Statistics of Stellar Variability from Kepler - I: Revisiting Quarter 1 with an Astrophysically Robust Systematics Correction

TL;DR

This study analyzes stellar variability in Kepler data, revealing that most stars are more variable than the active Sun, with variability characteristics differing across spectral types and linked to stellar activity and location.

Contribution

Introduces a new astrophysically robust systematics correction for Kepler data, enabling detailed analysis of stellar variability and its dependence on spectral type and activity.

Findings

60% of stars are more variable than the active Sun

Periodic variability found in 16% of stars, with spectral type differences

Stochastic variability amplitude and timescale increase towards later spectral types

Abstract

We investigate the variability properties of main sequence stars in the first month of Kepler data, using a new astrophysically robust systematics correction, and find that 60% of stars are more variable then the active Sun. We define low and high variability samples, with a cut corresponding to twice the variability level of the active Sun, and compare the properties of the stars belonging to each sample. We show tentative evidence that the more active stars have lower proper motions and may be located closer to the galactic plane. We also investigate the frequency content of the variability, finding clear evidence for periodic or quasi-periodic behaviour in 16% of stars, and showing that there exist significant differences in the nature of variability between spectral types. Of the periodic objects, most A and F stars have short periods (< 2 days) and highly sinusoidal variability, suggestive of pulsations, whilst G, K and M stars tend to have longer periods (> 5 days, with a trend towards longer periods at later spectral types) and show a mixture of periodic and stochastic variability, indicative of activity. Finally, we use auto-regressive models to characterise the stochastic component of the variability, and show that its typical amplitude and time-scale both increase towards later spectral types, which we interpret as a corresponding increase in the characteristic size and life-time of active regions.