Space & astronomyPreprintData analysis3 min read

A BUTTERFLY COLLECTOR'S FORMULA SAYS COSMIC RADIO REPEATERS FADE FAST

Fast radio bursts are radio flashes lasting a few milliseconds, coming from other galaxies. Most of their sources have been seen only once. A small fraction repeat. How long such a source stays active is a key clue to what it is, but for the population as a whole it was poorly known: the few well-studied repeaters are nearby or unusually active.

Counting the butterflies nobody has caught

Shotaro Yamasaki, at National Chung Hsing University in Taichung, Taiwan, turned to an old branch of statistics: unseen species estimation. In a 1943 study, researchers looked at the butterflies a collector had caught in Malaya over two years. From the number of species caught once, twice, and so on, they asked how many new species one more trip would find. The same reasoning was later used to estimate how many words Shakespeare knew but never used, and how many flare stars in the Pleiades had not yet flared.

Its central identity is the Robbins formula (1956). If every species is caught at a steady rate, the number of species caught k + 1 times tells you how many specimens the species caught k times will give on the next trip — whatever the mix of common and rare species. For instance, the species caught twice tell you how many specimens the species caught only once will give.

A telescope as a butterfly net

A radio survey works the same way. The telescope collects bursts; each source plays the role of a species. The catch: a butterfly species is still there on the next trip, while a radio source may switch off.

So Yamasaki turned the formula’s assumption into the thing to measure. Take two windows of 180 days separated by a gap. Count the sources seen once, twice or three times in the first window, and use the formula to predict how many bursts those same sources should give in the second. Then divide what is observed by what is predicted. Steady sources give a ratio of 1, whatever the gap. A ratio that falls as the gap grows means the sources are going quiet.

The method needs no assumption about how often the sources burst or how far away they are. The data: the second catalogue of the CHIME/FRB survey, which watches the whole northern sky every day — 4,539 bursts from 3,641 sources, 83 of them known repeaters, over five years. After cleaning, 2,686 sources remain, 62 of them detected more than once.

Half the bursts missing after a year

  • For gaps up to 60 days, the ratio stays close to 1: the formula works.
  • At one year, it falls to 0.57. It stays near one half out to 900 days and reaches 0.36 at about three years.
  • Among 10,000 simulated catalogues of steady sources, none falls that far.
  • Sources that never stop but flicker at random fail too — unless they spend most of their time silent.

The drop is not an instrument effect: the method measures changes in the telescope’s sensitivity from the number of brand-new sources, and corrects for them.

Three ways to go quiet

Three simple pictures fit the data equally well:

  1. every source lives about 2.9 years, then stops for good;
  2. sources switch on and off, with active phases of about 800 days;
  3. most sources are short-lived (about 500 days) while about 30% keep going.

The test that convinces: models fitted only to the first 1,100 days of the survey correctly predict how the ratio keeps falling in the last 738 days, which they never saw. Steady sources predict no fall at all.

All three point to activity that declines over one to three years. Magnetars, a kind of young neutron star, have been proposed as the engines of these bursts. One to three years is far shorter than the ages of the magnetars in our own Galaxy, the author notes — so a repeater is more likely a magnetar caught in one short active episode than for its whole life. It also means some “one-off” sources may be repeaters that have already gone quiet.

Catch them while they last

A ten-year CHIME survey, or a more sensitive telescope, should start to tell the single-lifetime picture from the other two. Only the on/off one predicts that long-silent repeaters will come back at their old positions. Meanwhile, the paper ends on practical advice: follow new repeaters soon after they are found, because many sources seen today “may have gone quiet by the time we return.”