Search for exoplanets with the radial-velocity technique: quantitative diagnostics of stellar activity

Abstract

Aims: Stellar activity may complicate the analysis of high-precision radial-velocity spectroscopic data when looking for exoplanets signatures. We aim at quantifying the impact of stellar spots on stars with various spectral types and rotational velocities and comparing the simulations with data obtained with the HARPS spectrograph. Methods: We have developed detailed simulations of stellar spots and estimated their effects on a number of observables commonly used in the analysis of radial-velocity data when looking for extrasolar planets, such as radial-velocity curves, cross-correlation functions, bisector spans and photometric curves. The computed stellar spectra are then analyzed in the same way as when searching for exoplanets. Results: 1) A first grid of simulation results is built for F-K type stars, with different stellar and spot properties. 2) It is shown quantitatively that star spots with typical sizes of 1% can mimic both radial-velocity curves and the bisector behavior of short-period giant planets around G-K type stars with a vsini lower than the spectrograph resolution. For stars with intermediate vsini, smaller spots may produce similar features. In these cases, additional observables (e.g., photometry, spectroscopic diagnostics) are mandatory to confirm the presence of short-period planets. We show that, in some cases, photometric variations may not be enough to clearly rule out spots as explanations of the observed radial-velocity variations. This is particularly important when searching for super-Earth planets. 3) It is also stressed that quantitative values obtained for radial-velocity and bisector span amplitudes depend strongly on the detailed star properties, on the spectrograph used, on the set of lines used, and on the way they are measured.