Filtering solar-like oscillations for exoplanet detection in radial velocity observations

Cool main-sequence, sub-giant and red-giant stars all show solar-like oscillations, pulsations that are excited and intrinsically damped by near-surface convection. Many overtones are typically excited to observable amplitudes, giving a rich spectrum of detectable modes. These modes provide a wealth of information on fundamental stellar properties. However, the radial velocity shifts induced by these oscillations can also be problematic when searching for low-mass, long-period planets; this is because their amplitudes are large enough to completely mask such minute planetary signals. Here we show how fine-tuning exposure times to the stellar parameters can help efficiently average out the solar-like-oscillation-induced shifts. To reduce the oscillation signal to the radial velocity precision commensurate with an Earth-analogue, we find that for cool, low-mass stars (near spectral type K) the necessary exposure times may be as short as 4 minutes, while for hotter, higher-mass stars (near spectral type F, or slightly evolved) the required exposure times can be longer than 100 minutes. We provide guideline exposure durations required to suppress the total observed amplitude due to oscillations to a level of $0.1\,\rm m\,s^{-1}$, and a level corresponding to the Earth-analogue reflex amplitude for the star. Owing to the intrinsic stochastic variability of the oscillations, we recommend in practice choosing short exposure durations at the telescope and then averaging over those exposures later, as guided by our predictions. To summarize, as we enter an era of $0.1\,\rm m\,s^{-1}$ instrumental precision, it is critical to tailor our observing strategies to the stellar properties.