographic
Centre at the University of Southampton.
The skeleton is being fleshed out. Satellites, for example, are becoming
increasingly important. Within the next five years, France, Japan and the US
plan to put a series of Earth-observation satellites in space that will be
invaluable for GOOS. At least three will have radar altimeters sensitive enough
to reveal topographical differences at the surface of the sea as small as 4
centimetres. 鈥淲ind, currents and barometric pressure distort the sea surface
over large areas for days at a time,鈥� Flemming notes. 鈥淎 column of water that鈥檚
warmer and less saline will stack up higher than an equivalent column of water
that鈥檚 colder and less saline. If you can measure the topography at the sea
surface, you can work out patterns of currents as well as the density of the
飞补迟别谤.鈥�
The same device also measures wave height. 鈥淲ave height tells you about wind
speed and direction,鈥� says Trevor Guymer, director of ocean circulation research
at Southampton. 鈥淭hat鈥檚 important for routeing ships around rough seas or
developing storms.鈥�
With the data from these devices, researchers have already discovered that
wave heights in the North Atlantic have been increasing over the last 25 to 30
years. 鈥淭hat means the seas there are getting rougher. In other parts of the
ocean, waves are getting calmer over time. That鈥檚 a useful indicator that some
significant change is taking place on a decadal time scale,鈥� adds Guymer.
Heat sensors on satellites can already measure the temperature of a square
kilometre of ocean to an accuracy of 0.5 掳C鈥攆rom 2000 kilometres. A
synthetic-aperture radar system aboard the European Space Agency鈥檚 ERS-2
satellite, which was launched in 1995, can even spot differences in the texture
of the sea surface and use this information to pinpoint oil slicks and other
pollutants. The radar system can measure weather conditions with resolutions as
fine as 30 metres. 鈥淚t鈥檚 causing some excitement,鈥� Guymer notes.
This latest generation of remote-sensing satellites can also measure the
colour of the oceans. 鈥淏y looking at several wavelengths of visible light, you
get information on water quality, and particularly on chlorophyll content, which
is an indicator of biological activity,鈥� says Guymer. 鈥淏ecause the biology of
the ocean is intimately involved in the carbon cycle and the carbon dioxide
question, colour becomes an important indicator of climate change.鈥�
But GOOS must also correlate these large-scale observations with localised
measurements of nutrient levels in the ocean. 鈥淥ne area where we鈥檙e very short
of sufficient capability is in measuring nutrients in the ocean, which are
fundamental markers of biological activity,鈥� says Flemming. 快猫短视频s have yet
to develop autonomous sensors that measure levels of nitrate, phosphate, ammonia
and oxygen in seawater. 鈥淭hese things exist in the laboratory, but not as
instruments which can be left alone at sea for a year,鈥� he says.
They soon could be, however. Nitrate is important in studies of biological
activity since it is a key food for plankton, which anchor the ocean鈥檚 food
chain. Now researchers in several labs are perfecting devices that measure
nitrate concentrations by adding chemicals that react with nitrate to produce a
colour change, or by illuminating a sample of seawater with ultraviolet light
which is reflected by nitrate molecules. Measuring the colour of the water or
the amount of reflected light gives an idea of the nitrate concentration.
Bacterial signature
While such methods are useful for inferring the level of biological activity,
researchers would prefer to measure it directly. An instrument that may allow
them to do this rests on Christopher Scholin鈥檚 workbench at the Monterey Bay
Aquarium Research Institute, 135 kilometres south of San Francisco. Scholin is a
marine biologist and his device identifies the presence of ribonucleic acids or
RNA鈥攖he molecular blueprint used to synthesise proteins and an indicator
that can identify organisms uniquely.
鈥淚n microbiology, classification of microorganisms has fallen almost entirely
to the analysis of ribosomal RNA,鈥� says Scholin. 鈥淪ome portions of the molecule
are common鈥攜ou and I share them with oak trees and bacteria.鈥� But other
parts are unique and so act as a signature of a particular group or species of
organism. 鈥淚n some cases, the molecule even identifies a specific population
within a species,鈥� says Scholin.
His prototype is an adaptation of a commercial biomedical device which draws
in a sample of fluid and carries it past a card that is chemically treated to
change colour in the presence of RNA from a particular organism. In Scholin鈥檚
version, the reaction signals the presence of two algae,
pseudo-Nitzschia and Alexandrium, which bloom in warm weather and
deplete waters of oxygen.
