Now You See Them, Now You Don't

James M. Cordes, T. Joseph W. Lazio, Carl Sagan

Cornell University

Over the past 30 years, searches for extraterrestrial intelligence have come across many signals that, for a brief moment, met all the requirements for an alien transmission --- except that they were never seen again. The lack of redetection has not been for want of trying. The META search, described below, has tried hundreds of times to relocate a tantalizing signal, up to five years after its original detection, without success.

The usual conclusion has been that the signals were actually radio interference from here on Earth --- somebody's malfunctioning walkie-talkie or a wayward satellite. But recent research into how signals propagate through space has raised another possibility: that the bleeps were weak extraterrestrial signals momentarily amplified by the interstellar medium. In that case, there would be a simple reason why scientists haven't been able to find those signals again: the conditions that amplified them have yet to repeat.

The amplification occurs because of a process known as scintillation. Density variations in the interstellar medium --- in particular in the ionized portion of the thin gas in the space between the stars --- refract radio waves. Just as refraction in the Earth's atmosphere causes stars to twinkle, refraction in the interstellar medium causes radio signals to fade in and out. The effect is pronounced for signals from compact sources, such as pulsars. In fact, astronomers first noticed interstellar scintillation in their observations of pulsars during the late 1960s.

What is the impact of scintillation on SETI? Most searches work by looking for signals that are many times stronger than the background noise. Scintillation can amplify a previously undetectable signal above the noise, effectively increasing the sensitivity of the search. If there are extraterrestrial signals, it is more likely that they will be discovered because of scintillation. This is because SETIs make extremely large numbers of independent observations. Even if amplification is rare, sooner or later the search will encounter it.

Conversely, scintillations hamper reobservations. Within a few minutes of the initial observation, the amplification comes to an end, and the source sinks back into the noise. A huge number of reobservations would be required before the signal would once again, by chance, scintillate to a detectably bright strength.

Fortunately, the statistics of strong scintillation are well known. (If the intensity of a signal in the absence of scintillation is I, then in the presence of scintillation the probability that the scintillating signal will exceed a threshold S is exp[-S/I].) About 14 percent of the time,scintillation more than doubles the signal strength. About 40 percent of the time, it dims the signal by at least one-half. Notice that scintillation is more likely to dim rather than amplify the signal.

We have reanalyzed the results of one of the most comprehensive searches to date, the META search by Paul Horowitz and his students at Harvard, and discussed in detail in Horowitz & Sagan (1993). It scanned the northern sky for five years for exceptionally narrowband signals at the frequencies of 1420 and 2840 megahertz and investigated more than 10 trillion combinations of sky position and frequency. After weeding out terrestrial radio interference, META was left with 11 signals it could not explain. Some of these signals lay close to the Galactic plane, which is exactly where we expect most extraterrestrial signals to originate. The chances of this happening randomly is less than 2% according to Horowitz & Sagan. Yet, despite hundreds of reobservations, none of the candidates has been seen a second time.

We evaluated three models for the signals:

  1. Noise from the sky and the radio receivers;
  2. Alien signals temporarily amplified by scintillation; and
  3. Radio interference from human sources.

We found that the first model cannot explain the signals, but either of the other two models is a satisfactory explanation, statistically speaking. The META candidates could have been real alien signals. Or they could have been signals from mundane, but rare, terrestrial sources that had a statistically improbable correlation with the Galactic plane and that mimicked extraterrestrial signals in all other respects but one: repeatability.

Figures

  1. hypothetical signal
    A hypothetical signal from an ASP member on a planet orbiting a distant star. On this graph, the amplitude of the antenna output varies with time because of noise from the electronics, background astronomical objects, leakage from radio stations, and other sources of interference. If the antenna output was connected to a television, most of the time only static would be visible (box at upper left). A signal might be buried in the noise, but if it is weaker than the detector threshold (dotted line), it cannot be seen. Interstellar scintillation may allow brief periods of detectability, during which the signal towers above the noise. In this case, the television screen would show the alien transmission (box at upper right). Diagram by Joseph Lazio and David Kaplan.

  2. The Wow signal
    In 1977, Ohio State University radio astronomers John Kraus and Robert Dixon saw this fleeting blip on their printouts. It sure looked like an alien signal, yet it never recurred. Was it an extraterrestrial transmission that the interstellar medium temporarily boosted? We may never know. Photo courtesy of Ohio State University Radio Observatory in Columbus.

References

JAMES M. CORDES is an astronomy professor at Cornell University. His research interests include neutron stars, gamma-ray bursts, the interstellar medium, the search for extraterrestrial intelligence, wave propagation, and signal-processing techniques. Both he and Lazio regularly observe at Arecibo, the Very Large Array, and the Very Large Baseline Array.

T. JOSEPH W. LAZIO is a graduate student in the Department of Astronomy at Cornell University in Ithaca, N.Y. His research interests include the interstellar medium, the search for extraterrestrial intelligence, wave propagation, and signal-processing techniques. His email address is lazio@spacenet.tn.cornell.edu.

CARL SAGAN is the David Duncan Professor of Astronomy and Director of the Laboratory for Planetary Studies at Cornell University. His research interests include the origins of life, the search for extraterrestrial intelligence, and the chemistry of planetary atmospheres and surfaces, and the evolution of the Earth's environment.

Copyright

The copyright on this paper is held by the Astronomical Society of the Pacific, © 1996 Astronomical Society of the Pacific. Originally published in Mercury, March/April issue.

WebTechs HTML 3.0 (Beta) Checked!