First Look at Kepler’s Complete Primary Mission Data Set

For half a decade, the world has witnessed an unprecedented stream of planetary discoveries from NASA’s Kepler mission which started its primary mission on May 13, 2009. While most of the previous extrasolar planets had been found by looking for the reflex motion of stars orbited by a planet using precision radial velocity measurements, Kepler looked for tiny dips in a star’s brightness from an orbiting planet transiting the star from our point of view. Since such transits only occur if the planet’s orbit is aligned nearly edge-on to our line of sight, they are relatively rare. For example, a planet in an Earth-like orbit around a Sun-like star has only about a 0.5% chance of producing an observable transit. Planets in smaller orbits have proportionally higher chances of producing transits.

In order to detect a significant number of these relatively uncommon planetary transits, Kepler observed the brightness of almost 200,000 stars in a single patch of sky straddling the border of the constellations Cygnus and Lyra covering 115 square degrees over a four year period. The only scheduled breaks in Kepler’s nearly continuous observations occurred every 93 days when the spacecraft was rotated by 90° to keep its solar panels and thermal radiator properly oriented in relation to the Sun. These events conveniently broke up Kepler’s data set into units called quarters.

Diagram showing the major components of NASA’s Kepler spacecraft. (NASA/Kepler Mission/Ball Aerospace)

To date, the majority of the extrasolar planets found by Kepler were discovered by processing an ever-growing subset of the data available from Kepler’s ongoing mission. With the announcement during the first week of January 2015 of the discovery of eight new planets allegedly in the habitable zones of the stars they orbit, Kepler had found 1,013 confirmed planets with another 4,175 planet candidates waiting for their planetary status to be confirmed with follow-up observations (see “Habitable Planet Reality Check: 8 New Habitable Zone Planets”).

New Analysis Results

With the end of Kepler’s primary mission on May 11, 2013 when a second reaction wheel failure left the spacecraft unable to accurately point at its predetermined patch of the celestial sphere, a total of 17 quarters of data were available for analysis by the Kepler project’s data processing pipeline. A team led by Shawn Seader (SETI Institute/NASA Ames Research Center) has now submitted for publication the first results of an initial analysis of the entire data set from Kepler’s primary mission.

Seader et al. used the latest version of the Transiting Planet Search (TPS) pipeline module to search the photometric data of 198,675 Kepler targets. Of these targets, 112,001 were observed in every quarter with 86,645 targets observed in a subset of the 17 quarters. They found a total of 12,669 targets that contained at least one signal that meets the following criteria for being a transit event:

– The transit events occurred at regular intervals
– A minimum of three transit events are observed
– The signal-to-noise ratio of the events is greater than a preset threshold
– The transit events pass four consistency tests designed to weed out false positives

Seader et al. repeatedly searched targets that contained at least one sequence of transit-like events looking for other signals that met these criteria indicating possible multiple planet systems. This search added an additional 7,698 transit-like signals for a total of 20,367 – nearly four times the current official tally of confirmed planets and planet candidates. The team compared these prospective finds against the set of currently known and vetted planets discovered in the Kepler data and were able to recover 90.3% of them. The difference is believed to be the result of changes made to the TPS algorithm over the years to make it more conservative and less prone to false positives.

Looking at just the subset of the best 1,752 “golden KOIs” (Kepler Objects of Interest) which represent the most secure planet detections to date, Seader et al. were able to recover 99.1% during their processing of 17 quarters of data using the latest version of TPS. The team believes that further refinements to the TPS algorithm should improve the recovery rate while continuing to suppress false positives. With these improvements and factoring in that only about 10% of such detections prove to be false positives, it appears that Kepler’s eventual tally of planetary finds from its primary mission could reach on the order of at least 20,000 extrasolar planets!

Seader et al. go on to report some details of eight new transit event sequences that had never been reported previously. Based on an initial analysis, six of these appear to be nearly Earth-size planets that seem to be in or near the habitable zones (HZ) of their sun. The other two candidates are of interest because they are smaller than the Earth. The properties of these eight planet candidates calculated using the data in Kepler’s current target data base are listed below in Table 1. The data are taken directly from Seader et al. except for the effective stellar flux value, S_eff, which I calculated using information from the paper and Kepler’s target data base.

Table 1: Summary of New Planets Described in Seader et al.

While these results are very preliminary and will be subject to updates, two of these unconfirmed planets immediately caught my attention as orbiting Sun-like stars inside the conservative definition of the HZ by Kopparapu et al.. By this definition, the inner edge of the HZ is defined as the S_eff where a runaway moist greenhouse effect sets in. The outer edge of the HZ is defined by the S_eff of the maximum greenhouse effect limit beyond which a CO₂-dominated greenhouse can no longer maintain surface temperatures above freezing.

