Modeling the quasiperiodic radial velocity variations of gamma Draconis

TL;DR

This paper investigates the complex radial velocity variations of gamma Draconis, proposing a two-component quasi-periodic Gaussian process model to explain observed signals, including potential magnetic activity and other phenomena.

Contribution

It introduces a novel two-component quasi-periodic Gaussian process model to better fit gamma Draconis RV data, advancing understanding of stellar activity signals.

Findings

A single quasi-periodic component is insufficient to model the data.

The best-fit model includes signals at 705 days and 15 days.

The 705-day signal may be related to magnetic activity.

Abstract

gamma Draconis, a K5III star, showed radial velocity (RV) variations consistent with a 10.7 Jupiter mass planet from 2003-2011. After 2011, the periodic signal decayed, then reappeared with a phase shift. Hatzes et al. (2018) suggested that gamma Dra's RV variations could come from oscillatory convective modes, but did not fit a mathematical model. Here we assess whether a quasi-periodic Gaussian process (GP)---appropriate when spots with finite lifetimes trace underlying periodicity---can explain the RVs. We find that a model with only one quasiperiodic signal is not adequate: we require a second component to fit the data. The best-fit model has quasi-periodic oscillations with P1 = 705 days and P2 = 15 days. The 705-day signal may be caused by magnetic activity. The 15-day period requires further investigation.