Identification and Removal of Noise Modes in Kepler Photometry

We present the transiting exoearth robust reduction algorithm (TERRA)—a novel framework for identifying and removing instrumental noise in Kepler photometry. We identify instrumental noise modes by finding common trends in a large ensemble of light curves drawn from the entire Kepler field of view. Strategically, these noise modes can be optimized to reveal transits having a specified range of timescales. For Kepler target stars of low photometric noise, TERRA produces ensemble-calibrated photometry having 33 ppm rms scatter in 12 hr bins, rendering individual transits of Earth-size planets around Sun-like stars detectable as ∼3σ signals.