KIC 8311864 (Kepler 452)

The planet candidate found orbiting KIC 8311864 has an orbital period of 384.85 days. Based on the initial assessment of the properties of this star, this planet candidate would have a radius 1.19 times that of the Earth (or R_E) and an S_eff I estimate to be 0.56. Based on the work by Rogers, which indicates that planets transition from being predominantly rocky to non-rocky at a radius no greater than 1.6 R_E, it seems likely that this candidate is a rocky planet like the Earth (see “Habitable Planet Reality Check: Terrestrial Planet Size Limit”). With an effective temperature of 5,587 K (hinting that the star is a G-type star slightly cooler than the Sun), the HZ for KIC 8311864 as defined by Kopparapu et al. corresponds to an S_eff from 1.08 to 0.34 assuming an Earth-mass planet orbiting this star indicating that this planet candidate is comfortably inside the HZ.

Even though this initial assessment seems very promising, a more in depth analysis currently underway by a team led by Jon Jenkins (NASA Ames Research Center) indicates that the stellar properties that they have derived from spectroscopic analysis of KIC 8311864 differ significantly from those currently in the Kepler target data base. Their data, which will be the subject of a discovery paper currently being prepared, indicate that this star is actually ~50% larger than originally thought. As a result, the radius of the planet candidate will also be larger – on the order of 1.8 R_E. With this larger radius estimate, this planet candidate now has about a one-in-three chance of being a rocky planet and is more likely to be a mini-Neptune instead. The update in stellar properties also implies that the S_eff might be about double the earlier estimate placing this world near the inner edge of the HZ.

While we will have to wait until the paper by Jenkins et al. is released to get all of the details and perform a better habitability assessment, this world’s prospects for being potentially habitable are not as good as first believed but it is still worth additional consideration. The situation with this new find also illustrates the need for follow-up observations of host stars to refine important stellar parameters which in turn affect the derived properties of planetary finds.

(Note: Since the original publication of this post, this object was subsequently confirmed as Kepler 452b. A full discussion of this find can be found in “Habitable Planet Reality Check: Kepler 452b” and “Is Kepler 452b a Rocky Planet or Not?“)

KIC 9674320

The planet candidate found orbiting KIC 9674320 has an orbital period of 317.05 days. Assuming that the stellar parameters for KIC 9674320 in the Kepler target data base are accurate, this world has a radius of 1.66 R_E and an S_eff that I estimate to be 0.57. With an effective temperature of 5,370 K (hinting this star is a late G-type or maybe an early K-type star), the HZ for KIC 9674320 as defined by Kopparapu et al. has S_eff values ranging from 1.06 to 0.33 for an Earth-mass planet. While the size of this planet makes it a bit more likely to be a mini-Neptune instead of a terrestrial planet, its S_eff value seems to place it comfortably inside the HZ. As the episode with KIC 8322864 has shown us, however, detailed follow-up observations will be needed to pin down the properties of this planet candidate and its host star more accurately. But at this point, it seems to be a promising potentially habitable planet candidate orbiting a Sun-like star.

Conclusions

The first look at all 17 quarters of data from Kepler’s primary mission strongly suggests that on the order of another 15,000 planets will be eventually added to the current official tally of confirmed planets and planet candidates. In their analysis, Seader et al. have identified another half dozen roughly Earth-size planet candidates that might be in or near the HZ of their systems. I find two of these candidates, one each orbiting the Sun-like stars KIC 8311864 and KIC 9674320, to be more promising than the other four which seem to have S_eff values too high to be in a conservatively defined HZ.

Of course, this assessment is subject to change based on follow-up observations needed not only to confirm the planetary status of these finds but also refine their properties. These promising worlds can be added to the growing list of similar potentially habitable planet candidates orbiting Sun-like stars currently subject to follow-up investigation (see “Earth Twins on the Horizon?”). While much more work remains to be done, these candidates hint at the discoveries yet to come from the continuing analysis of the data from Kepler’s primary mission as well as its ongoing extended K2 mission which has just started to produce discoveries (see “Habitable Planet Reality Check: Kepler’s New K2 Finds”).

Follow Drew Ex Machina on Facebook.

Related Reading

In addition to the articles cited above, there is an ever-growing list of articles on Drew Ex Machina related to the results from NASA’s Kepler mission. A complete list of these articles can be found on this web site’s Kepler mission page.

General References

R. K. Kopparapu et al., “Habitable zones around main-sequence stars: new estimates”, The Astrophysical Journal, Vol. 765, No. 2, Article ID. 131, March 10, 2013

Ravi Kumar Kopparapu et al., “Habitable zones around main-sequence stars: dependence on planetary mass”, The Astrophysical Journal Letters, Vol. 787, No. 2, Article ID. L29, June 1, 2014

Leslie A. Rogers, “Most 1.6 Earth-Radius Planets are not Rocky”, arVix 1407.4457 (submitted to The Astrophysical Journal), July 16, 2014 [Preprint]

Shawn Seader et al., “Detection of Potential Transits Signal in 17 Quarter of Kepler Mission Data”, arVix 1501.03586, submitted January 15, 2015 [Preprint]