ÐÓ°ÉÔ­´´

FOR ALMOST 30 years, solar physicists have known that the surface of
the Sun is shaking. The study of solar vibrations, helioseismology, is the
best source of information that researchers have about the internal properties
of our nearest star. But before physicists can polish up their theories
of the interior of the Sun, as Jaymie Matthews described last week in ‘Good
vibrations from the stars’, they need to observe and measure the oscillations.

There are several methods available. One arises from the fact that,
as the Sun expands and contracts during an oscillation, its temperature
changes slightly, making its brightness vary by a few parts per million.
Such changes in intensity have been detected and studied from space, most
notably by the Solar Maximum Mission satellite in the 1980s. In the past
year beautiful results have come from a device carried by the Soviet probe
Phobos on its flight to Mars, and run by a team headed by Claus Frohlich
of the Physical Meteorology Observatory of Davos, Switzerland.

Another approach is to study the minute changes which the oscillations
produce in the diameter of the Sun. Henry Hill and his colleagues at the
University of Arizona have carried out this kind of work since the 1970s.
However the most common and productive approach, to date, is to find out
the speed of the Sun’s surface as it vibrates to and fro. By studying precise
details of the spectrum of sunlight, solar physicists can measure this velocity
very accurately.

Measurements of velocity in astronomy depend on the doppler effect,
the change in a wave’s frequency when its source is moving (or when the
observer or listener is). The frequency shift is greater the higher the
relative speed: it could make a red traffic light appear green, but you
would have to drive towards it at about one-third of the speed of light.

In the case of the Sun, the doppler effect means that as its surface
moves to and fro, the light detected by an observer on Earth is shifted
alternately towards blue and then red. However, a typical velocity during
a solar oscillation is 1 metre per second, which is about 3 kilometres per
hour, a slow walking pace. At such speeds, the doppler effect changes the
wavelength by only 1 part in 300 million – much too tiny to see directly
as a change in the colour of sunlight. Fortunately, a feature known as Fraunhofer
lines allows astronomers to detect these small shifts indirectly.

In the early 1800s, Joseph Fraunhofer, the pioneering German optical
scientist, noticed that when he split sunlight into its spectrum of colours,
there were dark narrow lines across the bright colours of the spectrum.
As light streams out from the Sun, it passes through the gases of the solar
atmosphere, which absorb certain wavelengths of light, and transmit all
others. Each Fraunhofer line corresponds to one of the absorption wavelengths
characteristic of a particular element. Solar physicists use the position
of these narrow lines, which are shifted up and down in wavelength by the
doppler effect, to observe the tiny changes which occur during an oscillation.

Researchers can then look at the average velocity over the Sun’s surface
for a day at a time. The velocity follows a sweeping curve through the day,
changing by about 800 metres per second from dawn to dusk. This large change
is almost entirely due to the spin of the Earth; this and a few other effects
have no relation to solar oscillations, so physicists remove them from the
raw data. They are left with a complicated trace which varies by about 3
metres per second either side of zero. At times the trace is quite strong
and appears to vary in a regular way, at others it is very nearly zero.
Bearing in mind that, as it rotates, the edges of the Sun move at some 2
kilometres per second, it might seem that the results are no more than an
artefact of the measuring techniques, or an effect from the Earth’s atmosphere.
But careful analysis of results obtained simultaneously at different observatories
shows conclusively that the oscillations are of solar origin.

To put the data in a clearer form researchers produce a ‘power spectrum’.
In theory, any oscillation can be built up by adding together several simpler
vibrations with a range of amplitudes and wavelengths. When researchers
analyse a complex wave, they break it down into simpler component oscillations,
usually sine waves. The power spectrum is simply a plot of the power (the
square of the amplitude) against frequency, for each component oscillation,
showing the tallest spikes at the frequencies of the oscillations that are
strongest. The researchers use a mathematical technique called a Fourier
transform to produce such a spectrum.

Solar oscillations have a power spectrum with a few pronounced peaks
at frequencies of about 3 thousandths of a cycle per second. These correspond
to oscillations with cycle times of about 5 minutes. The fact that the power
spectrum contains many peaks tells researchers that the Sun is vibrating
at many well-defined frequencies, called the normal modes of vibration,
or p-modes. These are similar to the well-defined harmonic frequencies of
a musical instrument. During such an oscillation, the Sun’s surface will
move in and out by a few tens of metres, which is a tiny distance compared
with the Sun’s diameter – about a million kilometres. Individual modes
have velocities of, at most, 0.25 metres per second when averaged over the
visible disc of the Sun; sometimes several modes are vibrating strongly
at the same time and combine to give the maximum speeds seen, a few metres
per second.

Peak power

George Isaak and his colleagues from the University of Birmingham and
the Astrophysical Institute of the Canaries first published results of this
kind in 1979. Almost immediately, astronomers realised that these spectra
could be concealing more detailed information; theory predicts that each
of the peaks in the power spectrum should be multiple, provided that the
solar oscillations do not die away too quickly. In order to detect such
splitting in the power spectrum, however, astronomers need to record sharper
power spectra. A string of data from the Sun over a single day produces
peaks which are too broad to be of direct use.

