Re-Evaluating Small Long-Period Confirmed Planets From Kepler

Abstract

We re-examine the statistical confirmation of small long-period Kepler planet candidates in light of recent improvements in our understanding of the occurrence of systematic false alarms in this regime. Using the final Data Release 25 (DR25) Kepler planet candidate catalog statistics, we find that the previously confirmed single planet system Kepler-452b no longer achieves a 99% confidence in the planetary hypothesis and is not considered statistically validated in agreement with the finding of Mullally et al. (2018). For multiple planet systems, we find that the planet prior enhancement for belonging to a multiple planet system is suppressed relative to previous Kepler catalogs, and we identify the multi-planet system member, Kepler-186f, no longer achieves a 99% confidence in the planetary hypothesis. Because of the numerous confounding factors in the data analysis process that leads to the detection and characterization of a signal, it is difficult to determine whether any one planetary candidate achieves a strict criterion for confirmation relative to systematic false alarms. For instance, when taking into account a simplified model of processing variations, the additional single planet systems Kepler-443b, Kepler-441b, Kepler-1633b, Kepler-1178b, and Kepler-1653b have a non-negligible probability of falling below a 99% confidence in the planetary hypothesis. The systematic false alarm hypothesis must be taken into account when employing statistical validation techniques in order to confirm planet candidates that approach the detection threshold of a survey. We encourage those performing transit searches of K2, TESS, and other similar data sets to quantify their systematic false alarms rates. Alternatively, independent photometric detection of the transit signal or radial velocity measurements can eliminate the false alarm hypothesis.