¿ìè¶ÌÊÓƵ

‘When the NASA people said they were going to put up a satellite that
could tell you wave height, wind speed and direction and provide images
of the waves day and night, irrespective of clouds, we just laughed,’ recalls
Peter Taylor, an oceanographer from the James Rennell Centre for Ocean Circulation
in Southampton. Oceanographers usually face long, unglamourous cruises on
research ships, often in inhospitable conditions, to collect the data they
need. Sitting back and letting a satellite do the laborious work seemed
too good to be true.

But NASA was not joking. In July 1978, it launched Seasat, a remote
sensing satellite to study the oceans and their interaction with the atmosphere.
The only trouble was that Seasat stopped working after just 100 days when
its power system failed-and oceanographers were again forced to rely on
long cruises during which they would periodically dip a sampler or sensor
into the water.

Next week’s planned launch by the European Space Agency of its first
remote sensing satellite, ERS-1, promises to give oceanographers another
chance to study the seas from space. Their hopes are high. The data Seasat
collected during its brief life surpassed all expectations.

Unlike most remote sensing satellites whose sensors just passively receive
visible or infrared light reflected from the Earth, Seasat used radar. The
satellite emitted pulses of microwave radiation that bounced back to it
from the sea’s surface. These reflections allowed scientists to calculate
the direction and the speed of the wind, the height of waves and the overall
height of the surface itself. The system could operate at night because
radar is not dependent on the Sun’s rays, and the longer wavelength of the
radiation meant it could penetrate clouds and rain. Though the system was
not intrinsically as sensitive as one based on radiation of a shorter wavelength,
such as visible or infrared light, intensive processing of the data compensated
for this deficiency and enabled scientists to produce images of photographic
quality.

But analysing microwave data is more complicated than assessing optical
information. Radar satellites produce an enormous amount of data and its
processing is laborious. This helps to explain why radar was not used before
and why scientists are still working on Seasat’s results. ‘It is probably
the most pored over set of data of any remote sensing satellite,’ says Trevor
Guymer, head of the team that analyses satellite data at the James Rennell
Centre. The JRC was inaugurated earlier this year to coordinate the British
contribution to the World Ocean Circulation Experiment, a seven-year project
of the World Climate Research Programme, which was initiated last year by
the International Council of Scientific Unions and the World Meteorological
Organisation.

The WOCE, which will gather data on ocean circulation and its influence
on climate, was timed to make the most of ERS-1. The satellite’s primary
aim is to provide oceanographic information using instruments similar to
those aboard Seasat. In addition it will be able to make the most accurate
measurements to date of the temperature of the sea surface. ERS-1 is scheduled
to lift off from Kourou in French Guiana, on top of an Ariane 4 rocket,
on the night of 3/4 May. ¿ìè¶ÌÊÓƵs around the world are poised to begin
unravelling the nature of the oceans from the streams of data expected.

¿ìè¶ÌÊÓƵs already know much about the atmosphere’s influence on climate.
Understanding the oceans is the next important step in comprehending the
processes that govern climatic changes.

The oceans have an enormous capacity for storing heat; the ease with
which they release it to the atmosphere is what gives them such an influence
on climate. As warm water flows from the tropical waters of the Atlantic
towards the North Pole, about 1015 watts of heat is released into the atmosphere,
equivalent to the output of a million power stations. This heat warms the
eastbound winds across the Atlantic and explains why Europe experiences
much milder winters than the East coast of North America. The oceans heat
the atmosphere more than the Sun’s rays do and the top 3 metres of the ocean
contain as much heat as the 100 kilometres or so over which the atmosphere
extends.

But it is not only the surface of the oceans that is important. Circulation
currents, which can extend over thousands of kilometres, mix warm water
at the surface with cold water down to depths of hundreds or even thousands
of metres. These large circulations, such as the Gulf Stream, are driven
by the force of wind moving over the ocean surface. But while the land and
the air can change temperature in a matter of hours, the oceans take their
time. Deep, slow-moving currents can take centuries to transfer heat from
one side of the world to the other.