The simplest way to increase the sharpness of the peaks – the resolution
– is to record data at one site for several consecutive days, and then string
them together, keeping track of when night-time and bad weather interrupted
observations. Strings of data lasting for weeks or months are needed to
study the details of the Sun’s vibrations. Although this does produce higher-resolution
spectra that show up the predicted extra structure, the processing itself
seriously affects the results.

The mathematical analysis which converts the velocity data into a power
spectrum picks up any regular pattern in the strings of data. The gaps in
data look like another component of the overall oscillation, with the result
that every genuine peak on the Sun’s power spectrum is accompanied by ‘sidebands’.
As (bad) luck would have it, the spacing of the sidebands, related to the
length of a day, is almost exactly the same as that between some of the
peaks that solar physicists wish to study. One sideband of each genuine
solar peak lies on top of a neighbouring peak.

The problems caused by sidebands are not merely cosmetic. Astronomers
need to measure frequency differences accurately in their investigations
of the interior of the Sun. For example, to study the rotation and magnetism
of the Sun they must determine the structure of individual peaks, an exercise
severely hampered by sidebands superimposed on genuine peaks. One of the
goals of helioseismologists, therefore, is to rid their spectra of sidebands,
as far as possible. To achieve this they need to make continuous measurements
of the Sun’s oscillations.

There are three possible solutions. Eric Fossat and colleagues from
the University of Nice in France and the Bartol Institute in the US obtained
a stretch of data spanning about five days in 1980 by observing from the
South Pole. Their spectrum was free of sidebands, but lacked the high resolution
of data sets taken over longer periods of time. Weather conditions at the
Pole make the chances of recording continuous stretches of data over significantly
longer periods than this very unlikely.

A second approach is to put measuring equipment on board a satellite
in an appropriate orbit, one that receives unbroken sunlight. In 1995, the
European Space Agency and NASA will launch a joint mission known as SOHO
(Solar Heliospheric Observatory), which will carry equipment to measure
solar oscillations. SOHO should produce the highest quality data yet, but
only for a few years. The Sun shows interesting effects which vary on timescales
of decades, the most famous being the 11-year cycle of sunspots, which reaches
its peak this winter. Our best hope for the short-term is to set up ground-based
networks of instruments around the world, each with at least one site in
sunlight at all times. Three main networks are under way.

At Birmingham University, in collaboration with the Astrophysical Institute
of the Canaries, we are setting up a network of six instruments at good
sites for astronomy. We are already studying the unresolved Sun, that is,
we average the velocity readings over the whole of the Sun’s visible surface.
The instruments are, therefore, sensitive to the simplest modes of oscillation
which, it turns out, penetrate the Sun most deeply. Although some of the
instruments are run by hand, we are building apparatus that is mainly automatic,
and have also built a testing and training facility in Birmingham.

At present, we have three stations, at Izana on Tenerife, Haleakala
on Hawaii, and Carnarvon, Western Australia. When all three instruments
work well and the weather is good, we can take data 24 hours a day. Of course,
failure of instruments and bad weather mean that strings of data are rarely
unbroken for long, but we do have some results which are complete enough
to reduce the problem of sidebands considerably. The power spectrum of results
taken during two months of 1987 gives some indication of the resolution
and cleanliness we hope will become commonplace. We plan to set up the remaining
three instruments over the next few years.

Fossat’s group based in Nice is deploying a second network of unresolved
Sun instruments, known as IRIS. These instruments are operated by hand by
collaborating groups at the chosen sites. In recent years the group has
set up instruments at Stanford in California, La Silla in Chile and Kum-Bel
in the USSR, with four more planned to complete the network.

Finally, a group based at the National Solar Observatory in Tucson Arizona
led by John Leibacher and John Harvey is working on GONG, the Global Oscillation
Network Group. The instruments in this network look at the Sun in a slightly
different way. By resolving images of the Sun into many segments, the team
will be able to study modes of oscillation that vary across the Sun’s disc,
something not possible with the other two networks (see Box ).

The resolved measurements of the GONG network and the unresolved results
of the other two groups will be complementary. GONG and IRIS are organising
regular international workshops throughout the duration of the projects,
so that we can pool ideas and results. The results of the ground-based networks
will also be an invaluable resource when the SOHO data, which should be
of higher quality, though of shorter duration, are available in the mid-1990s.
With all this data, researchers should see the Sun in a new light.

Helioseismology is still in its infancy, and we have only started to
study the interior of our nearest star. In truth, the early results have
posed more new questions than they have answered. Close investigations of
properties such as the rotation, thermal convection and chemical composition
of the inside of the Sun are, however, under way. We hope that the experimenters’
efforts to construct observing networks will ease the subject through its
adolescence into a healthy and rewarding maturity.

Roger New is a senior lecturer in Applied Physics at Sheffield City
Polytechnic working in collaboration with the University of Birmingham.