Four planets in 'habitable zones' spotted within spitting distance of Earth

Astronomers say they used a new statistical technique to find four possible super-Earths orbiting in the habitable zone of two stars within 22 light-years of Earth, Gliese 667C and tau Ceti.

Astronomers say they have uncovered evidence for what could be four super-Earth planets orbiting within the habitable zones of two stars within 22 light-years of Earth.

Three of those candidate planets are among a tightly packed clutch of five that orbit Gliese 667C, part of a triple-star system 22 light-years away in the constellation Scorpius. The other possible planet is one of five orbiting tau Ceti, a sun-like star 12 light-years away in the constellation Cetus.

Taken together, the detections not only add to accumulating evidence that planets look to be more common than stars – and that planets in habitable zones could be more common than previously thought, some of researchers reporting the finds say.

The finds also illustrate the power of improved statistical tools to boldly uncover candidate planets where no planet had been found before.

The evidence for these candidate planets requires independent confirmation, the researchers caution. Still, the tools represent "a real breakthrough," says Steven Vogt, an astronomer at the University of California at Santa Cruz and a member of the team reporting the results for tau Ceti. The approach the team took leaves only about one chance in 3 million that the detections could herald something other than a planet.

Since the mid-1990s, astronomers have bagged more than 850 extra-solar planets. The ultimate goal is to find rocky planets with Earth-like masses orbiting within their stars' habitable zones – a region where under the right conditions, liquid water can form stable pools on the surface. Liquid water is considered an essential ingredient for organic life.

Different groups of astronomers had aimed three telescopes for various lengths of time at tau Ceti and found nothing. Led by Mikko Tuomi at the University of Hertfordshire in Britain, the team reporting this latest analysis applied relatively new statistical tools to the combined data from these telescopes.

The result: "Five planets came out: boom, boom, boom, boom, boom ... as clear as a bell," Dr. Vogt says.

Tau Ceti has about 78 percent of our own sun's mass, but its composition is quite similar, Dr. Tuomi's team reports. Its candidate planets range from 2 to 6.6 times Earth's mass.

The innermost object orbits the star once every 14 days, while the outermost takes 642 days to make its circuit. The fourth planet from the star, with a 168-day orbit, travels well within a zone where liquid water could remain stable on the planet's surface, the team estimates. However, the results don't speak directly to what the planets are made of. 

A similar story has played out for Philip Gregory, an astronomer at the University of British Columbia in Vancouver. Previous researchers had found two planets orbiting Gliese 667C, a red dwarf with 31 percent of the sun's mass. Using a broadly similar statistical approach, he reports detecting the initial two, plus three more planets. Three of the five fall within the star's habitable zone, he estimates.

There, the planets range in mass from twice to five time's Earth's mass. Orbit times for the planets range from seven days for the innermost to 91 days for the outermost. Here again, the fourth planet out – an object with twice Earth's mass and a travel time of one orbit every 39 days – is the habitable-zone winner.

The results from Tuomi's team were published online Wednesday by the journal Astronomy and Astrophysics. Dr. Gregory has submitted a formal paper reporting his results to the Monthly Notices of the Royal Astronomical Society in Britain and has posted a draft of the paper on the research server arXiv.

The two efforts were based on so-called radial-velocity measurements of the host stars. In essence, these are measurements of a star's back and forth motion as a planet orbits and its gravity tugs on the star. The motion appears as a back-and-forth wobble in the star's spectrum.

A big planet close in has a more pronounced effect than either a small planet close in or more-distant planets. For Earth-scale planets in habitable zones, the effect can be quite weak. This opens the door to confusion, because the star's own activity – star spots, flares, prominences, and other features – can also produce weak wobbles in the star's spectrum.

A star's "noise" lacks an important feature: the consistent, regular repetition of a planet's orbit. But that feature can be hard to detect amid a star's noise, dubbed "star jitter."

Rather than using a traditional approach of starting with the hunt for a planet-imparted wobble, Vogt explains, Tuomi devised a mathematical approach that focused on a star's noise.

To test the approach, the team needed a very bright star that had been a target of lots of radial-velocity measurements from different observatories and that had yielded no evidence of a planet. No planet must mean all jitter. Tau Ceti fit the bill.

The team combined 5,943 radial-velocity measurements gathered at observatories in Chile, Australia, and Hawaii over time spans of up to 13-1/2 years. The largest number of measurements came from a high-precision HARPS spectrometer at the European Southern Observatory's 3.6 meter telescope at La Silla. The team then looked for spectral signatures of the star's jitter across all three sets of observations and how that jitter changed over time.

Tuomi then developed a set of competing mathematical simulations of the star, each of which yielded the jitter patterns the team saw in the star. Then he introduced signatures of faux planets in each model to see which model had the highest probability of mimicking the star's jitter – and of revealing the faux planets.

The team applied the winning model of tau Ceti's star jitter to the full range of data coming from the star, and out popped the five candidate planets.

"These were signals underneath the noise that none of us had really known how to deal with," Vogt says. "These are very weak signals, and out they came."

"We have only now started to realize how important it is to model the stellar jitter as realistically as possible," Tuomi adds in an e-mail exchange.