Oceanographers know something about these processes from studies done
on merchant ships and aircraft, but there is no substitute for the volume
of data that a remote sensing satellite can produce every day. ERS-1 will
also provide more information about the Southern Ocean-the only unbroken
ring of water that links the Atlantic, Pacific and Indian oceans-where few
ships go.

Besides filling the gaps in our knowledge, ERS-1 will also allow scientists
to estimate the extent and age of polar ice and the thawing and freezing
cycles of the polar caps. The satellite will be able to detect oil slicks
and other forms of pollution, and, moving inland, study patterns of land
use. ESA intends to make some forms of raw data available within a few hours
so that meteorological offices can improve the accuracy of their forecasts
and provide shipping with warnings of high winds and waves and positions
of ice floes.

The largest sensor on board ERS-1 is called the active microwave instrument.
This consists of two radars: a synthetic aperture radar (SAR) that produces
images of the sea surface with a resolution of 25 metres, and a wind scatterometer
that measures wind speed and direction.

The SAR looks down from the satellite sideways at an angle of 23 degrees
from the nadir and records a swathe of the surface 100 kilometres across.
To produce images of the quality required from the satellite’s altitude
of around 780 kilometres using just one set of reflections, as a photographic
camera would, the instrument would need a receiving antenna about 10 kilometres
long. In fact, it manages with an antenna just 10 metres long by accumulating
reflected signals as it moves over the ground and then sorting them out
electronically to give an image of each point. Because the distance the
satellite moves while it is collecting all the microwave signals reflected
from a single point is about 10 kilometres, the antenna is said to have
a ‘synthetic aperture’ of 10 kilometres.

The SAR uses the Doppler effect to sort out the data it receives. As
the satellite speeds away from a point, the microwave reflections of constant
wavelength will appear to the SAR antenna to be stretched to a longer wavelength;
as it speeds towards a point, the reflections will appear to be condensed
to a shorter wavelength. This ‘wavelength shift’ allows the instrument to
pinpoint the location of the source of a reflection and sort out which signals
belong to which points.

Before Seasat, no one was sure that a SAR could produce images of waves
from space. Seasat proved it was possible, though the radar images are more
difficult to interpret than photographs. One of the main problems is that
objects moving on the Earth’s surface distort the Doppler reflections and
confuse the SAR’s image processor. For instance, when the SAR produces an
image of a (moving) ship and its (non-moving) wake, the ship appears no
longer at the head of the wake but to one side of it. As far as ships are
concerned, this is not a problem because we know where a ship should be
in relation to its wake. Besides, for surveillance purposes, these disjointed
images are useful for giving an indication of a vessel’s speed. For waves,
however, the distortion is more difficult to contend with.

Waves cause particles on the sea surface to move in a circular path
in the vertical plane. This means that if the satellite’s image of the peak
of a wave, where the particles are moving in one direction, is shifted forward,
then the image of the trough of that wave, where the particles move in the
opposite direction, is shifted back. The two images of peak and trough may
then end up in the same position in the SAR image and cancel each other
out leaving an apparently flat sea. Researchers at the British electronics
company Marconi, which leads the consortium that developed the SAR, are
working on techniques to overcome interpretation problems.

SAR images have many uses. They can detect oil slicks because slicks
alter the shapes of waves. They can map the seabed in shallow coastal waters
because this influences the profile of the surface. According to Guymer,
SAR images can also identify ocean fronts, analogous to weather fronts in
meteorology, where bodies of warm and cold water meet. The SAR spots the
greater number of waves generated by the increased convective mixing of
the air above the warm water. Knowing the location of ocean fronts is particularly
useful for fishing fleets.

