# Photometric detection of internal gravity waves in upper main-sequence stars. III. Comparison of amplitude spectrum fitting and Gaussian process regression using celerite2

**Authors:** Dominic M. Bowman, Trevor Z. Dorn-Wallenstein  
**arXiv:** [2211.08347](https://arxiv.org/abs/2211.08347)  
**Date:** 2022-11-15  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2211.08347v1/

## Key points

- Space telescopes such as CoRoT, Kepler/K2, and TESS have provided a wealth of time series photometric data on massive stars
- Stochastic low-frequency (SLF) variability has been observed in essentially all early-type main-sequence stars and has also been reported in later evolutionary stages of massive stars
- The authors aim to compare the properties of SLF variability in massive main-sequence stars observed by the TESS mission using different statistical methods
- There is a correlation between the morphology of SLF variability and a star's location in the HR diagram
- GP regression is a useful methodology to efficiently characterize SLF variability in massive stars compared to previous techniques used in literature
- There exists a correlation between a star's SLF variability, particularly its characteristic frequency, and its location on the HR diagram for main-sequence massive stars
- Coherency is also dependent on mass and age

## Summary

In recent years, space telescopes such as CoRoT, Kepler/K2, and TESS have provided a wealth of time series photometric data on massive stars. These observations have revealed a diverse range of variability mechanisms in these stars, including stochastic low-frequency (SLF) variability. SLF variability has been observed in essentially all early-type main-sequence stars and has also been reported in later evolutionary stages of massive stars. In this study, the authors aim to compare the properties of SLF variability in massive main-sequence stars observed by the TESS mission using different statistical methods. They confirm the correlation between the morphology of SLF variability and a star's location in the HR diagram. The authors analyze a sample of 30 massive stars observed by TESS and compare the resultant parameters of SLF variability obtained from fitting the amplitude spectrum of the light curve with those inferred from fitting the covariance structure of the light curve using celerite2 Gaussian process (GP) regression software and a damped SHO kernel. The authors find that there is a difference in the characteristic frequency obtained from amplitude spectrum fitting and from fitting the covariance structure of the light curve using GP regression with celerite2 for only a minority of their sample. However, they note that trends among mass, age, and properties of SLF variability previously reported remain unaffected. The authors conclude that GP regression is a useful methodology to efficiently characterize SLF variability in massive stars compared to previous techniques used in literature. The study confirms that there exists a correlation between a star's SLF variability, particularly its characteristic frequency, and its location on the HR diagram for main-sequence massive stars. Additionally, there is also a distribution in stochasticity indicating that coherency is also dependent on mass and age. Overall, this study highlights how time series photometry can provide detailed insights into stellar interiors through asteroseismology when coupled with spectroscopy. The investigation of SLF variability has opened up a novel method to understanding the interiors of massive stars, and this study provides an effective methodology for characterizing such variability which could be applied to future studies involving similar objects or phenomena.