Are We Really Alone? NASA’s Kepler Mission Discovers 1,284 New Planets

13th May, 2016 by

This week 100TB brings you the highlights from the world of science and technology: the Kepler Space Telescope and a new false positive probabilistic analysis confirms the discovery of 1,284 new planets.

Kepler’s Discovery of 1,284 New Planets

NASA’s announcement took the world by storm once again: not only has their Kepler space telescope discovered 1,284 new planets, but chances are that 9 of these could be habitable. Ellen Stofan, chief scientist at NASA Headquarters in Washington sums it all up:

“This gives us hope that somewhere out there, around a star much like ours, we can eventually discover another Earth.”

What do we know about the 1,284 new planets? NASA believes that around 550 are Earth-like and rocky, 9 of which sit in the “habitable zone”. This is the precious orbit around a star where the planet is far enough from the star not to have a molten boiling liquid surface with extremely high temperatures, yet close enough to the star where water remains liquid and there is enough warmth for life not to freeze.

The current number of known planets outside our Solar System is 3,264, and the number of known planets which potentially could support life tally at 21. How does Kepler actually detect planets? “Kepler spots planets by looking at stars and watching for any dimming, which can happen when a planet passes in front of a star.” And how are these findings analyzed and confirmed? Until now it was an arduous procedure using observations to do just that, but there is a new statistical and simpler way, thanks to Timothy Morton and his team.

100TB, SciTech, probability, NASA, Kepler Mission

False Positive Probabilities: A New Method to Confirm Planets

The original method of planet validation through observation and radial-velocity measurements of a planet’s wobble required the planets under consideration to be of a certain size. The planets Kepler is discovering are often small and are in orbit around faint stars, which don’t allow the measurements needed for the classical approach. Hence, a new method of validation had to be created: the so called “probabilistic validation” analysis technique.

To take things to another level, Timothy Morton and his team of Princeton University have developed the first code to analyze and process batches of data of large numbers of potential planets automatically.

In their paper, Morton et al state: “The principle of probabilistic validation is to demonstrate that all conceivable astrophysical false positive scenarios are negligibly likely to be the cause of a transit candidate signal compared to the explanation of a planet transiting the presumed target star.” This basically means that all possible scenarios of the candidate not being a planet are calculated to be so small that they are meaningless, hence making it probabilistically true that the candidate is a planet. Basically in this approach probabilities are assigned to various hypotheses “that might describe a transiting planet candidate signal”. The likelihood is determined by how similar the observed transit signal is to the expected outcome calculated by the hypothesis. Using a Markov Chain Monte Carlo method, the likelihood is calculated. The results then fall into three categories: confirmed, candidate and false positive.

The results of the code analysis of 7,056 candidates showed 428 candidates with a “false-positive probability (FPP)” of over 90%, meaning they are likely not to be planets. 1,935 candidates were calculated to have a FPP of less than 1% meaning they are likely to be planets. 1,284 of these planets were previously unknown, hence the amazing discovery of Kepler. To all physics lovers there is some more good news: Morton’s code is publicly available to use on any exoplanet candidate!

Wishing you an out-of-this-world weekend!

(Visited 10 times, 1 visits today)