The amount of data the SAR gathers, about 100 million bits per second,
is far too much for the satellite’s tape recorders to store. As a result,
the SAR can only be used when it is in direct contact with a ground station
equipped to receive its signals, such as ESA’s station at Kiruna in northern
Sweden. (The agency says there will be about 10 stations in operation when
the satellite is launched.) Also, the SAR’s microwave imaging signal demands
so much power that it can be used for only 10 minutes in each 100-minute
orbit of the Earth. Alternatively, the SAR can sample the ocean rather than
record it in its entirety, which demands less power and storage capacity.
Every 200 or 300 kilometres, the instrument analyses the reflections from
an area of about 5 square kilometres. It does not produce an image but measures
the dominant wavelength and direction of ocean waves. The samples are stored
and transmitted once every orbit. Taken together, they allow scientists
to derive a crude measure of the state of the oceans.

The other radar component of the active microwave instrument is the
wind scatterometer. Winds moving across water roughen its surface, which
then reflects the radar signals in a pattern, or backscatter, that depends
on the speed and direction of the wind. The scatter-ometer has three flat
antennas that analyse a swathe 500 kilometres wide to one side of the satellite’s
track. The antennas look at the swathe from different angles: one looks
perpendicular to the satellite’s track, another looks 45 degrees forward
and the third 45 degrees backwards. Each antenna looks at the backscatter
inside individual ‘cells’, 25 kilometres across, inside the swathe. The
three backscatter measurements are then compared to determine the wind speed
and direction in each cell to an accuracy of 2 metres per second and – 20
degrees. This is a very broad range though it is an improvement on Seasat’s
performance. Seasat’s scatterometer, which had only two antennas, produced
four possible wind directions for each cell analysed and meterologists often
could not choose between them. ERS-1’s three antennas narrow the choice
to two, in roughly opposite directions. This should enable meteorologists
to pick the correct one by comparing the results of adjacent cells. They
could also use other sources of wind data, such as weather ships, to help
them determine the true direction of the wind.

The problem arises because the scatterometer measures ripples on the
surface rather than the wind directly, says Andrew Lorenc, head of data
assimilation research at Britain’s Meteorological Office. This means that
mathematical processing, which can introduce arithmetical confusion, is
necessary before scientists can derive useful information about the wind.
Nevertheless, the Met Office hopes to be using the information for its weather
forecasting before the end of this year.

The second major instrument on ERS-1 is the radar altimeter. This uses
a dish antenna pointing directly downwards to send out pulses of microwaves
and then detect the reflected signal. Far from simply measuring the distance
between the satellite and the ground to an accuracy of about 10 centimetres,
the radar altimeter can also determine the height of waves and wind speed
by analysing the shape of the returned pulse (see Figure). It is only possible
to use radar altimetry on the sea surface because the height of waves is
roughly symmetrical, or near-Gaussian, about the mean sea height. According
to John Powell, head of the radar altimetry group at the Rutherford Appleton
Laboratory (RAL), this gives the return pulse a roughly predictable shape
that can be analysed.

Reflections on radar

The time delay between the transmitted pulse and the return pulse gives
the distance between the satellite and surface. But as the microwaves reflected
off the peaks of waves will return to the satellite sooner than those reflected
off troughs, the return pulse is spread over time to give an inclined rather
than a vertical response line. The shallower the inclination, the higher
the waves. Also, when high winds roughen the surface, more of the transmitted
pulse will be randomly scattered and less will get back to the satellite.
The stronger the return pulse, the lower the wind speed. These measurements
of the returned pulse are carried out 20 times a second and processed on
board to give an average value roughly once per second.

