Investigating stellar RV behavior in Fourier space
Accurately modeling stellar activity signals in radial velocity (RV) data is now an essential step in detecting and measuring masses of very low-mass planets. Effects from the rotation of the stellar surface dominate stellar activity in RVs on a timescale similar to the stellar rotation period. A common assumption is that effects from stellar rotation will present as the maximum peak in a star's RV periodogram, with an uncertainty equal to the FWHM of the peak. However, stellar rotation signals are not perfectly periodic; spot groups appear, move, evolve, and disappear on the stellar surface. Gaussian processes (GP) can model quasi-periodic signals typical of these evolving, rotating spots, and therefore provide a tool for simulating stellar activity. Here I present details of my simulations investigating the behavior of stellar RVs in Fourier space, with signals modeled by quasi-periodic GP kernels and real HARPS-N observation schedules. My results thus far indicate that stellar activity rarely peaks at the stellar rotation period in Fourier space and can impersonate planetary companions orbiting at periods of many days separation from the stellar rotation period.