≡ Menu

Allen Telescope Array Begins Work

The Allen Telescope Array, devoted both to SETI and astronomical observations, has begun operations. With 42 radio dishes now active, the array ultimately will be used to scan several billion stars in the Milky Way looking for the signals of an extraterrestrial civilization. That’s a staggeringly broad survey, and one that will be followed up by detailed examinations of a million star sample. The ATA is known as Paul Allen’s project, but he’s joined in philanthropy by the SETI Institute and UC Berkeley, among others.

Says Seth Shostak (SETI Institute):

“For SETI, the ATA’s technical capabilities exponentially increase our ability to search for intelligent signals, and may lead to the discovery of thinking beings elsewhere in the universe. It is the first major telescope in the world built specifically for undertaking a search for extraterrestrial intelligence.”

Allen Telescope Array

It’s always interesting to track how the press handles such stories, and this Seattle Timesarticle plays it straight, with a description of the installation’s 42 radio dishes northeast of San Francisco (at Hat Creek in the Cascade Mountains north of Lassen Peak) and their potential use for the study of astronomical phenomena — pulsars, supernovae, black holes, gravity waves — as well as the ongoing SETI work. Allen, who also backed the winning entry in the X-prize competition, recalls reading Heinlein’s Rocket Ship Galileo as the defining moment in creating his youthful interest in space.

Image: The Allen Telescope Array now begins observations. Hat Creek will eventually see 350 such dishes mining the sky for radio astronomy and SETI investigations. Credit: ATA/SETI Institute.

With the ATA now engaged in active research, we can ponder the differences between it and other observing programs. Time on Arecibo’s huge dish was necessarily limited, with the SETI Institute’s Project Phoenix managing to use the instrument for just three weeks each spring and fall. That added up to a total of 100 observing days between September of 1998 and March of 2004. The Allen Telescope Array offers 24-hour a day access seven days a week.

Moreover, the frequencies available to the ATA — between 1000 and 10,000 MHz — cover five times the range of Project Phoenix. With 350 individual dishes planned for the site, the ATA should be able to examine multiple stars simultaneously, another step up from Arecibo. Thus Project Phoenix’s observations (covering roughly 800 stars) can be upped to the huge numbers Allen’s team envisions, a major step forward for SETI work.

But will the new array produce better results than Project Phoenix itself, which found no sign of intelligent extraterrestrials? As always, that’s the biggest SETI question, and those of us who doubt that advanced civilizations are numerous in the galaxy won’t be surprised if a detection takes a long time, if and when it comes. But the ATA model, using mass-produced 6.1-meter radio dishes and commercial telecommunications technologies, seems the way to go, particularly since it’s backed by philanthropic funding.

$50 million is buying a lot of observing power (see this news release for funding details). We can hope for a breakthrough reception, but even if it doesn’t come, we’re gaining an instrument whose multi-tasking capabilities allow radio astronomy to be conducted on a daily basis even as the search continues. In those terms, this is a gamble well worth taking, and one with significant payout no matter what the SETI outcome.

Comments on this entry are closed.

  • Lubo October 12, 2007, 17:11

    What about laser or infrared form of communication?
    What if the extraterrestrials are using laser instead of radio waves?
    SETI is doing fine, but it alone won’t achieve any success, not any time soon. What is needed is another SETI like organization(or the same with another section) which will emphasize on light….

  • Administrator October 12, 2007, 18:04

    One organized search using optical SETI is Harvard’s:


    You can read more on optical SETI here, among other places:


    Berkely is also doing optical SETI work, so there is interest out there.

  • andy October 12, 2007, 19:06

    Good to have these facilities. While I am sceptical that SETI will work (though I wouldn’t complain if they do find aliens… provided the aliens aren’t inimical to humanity), radio could potentially yield a lot of information about extrasolar planets: the astrometric precision is very good and could potentially yield the parameters of Earth-mass planets, and since this is an astrometric method it removes the sin(i) degeneracy that is the major deficiency of the radial velocity method.

    Furthermore, the detection of exoplanetary radio emissions may be the easiest way to measure rotation periods of exoplanets (provided the planet is magnetic, of course), or even detect moons: some of Jupiter’s radio emission is a result of the Io-Jupiter flux tube interaction, though space-based facilities would probably be necessary because the ionosphere absorbs low-frequency radiation.