Altimetry measurements are vital to the study of ocean circulation.
Major currents cause variations in the height of the sea surface; there
is a height difference of 1.5 metres across the Gulf Stream. Eddies with
diameters of about 100 kilometres spin off the sides of major currents;
they contribute to dissipating the currents’ energy and mixing warm and
cold water. Eddies are crucial to the transfer of energy in the oceans and
can be spotted by the altimeter because they cause depressions in the sea
surface. The British arm of the WOCE is using an ocean circulation model
called the Fine Resolution Antarctic Model, or FRAM, to study the Southern
Ocean. Information from ERS-1’s radar altimeter on the extent and location
of eddies should enhance FRAM, which can resolve ocean eddies that are 100
kilometres across. Eventually, the researchers hope to use FRAM to study
all the world’s oceans. For the moment, however, the altimeter’s performance
is limited by its ability to distinguish between the effects of current
and the effects of gravitation on the shape of the sea’s surface.

One of Seasat’s greatest achievements was to map the deep seabed by
recording the undulations of the sea’s surface. Seasat’s altimeter images
showed variations in height of as much as 100 metres. These were not produced
by currents but by large features on and below the ocean floor such as trenches,
ridges and faults, which influence the Earth’s gravitational field and hence
affect the elevation of the sea’s surface. The ‘gravitational shape’ of
the Earth is known as the geoid and a sea devoid of currents and tides would
follow it exactly. But it is difficult to separate those height changes
caused by the geoid from those caused by currents, but not impossible. Changes
caused by the geoid do not move and so any change in elevation from day
to day can be attributed to currents. In order to obtain a global picture
of currents it is necessary to determine the geoid by other means and then
subtract the geoid from the altimeter image. A joint NASA/ESA satellite
specifically designed to map the geoid, called Aristoteles, is due to be
launched in 1997 and will make mapping the ocean currents with the radar
altimeter much more accurate.

One of ERS-1’s most valuable assets is its ability to calculate the
size and direction of wind and waves while it also measures the temperature
of the surface of the sea using the satellite’s only passive instrument,
the along-track scanning radiometer. RAL’s atmospheric science group, which
developed the ATSR, expects the instrument to be accurate to at least 0.3
°C, more accurate than any other radiometer in orbit.

Measuring sea surface temperature is normally difficult because moisture
in the air absorbs some of the radiation returning to the satellite. The
ATSR gets around this problem in a number of ways. First, it measures radiation
at four separate wavelengths in the infrared part of the spectrum. Moisture
in the atmosphere attenuates radiation of different wavelengths at different
rates. By comparing the brightness detected at the four wavelengths, scientists
can estimate the degree of attenuation. Secondly, the ATSR looks twice at
each point in its 500-kilometre-wide swathe. It looks ahead along the track
of the satellite at an angle of 52 degrees and also directly below the satellite.
By comparing the radiation received from each point through two different
paths, scientists can make another estimate of the attenuation. The ATSR
also contains a separate microwave radiometer that looks at the surface
directly below the satellite using two frequencies particularly sensitive
to water vapour in the air.

The ATSR views the surface using a circular, parabolic mirror that focuses
the incoming radiation through filters to direct the four wavelengths onto
four detectors, two of which are alloys of cadmium, mercury and tellurium,
and two are alloys of indium and antimony. To protect the detectors from
thermal noise, they are cooled to a temperature of 80 K using a mechanical
heat pump based on the Stirling cycle .

As the mirror spins rapidly, two apertures restrict its circular view
of the Earth’s surface to two arcs, 500 kilometres across. One gives a view
directly below the satellite and the other 52 degrees ahead. Inside the
ATSR, two so-called black bodies, one hot and one cold, are positioned where
the mirror’s view of the surface is obscured. The mirror scans the two black
bodies of known temperature every revolution, using them as references for
the sensors to gauge the temperature of the incoming radiation.

All the data from the four sensors, including the black body readings,
plus the data from the microwave radiometer are transmitted down to Earth
once per orbit for processing into maps of the sea surface temperature.
Because clouds obscure the surface for infrared instruments like the ATSR,
it will take a few weeks to build up a global map of sea surface temperature,
says David Llewellyn-Jones, head of the atmospheric science division at
RAL.

