Modeling of Granulation in Red Supergiants in the Magellanic Clouds with the Gaussian Process Regressions

Abstract

The granulation of red supergiants (RSGs) in the Magellanic Clouds are systematically investigated by combining the latest RSGs samples and light curves from the Optical Gravitational Lensing Experiment and the All-Sky Automated Survey for Supernovae. The present RSGs samples are firstly examined for foreground stars and possible misidentified sources, and the light curves are sequentially checked to remove the outliers by white noise and photometric quality. The Gaussian Process regression is used to model the granulation, and the Markov Chain Monte Carlo is applied to derive the granulation amplitude $\sigma$ and the period of the undamped oscillator $\rho$, as well as the damping timescale $\tau$. The dimensionless quality factor $Q$ is then calculated through $Q=\pi \tau/\rho$. RSGs around $Q = 1/\sqrt{2}$ are considered to have significant granulation signals and are used for further analysis. Combining granulation parameters with stellar parameters, robust scaling relations for the timescale $\rho$ are established, while the scaling relations for amplitude $\sigma$ are represented by a piecewise function, possibly related to the tendency of amplitudes in faint RSGs to converge towards a certain value. Comparing results between the SMC and LMC confirms that amplitudes and timescales become larger with metallicity. In examining the scaling relations between the two galaxies, it is found that $\rho$ is nearly independent of metallicity, whereas $\sigma$ is more significantly affected by metallicity. The Gaussian Process method is compared with the periodogram fitting of the granulations, and the advantages of either are discussed.