Exoplanet Imitators: A test of stellar activity behavior in radial velocity signals

Abstract

Accurately modeling effects from stellar activity is a key step in detecting radial velocity signals of low-mass and long-period exoplanets. Radial velocities from stellar activity are dominated by magnetic active regions that move in and out of sight as the star rotates, producing signals with timescales related to the stellar rotation period. Methods to characterize radial velocity periodograms assume that peaks from magnetic active regions will typically occur at the stellar rotation period or a related harmonic. However, with surface features unevenly spaced and evolving over time, signals from magnetic activity are not perfectly periodic, and the effectiveness of characterizing them with sine curves is unconfirmed. With a series of simulations, we perform the first test of common assumptions about signals from magnetic active regions in radial velocity periodograms. We simulate radial velocities with quasi-periodic signals that account for evolution and migration of magnetic surface features. As test cases, we apply our analysis to two exoplanet hosts, Kepler-20 and K2-131. Simulating observing schedules and uncertainties of real radial velocity surveys, we find that magnetic active regions commonly produce maximum periodogram peaks at spurious periods unrelated to the stellar rotation period: 81% and 72% of peaks, respectively for K2-131 and Kepler-20. These unexpected peaks can potentially lead to inaccuracies in derived planet masses. We also find that these spurious peaks can sometimes survive multiple seasons of observation, imitating signals typically attributed to exoplanet companions.