Commentary

Observational Ambiguity

The frontier of science is marked by the ambiguity of the data. An instrument's technical limitations, the limited number of events or sources, and the complexity of phenomena are the roadblocks that confound theorists attempting to develop a theory. At the frontier, only some data show the physical signatures of the source; more often, one finds signatures that are either artifacts of the instrument or statistical flukes, leading to a body of evidence that supports a range of contradictory theories. Part of the art of theory is deciding which observations to believe outright, which to give more weight to, and which to discount. Another part of the art is to learn when to reevaluate the significance of data when new observations are made.

The phenomena that I spent most of my career studying, gamma-ray bursts, has a peerless history of ambiguous data generating numerous contradictory theories. Much of this is a consequence of the difficulty of conducting gamma-ray observations in astronomy. Gamma-ray bursts were first seen by U.S. satellites operated by Los Alamos National Laboratory that were monitoring the Nuclear Test Ban Treaty. These satellites were designed to monitor Earth's atmosphere for above-ground nuclear explosions. After they were placed in orbit, they began observing short outbursts of gamma-ray from sources in space that lasted from several seconds to hundreds of seconds in duration. These observations were not presented by Los Alamos scientists until after the bursts were observed by Russian spacecraft designed to observe astronomical sources.

The question was what types of sources produce these events. The best information for answering this question is a precise location on the sky for the event, so that you can observed its source with optical, radio, and x-ray telescopes. But this information was absent from the first observations. Gamma-ray sources are difficult to localize on the sky because of the nature of gamma-rays. Unlike optical radiation, or even x-ray, gamma-rays cannot be manipulated with optics. This limits the methods of determining a direction to measuring the time of arrival of widely-separated detectors, the measurement of a sources brightness in detectors of different orientation relative to the source, and the determination of a direction of motion of a gamma ray through its interactions with the detector. The Los Alamos researchers knew that the gamma-ray bursts were from space rather than from Earth because of the shielding of some satellites by Earth and from the time that the radiation arrived at different satellites. But for finding the gamma-ray source with a telescope the data was useless.

Campaign buttons were printed for a debate held in honor of the debate between Curtis and Shapley about the nature of galaxies. This debate concerned the origin of gamma-ray bursts. Are they from objects within the Milky Way, are they from objects at high cosmological redshift, or are they from other regions, such as from within our own solar system? Lamb argued a galactic origin and Paczynski argued a cosmological origin. Consensus tilted towards the cosmological origin at that time.

In the absence of a source, theorists were free to associate gamma-ray bursts with virtually any source they chose. Most researchers associated gamma-ray bursts with neutron star systems within the Galaxy, because they resembled other neutron star systems within the galaxy such as the x-ray pulsars, a binary-star system with a neutron star that pulls material from its companion. These systems produce both x-rays and gamma-rays, so the association of gamma-ray bursts with them was natural. The rapid start of the gamma-ray burst, which can be much less than a second, also suggested a very compact and energetic source, which again made the association with galactic neutron stars plausible.

The observation that nailed this association for many theorists was the detection of a line in the gamma-ray spectra by the Russian Venera spacecraft and the subsequent observations of pairs of lines by a Los Alamos experiment on the Japanese Ginga spacecraft. A lines of this type could be produced by electrons moving in a magnetic field of the strength expected for neutron stars.

But could the line be something else? Gamma-ray detectors do not produce spectra; researchers produce spectra by finding a simple model spectrum that, when folded through the model for the detector's response, reproduces the raw count data returned by the detector. The problem one has is that the detector itself creates lines in the raw count spectrum for a source without lines. The reason is that a gamma-ray above a threshold energy will unbind an electron from the atom its bound to, while a gamma-ray below this threshold energy will not unbind the electron. The practical consequence is that the detector suddenly becomes more opaque to gamma-rays with energy above an ionization threshold. If this characteristic is not well modeled, this effect will appear as a line in the spectrum.

So lines in gamma-ray spectra present one ambiguity. A second arrived on March 5, 1979, when an extremely bright gamma-ray burst was observed by many spacecraft, including interplanetary spacecraft. This event was well-localized through precise measurement of the arrival time at each spacecraft. To everyone's surprise, this gamma-ray burst was located on a supernovae remnant in the Large Magellanic Cloud, a satellite galaxy of the Milky Way. What was the brightest gamma-ray burst doing in another galaxy if all the other dimmer gamma-ray bursts are nearby and within our own galaxy? To make the problem more vexing, how can a compact source produce gamma-rays of that intensity? When the density of gamma-rays becomes high enough, they interact to produce electron-positron pairs. This is the inverse process of electron-positron pairs annihilating to produce two gamma-rays. If electron-positron pairs are produced, the remaining gamma-rays would scatter with the pairs, and the gamma-ray bursts would acquire the spectrum characteristic of a thermal source. The March 5th gamma-ray burst spectrum had no hint of a thermal spectrum. On the otherhand, the spectrum was atypical of other gamma-ray bursts. So what does this mean? Is the association of the burst with the supernovae remnant a coincidence, with the gamma-ray burst nearby, or is that burst in another galaxy, implying that nature has a way to avoid a thermal gamma-ray burst, and implying that other gamma-ray bursts are even more distant? Or is this bursts with its somewhat different spectrum a different type of burst, with a different source?

So how does one proceed in such an environment? In truth, theory cannot advance under these circumstances. Many theories were advanced, with some placing gamma-ray bursts in other galaxies, other keeping them within the galactic plane, and others placing them in the galactic halo. They were associated with every known and conceivable source. Without definite observations that constrain this complex physical phenomenon, theory is too unconstrained to provide useful results.

With time the problem was resolve observationally. Researchers with improved instruments were able to show that gamma-ray bursts were not in the galactic plane. Later, the observation of optical counterparts to highly-localized gamma-ray bursts within hours of the burst finally settle the issue of what produced gamma-ray bursts: gamma-ray bursts are from the most distant supernovae, an idea that few considered and virtually all rejected because of the inability to develop a priori a plausible mechanism for creating the burst.

Which observations proved correct? The observations of lines are now universally believed to be instrumental effects. On the other hand, it appears that the event of March 5, 1979 is not a gamma-ray burst at all, but a different type of source called a soft gamma-ray repeater. And what of the theoretical objection concerning pair creations? A source moving at close to the speed of light can create a high density of gamma-rays without pair creation occurring. A simple solution, and one supported by current radio observations of gamma-ray burst afterglow.

At heart astronomy is an observational science, and without a sufficient observational base, theory has little to offer. The value of theory is not realized until the observations are sufficient to tightly constrain the bounds of theoretical research.