Title:Contamination in the Kepler Field. Identification of 685 KOIs as False Positives Via Ephemeris Matching Based On Q1-Q12 Data

Abstract:The Kepler mission has to date found almost 6,000 planetary transit-like signals, utilizing three years of data for over 170,000 stars at extremely high photometric precision. Due to its design, contamination from eclipsing binaries, variable stars, and other transiting planets results in a significant number of these signals being false positives. This directly affects the determination of the occurrence rate of Earth-like planets in our Galaxy, as well as other planet population statistics. In order to detect as many of these false positives as possible, we perform ephemeris matching among all transiting planet, eclipsing binary, and variable star sources. We find that 685 Kepler Objects of Interest - 12% of all those analyzed - are false positives as a result of contamination, due to 409 unique parent sources. Of these, 118 have not previously been identified by other methods. We estimate that ~35% of KOIs are false positives due to contamination, when performing a first-order correction for observational bias. Comparing single-planet candidate KOIs to multi-planet candidate KOIs, we find an observed false positive fraction due to contamination of 16% and 2.4% respectively, bolstering the existing evidence that multi-planet KOIs are significantly less likely to be false positives. We also analyze the parameter distributions of the ephemeris matches and derive a simple model for the most common type of contamination in the Kepler field. We find that the ephemeris matching technique is able to identify low signal-to-noise false positives that are difficult to identify with other vetting techniques. We expect false positive KOIs to become more frequent when analyzing more quarters of Kepler data, and note that many of them will not be able to be identified based on Kepler data alone.