Discovering planets on year-long orbits by mitigating local background variations

[christinahedges]

Discovering small planets on year-long orbits informs our understanding of our place in the universe by helping to establish how common Earth-like planets may be in the universe. Kepler’s ability to discover these planets has been limited by the presence of systematics with similar year-long periodicities. Certain CCD channels (in particular channels 26, 44 and 58) experience the rolling band effect, where the background shows a strong time-varying component appearing as bands moving across the detector (see §6.7 of Van Cleve & Caldwell 2016).

The rolling band artifact often adds spurious signals which mimic small planet transits. Because temperature variations trigger the artifact, the presence and characteristics of the rolling bands are correlated with Kepler’s own ∼year-long orbit around the Sun, leading to a significant excess of false positive planet candidates on ∼year-long orbits (Thompson et al. 2018).

Local background estimation techniques may help to remove the systematic10 , but the Kepler pipeline only applied global background models owing to the limited number of pixels that were downlinked from the spacecraft. Detection and confirmation of small planet candidates on year-long orbits, and thus estimates of the occurrence rates, would particularly benefit from additional research into the removal of this noise component, or its probabilistic modeling.