Vanouver astronomer Phil Gregory says his research indicates it is most likely four planets orbit the star Gliese 581, one of which is shown in this artist's concept. (European Space Observatory)

An Earth-like planet orbiting a star 20 light years away may not exist after all, according to a new study by a B.C. astronomer.

Last September, a team of astronomers led by Steven Vogt of the University of California, Santa Cruz announced the discovery of a fifth planet orbiting the star Gliese 581.

Known as Gliese 581g, it was found orbiting in a region of space known as the Goldilocks zone, where temperatures on the surface may allow liquid water to exist.

The astronomers detected the planet in data collected by the Keck telescope's HIRES instrument and the HARPS instrument attached to the European Southern Observatory's La Silla telescope.

But two weeks after the announcement, Swiss astronomers claimed they could not detect Gliese 581g in the HARPS data.

'The extra two planets that Vogt et al claimed … the grounds don't exist any longer for me to support that.'— Phil Gregory, University of British Columbia

A new analysis of data, appearing on the pre-press website arXiv.org and submitted to the journal Monthly Notices of the Royal Astronomical Society, backs the Swiss finding.

Phil Gregory, professor emeritus at the University of British Columbia in Vancouver, says that according to his research it is most likely four planets orbit the star, which would rule out Gliese 581g.

"If you combine the two data sets, then the most probable hypothesis is four planets," he said.

"The extra two planets that Vogt et al claimed … the grounds don't exist any longer for me to support that."

Uncertainty underestimated: Gregory

Gregory believes one of the reasons Vogt and his colleagues "detected" Gliese 581g was that they underestimated the level of uncertainty in their measurements.

Gregory calculated the uncertainty to be 1.8 metres per second, several times greater than the 0.5 metres per second quoted by Vogt.

"I can't make sense of their data unless I find there is an additional noise they haven't accounted for," Gregory said.

Another assumption made by the U.S. researchers was that the orbits of all the planets are circular. But according to Gregory, and the Swiss team of astronomers, one of the planets appears to have a highly elliptical orbit, which may be skewing the data.

Steve Vogt says he and his colleagues "stand solidly" by their original findings.

"I have studied [the paper] in detail and do not agree with his conclusions," Vogt said.

Gregory accused of manipulating data

Vogt is concerned that Gregory has unfairly manipulated the HIRES data.

"By doing so, he finds a solution that is more consistent with the HARPS data only," he said.

Vogt also dismisses the claim that one of the planets orbiting Gliese 581 has a highly eccentric orbit. He says this inflates the uncertainties in the HIRES data, making it impossible to detect Gliese 581g.

Gregory says while his study failed to detect Gliese 581g, future data may support it.

"I don't want to take away or underestimate the challenge of disentangling six-planet signals all buried in a noisy, very sparse collection of data," he said. "It's really an amazing challenge. With more data it [Gliese 581g] may, sometime in the future, become reality."