The method is ideal for a hospital lab but has limited application at sea
since the cards can be used only once. However, Scholin is working on a reusable
card and wants to extend its sensitivity to as many forms of life as he can. He
also wants to automate testing. 鈥淚n tests over the next year, we鈥檒l find out if
we鈥檝e been successful in creating a reusable dipstick鈥攕omething that could
process multiple samples and be reused for some extended period,鈥� he says. 鈥淥ur
goal is to be able to measure and, ultimately, to predict the cycles and reasons
for the growth and distribution of organisms.鈥�
Cruising for data
With GOOS, the immediate problem is not a lack of technologies, but the lack
of means to deploy enough of them. 鈥淭he sheer number of measurements we can make
is vitally important to GOOS, so we need to reduce the cost of each one,鈥� says
Gwyn Griffiths, director of ocean technology development at the Southampton
centre. 鈥淭he problem is making instruments cheap enough to use in their hundreds
and thousands.鈥�
Today, about 500 sensor-laden buoys ply the world鈥檚 waves. Some are moored;
others drift with surface flows, or dive to follow subsea currents and only to
rise to the surface to broadcast their position and the results of temperature,
salinity and atmospheric pressure measurements to orbiting satellites. 鈥淭o
gather the density of data we really need, we should have one in each
100-kilometre-square section of ocean,鈥� notes Griffiths. But such a plan would
require 43 000 buoys, not 500. 鈥淎nd the results would still be sparse in
coverage,鈥� he adds.
To achieve anything near this figure, researchers must reduce the price of
these buoys, which cost $15 000 each. Engineers at Southampton and
elsewhere hope to do this by reducing the complexity and size of the buoys. With
mass production, the next generation of floats could cost as little as
$2000 apiece
But these buoys are essentially dumb, drifting where the ocean takes them.
Researchers would like to pick their targets more accurately. 鈥淭here are parts
of the ocean where processes are repetitive, consistent, or low in variability,
so you don鈥檛 need to measure them every hour or every kilometre,鈥� points out
Flemming. 鈥淵ou鈥檇 be wasting money if you did. In other places, very highly
energetic events or discontinuities occur. If you miss these, you could get your
forecasts very badly wrong.鈥�
To help study parts of the ocean of particular interest, Griffiths heads a
team that is developing an autonomous underwater vehicle called Autosub, which
can navigate the ocean depths laden with sensors. Seven metres long and 90
centimetres wide, the probe will carry 150 kilograms of sensing equipment and
travel perhaps 800 kilometres on internal batteries鈥攁 journey that would
take about 100 hours.
The sub will use up most of its energy commuting between the ocean depths and
sea surface. Along the way, it will 鈥渋nhale鈥� water through a port on the tip of
its snout. The water will flow among a series of sensors that will measure,
among other things, salinity, chlorophyll and nitrate concentrations, and then
be expelled through a vent on its side.
The sub will cruise at depths of up to 2000 metres and will surface
periodically to transmit its data via satellite to a ground station. Bobbing in
the water, it will fix its position by taking a reading from the global
positioning satellite system. Submerged again, it will navigate by dead
reckoning鈥攃locking its speed and tracking its heading with a magnetic
sensor. 鈥淐urrents affect trajectory,鈥� Griffiths notes. 鈥淭ypically, the Autosub
will run about one and one-half hours submerged. With a maximum expected current
of 1 metre per second, the Autosub might be 3 kilometres adrift between
readings. But it corrects its positioning data every time it surfaces.鈥� At the
end of each mission, the probe will rendezvous with a ship, swap its spent
batteries for new ones, perhaps be refitted with different sensors, and set off
again in search of fresh data.
This summer, Griffiths and his team tested the sub in waters off Weymouth on
the south coast of England. The vehicle took 19 short trips and performed
without a flaw. 鈥淚nitially, we plan to run autosubs between Scotland and
Iceland,鈥� Griffiths explains. 鈥淭hat鈥檚 a critically important transect, with
water flowing between the Arctic Ocean, the North Atlantic, and the Norwegian
Sea鈥攁nd the range is one which the Autosub can handle now.鈥�
With the data from vehicles like Autosub, researchers hope to be able to
fine-tune their models of how the ocean works鈥攁n essential for accurate
global forecasts. Modelling the ocean on computer is hugely complex. It begins
with a model of the simplest ocean processes, such as the tide, and gradually
adds more complex phenomena, such as wind, atmospheric pressure and currents in
three dimensions. Only when this model has been perfected can researchers begin
to look at the way the ocean moves sediment and erodes coasts, says Flemming.
