2015A&A...583A.118M

Stellar variability, at a variety of timescales, can strongly affect the ability to detect exoplanets, in particular when using radial velocity (RV) techniques. Accurately characterized solar variations are precious in this context to study the impact of stellar variations on planet detectability. Here we focus on the impact of small timescale variability. The objective of this paper is to model realistic RV time series due to granulation and supergranulation and to study in greater detail the impact of granulation and supergranulation on RV times series in the solar case. We have simulated a collection of granules and supergranules evolving in time to reproduce solar photometric and RV time series. Synthetic time series are built over the full hemisphere over one solar cycle. We obtain intensity and RV rms due to solar granulation of respectively 0.8m/s and 67 ppm, with a strong variability at timescales up to more than 1 h. The rms RV due to supergranulation is between 0.28 and 1.12m/s. To minimize the effect of granulation, the best strategy is to split the observing time during the night into several periods instead of observing over a consecutive duration. However, the best strategy depends on the precise nature of the signal. The granulation RV remains large after even an hour of smoothing (about 0.4m/s) while the supergranulation signal cannot be significantly reduced on such timescales: a reduction of a factor 2 in rms RV can for example be obtained over 7 nights (with 26min/night). The activity RV variability dominates at larger timescales. Detection limits can easily be as high as 1M_Earth or above for periods of tens or hundreds of days. The impact on detection limits is therefore important and may prevent the detection of 1M_Earth planets for long orbital periods, while the impact is much smaller at small orbital periods. These results do not take the presence of pulsations into account.