  • Adam October 12, 2007, 21:31

    Hi All

    Another possibility is the detection of microwave emissions from the putatively warm (~300K) surfaces of interstellar planets wrapped in 1,000 bar hydrogen atmospheres. We might find several such planets closer than the nearest stars. Way-stations to Alpha Cen.

  • Administrator October 12, 2007, 21:38

    Interesting idea, Adam. And interesting to speculate on just how many such planets might be out there. Steinn Sigurðsson at Penn State has been doing interesting work on such possibilities.

  • Administrator October 12, 2007, 21:44

    Andy, if you get the chance, help me to understand the sin(i) degeneracy business re radial velocity methods. I need to get up to speed on this.

  • andy October 13, 2007, 11:52

    The sin(i) degeneracy arises from how the radial velocity method works.

    Treat the planet-star system as a two-body system: the planet and star orbit their centre of mass. The velocity of the star is a vector in the orbital plane. The radial velocity method measures the component of the velocity along the line of sight. The inclination is the angle between the line of sight and the normal to the orbital plane (so inclination values sufficiently close to 90 degrees result in a transit).

    So we have a situation with the true and measured velocity vectors forming a right-angled triangle: the true velocity is the hypotenuse, the inclination angle is opposite to the measured velocity. Simple trigonometry gives the following relationship between measured velocity, true velocity and inclination:

    v_measured = v_true * sin(i)

    It can be shown that in the limit that the planet is much less massive than the star, the velocity amplitude of the star and the mass of the planet are directly proportional. The velocities in the above expression can then be replaced with masses:

    m_measured ≈ m_true * sin(i)

    where I’ve used the approximately-equal symbol because I’ve neglected higher order terms in m_planet/m_star in the relationship between velocity amplitude and mass.

  • Administrator October 13, 2007, 14:29

    Extremely helpful, Andy. Thanks!

  • David October 14, 2007, 8:32

    It is also useful to keep in mind that for a randomly chosen inclination value i, there is a 90% probability that the true value for M and the measured value for M will differ by by no more than a factor of 2.3.

  • Spacefiles October 18, 2007, 10:35


    Removing the sin(i) degeneracy with a radio telescope – how? Does it essentially see the track of those planets? What resolution will this telescope have? (the article you linked mentions a 100 uarcsec value, which would be enough within 100 lightyears… or even 1000)

  • Spacefiles October 18, 2007, 10:36

    And also, why isn’t this method used to discover exasolar planets? Is it too slow? Too expensive?

  • andy October 18, 2007, 11:02

    Spacefiles: the radio astrometric method is not a Doppler method, but an astrometric method. What the astrometric method does is monitors the star’s position in the sky (it does not measure motion back and forth along the line-of-sight), which gives 2 spatial dimensions+time, as opposed to 1 spatial dimension+time, and that extra dimension breaks the inclination degeneracy.

    (On the other hand, astrometric measurements do not necessarily tell you which of the orbital nodes is ascending and which is descending, but this degeneracy in orbital orientation does not affect the mass determination – this degeneracy can be overcome by combining astrometry with radial velocities to give a full 3D orbit)

    Main problem with astrometric measurements is that it requires very precise position measurements, which means the atmosphere is a real nuisance. Contrast the radial velocity measurements, which only require frequencies of spectral lines, for which the atmosphere is less of a problem.

    Furthermore, astrometric wobble increases with semimajor axis, so the planets causing the largest astrometric wobbles are going to have long orbital periods, which doesn’t help the detection rate. Contrast with radial velocity wobbles which are largest for planets close to their stars. Both methods are more sensitive to massive planets.

    Astrometry has been used on some extrasolar systems: the planet candidate HD 33636b was shown to be a red dwarf star through astrometry (wasn’t radio astrometry, but principles are the same), the orbit of Epsilon Eridani b has been constrained astrometrically. There was a study which claimed that most “planetary” systems were being viewed pole-on (and hence the objects in question were brown dwarfs or even stars), but this was probably an artefact of the data reduction process.