Data analysis and mathematics used to predict Earth-like planets

Astrophysicists analyse data from NASA’s Kepler space telescope and apply a generalised mathematical rule to find new planets

Astrophysicists at the Australian National University (ANU) have predicted most stars in our galaxy have two habitable or Earth-like planets near them by analysing data from NASA’s Kepler space telescope and applying a mathematical rule.

The astrophysicists analysed data of 151 multiple exoplanet systems (planets outside solar system) coming from Kepler, which observes stars, and used a generalised Titius-Bode (TB) relation, a mathematical rule for predicting distances of planets from the sun. They were able to predict 228 planets Kepler is unable to see because it is biased towards capturing hot, huge planets close to stars.

“What we are doing is using computers for heavy duty crunching of data that’s coming from the Kepler space telescope. It’s looking for stars that would dim briefly for a few hours and then get bright again.

"That’s a tell-tale signal of what’s called a transit of a planet transiting in front of a star,” said associate professor Charley Lineweaver at the ANU Research School of Astronomy and Astrophysics.

The team looked at the distance of the planets from their stars where liquid water or H20 can exist.

“There seems to be a significant fraction of these [planets] that are small enough to be Earth like and have rocky surfaces with liquid water on them.

“But there’s one other thing you need and that is a planet that’s probably a little bit bigger than Mars. The reason for that is you have to have a planet that’s big enough to hold an atmosphere to have sufficient surface pressure to keep the water as a liquid. Otherwise it is like the moon, which is in the habitable zone but it has no atmosphere,” Lineweaver said.

The team generalised the TB relation by making less parameters (fixed values) so that they could make predictions that weren’t specific to the solar system.

“The TB relation was originally defined with only our solar system in mind. We wanted to not condition on anything, we didn’t want anything specific about the solar system to be included in the parameterisation, and we didn’t want the planet with the closest orbit to have more of a weighting in the fitting. So that’s why we moved one of the parameters and only fit two,” said Lineweaver.

Read: Machine learning used to predict hazardous solar flares

By predicting quite a large amount of potential “habitable real estate” in the galaxy, it’s easy for people to think that means the chance of crossing paths with aliens or other intelligent life forms on these planets increases, Lineweaver said. But that’s not necessarily the case, he said.

“The emergence of life might be rare, not based on our lack of ingredients but on a rare recipe, for example.

“Another bottleneck or explanation for the rarity of intelligent aliens might be that human-like intelligence to build the ability to send radio signals and spaceships might be something that is very, very rare and evidence from Earth suggests that we are the only species who can do it so therefore it can’t be a very common thing.

“The other possibility might be that once you do get human-like intelligence you are powerful enough to destroy yourself. Advance civilisations always self-destruct.”

The team’s work is published in Monthly Notices of the Royal Astronomical Society.

Follow Rebecca Merrett on Twitter: @Rebecca_Merrett

Join the TechWorld newsletter!

Error: Please check your email address.

Tags data analyticsmathematicsexpolanetexoplanet systemsAustralian National University (ANU)big dataCharley Lineweaverastrophysicsdata analysis

More about Australian National UniversityNASAResearch School of Astronomy

Show Comments
[]