Skip to main content
SearchLoginLogin or Signup

Testing the limits of multidimensional Gaussian process approaches in the detection of RV planet signals

Presentation #102.402 in the session Poster Session.

Published onJun 20, 2022
Testing the limits of multidimensional Gaussian process approaches in the detection of RV planet signals

Active regions on stellar surfaces generate radial velocity variations that can mimic or conceal planet-induced signals. A common approach to dealing with stellar activity in radial velocity data is to model the stellar signal as a stochastic Gaussian Process (GP). In this contribution, we test the detectability of single planets orbiting stars exhibiting different levels of activity by employing two-dimensional GP regression on simulated radial velocity and activity indicator data sets. We will show some practical examples of how this approach can be used to detect planet signals in active stars. We investigate the interaction between the planet- and stellar-induced signals by analysing the effects the different Keplerians have on constraining any combination of the tested GP hyperparameters, and thus on our ability to characterise the stellar behaviour. We explore the detectability of planets using different observational sampling scenarios and determine a 3-sigma detection threshold for a set of discrete semi-amplitudes and diagnose certain edge cases.

Comments
0
comment
No comments here