This is the stage at which the researchers now find themselves.
Fiendishly difficult
The next step is to work out which chemical processes occur, and incorporate
these into the model. Chemical processes rely crucially on the transport of
sediment since pollutants and radioactive molecules tend to become adsorbed on
particles of mud and sand. Finally, researchers hope to be able to model the way
these chemical processes effect biological activity.
At the Sir Isaac Newton Institute of Mathematical Sciences at Cambridge
University, researchers met this summer to begin constructing numerical models
of the ocean鈥檚 biological processes and how they fit in with the physical and
chemical processes. 鈥淲e know the basic laws that govern physical processes in
the ocean, so we know we鈥檙e starting with the right equations,鈥� says Allan
Robinson, a specialist in geophysical fluid dynamics at Harvard University.
鈥淲e鈥檙e just now beginning to understand how to assimilate chemical and
biological data into the models,鈥� he adds. Constructing models of biological
activity such as the interaction between species and the food chain is
fiendishly difficult, and nobody is quite sure how they will achieve it.
But before GOOS鈥檚 models can even begin to make forecasts of physical,
chemical and biological activity, the floods of data their designers receive
must be collated, confirmed and compared, and then dispatched instantly to the
people who need them.
The task is neither easy nor cheap. But an infrastructure is emerging from
discussions among GOOS鈥檚 participants as they attempt to agree on what to
measure, where and how often to measure it, what protocols and standards to use,
how to express the results and where to send the compiled data. Wherever the
numbers arrive, complex software must then scan them for accuracy by checking
that the figures make sense. 鈥淔or example, it will see whether temperature and
salinity bear logical relationships to each other,鈥� Flemming explains. And since
the amount of data will be vast, the collection and checking must be
automated.
Researchers in Canada, Germany and Australia are already at work creating the
software that will automate and integrate these chores. 鈥淚t鈥檚 by no means beyond
the kind of tasks performed by banks or tax offices all the time,鈥� Flemming
points out. 鈥淢eteorological offices know how to do it, although they deal with
fewer parameters and much larger geographical scales. It鈥檚 totally within the
state of the art, but it means a lot of very, very good logistical design.鈥�
It also means investment from sponsoring industries and governmental
agencies鈥攁n investment which GOOS鈥檚 champions insist will pay huge
dividends. Flemming calculates that the world鈥檚 marine industries and services,
GOOS鈥檚 primary customers, could save at least $20 billion each year.
Determining how to make best use of transient ocean currents, for example, could
be very useful for shipping companies. 鈥淯sing currents to add half a knot to a
ship鈥檚 speed and reduce transit time and fuel costs by even a few per cent can
add up to significant value across the industry. Being able to route ships away
from storms is of value,鈥� Griffiths muses.
Cutting crop losses
GOOS also has other clients鈥攁griculture, aviation, power producing
companies and communities based near the sea鈥攚ho could save as much again.
Researchers hope to learn to forecast events like El Ni帽o, the disruption
of Pacific currents which causes torrential storms on the western American coast
from Oregon to Peru and produces droughts in Asia. 鈥淕iving farmers that
knowledge will help them decide on planting strategies,鈥� says Griffiths. One
study sponsored by the US National Oceanic and Atmospheric Administration
reports that accurate forecasts of severe weather systems that develop at sea
could save $200 million in annual crop losses in the southeastern US
alone.
So what will GOOS cost? 鈥淚f you add up what鈥檚 already being spent to monitor
and study the ocean through satellites, it might be $5 million a day,鈥�
says Woods. 鈥淎n inspired guess might be that GOOS will require as much
again鈥� another $5 million for Autosubs, moorings, drifters and
sensors.鈥� The total approaches $4 billion a year. 鈥淚f you add it all up,
you find that it鈥檚 not much different from what weather forecasters now spend to
maintain their systems.鈥�
Indeed, GOOS鈥檚 backers yearn to achieve a system as efficient as the global
meteorological version over the next two decades. Like weather forecasters,
GOOS鈥檚 researchers hope to unite many different kinds of information in an
attempt to view the world鈥檚 oceans as a dynamic, living whole. If they get their
way, GOOS could lead to forecasts that are as widespread as weather predictions
are today. Shipping forecasts will never be the same again.