The Whole Earth Telescope is a collection of small telescopes at various locations around the Earth. They combine forces to perform uninterrupted, 24-hour observations of variable stars. The project was originally designed to record the light curves of pulsating white dwarf stars, but can be used to monitor any variable star which changes on short time scales. Its main purpose is to provide data with which to perform asteroseismology -- the measurement of stellar oscillations to determine the interior structure of stars.

Background

Optical astronomy from the ground is limited by (at least) two important things -- the cycle of day and night, and local weather conditions. If you are attempting to measure how the brightness of stars change over time, you will have daily gaps in the data while the object is invisible because the Sun is up, the star has set, or the weather is poor. When you then perform a Fourier transform on the data, the variability spectrum becomes corrupted -- the real frequency spectrum is convolved with a window function corresponding to the frequency of the data gaps.

Here's an example. Suppose you have a star which varies with a pulsation frequency of four cycles per day (i.e. the star goes from bright to faint to bright again four times per day -- once every six hours). In a perfect universe, if we can constantly measure the brightness and perform a Fourier transform of the data from the time domain to the frequency domain, we would see a peak in the Fourier spectrum at four cycles per day,

```

|                       *
|                       *
|                       *
|                       *
amplitude|                       *
|                       *
|                      * *
|                      * *
|**********************   **********************
-------------------------------------------------
0     1     2     3     4     5     6     7    8
frequency (cycles/day)

```

However, we don't live in a perfect universe. Suppose we can only record the data twelve hours per day because of daylight. We will then have data gaps of twelve hours in the light curve. The Fourier transform of the data gaps results in the window function shown below. (Specifically, it is a modified shah function, with a taper caused by the finite width of the data windows.) The spacing of the peaks corresponds to one cycle per day, and its integer multiples:

```

*
*     *
*     *     *
*     *     *     *
window  *     *     *     *     *
function *     *     *     *     *     *
*     *     *     *     *     *     *
**   * *   * *   * *   * *   * *   * *    *
| ***   ***   ***   ***   ***   ***   **** *****
-------------------------------------------------
0     1     2     3     4     5     6     7    8
frequency (cycles/day)

```

So when we go to transform the observational data, we find the true signal from the star convolved with the window function, resulting in a spectrum like that shown below.

```

|
|                       *
|                       *
|                       *
|                 *     *     *
|                 *     *     *
observed |                 *     *     *
amplitude|           *     *     *     *     *
|           *     *     *     *     *
|     *     *     *    * *    *     *     *
|    * *   * *   * *   * *   * *   * *   * *
|****   ***   ***   ***   ***   ***   ***   ****
-------------------------------------------------
0     1     2     3     4     5     6     7     8
frequency (cycles/day)

```

Instead of seeing a single peak at four cycles per day, we see a forest of peaks separated by one cycle per day -- the separation of the data windows. This is known as aliasing. This isn't too serious for stars which only have one pulsation frequency -- you just pick the highest peak of the bunch, and you have your answer. But suppose the star pulsates with two superimposed frequencies? You'd then have the same thing as above, but with two separate sets of peaks centered on the two real frequencies. But then suppose the star pulsates in three frequencies? Five? Fifty? It obviously gets messier and messier as the star's behavior becomes more and more complex, and harder to disentangle what is physical reality and what is an artifact of your measurement sampling. The best way to fix this problem is to lessen or eliminate the effects of the window function. You can do this by eliminating gaps in the data. On Earth, the only way to do that is to distribute many telescopes longitudinally around the Earth, and combine the data from different telescopes into a single light curve.

The Whole Earth Telescope

The idea behind the Whole Earth Telescope originated in the mid-1980's at the University of Texas -- Austin Astronomy Department by several astronomers interested in white dwarf stars. White dwarf pulsations had been known about for years, and those who studied them were using the principles of asteroseismology to learn about their composition and internal structure. The difficulty with their work was that photometry conducted at a single telescope suffers from the aliasing described above. This wouldn't be so bad except some white dwarfs have dozens of pulsation modes, and the aliasing made it difficult to disentangle the real pulsation frequencies from the alias frequencies. So they hit upon the idea of asking several observatories around the world to observe the same star that astronomers at UT Austin were. Lo and behold, you could mesh all the data together into one single light curve, and the aliases were gone (for the most part). It was then easy to pick out the real signals from the data artifacts, and the field of white dwarf asteroseismology took a big leap forward. Their first observing campaign was undertaken in 1988 to study CR Bootes, a cataclysmic variable system with a pulsating white dwarf in the center. Since then, they've had twenty more observing campaigns on other white dwarf stars to study things like how magnetic fields and rotation affect pulsations, and crystallization of white dwarf cores.

While the project was run from Texas for its first decade of operation, it is now run from Iowa State University as part of its International Institute of Theoretical and Applied Physics, a member of UNESCO. WET doesn't own or run observatories. They work with local astronomers at 21 observatories around the world to coordinate observations of targeted stars. Any members who participate in observations are then given co-authorship credit on any publications written from the data, behind the Principal Investigator who actually organizes the campaign and analyzes the data. There are member observatories on every continent except Antarctica, including sites in New Zealand, Australia, (mainland) China, India, Poland, France, Spain, Italy, Israel, South Africa, Brazil, Chile, and several observatories in the United States.

Several other groups have followed the lead of the Whole Earth Telescope to set up their own networks. These include the Global Oscillation Network Group (GONG) to study the Sun (based in Tucson, Arizona), and the Delta Scuti Network to study delta Scuti and related stars (based in Vienna, Austria).

Sources:
http://wet.iitap.iastate.edu/
http://www.gong.noao.edu/
http://www.deltascuti.net/
And personal experience (came close to getting a job with WET a few years ago).