Some Of The Exoplanets Discovered By Kepler Space Telescope Might Be False Positives


A recent study pointed out the main flaw of NASA’s current technique in confirming exoplanet candidates spotted by Kepler – the agency’s formulas do not take potential instrument errors into account.

The Kepler Space Telescope has had a very successful run so far in the nine years or so since it was launched. Since 2009, NASA has used data from Kepler to confirm close to 1,300 exoplanets as of 2017, as it continues to detect more unexpected dips in brightness in stars, and possibly more planets well beyond our solar system, as signified by those dips. However, a new study suggests that there might be some exoplanets that aren’t quite confirmed after all — false positives that suggest Kepler isn’t 100 percent accurate when confirming a brightness dip as a planet.

In a study published earlier this month in the Astronomical Journal, researchers took a more in-depth look at how astronomers decide whether the Kepler Space Telescope has found an exoplanet or not. As noted on the paper, astronomers used a second telescope to double-check whether Kepler had indeed found a new planet after detecting a dip in brightness. In some instances, the brightness dip was actually caused by another object causing the light from the star to dim, or by noise in the data sent back to Earth, according to Popular Mechanics.

As further summarized by Popular Mechanics, astronomers changed their methods in confirming the potential exoplanets spotted by Kepler when the space telescope began spotting a growing number of possible leads. Outer Places added in its own summary of the new study that Kepler had become “so prolific” in detecting planets that NASA no longer was able to keep up and double-check all of those exoplanet candidates. With that in mind, scientists switched to a technique where they used mathematical formulas to determine the likelihood that Kepler found a planet, only making confirmations if the chances were at 99 percent or greater.

According to the scientists behind the new study, that’s where the problem with the new technique lies, as NASA’s formulas do not take possible inconsistencies or errors in instrument readings into account. More tellingly, the study noted that the potential errors are at their most crucial when NASA is trying to confirm exoplanets similar to Earth, meaning those that are of a similar size and distance from their host star. Popular Mechanics wrote that it’s just as easy for the signals Kepler detects from those planets to be little more than anomalies with the observatory’s instruments, especially now that it’s been close to a decade since Kepler launched.

In a statement quoted by NPR, researcher Fergal Mullally of the SETI Institute said that he finds it hard to suggest Kepler might not be as accurate as once thought because it remains an “absolutely wonderful” instrument.

“It was exquisite in the quality of the data it could detect. But nothing is perfect.”

The researchers specifically analyzed the exoplanet Kepler 452b, which was discovered in 2015 and thought to be an “older cousin” or “older brother” to Earth, but was one of the exoplanets confirmed via statistical formulas, as opposed to observations from a second telescope. With potential instrument errors figured into the equation, the researchers estimate that there’s a 16 to 92 percent chance that Kepler 452b is a planet — well below the 99 percent requirement for Kepler exoplanet confirmation, despite the high upside.

All told, the study authors believe that about nine out of 10 Kepler Space Telescope exoplanets are legitimate, with the rest potentially being false positives, assuming all candidates are similar to Kepler 452b. This could be further corroborated by NASA’s recently launched Transiting Exoplanet Survey Satellite (TESS), but as Outer Places noted, the best form of corroboration might come from the James Webb Space Telescope, which is not expected to launch until 2020 at the earliest.



Source link

Leave a Reply

Your email address will not be published.