ERS-1 is not a glamorous piece of hardware that is going to give researchers
all the answers on a plate. It provides raw material that they must learn
how to interpret. Many research cruises and aircraft flights are planned
for the next few years to calibrate the data from the satellite and to see
how the processed information compares with the conditions at the sea surface.
The ATSR detects the temperature of only the top 1 millimetre of the sea;
how this relates to the bulk temperature of the water below will be closely
examined.

ERS-1 has a design life of three years. It may last for much longer
but the stresses on the microwave instruments from continually generating
high-powered signals will certainly take their toll. So that all the work
done in developing processes for handling ERS-1 data do not simply come
to an end when the satellite fails, ESA has approved a successor, ERS-2,
which is due to be launched in 1994 and will carry similar instruments plus
a new one to monitor atmospheric ozone. This will be followed towards the
end of the decade by the Polar Platforms, two remote sensing satellites
to be launched by ESA and NASA as part of the space station programme, which
should put remote sensing on a firm

* * *

A Stirling way to keep your sensors cool in space

The infrared sensors of ERS-1’s temperature instrument, the ATSR, need
to be kept at a temperature of -193 degres C so that their own heat does
not generate extraneous radiation that would interfere with the incoming
signal. Normally, such sensors on satellites are kept cool using passive
radiators emitting heat into the cold of space, but these cannot achieve
the low temperatures needed by the ATSR. Tanks of liquid nitrogen are another
option but these are very heavy.

Instead, the ATSR (along-track scanning radiometer) developed at the
Rutherford Appleton Laboratory uses a mechanical heat pump based on the
Stirling cycle, a thermodynamic cycle devised by the Scottish clergyman
and inventor Robert Stirling in 1816. The cycle can work as an engine, using
heat generated externally to produce mechanical energy, or it can work as
a pump, using mechanical energy to transfer heat.

RAL’s Stirling cycle cooler consists of a sealed chamber filled with
helium at a pressure of 10 atmospheres. The chamber contains a tube known
as the displacer that can move up and down inside it. The middle of the
displacer is filled with a porous material; in this case tightly packed
layers of phosphor-bronze gauze, which is known as the regenerator and acts
as a temporary heat sink. At one end of the chamber, from where heat is
ejected into space, is a piston. The other end of the chamber, the end that
draws heat from the sensor, is known as the cold tip.

The cooling cycle begins with one end of the displacer set back from
the cold tip and the piston up against its other end (see Figure). As the
piston moves back away from the displacer, the helium expands and absorbs
heat from the cold tip. The displacer then moves towards the cold tip, which
forces the helium through the regenerator where the gas picks up any heat
temporarily stored there.

With the displacer up against the cold tip, the piston moves forward
to compress the helium in the space between the piston and the displacer.
As the helium is compressed, heat is ejected from the gas, through the casing
and into space.

To complete the cycle, the displacer moves back towards the piston and
the helium is forced through the regenerator into the space between the
dis-placer and the cold tip. The helium temporarily deposits heat in the
regenerator as it passes through.

By driving the piston and the displacer, the cycle absorbs heat at the
cold tip, where the sensor is located, and ejects it at the other end of
the chamber, into space.

In RAL’s Stirling cycle cooler, both the displacer and compressor are
mounted on shafts that are held in position by two sets of flat discs made
of beryllium copper alloy. The shafts pass through the centres of the discs,
which have spiralling incisions that enable the discs to act as springs.
These let the shafts move easily back and forth along the axis of the chamber
but prevent them from moving from side to side and touching the inside of
the chamber. This makes the chamber free of mechanical friction, which should
help to extend its working life.

The cycle is performed very quickly, 44 times per second, with the compressor
and displacer driven by a magnet and coil mechanism similar to that in a
loudspeaker.

British Aerospace is making versions of RAL’s design for other space
uses and for military applications, such as cooling the sensors of night-vision
devices.